ECSS E HB 31 01 - Part13A Fluid Loops

ECSS-E-HB-31-01 Part 13A
5 December 2011
Space engineering
Thermal design handbook - Part 13:
Fluid Loops
ECSS Secretariat
ESA-ESTEC
Requirements & Standards Division
Noordwijk, The Netherlands
ECSSEHB3101Part13A
5December2011
Foreword
ThisHandbookisonedocumentoftheseriesofECSSDocumentsintendedtobeusedassupporting
material for ECSS Standards in space projects and applications. ECSS is a cooperative effort of the
EuropeanSpaceAgency,nationalspaceagenciesandEuropeanindustryassociationsforthepurpose
ofdevelopingandmaintainingcommonstandards.
The material in this Handbook is a collection of data gathered from many projects and technical
journals which provides the reader with description and recommendation on subjects to be
consideredwhenperformingtheworkofThermaldesign.
Thematerialforthesubjectshasbeencollatedfromresearchspanningmanyyears,thereforeasubject
mayhavebeenrevisitedorupdatedbyscienceandindustry.
The material is provided as good background on the subjects of thermal design, the reader is
recommended to research whether a subject has been updated further, since the publication of the
materialcontainedherein.
This handbook has been prepared by ESA TEC MT/QR division, reviewed by the ECSS Executive
SecretariatandapprovedbytheECSSTechnicalAuthority.
Disclaimer
ECSSdoesnotprovideanywarrantywhatsoever,whetherexpressed,implied,orstatutory,including,
butnotlimitedto,anywarrantyofmerchantabilityorfitnessforaparticularpurposeoranywarranty
that the contents of the item are errorfree. In no respect shall ECSS incur any liability for any
damages,including,butnotlimitedto,direct,indirect,special,orconsequentialdamagesarisingout
of,resultingfrom,orinanywayconnectedtotheuseofthisdocument,whetherornotbasedupon
warranty,businessagreement,tort,orotherwise;whetherornotinjurywassustainedbypersonsor
propertyorotherwise;andwhetherornotlosswassustainedfrom,oraroseoutof,theresultsof,the
item,oranyservicesthatmaybeprovidedbyECSS.
Publishedby: ESARequirementsandStandardsDivision
ESTEC,P.O.Box299,
2200AGNoordwijk
TheNetherlands
Copyright: 2011bytheEuropeanSpaceAgencyforthemembersofECSS
2
ECSSEHB3101Part13A
5December2011
Table of contents
1 Scope.....................................................................................................................30
2 References ............................................................................................................31
3 Terms, definitions and symbols..........................................................................32

3.1 Terms and definitions ............................................................................................... 32
3.2 Abbreviated terms .................................................................................................... 32
3.3 Symbols....................................................................................................................34
4 General introduction ............................................................................................46

4.1 Fluid loops ................................................................................................................47
4.2 Comparison between fluid loops and alternative systems........................................ 48
4.2.1 Passive thermal insulations ........................................................................48
4.2.2 Thermoelectric devices............................................................................... 48
4.2.3 Phase change materials (pcm)...................................................................49
4.2.4 Heat pipes ..................................................................................................50
4.2.5 Short-term discharge systems.................................................................... 50
5 Analysis of a fluid loop ........................................................................................52

5.1 General.....................................................................................................................52
5.2 Thermal performance ...............................................................................................53
5.3 Power requirements .................................................................................................56
6 Thermal analysis ..................................................................................................58

6.1 General.....................................................................................................................58
6.2 Analytical background .............................................................................................. 58
6.2.1 Heat transfer coefficient..............................................................................58
6.2.2 Dimensionless groups ................................................................................60
6.2.3 Simplifying assumptions .............................................................................61
6.2.4 Temperature-dependence of fluid properties ............................................. 61
6.2.5 Laminar versus turbulent fluid flow ............................................................. 63
6.2.6 Heat transfer to internal flows..................................................................... 63
6.2.7 Heat transfer to external flows....................................................................65
3
ECSSEHB3101Part13A
5December2011
6.3 Thermal performance data .......................................................................................68
6.3.1 Heat transfer to internal flow....................................................................... 68
6.3.2 Heat transfer to external flows....................................................................83
7 Frictional analysis ................................................................................................92

7.1 General.....................................................................................................................92
7.2 Analytical background .............................................................................................. 92
7.2.1 Introduction.................................................................................................92
7.2.2 Fully developed flow in straight pipes......................................................... 93
7.2.3 Temperature-dependence of fluid properties ............................................. 97
7.2.4 Several definitions of pressure loss coefficient........................................... 98
7.2.5 Entrance effects........................................................................................ 100
7.2.6 Interferences and networks ...................................................................... 101
7.2.7 Flow chart .................................................................................................102
7.3 Pressure loss data..................................................................................................105
7.3.1 Straight pipes............................................................................................105
7.3.2 Bends .......................................................................................................106
7.3.3 Sudden changes of area ..........................................................................113
7.3.4 Orifices and diaphragms........................................................................... 116
7.3.5 Screens ....................................................................................................119
7.3.6 Valves.......................................................................................................120
7.3.7 Tube banks............................................................................................... 121
7.3.8 Branching of tubes.................................................................................... 124
8 Combined thermal and frictional analysis........................................................125

8.1 General...................................................................................................................125
8.2 Analogies between momentum and heat transfer .................................................. 125
8.2.1 The Reynolds analogy.............................................................................. 125
8.2.2 The Prandtl analogy .................................................................................128
8.2.3 The Von Karman analogy......................................................................... 129
8.2.4 Other analogies ........................................................................................129
9 Heat transfer enhancement ...............................................................................130

9.1 General...................................................................................................................130
9.1.1 Basic augmentation mechanisms.............................................................131
9.1.2 Criterion for the evaluation of the several techniques .............................. 132
9.1.3 Index of the compiled data. ...................................................................... 133
9.1.4 Validity of the empirical correlations ......................................................... 133
9.2 Single-phase forced convection data ..................................................................... 136
4
ECSSEHB3101Part13A
5December2011
10 Working fluids...................................................................................................170
10.1 General...................................................................................................................170
10.2 Cooling effectiveness of a fluid...............................................................................170
10.2.1 Simplified fluid loop configuration ............................................................. 172
10.2.2 Thermal performance of the simplified loop ............................................. 172
10.2.3 Power requirements of the simplified loop................................................ 173
10.2.4 Several examples .....................................................................................173
10.3 Properties of liquid coolants ................................................................................... 178
10.4 Properties of dry air ................................................................................................213
11 Heat exchangers...............................................................................................215
11.1 General...................................................................................................................215
11.2 Basic analysis.........................................................................................................218
11.2.1 Introduction............................................................................................... 218
11.2.2 Analytical background ..............................................................................219
11.2.3 Exchanger performance ...........................................................................222
11.3 Exchanging surface geometries ............................................................................. 237
11.3.1 Tubular surfaces.......................................................................................238
11.3.2 Plate-fin surfaces......................................................................................241
11.3.3 Finned tubes............................................................................................. 247
11.3.4 Matrix surfaces .........................................................................................249
11.4 Deviations from basic analysis ............................................................................... 250
11.4.1 Introduction............................................................................................... 250
11.4.2 Longitudinal heat conduction.................................................................... 251
11.4.3 Flow maldistribution.................................................................................. 254
11.5 Manufacturing defects ............................................................................................ 264
11.5.1 Introduction............................................................................................... 264
11.5.2 Variations of the flow passages................................................................ 264
11.5.3 Fin leading edge imperfections................................................................. 268
11.5.4 Brazing .....................................................................................................268
11.6 In service degradation ............................................................................................ 272
11.6.1 Introduction............................................................................................... 272
11.6.2 Fouling......................................................................................................272
11.7 Existing systems.....................................................................................................276
12 Pumps................................................................................................................284
12.1 General...................................................................................................................284
12.2 Specified speed......................................................................................................288
5
ECSSEHB3101Part13A
5December2011
12.3 Net suction energy ................................................................................................. 290
12.4 Requirements for spaceborne pumps .................................................................... 291
12.5 Commercially available pumps...............................................................................292
12.6 European pump manufacturers .............................................................................. 298
13 System optimization.........................................................................................299
13.1 General...................................................................................................................299
13.2 Basic analysis.........................................................................................................299
13.2.1 Interface heat exchanger.......................................................................... 300
13.2.2 Supply and return plumbing...................................................................... 301
13.2.3 Radiator ....................................................................................................302
13.3 Special examples ...................................................................................................302
13.3.1 Constraints based on source temperature ............................................... 303
13.3.2 Constraints imposed by the integration .................................................... 306
14 Two-phase flow.................................................................................................310
14.1 General...................................................................................................................310
14.2 Pressure loss..........................................................................................................313
14.2.1 Lockhart-martinelli correlation .................................................................. 313
14.2.2 Improvements upon martinelli correlation................................................. 318
14.3 Annular flow............................................................................................................319
14.3.1 Ideal annular flow model........................................................................... 320
14.3.2 Annular flow with entrainment model........................................................ 329
14.4 Condensation in ducts............................................................................................343
14.4.1 Condensing flow model ............................................................................ 343
14.4.2 Variation of the vapor quality along the duct in the stratified model ......... 349
14.4.3 Limits of validity of the stratified model..................................................... 351
14.4.4 Annular flow model ...................................................................................352
14.4.5 Variation of the vapor quality along the duct in the annular model........... 356
15 Two-phase thermal transport systems ...........................................................359

15.1 General...................................................................................................................359
15.1.1 Evolution of thermal transport systems .................................................... 359
15.1.2 Two-phase loop general layout ................................................................ 360
15.1.3 About the nomenclature of this clause ..................................................... 363
15.2 Tms trade-off study ................................................................................................363
15.2.1 TMS study baseline ..................................................................................366
15.2.2 TMS design concepts ............................................................................... 366
15.2.3 Evaluation of tms concepts....................................................................... 369
6
ECSSEHB3101Part13A
5December2011
15.3 Design for orbital average load .............................................................................. 372
15.3.1 Phase change capacitor performance...................................................... 372
15.4 Off-design operation...............................................................................................378
15.4.1 Temperature control .................................................................................380
15.4.2 Instrumentation requirements................................................................... 383
15.5 Radiator-loop interaction ........................................................................................384
15.5.1 Boosting radiator temperature with a heat pump...................................... 385
15.5.2 Thermal-storage assisted radiator............................................................ 390
15.5.3 Steerable radiators ................................................................................... 393
15.5.4 Radiators coupling.................................................................................... 404
15.6 Capillary pumped loop (cpl) technology ................................................................. 406
15.6.1 Advantages of cpl systems....................................................................... 410
15.6.2 CPL performance constraints ...................................................................410
15.6.3 CPL basic system concept ....................................................................... 410
15.7 Components ...........................................................................................................413
15.7.1 Pumping systems .....................................................................................413
15.7.2 Mounting plates ........................................................................................416
15.7.3 Vapour quality sensors ............................................................................. 418
15.7.4 Fluid disconnects......................................................................................422
16 Control technology...........................................................................................424
16.1 Basic definitions .....................................................................................................424
16.2 General description of control systems .................................................................. 425
16.2.1 Introduction............................................................................................... 425
16.2.2 Closed-loop control systems ....................................................................426
16.2.3 Open-loop control system......................................................................... 426
16.2.4 Adaptative control systems....................................................................... 427
16.2.5 Learning control system ...........................................................................428
16.2.6 Trade-off of open- and closed-loop control systems ................................ 428
16.3 Basic control actions ..............................................................................................433
16.3.1 Introduction............................................................................................... 433
16.3.2 Two-position or on-off control action ........................................................ 434
16.3.3 Proportional control action (p controller)................................................... 435
16.3.4 Integral control action (i controller). .......................................................... 436
16.3.5 Proportional-integral control action (pi controller)..................................... 437
16.3.6 Proportional-derivative control action (pd controller) ................................ 438
16.3.7 Proportional-integral-derivative control action (pid controller) .................. 439
16.3.8 Summary ..................................................................................................440
7
ECSSEHB3101Part13A
5December2011
16.4 Implementation techniques of control laws ............................................................ 441
16.4.1 Introduction............................................................................................... 441
16.4.2 Devices characterization .......................................................................... 443
16.4.3 Analog-controller implementation techniques........................................... 447
16.4.4 Summary ..................................................................................................458
16.5 Hardware description ............................................................................................. 460
16.5.1 Introduction............................................................................................... 460
16.5.2 Controllers ................................................................................................462
16.5.3 Sensors ....................................................................................................467
16.5.4 Actuators. Control valves..........................................................................470
16.6 Control software .....................................................................................................471
16.7 Existing systems.....................................................................................................474
16.7.1 Space radiator system.............................................................................. 474
Bibliography...........................................................................................................478
Figures
Figure 5-1: Schematic representation of the fluid loop........................................................... 52
Figure 6-1: Nusselt numbers, Nu, for fully developed laminar flow through straight pipes
of several cross-sectional shapes. Nuq is the Nusselt number for constant
heat transfer rate along the duct, and NuT that for constant wall temperature
along the duct. From Kays & London (1964) [102]. ............................................. 70
Figure 6-2: Nusselt numbers, Nu, vs. ratio, a/b, of short side to long side for fully
developed laminar flow through straight pipes of rectangular cross section.
From Kays & London (1964) [102]....................................................................... 70
Figure 6-3: Nusselt numbers, Nu, vs. ratio of inner to outer diameter, r1/r2, for fully
developed laminar flow in concentric- circular-tube annuli. Constant heat
transfer rate. From Kays & London (1964) [102]. ................................................ 70
Figure 6-4: Influence of coefficients, Z, vs. ratio of inner to outer diameter, r1/r2, for fully
developed laminar flow in concentric-circular-tube annuli. Constant heat
transfer rate. From Kays & London (1964) [102]. ................................................ 71
Figure 6-5: Nusselt number, Nu, vs. Dean number, K, for fully developed laminar flow
in curved pipe of circular cross section. Constant heat transfer rate. Results
are shown for different Prandtl numbers, Pr. Calculated by the compiler
after Mori & Nakayama (1965) [128]....................................................................71
Figure 6-6: Thermal entry length Nusselt numbers, Nu, vs. non-dimensional axial
distance, x+, for laminar flow through straight pipes. Constant wall
temperature. Calculated by the compiler after Kays (1966) [101]........................ 72
Figure 6-7: Thermal entry length Nusselt number, Nux, vs. non-dimensional axial
distance, x+, for laminar flow through straight pipes. Constant heat transfer
rate. Also shown the influence coefficient, Z, for laminar flow between
parallel plates with one side insulated. Calculated by the compiler after
Kays (1966) [101]. ...............................................................................................72
8
ECSSEHB3101Part13A
5December2011
Figure 6-8: Thermal entry length Nusselt numbers, Nux, and influence coefficients, Z,
vs. dimensionless axial distance, x+, for laminar flow in concentric-circular-
tube annuli. Constant heat transfer rate. Calculated by the compiler after
Kays (1966) [101]. ...............................................................................................73
Figure 6-9: Thermal entry length Nusselt number, Nux, vs. non dimensional distance
along the coil centerline, x+, for laminar flow through a coil. The results are
given for two values of the ratio, r/R, between the cross-sectional radius
and the coil radius. Constant wall temperature. Calculated by the compiler
after Kubair & Kuloor (1966) [111]. ...................................................................... 73
Figure 6-10: Nusselt numbers, Nu, vs. non-dimensional axial distance, x+, for the
combined hydrodynamical and thermal entry length. Laminar flow through
straight pipes of circular cross section. Constant wall temperature. Pr = 0.7.
Replotted by the compiler after ESDU 68006 (1968) [48]. .................................. 74
Figure 6-11: Local Nusselt number, Nux, vs. non-dimensional axial distance, x+, for the
combined hydrodynamical and thermal entry length. Laminar flow through
straight pipes of circular cross section. Constant heat transfer rate. Results
after Heaton et al. (1964) [82].............................................................................. 74
Figure 6-12: Local Nusselt number, Nux, and influence coefficient, Z, vs. dimensionless
axial distance, x+, for the combined hydrodynamical and thermal entry
length. Laminar flow between parallel plates, one of them insulated.
Constant heat transfer rate. Results are shown for different Prandtl
numbers, Pr. Calculated by the compiler after Heaton et al. (1964) [82]. ............ 75
Figure 6-13: Local Nusselt number, Nux, vs. Reynolds number, Re, for fully developed
transitional flow through cylindrical ducts of circular cross section. Constant
wall temperature. Gas Flow (Pr 0.7). From ESDU 68006 (1968) [48]. ............. 75
Figure 6-14: Nusselt number, Nu, vs. Reynolds number, Re, for fully developed
turbulent flow through cylindrical ducts. Constant heat transfer rate. Results
after Petukhov & Roizen (1975) [143].................................................................. 76
Figure 6-15: Ratio of Nusselt number at constant heat transfer rate, Nuq, to Nusselt
number at uniform wall temperature, NuT, vs. Reynolds number, Re, for
fully developed turbulent flow through a straight pipe of circular cross
section. Results are shown for different Prandtl numbers, Pr. From Sleicher
& Tribus (1957) [167]. .......................................................................................... 76
Figure 6-16: Nusselt number, Nu, vs. Reynolds number, Re, for fully developed
turbulent flow between parallel plates, one of them insulated. Constant heat
transfer rate. Results are shown for different Prandtl numbers, Pr.
Calculated by the compiler after Kays (1966) [101]. ............................................ 77
Figure 6-17: Influence coefficient, Z, vs. Reynolds number, Re, for fully developed
turbulent flow between parallel plates. Constant heat transfer rate. Results
after Kays (1966) [101]. .......................................................................................77
Figure 6-18: Nusselt number, Nu11, and influence coefficient, Z1, vs. Reynolds number,
Re, for fully developed turbulent flow in concentric-circular-tube annuli. r1/r2
= 0,2. Constant heat transfer rate. Results are shown for different Prandtl
numbers, Pr. Calculated by the compiler after Kays (1966) [101]. ...................... 78
9
ECSSEHB3101Part13A
5December2011
Figure 6-22: Nusselt number times Prandtl number to the minus 0.4 power, NuPr-0.4, vs.
Reynolds number, Re, for fully developed turbulent flow in helically coiled
tubes. The results are given for two values of the ratio, r/R, between the
cross-sectional radius and the coil radius. Constant heat transfer rate.
Calculated by the compiler after an experimental correlation obtained by
Seban & McLaughlin (1963) [162] from data for water. ....................................... 80
distance, x/D, for fully developed turbulent flow through a straight pipe of
circular cross section. Constant wall temperature. Pr = 0.01. Results are
shown for different Reynolds numbers, Re. Calculated by the compiler after
Kays (1966) [101]. ...............................................................................................80
distance, x/D, for fully developed turbulent flow through a straight pipe of
circular cross section. Constant wall temperature. Pr = 0.7. Results are
shown for different Reynolds numbers, Re. Calculated by the compiler after
Kays (1966) [101]. ...............................................................................................81
Figure 6-25: Ratio of thermal entry length Nusselt number, Nux, to Nusselt number for
fully developed turbulent flow, Nu, vs. non-dimensional axial distance, x/D.
Straight pipe of circular cross section. Constant heat transfer rate. Pr =
0.01. Results are shown for different Reynolds numbers, Re. Calculated by
the compiler after Kays (1966) [101].................................................................... 81
fully developed turbulent flow, Nu, vs. non-dimensional axial distance, x/D.
Straight pipe of circular cross section. Constant heat transfer rate. Re =
105. Results are shown for different Prandtl numbers, Pr. Calculated by the
compiler after Kays (1966) [101].......................................................................... 82
fully developed turbulent flow, Nu, vs. non-dimensional axial distance, x/DE.
Parallel plates at distance 2DE, one of them insulated. Constant heat
transfer rate. Also shown the influence coefficient, Z. Results are shown for
three different Prandtl numbers, Pr, and two Reynolds numbers, Re.
Calculated by the compiler after Kays (1966) [101]. ............................................ 82
Figure 6-28: Nusselt number, Nu, vs. Reynolds number, Re. Flow of a fluid having
constant physical properties over a constant temperature circular cylinder
whose axis is normal to the incoming flow. From ESDU 69004 (1969) [50]. ....... 84
Figure 6-29: Effect of variable fluid properties, (a) and (b), and of inclination angle, (c),
on the Nusselt number corresponding to the flow of a fluid over a constant
temperature cylinder. Nub (Nu90) can be deduced from Figure 6-28. From
ESDU 69004 (1969) [50]. .................................................................................... 85
10
ECSSEHB3101Part13A
5December2011
Figure 6-30: Guide for the selection of the curves given in Figure 6-31 and Figure 6-32
concerning in-line tube banks of different relative pitches. From ESDU
73031 (1973) [57]. ...............................................................................................86
Figure 6-31: Reference Nusselt number, Nur, for Prb = 1, as a function of Reynolds
number, Re. In-line tube banks. See Figure 6-30 for the meaning of the
numbers which appear on the curves. From ESDU 73031 (1973) [57]. .............. 87
number. Re. In-line tube banks. See Figure 6-30 for the meaning of the
numbers which appear on the curves. From ESDU 73031 (1973) [57]. .............. 88
number. Re. In-line tube banks. Staggered tube banks. From ESDU 73031
(1973) [57]. ..........................................................................................................89
Figure 6-34: Effect of the Prandtl number, Prb, on the reference Nusselt number, Nur,
for both in-line and staggered tube banks. From ESDU 73031 (1973) [57]......... 89
Figure 6-35: The factor F1 to account for variable fluid properties. From ESDU 73031
(1973) [57]. ..........................................................................................................90
Figure 6-36: The factor F2 accounting for abnormal number of rows vs. that number, N.
From ESDU 73031 (1973) [57]. ........................................................................... 90
Figure 6-37: The factor F3 accounting for the effect of yaw vs. the inclination angle, .
From ESDU 73031 (1973) [57]. ........................................................................... 90
Figure 6-38: The factor F4 for estimating the Nusselt number of the n-th row. From
ESDU 73031 (1973) [57]. .................................................................................... 91
Figure 7-1: Friction characteristics associated with four types of roughness geometry.
Notice that the equivalent roughness is different in every case. From
Reynolds (1974). .................................................................................................96
Figure 7-2: Friction factor, c, as a function of Reynolds number, Re, for different
values of the relative roughness, e/D: Cylindrical tubes of circular cross
section. From Idel'cik (1969) [97]....................................................................... 105
Figure 7-3: Correction factor, K, to be used when the cross section of the duct is not
circular. Laminar flow. K = 1 for turbulent flow through hydraulically smooth
ducts. From ESDU 66027 (1966) [46]. ..............................................................105
Figure 7-4: Boundary between short and long circular arc bends. From ESDU 67040
(1967) [47]. ........................................................................................................106
Figure 7-5: Boundaries between laminar, transitional and turbulent flows in long circular
arc bends. From ESDU 67040 (1967) [47]. ....................................................... 106
Figure 7-6: Pressure loss coefficient per unit bend angle, cK/, as a function of the
dimensionless radius of curvature of bend centerline, R/D, for different
values of Reynolds number, Re. Either circular or square cross section.
From ESDU 67040 (1967) [47]. ......................................................................... 107
Figure 7-7: Pressure loss coefficient, cK, as a function of the dimensionless radius of
bend centerline, R/D, for different values of Reynolds number, Re. Laminar
flow through short circular arc bends. From ESDU 67040 (1967) [47]. ............. 108
Figure 7-8: Pressure loss coefficient, cK, as a function of the dimensionless radius of
bend centerline, R/D, for different values of bend angle, . Turbulent flow
through short circular arc bends. Either circular or square cross section.
From ESDU 67040 (1967) [47]. ......................................................................... 109
11
ECSSEHB3101Part13A
5December2011
Figure 7-9: Pressure loss coefficient, cK, for short circular arc bends, having a short
downstream tangent of length, Ld, as a function of Ld/D, for different values
of the dimensionless radius of bend centerline, R/D. Turbulent flow. Either
circular or square cross section. From ESDU 67040 (1967) [47]. ..................... 110
Figure 7-10: The factor 1 to account for the aspect-ratio of the bend cross section.
From ESDU 67040 (1967) [47]. ......................................................................... 110
Figure 7-11: The factor 2 to account for the bend angle. From ESDU 67040 (1967)
[47].....................................................................................................................111
Figure 7-12: Pressure loss coefficient, cK, for single mitre bends, as a function of bend
angle, , for different values of the dimensionless length, Ld/D, of the
downstream tube. Turbulent flow. Either circular or square cross section.
From ESDU 67040 (1967) [47]. ......................................................................... 111
Figure 7-13: Factor , which account for the interaction between two 90 -circular arc
bends-, as a function of the dimensionless distance between both bends,
La/D. From ESDU 68035 (1968) [49]. ................................................................ 112
Figure 7-14: Factor , which account for the interaction between two mitre bends, as a
function of the dimensionless distance between both bends, La/D. From
ESDU 68035 (1968) [49]. .................................................................................. 113
Figure 7-15: Total-pressure loss coefficient, cKt, as a function of Reynolds number,
ReD1, for different values of the area ratio, . Enlargement with a duct
downstream 4D2 long. Uniform incoming flow at low Reynolds number.
From ESDU 72011 (1972) [54]. ......................................................................... 113
Figure 7-16: Different velocity profiles upstream of a sudden enlargement. From ESDU
72011 (1972) [54]. .............................................................................................114
Figure 7-17: Total-pressure loss coefficient, cKt, as a function of area ratio, .
Enlargement with a duct downstream 4D2 long. Numbers on curves indicate
the velocity profile in Figure 7-22 for which the curve applies. From ESDU
72011 (1972) [54]. .............................................................................................114
Figure 7-18: Static-pressure loss coefficient, -cKs, as a function of area ratio, .
Enlargement with a duct downstream 4D2 long. Numbers on curves indicate
the velocity profile in Figure 7-22 for which the curve applies. From ESDU
(1972) [54]. ........................................................................................................115
Figure 7-19: Total-pressure loss coefficient, cKt, as a function of Reynolds number,
ReD2, for different values of the area ratio, . The pressure loss coefficient
is expressed in terms of the dynamic pressure at clause 6. From Idel'cik
(1969) [97]. ........................................................................................................115
Figure 7-20: Reference values of the pressure loss coefficient, cK, as a function of the
ratio, , of the area available for fluid flow to the total area of the duct cross
section. Perforated plates and orifices. From ESDU 72010 (1972) [53]. ........... 116
Figure 7-21: The factor 3 to account for the effect of plate thickness when t/d < 0,8. cKo
is given in Figure 7-19. From ESDU 72010 (1972) [53]..................................... 117
Figure 7-22: The factor 4 to account for the effect of plate thickness when t/d 0,8.
cK0,8 is given in Figure 7-19. From ESDU 72010 (1972) [53]. ............................ 118
Figure 7-23: Comparison between the pressure loss coefficients, cK, in the intermediate
region calculated by assuming either of the two extreme cases, fully-
separated or reattached orifice flow. From ESDU 72010 (1972) [53]. ............... 119
12
ECSSEHB3101Part13A
5December2011
Figure 7-24: Reference pressure loss coefficient, cKr, as a function of porosity, .
Round-wire gauzes. From ESDU 72009 (1972) [52]. ........................................ 119
Figure 7-25: Factor 5 to account for low Reynolds number effects in round-wire
gauzes. Reynolds number based on the wire diameter. From ESDU 72009
(1972) [52]. ........................................................................................................120
Figure 7-26: Reference pressure loss coefficient, cKr, as a function of Reynolds
number, Re, for diaphragm and butterfly valves fully open. Prepared by the
compiler after ESDU 69022 (1969) [51]............................................................. 120
Figure 7-27: Factor 6, which accounts for the partial opening of the valve, as a
function of the degree of valve opening, . is defined as the ratio of valve
control travel from closed position to total valve control travel. From ESDU
69022 (1969) [51]. .............................................................................................121
Figure 7-28: Graphics for estimating the pressure loss coefficient, cK, for in-line tube
banks of several relative pitches, sl, st, and yaw angles, . The influence of
the heat exchange on the pressure loss is taken into account through the
tube bank inlet and exit temperatures, Ti and To, respectively. From Idel'cik
[97].....................................................................................................................122
Figure 7-29: Graphics for estimating the pressure loss coefficient, cK, for staggered
tube banks of several relative pitches, sl, st, and yaw angles, . The
influence of the heat exchange on the pressure loss is taken into account
through the tube bank inlet and exit temperatures, Ti and To, respectively.
From Idel'cik (1969) [97]. ................................................................................... 123
Figure 7-30: Pressure loss coefficient, cK, as a function of the ratio of lateral to total
mass flow rates in branching tubes. The mixed confluence-branching case
is not considered. From Idel'cik (1969) [97]. ...................................................... 124
Figure 8-1: The ratio 2St/f, for turbulent flow in constant wall temperature cylindrical
tubes, as calculated by use of several expressions, vs. the Reynolds
number, Re. E: Correlation of experimental results. From Goldstein (1950)
[73]. R: Reynolds Analogy. P: Prandtl Analogy. K: von Krmn Analogy.
Calculated by the compiler. ............................................................................... 127
Figure 9-1: Constant power heat transfer ratio, (ha-ho)P, vs. Reynolds number based on
non-augmentative conditions, Reo. From Bergles (1969) [8]. ............................ 136
Figure 9-2: Roughness function ue+(e+)for Nikuradse's sand roughness. (1)
Hydraulically smooth. (2) ue+ = 8,48, completely rough. From Schlichting
(1960) [157]. ......................................................................................................139
non-augmentative conditions, Reo. Curves A to D are from Bergles (1969)
[8], curves E and F have been calculated by the compiler after Webb,
Eckert & Goldstein (1971) [186]......................................................................... 139
Figure 9-4: Roughness function ue+(e+,b/e) for repeated-rib roughness. From Webb et
al. (1971) [186]...................................................................................................142
Figure 9-5: Flow pattern near the wall for different values of b/e. ........................................ 142
non-augmentative conditions, Reo. Curves A, B, C are from Bergles (1969)
[8], curves D to G have been calculated by the compiler after Sheriff &
Gumley (1966) [166]. .........................................................................................143
Figure 9-7: Roughness function, ue+(e+,b/e), for wire coil roughness. Plotted by the
compiler after Sheriff & Gumley (1966) [166]. ................................................... 146
13
ECSSEHB3101Part13A
5December2011
Figure 9-8: Velocity and Temperature distributions across the annulus. ............................. 147
Figure 9-9: Constant power heat transfer ratio, (ha/ho)P, vs. Reynolds number based on
non-augmentative conditions, Reo. From Carnavos (1974) [19]........................ 149
Figure 9-10: Constant power heat transfer ratio, (ha/ho)P, vs. Reynolds number based
on non-augmentative conditions, Reo. From Carnavos (1974) [19]................... 151
on non-augmentative conditions, Reo. From Carnavos (1974) [19]................... 153
on non-augmentative conditions, Reo. From Bergles (1969) [8]. ....................... 156
on non-augmentative conditions, Reo. Calculated by the compiler after
Hong & Bergles (1976) [91]. ..............................................................................157
on non-augmentative conditions, Reo. Curves A to I are from Bergles
(1969) [8], curves J to M have been calculated by the compiler after
Thorsen & Landis (1968) [178]. .........................................................................160
Figure 9-15: Isothermal Nusselt number, Nuab. divided by the ratio of friction factors, ,
vs. the Reynolds number, ReT, for different values of the Prandtl number,
Pr. Calculated by the compiler after Thorsen & Landis (1968) [178]. ................ 163
Figure 9-16: Constant power heat transfer ratio, (he/ho)P, vs. Reynolds number based
on non-augmentative conditions, Reo. From Bergles (1969) [8]. ....................... 164
on non-augmentative conditions, Reo. Curves A, B, C from Bergles (1969)
[8], curves D, E, F from Bergles, Lee & Mikic (1969) [9].................................... 167
Figure 10-1: Product of cooling effectiveness, F, of several fluids times the equivalent
length of the loop, LE, as functions of the difference between the heat
source and the inlet temperature, TS - Ti, for the following reference values:
Inner diameter of the duct, D = 10-2 m. Diabatic length of the duct, L = 1 m.
Heat flux, q = 250 W.m-2 for Air, Carbon Dioxide, Carbon Tetrachloride,
Hydrogen and Nitrogen, q = 1000 W.m-2 for Ethylene Glycol, Flutec PP50
and Water. Calculated by the compiler.............................................................. 171
Figure 10-2: Schematic representation of the fluid loop considered for estimating the
fluid cooling effectiveness.................................................................................. 172
Figure 10-3: Graphical method allowing for values of heat flux, q, and inner diameter of
the duct, D, different from those used in Figure 10-1. ....................................... 174
Figure 10-4: Graphic for estimating the product of the fluid cooling effectiveness, F,
times the equivalent length of the loop, LE, as a function of the difference
between the heat source and the inlet fluid temperature, TS - Ti. Fluid: Air.
Reference values: D = 10-2 m, L = 1 m, q = 250 W.m-2. Values FLE for
different D and q, yet L = 1, can be calculated graphically as is indicated in
the text. Prepared by the compiler..................................................................... 175
between the heat source and the inlet fluid temperature, TS - Ti. Fluid:
Ethylene Glycol. Reference values: D = 10-2 m, L = 1 m, q = 1000 W.m-2.
Values FLE for different D and q, yet L = 1, can be calculated graphically as
is indicated in the text. Prepared by the compiler. ............................................. 176
14
ECSSEHB3101Part13A
5December2011
Flutec PP50. Reference values: D = 10-2 m, L = 1 m, q = 1000 W.m-2.
Values FLE for different D and q, yet L = 1, can be calculated graphically as
is indicated in the text. Prepared by the compiler. ............................................. 177
Water. Reference values: D = 10-2 m, L = 1 m, q = 1000 W.m-2. Values FLE
for different D and q, yet L = 1, can be calculated graphically as is indicated
in the text. Prepared by the compiler. ................................................................178
Figure 10-8: Vapor pressure, psat, of Water vs. temperature, T. From Vargaftik (1975)
[183]...................................................................................................................186
Figure 10-9: Density, , of Water vs. temperature, T. From Vargaftik (1975) [183]. ............ 186
Figure 10-10: Specific heat, cp, of Water vs. temperature, T. From Vargaftik (1975)
[183]...................................................................................................................186
Figure 10-11: Thermal conductivity, k, of Water vs. temperature, T. From Vargaftik
(1975) [183]. ......................................................................................................187
Figure 10-12: Dynamic viscosity, , of Water vs. temperature, T. From Vargaftik (1975)
[183]...................................................................................................................187
Figure 10-13: Vapor pressure, psat, of Carbon Tetrachloride vs. temperature, T. From
Vargaftik (1975) [183]. .......................................................................................187
Figure 10-14: Density, , of Carbon Tetrachloride vs. temperature, T. From Vargaftik
(1975) [183]. ......................................................................................................188
Figure 10-15: Specific heat, cp, of Carbon Tetrachloride vs. temperature, T. From
Vargaftik (1975) [183]. .......................................................................................188
Figure 10-16: Thermal conductivity, k, of Carbon Tetrachloride vs. temperature, T.
From Vargaftik (1975) [183]...............................................................................188
Figure 10-17: Dynamic viscosity, , of Carbon Tetrachloride vs. temperature, T. From
Vargaftik (1975) [183]. .......................................................................................189
Figure 10-18: Vapor pressure, psat, of Coolanol 15, 25, 35 and 45 vs. temperature, T.
From Filippi & Guerra (1977) [64]. .....................................................................189
Figure 10-19: Density, , of Coolanol 15, 25, 35 and 45 vs. temperature, T. From Filippi
& Guerra (1977) [64].......................................................................................... 189
Figure 10-20: Specific heat, cp, of Coolanol 15, 25, 35 and 45 vs. temperature, T. From
Filippi & Guerra (1977) [64]. .............................................................................. 190
Figure 10-21: Thermal conductivity, k, of Coolanol 15, 25, 35 and 45 vs. temperature,
T. From Filippi & Guerra (1977) [64].................................................................. 190
Figure 10-22: Dynamic viscosity, , of Coolanol 15, 25, 35 and 45 vs. temperature, T.
Figure 10-23: Kinematic viscosity, , of DC 200 vs. temperature T. Numbers on curves
indicate the standard viscosity in cs. From DOW CORNING (1972) [38]. ......... 191
Figure 10-24: Freezing point, T, of Water/Glycol Solutions vs. Glycol mass fraction, s.
15
ECSSEHB3101Part13A
5December2011
Figure 10-25: Vapor pressure, psat, of Water/Glycol Solutions vs. temperature, T.
Numbers on curves indicate Glycol mass fraction, c. From Filippi & Guerra
(1977) [64]. ........................................................................................................192
Figure 10-26: Density, , of Water/Glycol Solutions vs. temperature, T. Numbers on
curves indicate Glycol mass fraction, c. From Filippi & Guerra (1977) [64]. ...... 192
Figure 10-27: Specific heat, cp, of Water/Glycol Solutions vs. temperature, T. Numbers
on curves indicate Glycol mass fraction, c. From Filippi & Guerra (1977)
[64].....................................................................................................................193
Figure 10-28: Thermal conductivity, k, of Water/Glycol Solutions vs. temperature, T.
(1977) [64]. ........................................................................................................193
Figure 10-29: Dynamic viscosity, , of Water/Glycol Solutions vs. temperature, T.
(1977) [64]. ........................................................................................................194
Figure 10-30: Vapor pressure, psat, of Flutec PP-2, PP-9 and PP-50 vs. temperature, T.
Data are from Dunn & Reay (1976) [40] except those corresponding to
Flutec PP-50 which are from Wyn-Roberts (1974) [193]. .................................. 194
Figure 10-31: Density, , of Flutec PP-2, PP-9 and PP-50 vs. temperature, T. Data are
from Dunn & Reay (1976) [40] except those corresponding to Flutec PP-50
which are from Wyn-Roberts (1974) [193]......................................................... 195
Figure 10-32: Specific heat, cp, of Flutec PP-50 vs. temperature, T. From Wyn-Roberts
(1974) [193]. ......................................................................................................195
Figure 10-33: Thermal conductivity, k, of Flutec PP-2, PP-9 and PP-50 vs. temperature,
T. Data are from Dunn & Reay (1976) [40] except those corresponding to
Figure 10-34: Dynamic viscosity, , of Flutec PP-2, PP-9 and PP-50 vs. temperature, T.
Data are from Dunn & Reay (1976) [40] except those corresponding to
Figure 10-35: Vapor pressure, psat, of Freon 11, 12, 13, 21, 22, 113, 114 and 142 vs.
temperature, T. From Vargaftik (1975) [183]. .................................................... 197
Figure 10-36: Density, , of Freon 11, 12, 13, 21, 22, 113, 114 and 142 vs.
temperature, T. From Vargaftik (1975) [183]. .................................................... 197
Figure 10-37: Specific heat, cp, of Freon 11, 12, 13, 21, 22, 113, 114 and 142 vs.
temperature, T. From Filippi & Guerra (1977) [64]. ........................................... 197
Figure 10-38: Thermal conductivity, k, of Freon 11, 12, 13, 21, 22, 113, 114 and 142
vs. temperature, T. From Vargaftik (1975) [183]................................................ 198
Figure 10-39: Dynamic viscosity, , of Freon 11, 12, 13, 21, 22, 113, 114 and 142 vs.
temperature, T. Data are from Vargaftik (1975) [183] except those
corresponding to Freon 13 which are from Filippi & Guerra (1977) [64]............ 198
Figure 10-40: Vapor pressure, psat, of Freon E1, E2, E3, E4 and E5 vs. temperature, T.
Figure 10-41: Density, , of Freon E1, E2, E3, E4 and E5 vs. temperature, T. From
Filippi & Guerra (1977) [64]. .............................................................................. 199
Figure 10-42: Specific heat, cp, of Freon E1, E2, E3, E4 and E5 vs. temperature, T.
16
ECSSEHB3101Part13A
5December2011
Figure 10-43: Thermal conductivity, k, of Freon E1, E2, E3, E4 and E5 vs. temperature,
T. From Filippi & Guerra (1977) [64].................................................................. 200
Figure 10-44: Dynamic viscosity, , of Freon E1, E2, E3, E4 and E5 vs. temperature, T.
Figure 10-45: Vapor pressure, psat, of FC 75 vs. temperature, T. From Filippi & Guerra
(1977) [64]. ........................................................................................................200
Figure 10-46: Density, , of FC 75 vs. temperature, T. From Filippi & Guerra (1977)
[64].....................................................................................................................201
Figure 10-47: Specific heat, cp, of FC 75 vs. temperature, T. From Filippi & Guerra
(1977) [64]. ........................................................................................................201
Figure 10-48: Thermal conductivity, k, of FC 75 vs. temperature, T. From Filippi &
Guerra (1977) [64]. ............................................................................................201
Figure 10-49: Dynamic viscosity, , of FC 75 vs. temperature, T. From Filippi & Guerra
(1977) [64]. ........................................................................................................202
Figure 10-50: Thermal conductivity, k, of Methanol/Water Solutions vs. temperature, T.
Numbers on curves indicate Methanol mass fraction, c. From Vargaftik
(1975) [183]. ......................................................................................................202
Figure 10-51: Dynamic viscosity, , of Methanol/Water Solutions vs. temperature, T.
Numbers on curves indicate Methanol mass fraction, c. From Vargaftik
(1975) [183]. ......................................................................................................203
Figure 10-52: Vapor pressure, psat, of Monsanto OS 59 vs. temperature, T. From Filippi
& Guerra (1977) [64].......................................................................................... 203
Figure 10-53: Density, , of Monsanto OS 59 vs. temperature, T. From Filippi & Guerra
(1977) [64]. ........................................................................................................204
Figure 10-54: Specific heat, cp, of Monsanto OS 59 vs. temperature, T. From Filippi &
Guerra (1977) [64]. ............................................................................................204
Figure 10-55: Thermal conductivity, k, of Monsanto OS 59 vs. temperature, T. From
Filippi & Guerra (1977) [64]. .............................................................................. 204
Figure 10-56: Dynamic viscosity, , of Monsanto OS 59 vs. temperature, T. From Filippi
& Guerra (1977) [64].......................................................................................... 205
Figure 10-57: Density, , of Air at a pressure of 105 Pa vs. temperature, T. From
Vargaftik (1975) [183]. .......................................................................................213
Figure 10-58: Specific heat, cp, of Air vs. temperature, T. From Vargaftik (1975) [183]....... 213
Figure 10-59: Thermal conductivity, k, of Air vs. temperature, T. From Vargaftik (1975)
[183]...................................................................................................................213
Figure 10-60: Dynamic viscosity, , of Air vs. temperature, T. From Vargaftik (1975)
[183]...................................................................................................................214
Figure 11-1: Typical regenerators. a) Rotary type. b) Valved type. From Kays & London
(1964) [102]. ......................................................................................................215
Figure 11-2: Typical recuperators. a) Counterflow heat exchanger. b) Crossflow heat
exchanger. From Welty, Wicks & Wilson (1969) [188]....................................... 216
Figure 11-3: Some typical examples of compact heat exchanger surfaces. From Kays
& London (1964) [102]. a) Circular tube bundle. b) Finned-circular-tube
surface. c) Finned-tube surface, flat tubes, continuous fins. d) Plate-fin
arrangement. e) Strip-fin surface. f) Regenerator compact matrix..................... 217
17
ECSSEHB3101Part13A
5December2011
Figure 11-4: a) Shell-and-tube exchanger with two shell passes and four tube passes.
b) Schematic representation of the exchanger which will be used in clause
11.2.3.................................................................................................................218
Figure 11-5: Liquid-coupled indirect-transfer type of heat exchanger. From Kays &
London (1964) [102]. ......................................................................................... 218
Figure 11-6: Thermal conductivity, k, of several metals vs. temperature, T. From Kays
& London (1964) [102]. ......................................................................................220
Figure 11-7: Heat transfer effectiveness, f, of trapezoidal fins, vs. dimensionless fin
length, L[2h/k(t+b)]1/2. Calculated by the compiler after Jakob (1958) [99]...... 221
Figure 11-8: Heat transfer effectiveness, f, of circular fins, vs. dimensionless fin
length, (rori)(h/k)1/2. Calculated by the compiler after Jakob (1958) [99]......... 221
Figure 11-9: Heat transfer effectiveness, , vs. number of heat transfer units, Ntu, in a
counterflow exchanger. From Kays & London (1964) [102]. ............................. 223
parallel flow exchanger. From Kays & London (1964) [102]. ............................. 223
crossflow exchanger with fluids unmixed. Calculated by the compiler after
Mason (1954) [124]............................................................................................224
crossflow exchanger with one fluid mixed. From Kays & London (1964)
[102]...................................................................................................................224
Figure 11-13: Heat transfer effectiveness, , vs. number of heat transfer units, Ntu, for
the case of n-pass counter-crossflow exchangers, when fluid A is unmixed
throughout and fluid B mixed throughout, and with passes connected in
reverse order. Calculated by the compiler after Stevens, Fernandez &
Woolf (1957) [170]. ............................................................................................225
the case of n-pass counter-crossflow exchangers, when fluid A is unmixed
throughout and fluid B mixed throughout, and with passes connected in
identical order. Calculated by the compiler after Stevens, Fernandez &
Woolf (1957) [170]. ............................................................................................226
multipass exchanger with 1 shell pass and 2 or more tube passes. From
Kays & London (1964) [102]. .............................................................................227
multipass exchanger with 2 shell passes and 4, 8, 12, .... tube passes.
From Kays & London (1964) [102]..................................................................... 227
multipass exchanger with 3 shell passes and 6, 12, 18, .... tube passes.
From Kays & London (1964) [102]..................................................................... 228
multipass exchanger with 4 shell passes and 8, 16, 24, ... tube passes.
From Kays & London (1964) [102]..................................................................... 228
multipass exchanger with 5 shell passes and 10, 15, 20, ... tube passes.
From Kays & London (1964) [102]..................................................................... 229
18
ECSSEHB3101Part13A
5December2011
different number of shell passes, in a multipass exchanger with R = 1. The
case R = 0 is also shown for comparison. Calculated by the compiler after
Kays & London (1964) [102]. .............................................................................229
Figure 11-21: Optimum liquid flow capacity rate, C1/CLopt, to maximize the heat transfer
effectiveness vs. ratio of the number of heat transfer units Ntu1/Ntu2, of
exchanger 1 to exchanger 2. Calculated by the compiler after Holmberg
(1975) [90]. ........................................................................................................230
Figure 11-22: Heat transfer effectiveness, , vs. liquid flow capacity rate, CL/CLopt, for
different values of the overall number of heat transfer units, Notu. Ntu1/Ntu2 =
1. Calculated by the compiler after Holmberg (1975) [90]. ................................ 231
Figure 11-23: Heat transfer effectiveness, , vs. liquid flow rate, CL/CLopt, for different
values of the ratio between the number of heat transfer units of exchanger
1 to exchanger 2, Ntu1/Ntu2. Notu = 2. Calculated by the compiler after
Holmberg (1975) [90].........................................................................................231
Figure 11-24: Overall heat transfer effectiveness, , vs. area ratio between exchanger 1
and 2, (A1/A2)/(A1/A2)opt, for the case of optimum liquid flow capacity rate
and R = 1. Calculated by the compiler after Holmberg (1975) [90].................... 232
Figure 11-25: Overall heat transfer effectiveness, t, of an assembly of n identical
exchangers in parallel, vs. effectiveness, , of a single exchanger.
Calculated by the compiler after Domingos (1969) [36]..................................... 233
exchangers in counterflow, vs. effectiveness, , of a single exchanger.
Calculated by the compiler after Domingos (1969) [36]..................................... 234
exchangers in parallel in the stream of lower capacity rate, vs.
effectiveness, , of a single exchanger. Calculated by the compiler after
Domingos (1969) [36]. .......................................................................................234
exchangers in parallel in the stream of higher capacity rate, vs.
effectiveness, , of a single exchanger. Calculated by the compiler after
Domingos (1969) [36]. .......................................................................................235
Figure 11-29: Heat exchanger ineffectiveness, 1, vs. number of heat transfer units,
Ntu, for a counterflow heat exchanger. Results are shown for R = 1 and
different values of the wall conduction parameter, . From Kroeger (1967)
[110]...................................................................................................................252
Ntu, for a counterflow heat exchanger. Results are shown for R = 0,95 and
[110]...................................................................................................................252
Ntu, for a counterflow heat exchanger. Results are shown for R = 0,90 and
[110]...................................................................................................................253
Figure 11-32: Heat exchanger ineffectiveness, 1-, vs. dimensionless wall conduction,
, for a counterflow heat exchanger. Results are shown for Ntu = 50 and
different values of the capacity-rate ratio, R. Calculated by the compiler
after Kroeger (1967) [110]. ................................................................................253
19
ECSSEHB3101Part13A
5December2011
Figure 11-33: Relative capacity, Qn/Q, of a simple two fluid heat exchanger vs. the
maldistribution parameter, , for several values of the nominal number of
heat transfer units, Ntu. From Weimer & Hartzon (1973) [187]. ......................... 256
Figure 11-34: Relative surface requirements, An/A, of a simple two fluid exchanger for
fixed total duty vs. the maldistribution parameter, . Results are shown for
several values of the nominal number of heat transfer units, Ntu. From
Weimer & Hartzog (1973) [187]. ........................................................................ 256
Figure 11-35: Flow distribution model for the non-uniform side. From Fleming (1967)
[65].....................................................................................................................257
Figure 11-36: Overall effectiveness, , and effective number of heat transfer units, Ntueff,
vs. the fraction, FLow, of channels which carry lower-than-average flow on
the nonuniform side of a "paired channels" heat exchanger. Results are
shown for different values of the ratio of the capacity rate, CLow, of a single
channel with lower-than-average flow to the capacity rate, CHigh, of a single
channel with higher-than-average flow. (a) is for a nominal number of heat
transfer units Ntu = 10 and (b) for Ntu = 100. From Fleming (1967) [65]............. 258
Figure 11-37: Overall effectiveness, , vs. the fraction, FLow, of channels which carry
lower-than-average flow on the nonuniform side of a heat exchanger with
uniform side mixed. Results are shown for different values of the nominal
number of heat transfer units Ntu, and of the ratio CLow/CHigh. From Fleming
(1967) [65]. ........................................................................................................258
Figure 11-38: Ideal flow distribution in the shell side of shell-and-tube heat exchangers. ... 259
Figure 11-39: Schematic of the temperature distribution along the heat exchanger. (a)
Balanced case. (b) Imbalanced case. From Cowans (1974) [28]. ..................... 263
Figure 11-40: Flow imbalance compensating technique for gas to gas heat exchangers.
From Cowans (1974) [28]. ................................................................................. 264
Figure 11-41: Plate spacing, (a), and fin spacing (b), type flow passage non-
uniformities. From London (1970) [118]............................................................. 265
Figure 11-42: Degradation in the heat exchanger thermal performance, measured by
CostNtu, and relative gain in pressure loss, 1-(p1/pn), as functions of the
deviation in channel size, 1-(DE1/DEn). Results in (a) apply to any cylindrical
passage provided that the non-uniformities are geometrically similar.
Results in (b) are for non-uniformities of the fin-spacing type. From London
(1970) [118]. ......................................................................................................267
Figure 11-43: Fin center of offset rectangular plat-fin surface. From Shah & London
(1970) [164]. ......................................................................................................269
Figure 11-44: Assumed core geometries. From Shah & London (1970) [164]..................... 270
Figure 11-45: Heat transfer, j, and friction, f, characteristics as functions of Reynolds
number, Re, for surfaces 501 and 501 MOD. From Shah & London (1970)
[164]...................................................................................................................271
Figure 11-46: Flow area goodness factor, j/f, as a function of Reynolds number, Re, for
surfaces 501 and 501 MOD. From Shah & London (1970) [164]. ..................... 272
Figure 12-1: Typical characteristic curves of a centrifugal pump for a given rotating
speed. ................................................................................................................288
Figure 12-2: Rotodynamic pump impellers. From Nekrasov (1969) [132]............................ 290
20
ECSSEHB3101Part13A
5December2011
Figure 12-3: Characteristic curves of SEALED MOTOR CONSTRUCTION Centrifugal
Pumps Cadet "Mini" and Cadet "S" pumping water. From Wyn-Roberts
(1973) [194]. ......................................................................................................295
Figure 12-4: Characteristic curves of EURAMO Centrifugal Pumps Cadet MX 32-E and
XA 15-R pumping water. From EURAMO-POMPES SALMSON (1977) [60]. ... 295
Figure 12-5: Characteristic curve of the Centrifugal Pump devised by Engel & Walter
Cadet "Mini" and Cadet "S" pumping water. From Wyn-Roberts (1973)
[194]...................................................................................................................296
Figure 12-6: Characteristic curves of Eastern-Iwaki Centrifugal Pumps MD-15T and
MDR-30T pumping water. From GELBER (1976) [71]. ..................................... 296
Figure 12-7: Characteristic curve of Centrifugal Pump AC-3C-MD pumping water. From
GELBER (1976) [71]..........................................................................................297
Figure 12-8: Characteristic curve of Positive Displacement Rotary Pump 413-7-1285
pumping water. From GELBER (1976) [71]....................................................... 297
Figure 13-1: Schematic representation of the fluid loop considered by Barker, Stephens
& Taylor (1967) [3]. ............................................................................................300
Figure 13-2: Results of the liquid coolant system optimization analysis. a) gives the
mass penalty, M+, of the system components vs. Freon 21 mass flow rate,
mc. The mass penalty includes the equivalent mass due to pumping power.
b) gives the system mass, M, vs. Freon 21 mass flow rate, mc. From Baker,
Stephens & Taylor (1967) [3].............................................................................302
Figure 13-3: Schematic representation of the fluid loop with EDHX. From Berner &
Schleicher (1976) [13]........................................................................................303
Figure 13-4: Mass penalty, M+, of the supply and return plumbing system vs. source
temperature, TS. M+ includes mass of tubes and fluid plus the equivalent
mass due to pumping. Heat load, Q = 103 W. From Berner & Schleicher
(1976) [13]. ........................................................................................................305
Figure 13-5: Source temperature, TS, vs. coolant mass flow rate, mh, for different
values of the heat transfer rate, Q. The interrupted lines are drawn through
the points for which the power requirements, for a given Di, equals 1 W.
From Berner & Schleicher (1976) [13]. .............................................................. 306
Figure 13-6: Schematic representation of basic ECLA system. The aim of the
accumulator, which is not mentioned in the text, is to accommodate
changes in coolant density with temperature. The coolant throttle is used
for adjusting the coolant mass flow rate, mh. From Berner & Schleicher
(1976) [13]. ........................................................................................................307
Figure 14-1: Flow-pattern map for vertical upward air-water flow. Prepared by the
compiler after Hewitt (1982) [86]. mG and mL are the gas and liquid mass
flow rates, respectively. AFL is the internal cross-sectional area of the duct. ..... 311
Figure 14-2: Baker flow-pattern map for horizontal air-water flow. From Hewitt (1982)
[86]. mG and mL are the gas and liquid mass flow rates, respectively. AFL is
the internal cross-sectional area of the duct. ..................................................... 312
Figure 14-3: Lockhart - Martinelli correlation for pressure loss multipliers. The figure
has been drawn by use of Chisholm analytical representation. See text........... 315
Figure 14-4: Gas phase pressure loss multiplier, G, vs. Lockhart - Martinelli
parameter, X, as deduced from Eqs. [14-9] and [14-10] and from
experimental data. a) tt case; b) vt case; c) vv case. Experimental points
are from Wallis (1969) [184]. ............................................................................. 317
21
ECSSEHB3101Part13A
5December2011
Figure 14-5: Liquid fraction, 1-, vs. Lockhart - Martinelli parameters, X. From Wallis
(1969) [184]. ......................................................................................................318
Figure 14-6: Comparison of Lockhart - Martinelli with available experimental data.
Points with C as per Eq. [14-13], with G/L = 0 are also shown. From Yang
& Palen (1977) [195]..........................................................................................319
Figure 14-7: The annular flow configuration......................................................................... 320
Figure 14-8: Liquid film friction factor, fL, as a function of liquid film Reynolds number,
ReL, for annular two-phase flow in cylindrical ducts of circular cross-section.
From Hewitt (1982) [86]. The full lines correspond, respectively, to Hagen-
Poiseuille formula and to Blasius formula. fL is equal to the friction factor fsL
which corresponds to the single phase flow along the duct at the same
Reynolds number except for an intermediate ReL. ............................................ 323
Figure 14-9: Pressure gradient multipliers, L and G, dimensionless film thickness,
/D, and liquid fraction, 1-, according to Eqs. [14-34], [14-35] and [14-36],
with L/G = 1000. Calculated by the compiler. .................................................. 325
Figure 14-10: Comparison of the liquid fraction, 1-, vs. Lockhart - Martinelli parameter,
X, as deduced from Eqs. [14-34] and [14-35], with L/G = 1000, with
experimental data from Wallis (1969) [184]. ...................................................... 326
Figure 14-11: Comparison of the gas pressure gradient multiplier, G = xL, vs.
Lockhart - Martinelli parameter, X, as deduced from Eqs. [14-34] and [14-
36], with L/G = 1000, with experimental data from Wallis (1969) [184]. tt
case. ..................................................................................................................326
Figure 14-12: The function F(/D) which appears in Eq. [14-62]. The effect of the liquid-
gas density ratio is negligible for the range of values given in the figure.
Compare the values given in this curve with those given by 2/D vs. X in
Figure 14-9. .......................................................................................................333
Figure 14-13: Martinelli parameter, X, vs. entrainment parameter, RE, as deduced from
air-water flow experiments. From Wicks & Duckler (1960) [190]. Note that
RE is not dimensionless. ....................................................................................334
Figure 14-14: Concentration of entrained droplets in the gas core, GmE/mG, vs.
dimensionless number i*/. Replotted by the compiler after Hutchinson &
Whalley (1973) [93]. Different symbols are used to indicate experiments by
different authors.................................................................................................335
Figure 14-15: Flow geometry when gravity is dominant (a) or negligible (b). The void
fraction (vapor fractional area in the figure) is the same in both cases. ............ 343
Figure 14-16: Geometry used in the model of stratified flow condensation. ........................ 344
Figure 14-17: Liquid fraction, 1-, vs. vapor quality, w. For stratified condensing flow of
several liquid along horizontal ducts. Calculated by the compiler...................... 350
Figure 14-18: Vapor quality, w, vs. dimensionless distance along the duct, x/D, for
stratified condensing flow for several liquids along horizontal ducts.
Calculated by the compiler. ............................................................................... 351
Figure 14-19: The Taitel and Dukler limit for stratified flow in the w vs. (1) plane.
Numbers on the curves represent constant values of the parameter
w2(1w)/2(1) which appears in the left hand side of Eq. [14-114]................ 352
Figure 14-20: Liquid-alone Nusselt number, NuSL, vs.liquid Reynolds number, ReL, as
given by different correlations in typical cases. Calculated by the compiler. ..... 356
22
ECSSEHB3101Part13A
5December2011
Figure 14-21: Liquid fraction, 1-, as a function of vapor quality, w, for annular flow of
several liquid along ducts. Calculated by the compiler. ..................................... 357
Figure 14-22: Vapor quality, w, as a function of dimensionless distance along the duct,
x/D, for annular flow of several liquids along ducts. Calculated by the
compiler. ............................................................................................................358
Figure 15-1: Schematic of ATCS fluid loop in the Module-Pallet mode. From Owen,
Sessions & Walker (1976) [139]. ....................................................................... 360
Figure 15-2: Schematic of three different types of two-phase flow loops. a) Parallel
circuit. b) Series circuit. c) Series/parallel circuit. .............................................. 361
Figure 15-3: Physical layout of 25 kW technology demonstrator (thermal bus). From
Carlisle / Nolan (1987) [18]. ............................................................................... 362
Figure 15-4: Space platform configuration. From Sadunas, Lehtinen & Parish (1985)
[153]...................................................................................................................363
Figure 15-5: Study baseline centralized liquid loop external bus. From Sadunas,
Lehtinen & Parish (1985) [153]. ......................................................................... 366
Figure 15-6: Concept III. Decentralized liquid loop. From Sadunas, Lehtinen & Parish
(1985) [153]. ......................................................................................................368
Figure 15-7: Concept IV. Decentralized two-phase transport. From Sadunas, Lehtinen
& Parish (1985) [153].........................................................................................368
Figure 15-8: Trade study mass summary. From Sadunas, Lehtinen & Parish (1985)
[153]...................................................................................................................372
Figure 15-9: Decentralized systems with TS. a) Pumped liquid loop. b) Two-phase
transport. From Lehtinen & Sadunas (1985) [114]. ........................................... 373
Figure 15-10: Radiator absorbed environmental heat flux, qe, vs. dimensionless time,
t/tp. Orbital period, tp = 90 min. Radiator limit: Black body emissive power.
From Lehtinen & Sadunas (1985) [114]............................................................. 373
Figure 15-11: Thermal performance of PC capacitors in different subsystems. a)
Metabolic. b) Equipment. c) Fuel cell. Thermal performance is given in
terms of: 1) energy storage rate, Qc; and 2) net stored energy, Jc. From
Lehtinen & Sadunas (1985) [114]. ..................................................................... 376
Figure 15-12: Required specific storage capacity. Jc/Qc, of typical TMSs vs.
temperature, T. From Lehtinen & Sadunas (1985) [114]. .................................. 378
Figure 15-13: Schematic of the fuel cell loop TMS. a) Pumped liquid loop. b) Two-
phase transport. From Sadunas, Lehtinen & Parish (1986) [153]. .................... 379
Figure 15-14: Respond-to-demand temperature control scheme From Sadunas,
Lehtinen, Nguyen & Parish (1986) [154]............................................................ 381
Figure 15-15: Orbital-average temperature control scheme. From Sadunas, Lehtinen,
Nguyen & Parish (1986) [154]. .......................................................................... 381
Figure 15-16: Instrumentation of the different control schemes. a) Respond-to-
demand pumped-liquid-loop scheme. b) Orbital-average pumped-liquid-
loop scheme. c) Two-phase transport loop scheme. From Sadunas,
Lehtinen, Nguyen & Parish (1986) [154]............................................................ 383
Figure 15-17: Evolution of the heat rejection rate and configuration of spacecraft
thermal control systems. From Dexter & Haskin (1984) [34]. ............................ 385
Figure 15-18: Schematic of the heat pump working principle. Adapted by the compiler
after Rye & Steen (1986) [152]. ......................................................................... 385
23
ECSSEHB3101Part13A
5December2011
Figure 15-19: Radiator mass, MR, of four typical systems (see text above) for different
values of the heat rejection rate, Q. Calculated by the compiler. ...................... 387
Figure 15-20: Mass breakdown of a heat pump augmented system for different values
of the heat rejection rate, Q. TR = 300 K. Calculated by the compiler................ 388
Figure 15-21: Mass breakdown of a heat pump augmented system for different values
of the heat rejection rate, Q. a) TR = 340 K. b) TR = 380 K. Calculated by the
compiler. ............................................................................................................389
Figure 15-22: Ratio of total mass, M, of the heat pump augmented system to the total
mass, Mr, of the reference classical pumped-liquid loop vs. the inverse
Carnot coefficient of performance [TRTc]/Tc for different values of the heat
rejection rate, Q. Calculated by the compiler..................................................... 390
Figure 15-23: Effect of s/ on radiator specific area, AR/Q. From Lehtinen & Sadunas
(1985). ...............................................................................................................391
Figure 15-24: Radiator specific area, AR/Q, vs. refurbishment frequency, t. From
Lehtinen & Sadunas (1985) [114]. ..................................................................... 392
Figure 15-25: Steerable to fixed radiator specific heat-rejection ration, (Q/AR)steer/(Q/AR),
vs. radiator temperature, TR. From Sadunas, Lehtinen & Parish (1985)
[154]...................................................................................................................394
Figure 15-26: Rotatable radial flow heat pipe. From Delil (1986a) [33]................................ 398
Figure 15-27: Thermal joint based on a radial flow heat pipe. From Hinderer & Savage
(1978) [88]. ........................................................................................................398
Figure 15-28: Heat transfer enhancement by flow oscillation demonstrator. From
Kurzweg & Zhao (1984) [113]............................................................................ 399
Figure 15-29: Hemispherical heat pipe junction. From Delil (1986a) [33]. ........................... 399
Figure 15-30: Cylindrical heat pipe junction. From Delil (1986a) [33]. ................................. 399
Figure 15-31: Heat pipe in heat pipe joint. This concept is an off-spring of the cylindrical
heat pipe joint. From Shaubach (1985) [165]. ................................................... 400
Figure 15-32: Felxible heat pipe consisting of a 0,15 m long rigid evaporator and a 0,20
m long rigid condenser. From Delil (1986)a [33]................................................ 400
Figure 15-33: Flexible cooper/acetone heat pipe. It features a 0,381 m long adiabatic
section between an evaporator and a condenser both 0,178 m long. Inside
diameter is 0,016 m. From Delil (1986a) [33]. ................................................... 400
Figure 15-34: Finned heat exchangers. Interstice either vacuum-filled or filled with a
gas, a low melting point or a grease. From French (1985) [68]. ........................ 400
Figure 15-35: Grease-filled heat pipe-heat pipe joint. Hinge joint-coaxial with the
internal heat pipe and parallel to the plane of the radiator heat pipe. From
Delil (1986a) [33]. ..............................................................................................401
Figure 15-36: Braided conductor flexible thermal joint. From Delil (1987a) [31]. ................. 401
Figure 15-37: Clamped joint contact conductor. Heat pipe penetrates heat exchanger.
Fluid pressure-enhanced thermal contact. From Ellis & Rankin (1983) [42]...... 402
Figure 15-38: Self deployed membrane heat pipe radiator. Both deployed and
undeployed configurations shown in the figure. From Delil (1986a) [33]. .......... 402
Figure 15-39: Internal details of the rotatable fluid transfer coupling. Dimensions are in
mm. From Heizer, Goo, Rhodes, Thoreson & Parish (1986) [83]...................... 403
24
ECSSEHB3101Part13A
5December2011
Figure 15-40: Rotatable fluid transfer coupling. a) General view showing the liquid and
vapour radial ports. b) Channel geometry. From Heizer, Goo, Rhodes,
Thoreson & Parish (1986) [83]........................................................................... 403
Figure 15-41: Pressure drop, p, along the duct at one position of the rotary coupling
for different values of the heat rejection rate, Q. From Heizer, Goo, Rhodes,
Thoreson & Parish (1986) [83]........................................................................... 404
Figure 15-42: a) Schematic of a three-axis stabilized spacecraft in geosynchronous
orbit. From Chalmers & Pustay (1986). b) Typical thermal load sharing of
east-west faces. From Wise (1986) [192]. ......................................................... 405
Figure 15-43: East-west radiator coupling. a) Based on HP technology. From Chalmers
& Pustay (1986). b) Based on CPL technology. From Chalmers, Pustay,
Moy & Kroliczek (1986) [23]...............................................................................406
Figure 15-44: Schematic of CPL engineering model. In a) the liquid fills most of the
loop, whereas in b) part of this liquid has been evaporated. From Chalmers,
Pustay, Moy & Kroliczek (1986) [23].................................................................. 407
Figure 15-45: Schematic of a basic CPL system. From Chalmers, Pustary, Moy &
Kroliczek (1986) [23]..........................................................................................411
Figure 15-46: a) Standard CPL evaporator pump. From Chalmers et al. (1986) [23]. b)
Heat flow in a cross section of a typical CPL evaporator. From Wise (1986)
[192]...................................................................................................................411
Figure 15-47: Prototype capillary cold plate (PCCP) design. Dimension in mm. From
Chalmers, Pustay, Moy & Kroliczek (1986) [23]. ............................................... 412
Figure 15-48: CPL technology radiators. a) Direct condensation radiator. b) Heat
exchanger-heat pipe radiator. From Chalmers, Pustay, Moy & Kroliczek
(1985) [23]. ........................................................................................................412
Figure 15-49: Monogroove heat pipe. From Alario, Haslett & Kosson (1981) [1]................. 414
Figure 15-50: Osmotically pumped heat transfer system. From Tanzer, Fleischman &
Stalmach (1982) [175]. ......................................................................................415
Figure 15-51: Biomorph (Biomorph) pump. From Peterson (1987) [142]............................. 415
Figure 15-52: Instrument-TMS interfaces. From Almgren et al. (1981) [2]........................... 417
Figure 15-53: Schematic of a grooved cold plate/hot plate in the cold plate mode. From
Hwangbo & McEver (1985) [94]......................................................................... 417
Figure 15-54: Two-Phase Mounting Plate (TPMP) development unit. The Lexan
window has been incorporated to observe the flow during testing. All
dimensions are in mm. From Grote & Swanson (1985) [77].............................. 418
Figure 15-55: Void fraction sensors of the capacitance type. a) Single coaxial capacitor.
b) Concave plate capacitor. c) Double helix capacitor. d) Film thickness
gage. From Delil (1986b) [30]. ........................................................................... 420
Figure 15-56: Dimensionless capacitance as a function of void fraction, . a) Annular
flow, single coaxial capacitor. b) Annular flow, concave plate capacitor.
From Delil (1986b) [30]. .....................................................................................421
Figure 15-57: Dimensionless capacitance as a function of dimensionless film thickness,
/d. Flat wall, film thickness gage. From Delil (1986b) [30]................................ 422
Figure 15-58: Engagement sequence of the RSO disconnect. Disengagement is
achieved through the reverse sequence. From MOOG [127]. ........................... 423
Figure 16-1: Closed-loop control system block diagram. ..................................................... 426
25
ECSSEHB3101Part13A
5December2011
Figure 16-2: Open-loop control system block diagram......................................................... 426
Figure 16-3: Typical block diagram of adaptative control systems.......................................427
Figure 16-4: Feedback system.............................................................................................428
Figure 16-5: Sketch of the gain of a system as a function of frequency............................... 429
Figure 16-6: Feedback system with two feedback loops...................................................... 430
Figure 16-7: Feedback system with a noise signal. ............................................................. 431
Figure 16-8: Feedback structure of instruments and regulators. From Ogata (1990)
[137]...................................................................................................................433
Figure 16-9: Block diagram of an industrial control system, which consists of an
automatic controller, an actuator, a plant, and a sensor (measuring
element). From Ogata (1990) [137]. .................................................................. 434
Figure 16-10: a) Block diagram of an on-off controller; b) block diagram of an on-off
controller with differential gap; c) output versus time curve. From Ogata
(1990) [137]. ......................................................................................................435
Figure 16-11: Block diagram of a proportional controller. From Ogata (1990) [137]............ 436
Figure 16-12: Block diagram of an integral controller. From Ogata (1990) [137]. ................ 437
Figure 16-13: a) Block diagram of a proportional-integral controller; b) and c) diagrams
depict a unit-step input and the controller output. From Ogata (1990) [137]. .... 438
Figure 16-14: a) Block diagram of a proportional-derivative controller; b) and c)
diagrams depict a unit-ramp input and the controller output. From Ogata
(1990) [137]. ......................................................................................................439
Figure 16-15: a) Block diagram of a proportional-integral-derivative controller; b) and c)
diagrams depict a unit-ramp and the controller output. From Ogata (1990)
[137]...................................................................................................................440
Figure 16-16: Block diagram of a digital control system. From Ogata (1987) [135]. ............ 442
Figure 16-17: Block diagram of a digital control system showing signals in binary or
graphic from. From Ogata (1987) [135]. ............................................................442
Figure 16-18: Schematic diagram of a pressure system. From Ogata (1990) [137]. ........... 443
Figure 16-19: Schematic diagram of a pneumatic actuating valve. From Ogata (1990)
[137]...................................................................................................................445
Figure 16-20: a) Dashpot; b) step change in x and the corresponding change in y
plotted versus t, c) block diagram of the dashpot. From Ogata (1990) [137]..... 446
Figure 16-21: a) Schematic diagram of a force-distance type pneumatic proportional
controller; b) block diagram; c) simplified block diagram. From Ogata (1990)
[137]...................................................................................................................448
Figure 16-22: Schematic diagram of a force-balance pneumatic proportional controller.
From Ogata (1990) [137]. ..................................................................................449
Figure 16-23: a) Servomoter that acts as a proportional controller; b) block diagram of
the servomotor. From Ogata (1990) [137]. ........................................................450
Figure 16-24: a) Pneumatic proportional-derivative controller; b) step change in e and
the corresponding changes in x and pc plotted versus t; c) block diagram.
From Ogata (1990) [137]. ..................................................................................451
Figure 16-25: a) Sketch of a hydraulic proportional-derivative controller; b) block
diagram. From Ogata (1990) [137]. ................................................................... 452
26
ECSSEHB3101Part13A
5December2011
Figure 16-26: Hydraulic integral controller. From Ogata (1990) [137].................................. 454
Figure 16-27: a) Pneumatic proportional-integral controller; b) step change in e and the
corresponding changes in x and pc plotted versus t; c) block diagram on the
controller; simplified block diagram. From Ogata (1990) [137]. ......................... 455
Figure 16-28: a) Schematic diagram of hydraulic proportional-integral controller; b)
block diagram. From Ogata (1990) [137]........................................................... 456
Figure 16-29: a) Pneumatic proportional-integral-derivative controller; b) block diagram
of the controller. From Ogata (1990) [137]. ....................................................... 457
Figure 16-30: Fluidloop modelled as a control system.........................................................460
Figure 16-31: Control block diagram of the fluid loop for cooling Spacelab experiments.
From Microtecnia (1977) [126]........................................................................... 461
Figure 16-32: Instrumentation and control system schematics. From Sadunas et al.
(1986) [154]. ......................................................................................................462
Figure 16-33: Block diagrams of automatic controllers with a) first-order sensor; b)
overdamped second-order sensor; c) underdamped second-order sensor.
From Ogata (1990) [137]. ..................................................................................467
Figure 16-34: Block diagram of a control system. ................................................................ 467
Figure 16-35: Space radiator system. From Baker et al. (1967) [3]. .................................... 474
Figure 16-36: Space radiator block diagram. From Baker et al. (1967) [3]. ......................... 476
Tables
Table 7-1: Effective roughness height of a number of common surfaces. ............................. 95
Table 7-2: Constants for Power Law Approximation. Hydraulically Smooth Regime............. 97
Table 7-3: Loading Factors Accounting for Temperature-Dependence of Diabatic
Friction .................................................................................................................98
Table 9-1: Heat Transfer Enhancement Techniques ........................................................... 130
Table 9-2: Ratio of heat transfer coefficients for constant wall temperature, NuT, and
constant heat flux, Nuq, for turbulent pipe flow. From Reynolds (1974) [149].... 134
Table 10-1: Physical Properties of Typical Liquid Coolants ................................................. 180
Table 10-2: Environmental Properties of Typical Liquid Coolants........................................ 183
Table 10-3: Properties of Dow Corning 200 Fluids (Dimethyl Siloxane Polymers) .............. 206
Table 10-4: Corrosion and Oxidation Text Data for Coolanol Liquids .................................. 208
Table 10-5: Toxicity of Several Freon Liquids ...................................................................... 209
Table 10-6: Swelling of Elastomers in Several Freon Liquids a ............................................ 209
Table 10-7: Compatibility of Freon E2 and FC-75 with Elastomers a ................................... 210
Table 10-8: Compatibility of Freon E3 with Elastomers, Plastics and Wire Coatings a ........... 210
Table 10-9: Compatibility of Oronite Flo-Cool 100 with Elastomers..................................... 212
Table 11-1: Flow Inside Circular and Flattened Circular Tubes ........................................... 239
Table 11-2: Flow Normal to Banks of Bare Tubes ............................................................... 240
Table 11-3: Plate-Fin surfaces, plain fins ............................................................................. 241
27
ECSSEHB3101Part13A
5December2011
Table 11-4: Plate-Fin surfaces, louvered fins....................................................................... 243
Table 11-5: Plate-Fin surfaces, strip fins.............................................................................. 244
Table 11-6: Plate-Fin surfaces, wavy fins ............................................................................245
Table 11-7: Plate-Fin surfaces, perforated fins .................................................................... 245
Table 11-8: Plate-Fin surfaces, pin fins................................................................................ 246
Table 11-9: Finned tubes, circular tubes, circular fins.......................................................... 247
Table 11-10: Finned tubes, circular tubes, continuous fins .................................................. 248
Table 11-11: Finned tubes, flat tubes, continuous fins......................................................... 248
Table 11-12: Crossed-Rod, woven-screen and sphere matrices ......................................... 249
Table 11-13: Fluid Paths in the Shell Side of Shell-and-Tube Exchangers a ....................... 260
Table 11-14: Engineering Practices for Reducing Maldistribution in the Shell Side of
Shell-and-Tube Exchangers a ............................................................................261
Table 11-15: Nominal Dimensions for Numerical Results.................................................... 265
Table 11-16: Basic assumptions for the theoretical analysis ............................................... 266
Table 11-17: Characteristics of Offset Rectangular Plate-Fin Surfaces............................... 269
Table 11-18: Core 501 MOD Geometries Derived from Different Models............................ 270
Table 11-19: Fouling Resistance of Several Common Materials a ....................................... 273
Table 11-20: Fouling Mechanisms a .....................................................................................274
Table 11-21: Effects of Different Physical Parameters on Fouling a..................................... 275
Table 12-1: Rotodynamic Pumps ......................................................................................... 285
Table 12-2: Displacement Pumps. Reciprocating ................................................................ 285
Table 12-3: Displacement Pumps. Rotary............................................................................286
Table 12-4: Main Features of Typical Pumps....................................................................... 286
Table 12-5: Conversion Factors in the Deduction of from ns ............................................ 289
Table 12-6: Characteristics of Several Commercially Available Pumps............................... 293
Table 13-1: Optimization of the Liquid to Air Heat Exchanger Case 1) mc = 0,0303
kg.s1, TS Tci = 20 K..........................................................................................308
Table 13-2: Optimization of the Liquid to Air Heat Exchanger Case 2) mc = 0,0817
kg.s1, TS Tci = 10 K..........................................................................................309
Table 14-1: Typical Pressure Losses in Air-Water Annular Flow with Entrainment (D =
25,4 x 10-3 m, p = 105 Pa, T = 293 K). ............................................................... 340
Table 14-2: Geometry and Flow-Dependent Terms in Eqs. [14-107] and [14-109] ............. 347
Table 14-3: Condensation in Ducts. Typical Fluid Properties. Assumed values: Tsat =
300 K, Tsat - Tw = 10 K........................................................................................349
Table 14-4: Condensation in Ducts. Parameters Depending on m and D. Assumed
values: mhfg = 1,5 x 103 W, D = 16,1 x 10-3 m. Stratified case........................... 350
Table 15-1: Evolution of Requirements ................................................................................359
Table 15-2: Two-Phase Loop Line Patterns.........................................................................361
Table 15-3: Power Dissipation by Module............................................................................364
28
ECSSEHB3101Part13A
5December2011
Table 15-4: TMS Design Requirements ............................................................................... 365
Table 15-5: TMS Design Concepts ......................................................................................367
Table 15-6: TMS Design Goals ............................................................................................367
Table 15-7: Evaluation of Concepts ..................................................................................... 370
Table 15-8: Characteristics of Single-Phase and Two-Phase TMSs with PC Capacitor...... 374
Table 15-9: PC Capacitor Performance ...............................................................................375
Table 15-10: Single-Phase and Two-Phase TMS Capacitor Specifications ........................ 377
Table 15-11: Fuel Cell Loop Design Requirements ............................................................. 380
Table 15-12: Orbital-Average Schema. Values of x for Off-Design Operation ..................... 382
Table 15-13: Two-Phase System. Radiator Temperatures for Off-Design Operation .......... 382
Table 15-14: Component Mass of a Typical Pumped Loop TMS......................................... 384
Table 15-15: Heat Pump Augmented vs. Classical Fluid Loop Trade-Off............................ 386
Table 15-16: EOL Radiator Area for 10-year Life.................................................................392
Table 15-17: Radiator Area vs. Design Life ......................................................................... 393
Table 15-18: Comparison of Fixed and Steerable Radiator Areas....................................... 394
Table 15-19: Requirements for a Rotary Coupling Onboard. Space Station. ...................... 395
Table 15-20: Joints for Steerable Radiators......................................................................... 396
Table 15-21: NASA development efforts in CPL technology................................................ 408
Table 15-22: Comparative Summary of Pumping Systems ................................................. 416
Table 16-1: Summary of basic control actions ..................................................................... 440
Table 16-2: Summary of Implementation Techniques.......................................................... 458
Table 16-3: Control unit philosophy trade-off. From Microtecnia (1977) [126]. .................... 463
Table 16-4: Space computers .............................................................................................. 465
Table 16-5: Available microprocessor options ..................................................................... 466
Table 16-6: Characteristics of several temperature sensors................................................ 468
Table 16-7: Characteristics of pressure sensors..................................................................469
Table 16-8: Characteristics of flow sensors. ........................................................................ 470
Table 16-9: Characteristics of control valves. From Liptak (1969) [115]. ............................. 470
Table 16-10: Some MATLAB-Driven CACSD Software .From Cellier and Rimvall
(1988) [21] .........................................................................................................472
Table 16-11: Some non-MATLAB CACSD Software Packages. From Cellier and
Rimvall (1988) [21].............................................................................................472
Table 16-12: A Brief Survey of 22 CACSD Packages. From Jamshidi et al. (1992) [100] ... 473
Table 16-13: System characteristic. From Backer et al. (1967) [3] ...................................... 475
Table 16-14: Control law for the block diagram in Figure 16-36 .......................................... 477
29
ECSSEHB3101Part13A
5December2011
1
Scope
Fluidloopsareusedtocontrolthetemperatureofsensitivecomponentsinspacecraftsystemsinorder
toensurethattheycanfunctioncorrectly.
While there are several methods for thermal control (such as passive thermal insulations,
thermoelectric devices, phase change materials, heat pipes and shortterm discharge systems), fluid
loopshaveaspecificapplicationarea.
ThisPart13providesadetaileddescriptionoffluidloopsystemsforuseinspacecraft.
TheThermaldesignhandbookispublishedin16Parts
ECSSEHB3101Part1A ThermaldesignhandbookPart1:Viewfactors
ECSSEHB3101Part2A ThermaldesignhandbookPart2:Holes,GroovesandCavities
ECSSEHB3101Part3A ThermaldesignhandbookPart3:SpacecraftSurfaceTemperature
ECSSEHB3101Part4A ThermaldesignhandbookPart4:ConductiveHeatTransfer
ECSSEHB3101Part5A ThermaldesignhandbookPart5:StructuralMaterials:Metallicand
Composite
ECSSEHB3101Part6A ThermaldesignhandbookPart6:ThermalControlSurfaces
ECSSEHB3101Part7A ThermaldesignhandbookPart7:Insulations
ECSSEHB3101Part8A ThermaldesignhandbookPart8:HeatPipes
ECSSEHB3101Part9A ThermaldesignhandbookPart9:Radiators
ECSSEHB3101Part10A ThermaldesignhandbookPart10:PhaseChangeCapacitors
ECSSEHB3101Part11A ThermaldesignhandbookPart11:ElectricalHeating
ECSSEHB3101Part12A ThermaldesignhandbookPart12:Louvers
ECSSEHB3101Part13A ThermaldesignhandbookPart13:FluidLoops
ECSSEHB3101Part14A ThermaldesignhandbookPart14:CryogenicCooling
ECSSEHB3101Part15A ThermaldesignhandbookPart15:ExistingSatellites
ECSSEHB3101Part16A ThermaldesignhandbookPart16:ThermalProtectionSystem
30
ECSSEHB3101Part13A
5December2011
2
References
ECSSSST0001 ECSSSystemGlossaryofterms
ECSSEHB3101Part8 ThermaldesignhandbookPart8:HeatPipes
ECSSEHB3101Part9 ThermaldesignhandbookPart9:Radiators
ECSSEHB3101Part10 ThermaldesignhandbookPart10:PhaseChangeCapacitors
ECSSEHB3101Part14 ThermaldesignhandbookPart14:CryogenicCooling
ECSSEHB3101Part15 ThermaldesignhandbookPart15:ExistingSatellites
AllotherreferencesmadetopublicationsinthisPartarelisted,alphabetically,intheBibliography.
31
ECSSEHB3101Part13A
5December2011
3
Terms, definitions and symbols
3.1 Terms and definitions

ForthepurposeofthisStandard,thetermsanddefinitionsgiveninECSSSST0001apply.
3.2 Abbreviated terms

ThefollowingabbreviatedtermsaredefinedandusedwithinthisStandard.
A/D analogtodigitalconverter
ARS atmosphererevitalizationsubsystem
ATCS activethermalcontrolsystem
BOL beginningoflife
CACSD computeraidedcontrolsystemdesign
CAD computeraideddesign
CAPL capillarypumpedheattransportloop
CPL capillarypumpedloop
CPU centralprocessingunit
D/A digitaltoanalogconverter
DCR directcondensationradiator
ECLA experimentcoolingsystemwithliquidtoairheat
exchanger
EDHX experimentdedicatedheatexchanger
EOL endoflife
FCL freoncoolantloop
FL fullydevelopedflow
GSE groundsupportequipment
32
ECSSEHB3101Part13A
5December2011
HP heatpipe
HPSTM highpowerspacecraftthermalmanagement
HTEL combinedhydrodynamicalandthermalentrylength
HX heatexchanger
I integralcontrollers
I/O input/output
Mac maximumallowableconcentration
MBA multipleberthingadaptor
MPU microprocessorunit
NPSH netpositivesuctionhead
OAO orbitingastronomicalobservatory
P proportionalcontrollers
PC phasechange
PCCP prototypecapillarycoldplates
PCM phasechangematerial
PD proportionalderivativecontrollers
PI proportionalintegralcontrollers
PID proportionalintegralderivativecontrollers
S/H sampleandhold
SLMS sealleakagemanagementsubsystem
STS spacetransportationsystem
TED thermoelectricdevice
TEL thermalentrylength
TMS thermalmanagementsystem
TS thermalstorage
VCHP variableconductanceheatpipe
33
ECSSEHB3101Part13A
5December2011
3.3 Symbols
A heattransfersurfacearea,[m2]
AFL internalcrosssectionalareaofaduct.[m2].Alsocalled
freeflowarea
AFR frontalareaofaheatexchangercore,[m2]
AG gasfilledpartofthecrosssectionalareaofaduct.[m2]
AL liquidfilledpartofthecrosssectionalareaofaduct,
[m2]
Af finarea,[m2]
Ak wallcrosssectionalareaforlongitudinalheat
conduction,[m2]
Aw Clause9:insidewallheattransfersurfacearea,[m2]
Clause11:averagewallarea,[m2]
C Chaps5and11:capacityrateofaflowstream,[W.K1]
C=mcp
Clause14:Chisholmparameter.Eq.[149]
Clause15:electricalcapacitance,[F]
C* capacityrateoftheinparallelstream,[W.K1]refersto
assembliesofheatexchangersinparallelinoneofthe
streams
CostNtu degradationintheheatexchangerthermal
performance,CostNtu=1(Ntueff/Ntu)
D diameter,[m]
DE equivalentorhydraulicdiameterofaductofNon
CircularCrossSection.[m].DE=4AFL/.
DEa equivalentorhydraulicdiameterofafinnedduct,[m]
DM innerdiameterofaroughtubedefinedona
volumetricbasis,[m]DM=(4v/L)1/2
Db diameterofthelargestcrosssectioninanaxial
displacementheattransferenhancementdevice.[m]
Dm nosheardiameterintheflowthroughanannular
duct,[m]
E emissivepower,[W.m2]
EL specificenergylostbyapump,[J.kg1]itisdefinedper
unitliquidmass
34
ECSSEHB3101Part13A
5December2011
En specificenergydevelopedbyapump,[J.kg1]itis
definedperunitliquidmass
Es specificenergysuppliedtoapump,[J.kg1]itis
definedperunitliquidmass
Ev specificsuctionenergyofapump,[J.kg1]itisdefined
perunitliquidmass
F Clause9:finfactor.ratioofthetotalconvectiveheat
transfertothattransferredthroughthetubewalls
alone,Clause10:coolingeffectivenessofafluid
FLow fractionofchannelscarryinglowerthanaverageflow
onthenonuniformsideofasplitcounterflowheat
exchanger(flemingsmodel)
Fr froudenumber,Fr=V2/gD
F1 factorintubebanks,itaccountsforvariablefluid
properties
F2 factorintubebanks,itaccountsforabnormalnumber
ofrows,(N10)
F3 factorintubebanks,itaccountsfortheeffectoftube
yaw
F4 factorintubebanks,itgivestherowtorowvariation
ofheattransfercoefficient
Gr gDE3 2 Tw Tb
Grashofnumber. Gr
2
Gz Graetznumber,Gz=DE/4LPrRe=mcp/kL.
H Clause12:head,[m]
Clause14:liquidlevelinstratifiedflow,[m]
Jc thermalcapacitorstoredenergy,[J]
K Clause6:deannumber,K=Re(r/R)1/2
Clause7:correctionfactor,itaccountsforthefactthat
thecrosssectionofthetubeisnotcircular
L length,[m]
LE equivalentlength,[m]
Lent entrancelength,[m]
M Clause7:machnumber,M=V/a
Clauses13and15:systemmass,[kg]
35
ECSSEHB3101Part13A
5December2011
MCP coldplatemass,[kg]
MP fluiddrivermass,[kg]
MSP solarpanelmass,[kg]
MwL (wet)plumbingmass,[kg]
M+ masspenaltyifafluidloop,[kg]itincludesthe
equivalentmassduetopumpingpower
N Clause6:numberofrowsintubebanks,therowsare
countedintheflowdirection
Clause11:passagecountofaheatexchangercore,
[m2]
NTD parameterintheTaitelDucklerlimitforstratified
flow,Eq.(10.42)
Ntu numberofheattransferunits
Notu overallnumberofheattransferunits,referstoa
liquidcoupledsystemofheatexchangers
1
1 1
N
o
tu
tu1
N N tu 2
Nu Nusseltnumber,Nu=hDE/k.
NuT Nusseltnumberforconstantwalltemperature
Nuq Nusseltnumberforconstantheatfluxthroughthe
wall
Nu11 Nusseltnumberfortheinnerwallofacylindrical
annuluswhentheouterwallisinsulated
Nu22 Nusseltnumberforouterwallofacylindricalannulus
whentheinnerwallisinsulated
P powersuppliedtoapump,[W]
P Pecletnumber,Pe=PrRe.
Pr Prandtlnumber,Pr=cp/k.
Q heattransferrate,[W]
Q(x) convectiveheattransferrateinatube,fromonsetof
heattransfertostationx,[W]
Qc thermalcapacitorenergystoragerate,[W]
R radiusofthecenterlineincurvedpipes,[m]
Clauses5,11and13:capacityratio,R=C1/C21
36
ECSSEHB3101Part13A
5December2011
RE WicksDuklerentrainmentparameter,[m2.s]Eq.
(10.29)
R Reynoldsnumber,Re=VDE/
S suctionspecificspeedofapumpatgivenoperating
m / 1/ 2
conditions, S
Ev3 / 4
Sf h fg
Stefannumber, Sf
c pL Tsat Tw
St Stantonnumber,St=Nu/PrRe=h/Vcp
T temperature,[K]
TR referencetemperature,[K]
Clause15:radiatortemperature,[K]
TS sourcetemperature,[K]
Tav
TdA FL
fluidaveragetemperature, Tav
AFL

AFL
Tb
uTdA FL
fluidbulktemperature, Tb
AFL

VAFL
Tbam arithmeticmeanbulktemperature[K],Tbam=(Tbi+
Tbo)/2.
Tc Clauses5,11,13and15:coldtemperature,[K]
Clause9:coolingwatertemperature,[K]
Tsat liquidvaportemperature,[K]
Tw walltemperature,[K]
T.5 fluidfilmtemperature,[K],T.5=(Tw+Tb)/2
T+ dimensionlesstemperature,T+=u*cp/h
U overallthermalconductance,[W.m2.K1],U=q/(Th
Tc).
V meanfluidvelocity,[m.s1]
We Webernumber,We=GV2D/.
X LockhartMartinelliparameter.Eq.[143]
37
ECSSEHB3101Part13A
5December2011
Z influencecoefficient,itrelatestheonesurfaceheated
NusseltnumberforthecylindricalannulustoNusselt
numbersonbothsurfaceswhentheyareheatedwith
anyrationofheatfluxes.
a Clause7:velocityofsound,[m.s1].
Clauses11and13:wallorplatethickness,[m]
a partingsheetthickness,[m]
b protrusionspacing,[m]itcanbemeasuredineither
axialorcircumferentialdirection
c massfractionofagivenspecies,whenseveralspecies
arepresent
cK pressurelosscoefficient,cK=(p1p2)/(1/2V2)
cp constantpressurefluidspecificheat,[J.kg1.K1]for
gasescpisusedindistinctiontocv,theconstant
volumespecificheat
d Clause7:diameterofanorifice,wireortubeplaced
insideaduct,[m]
Clause11:cellheight,[m]
Clause15:interplatedistanceofacapacitor[m]
e roughnessheight,[m]
e+ dimensionlessroughnessheight,e+=eu*/=
(e/D)Re(f/2)1/2
f fanningfrictionractor,f=/4
f*c fanningfrictionfactorforthehomogeneouscore
mixtureflowingaloneintheduct
fF fanningliquidfilmfrictionfactoratthewall,annular
flowwithentrainment
fGi fanningcorefrictionfactorattheinterface,ideal
annularflow
f*Gi fanningcorefrictionfactorattheinterface,annular
flowwithentrainment
fL fanningliquidfilmfrictionfactoratthewall,ideal
annularflow
fTP fanningfrictionfactoratthewall,twophasemixture
model
g accelerationduetogravity,[m.s2]
Clause5:dimensionlessheattransfersimilarity
function
38
ECSSEHB3101Part13A
5December2011
h convectiveheattransfercoefficient,[W.m2.K1]
hc thermaljointconductance,[W.m2.K1]
hfg heatofvaporizationorcondensation,[J.kg1]
hfg heatofcondensationcorrectedbypartialsubcooling
effect,[J.kg1]Eq.(10.34)
j colburnfactorforheattransfer,j=StPr2/3
k thermalconductivity,[W.m1.K1]
Clause7:incrementalpressurelossfactor,itaccounts
forentranceeffects
m fluidmassflowrate,[kg.s1]
Clauses6and9:exponent,itappearsintheloading
factoraccountingfortheinfluenceoffluid
temperatureonheattransfer
mE massflowrateofthefluidentrainedinthecore,
[kg.s1]annularflowwithentrainment
mF massflowrateoftheliquidremaininginthefilm,
[kg.s1]annularflowwithentrainment
n Clause6:numberoftubesintubebanks,alsonumber
assignedtothenthrowtubesintubebanks
Clause7:numberofwiresperunitlengthofwire
gauze,[m1]exponent
Clauses11and13:numberofassembledheat
exchangers
Clause12and15:rotatingspeed,[revolutionsper
minute]
ns (optimum)specificspeedofapump,[m3/4.s3/2]
nm /
1/ 2
ns
H 3/ 4 opt
p pressure,[Pa]
psat liquidvaporpressure,[Pa]
pt totalpressure,[Pa]
q heatflux,[W.m2]
r radiusorradialcoordinate,[m]
rf foulingresistance,[K.m2.W1]
s protrusionthickness,[m]itismeasuredinaxial
direction
39
ECSSEHB3101Part13A
5December2011
s1 longitudinalpitchintubebanks,[m]
s1 diagonalpitchinstaggeredtubebanks,[m]
st transversepitchineitherinlineorstaggeredtube
banks,[m]
t Clauses7and9:thickness,[m]
Clause11:temperature,[K]
Clause15.time,[s],[min],[h]or[yr]
u localfluidvelocity[m.s1]
u* frictionvelocity,[m.s1],u*=(w/)1/2.
u+c dimensionlessvelocity,u+=u/u*
u+ dimensionlessfrictionfactorsimilarityfunction
v volume,[m3]
w vaporquality
x axialdistance,[m]
Clause11:massflowfraction
x+ dimensionlessaxialdistancetoductentry,x+=
(x/r)/PrRe=/2Gz
y Clause9:tubediameterper180finortapetwist
Clause14:distancetothewalloftheduct,[m]
z verticaldistanceabovearbitrarydatumplane,[m]
Tm meanwalltofluidbulktemperaturedifferential,[K]
p pressureloss,[Pa]
radiatorareadensity,[kg.m2]
longitudinalwallconductionparameter,=(k/L)Ak/C1
wettedperimeterofaduct,[m]
Chaps5,7and11:ratiooffreeflowareatototal
frontalarea,alsocalledporosity
Clause14:LockhartMartinellipressurelossmultiplier
Eq.[142]
upstreamtodownstreamarearatio,referstosudden
areachangesinducts,factorintubebanks,itaccounts
fortheeffectoftubeyawonpressureloss
(optimum)specificspeedofapump,ECUACION
40
ECSSEHB3101Part13A
5December2011
Clause11:areadensity,[m2.m3],ratiooftotalheat
transferareaononesideofaclosedtypeheat
exchangerorrecuperatortototalexchangervolume,
inthecaseofperiodictypeexchangersorregenerators
isbasedontheareaofbothsides
Clauses14and15:voidfraction
1 factorinbends,itaccountsfortheaspectratioofthe
bendcrosssection
2 factorinbends,itaccountsforthebendangle
3 factorinorifices,itaccountsfortherelativethickness
ofanorificeplatewhent/d<0,8
4 factorinorifices,itaccountsfortherelativethickness
ofanorificeplatewhent/d0,8
a5 factorinscreens,itaccountsforlowReynoldsnumber
inroundwiregauzes
6 factorinvalves,itaccountsforpartialopening
s solarabsorptance
Clauses5and11:areadensity,[m2.m3]ratiooftotal
heattransferareaononesideofaplatefinclosedtype
heatexchangerorrecuperatortothevolumebetween
theplatesonthatside,inthecaseofperiodictype
exchangersorregeneratorsisbasedontheareaof
bothsides
Clauses6and9:coefficientofvolumetricthermal
expansion,[K1]
Clause7:factorinbends,itaccountsforthe
interactionbetweentwocirculararcbendshavinga
bendangleof90each
Clause15:90minustheanglebetweenthelinetothe
Sunandthenormaltoorbitplane
Ratioofspecificheatsinagas
Clause7:ratioofvalvecontroltravelfromclosed
positiontototalvalvecontroltravel
Clause8:laminarsublayerthickness,[m]
Chaps14and15:filmthickness,[m]
m correctiontoallowfordifferencesbetween
dimensionlesstemperatureanddimensionless
velocityinthefullyturbulentcoreinpipeflow,see
clause13.2.
decreaseineffectivenessofaheatexchangerbecause
oflongitudinalwallconduction
41
ECSSEHB3101Part13A
5December2011
+ dimensionlessfilmthickness,+=u*/
Clauses5,11and13:exchangerheattransfer
effectiveness
Clause9:radialgapbetweenthetwistedtapeinsert
usedforenhancingtheheattransferandthetubewall,
[m]
Clause15:hemisphericaltotalemittance
Clause15:dielectricpermittivity,[F.m1]
cp coldplateeffectiveness
0 dielectricpermittivityofvacuum,0=8,8542x1012
F.m1
t overallheattransfereffectivenessofanassociationof
heatexchangers
heattransfereffectiveness
f fineffectiveness
p overallefficiencyofapump
Clause6:yawangle,[angulardegrees]
Clause7:totalangleturnedthroughbyabend,
[angulardegrees]
frictionfactor
B bakerparameter
fluiddynamicviscosity,[Pa.s]itisalsoknownas
viscositycoefficient
fluidkinematicviscosity,[m2.s1]
fluiddensity,[kg.m3]
surfacetension,[N.m1]
Clause15:StefanBoltzmannconstant,=5,6697x108
W.m2.K4
shearingstress,[Pa]
Clause11:flowmaldistributionparameter,itis
defined,althoughnotinauniquefashion,intermsof
themassflowrate
Clause14:polarangleofstratifiedflowinterface,
[angulardegrees]
B bakerparameter
rotatingspeed,[radianspersecond]
42
ECSSEHB3101Part13A
5December2011
Subscripts
F liquidfilm,annularflowwithentrainment
G Clause7:grossvalue
Clause14:gas
Gi gasattheinterface
H affectedbyhallstransformation
High referstochannelswithhigherthanaverageflowon
thenonuniformsideofasplitcounterflowheat
exchanger(flemingsmodel)
L Clause7:referstolengthL
Clause11:referstoliquidstreaminliquidcoupled
heatexchanger
Clause14:liquid
Low referstochannelswithlowerthanaverageflowonthe
nonuniformsideofasplitcounterflowheatexchanger
(Flemingsmodel)
M phasechange
N netvalue
P evaluatedatconstantpumpingpower
R radiator
S referstoheatsource
SG gasflowingaloneintheduct
SL liquidflowingaloneintheduct
T referstototal(axialplusspiral)valuesinspiralling
flow
TP twophaseinthemixturetheory
a heattransferenhanceddate
Clause15:EarthAlbedo.
am arithmeticmeanvalue
av averagevalue
b fluidbulkproperties
c Clauses5,11,13and15:coldstream
43
ECSSEHB3101Part13A
5December2011
Clause7:cylindricaltubes
Clause14:core,annularflowwithentrainment
d downstreamtangentofabend
e spatialandtemporalmeanconditionsatthetiplevel
oftheroughnesselements
Clause15:spaceenvironmentcondition
equiv equivalent
f referstofinortwistedtape
f hotstream
i Clauses5,6and15:inletconditions
Clause14:interface
im inletmeanvalue
ir infrared
lm logarithmicmeanvalue
m meanvaluebetweencommencementofheattransfer
andstationx
max maximumvalue
n referstothenthcomponentofanensemble
o outletconditions
Clause9:heattransfernonenhanceddata
opt optimumconditions
r referencevalueforgiventhermaland/orgeometrical
characteristicsofasystem,itisomittedwhenno
correctionismade
s referstostaticpressure
Clause15:solar
t referstototalpressure
tt turbulentliquid,turbulentgas
tv turbulentliquid,laminargas
u upstreamtangentofabend
vt laminarliquid,turbulentgas
vv laminarliquid,laminargas
44
ECSSEHB3101Part13A
5December2011
w wallconditions
x localconditionsatsectionxofaduct
conditionsattheedgeofthelaminarsublayer
0 referstoanorificeplateofzerorelativethickness
0,8 referstoanorificeplateifrelativethicknesst/d=0,8
1 Clauses5,11and13:conditionsforthestreamwith
smallercapacityrate
Clauses6and9:innerwallconditionsinacylindrical
annulus
Clause7:upstreamreferenceconditions
2 Clauses6and9:outerwallconditionsinacylindrical
annulus
Clause7:downstreamreferenceconditions
Clause11conditionsforthestreamwithlarger
capacityrate
Superscripts
meanvalue
* entrainmenteffectsincluded
45
ECSSEHB3101Part13A
5December2011
4
General introduction
Fluidscanbeusedtoprotectfromthermalloadssensitivecomponentsofspacecraftsystems.
Thisprotectioncanbeachievedonthebasisoftwocompletelydifferentprocedures.
1)Thefluidisatrestalthoughchangingitsphase,storingheatthroughmelting(orevaporation)when
the temperature increases, and restoring it through freezing (or condensation), for dissipation at a
later time, when the temperature decreases. This is typical of phase change capacitors whose main
featuresarediscussedinECSSEHB3101Part10.
Singlephasefluidsatrestarenotusedforthermalcontrolpurposesbecauseofthefarbetterqualities
ofsolidmaterials,eitherasthermal)insulatorsorasthermalconductors.
2) The fluid is in motion, absorbing the heat at a relatively steady rate from the component whose
temperature is controlled, transferring it to a heat sink which can be placed fairly apart from the
source.
Heat transfer can be achieved either through sensible heat change or through phase change (latent
heat):
2.1)Sensibleheatchange.AliquidorgasphaseisusedtotransferheataccordingtotheequationQ=
mcpT.Inordertoincreasetheheattransferrate,Q,eitherthemassflowrate,m,orthetemperature
difference, T,isincreasedforaparticularfluid.Thespecificheat,cp,ismoreorlessthesameforall
liquidsandgasesatnormalconditions.
2.2) Phase change. The heat transferred is used to evaporate the working liquid according to the
equationQ=mhfg,wherehfgisthelatentheatofvaporization.Nowtheheattransferismuchhigher
thanin2.1forthesamemassflowrateandthesametypeoffluid,whereasthetemperaturegradient
alongtheductismuchsmaller(systemisothermalitywhichallowsthemaintenanceofisothermalheat
loadinterfaces).
Forcingofthefluidthroughtheductcanbeperformedbyuseeitherof:
2.3) A pumping device, centrifugal or positive displacement pump, normally driven by an electric
motor,asinthecaseofbothsingleandtwophasefluidloopswhicharethesubjectofthisPart.
2.4)Capillaryforces,asinHeatPipes.SuchdevicesarediscussedextensivelyinECSSEHB3101Part
8.
Acapillarypumpedloop(CPL)isatwophaseheattransfersystemwhichconsistsofmultipleparallel
evaporatorsandwherethepumpingfunctionisachievedthroughafineporesizecapillarystructure.
Unlikeconventional)evaporatorswhereheatistransferredthroughaliquidlayertotheevaporating
surface, in the CPL evaporator the heat is transferred directly to the vaporization surface. These
devicespresentthesameboiling,primingandcapillarypressurelimitsassociatedwithheatpipes(see
ECSSEHB3101Part8,clause5.2).
46
ECSSEHB3101Part13A
5December2011
2.5)Shorttermdischargeofastoredfluid.Thismaybeaccomplishedbyuseofeitherof(seeECSSE
HB3101Part14,clause5.3):
2.5.1)Suddenexpansionofagasstoredinahighpressurebottle,withtheassociateddecreaseintotal
temperature(JouleThomsonprocess).)
2.5.2) Boiling of liquids such as water or cryogenic fluids. These liquids are selected on the basis of
theirhighlatentheatsand(or)theirlowboilingpoints.
2.5.3)Sublimationofsolids.
Becausethelargevolumeratioassociatedwiththeseopencycleprocesses,thefluidisventedoutof
thevehicleonceithasabsorbedtheheat.
4.1 Fluid loops

Theaimofthefluidloop,asofanyotherthermalcontrolsystem,istokeepthetemperatureofagiven
component within the range which secures its correct functioning. To this goal, the heat flow rate
evacuated should be equal to the heat rate dissipated by the component plus that coming from
externalsources.
Accordingtotheirworkingmode,fluidloopsmaybeeitherconsideredas:
1. Thermal insulators, the aim of which is to protect the component against a thermally
hostileenvironment,or
2. Thermalacceptorstransferringtheexceedingthermalenergyfromthecomponenttothe
thermalsink.
The coolant may circulate through a singlewalled chamber enclosing the component, or through a
doublewalled component heat exchanger package. The singlewalled jacket affords more efficient
heattransferthantheheatexchanger,butthefluidcancontaminate,corrodeorchemicallyreactwith
thematerialsofthecomponentwhichisbeingthermallycontrolled.
Dualloop configurations, consisting of two circulating coolant loops coupled by an interloop heat
exchanger,areusedveryoften.Adualloopconfigurationoffersamoreprecisetemperaturecontrol
thancouldasingleloop;inaddition,itprovidesmoreflexibilitytoaccommodatechangesintheheat
dissipatingelements,spacecraftthermalloads,andcoolantinlettemperaturerequirements.
Theheatrejectiondependsonwhetherthecoolantisexpendableornonexpendable.Anexpendable
coolant is rejected from the vehicle once it has accomplished its mission, while a nonexpendable
coolant is recirculatedagain after losing its thermalenergy excess; in the lastcase the excess heatis
radiatedtothespaceviaaradiator.
Singlephase fluid loops incorporating space radiators are fairly simple but massive. Twophase
systemsarelessmassiveandmoreflexibleintheaccommodationofsystemgrowthorchangingheat
loads, but require careful design to work properly under reduced gravity conditions (phase
separation,goodheattransfer,appropriatecondensationrates,...).Masssavingoftwophasesystems
are confined to pumps, fluid inventory and plumbing. Pipe diameters are also smaller than those
correspondingtoasinglephaseliquidline,providingthepossibilityofretrofittinganexistingsingle
phasesystemtotwophaseoperation(Sadunas,Lehtinen&Parish(1985)[153].
Majordisadvantagesofthetwophasepumpedloopsystemarethelackoflonglifequalifiedpumps
andpotentiallycomplexvalving,meteringandcontrolalgorithms(Wise(1986)[192]).
When cryogenic cooling below 70 K is required, the direct radiation to the space is hardly possible,
thenthefluidloopincorporatesarefrigeratingsystem(Brayton,Stirling,orVuilleumiercycles)which
47
ECSSEHB3101Part13A
5December2011
compresses the gas at ambient temperatureand then expandsit at a lower temperature. During the
expansion,heatisaddedtothegasprovidingtherequiredcooling.Thisheatisradiatedtothespace.
SeveralrefrigeratorsusedonboardsatellitesaredescribedinECSSEHB3101Part14,clause5.2.
4.2 Comparison between fluid loops and alternative

systems
In order to define the field of application of fluid loops it should be convenient to compare them
brieflywithalternativeinsulatingorrefrigeratingsystems.
4.2.1 Passive thermal insulations

Fairlysophisticated,highlyreliable,andextremelyefficientpassivethermalinsulationsareavailable
for use in spacecraft. Data in ECSSEHB3101 Part 7 are referred to for details on the subject.
Althoughthesesystemscanbeusedinconjunctionwithcomponentprecooling,theycannotinsulate
for long times components having low thermal capacity, unless requiring prohibitively thick
insulations. Among these lowthermal capacity components, mention should be made to heavy
electronicdevices,suchasbatteriesandtaperecorders,whichhaveaspecificheatoftheorderofc=
850 J.kg1.K1, common to many electronic systems (Werth et al. (1966) [189]. In addition, thermal
insulationsarenotappropriateforcontrollingcomponentswhichoperateinrepeatingonoffcycles,
sinceinthesecasesitisrequiredbothtokeepofftheheatflowingoutsideandtoevacuatetheheat
dissipatedfrominside.
In addition to being well suited to thermal protection of electronic and similar components, fluid
loops afford a very precise control of the thermal energy being evacuated, and the flexibility to be
adaptedtoapreviouslydefinedgeometricalconfiguration.
4.2.2 Thermoelectric devices

Thermoelectricdevices(TEDs),alsocalledPeltiereffectdevices,arelowvoltage,highcurrentactive
thermalcontrolmodulesthatprovideheatingorcoolingwiththereliabilityassociatedwithapurely
electricaldevice.TherelativelylowperformanceefficiencyofTEDsinconjunctionwithitshighmass
havepreventedpreviousspacecraftapplicationasaprimeactivethermalcontrolsystem.
ThecomparisonofTEDswithfluidloopswithinthenormalrangeofapplicationoftheselastdevices
makes little sense, the fluid loops being far superior. Nevertheless TEDs present several features
whichfluidloopscannotmatch.Afewtypicalvaluescouldsubstantiatetheseassertions.
1. A typical thermoelectricmodule having an area of 0,03 m x 0,03 m pumps 5 W of heat
through a 50 K temperature differential (Scott (1974), p. 216). Assuming that the
thermallycontrolledareaequalsthethermoelectricpelletarea,theheatfluxwouldbe5,5
x 103 W.m2. This figure looks, however, very biased in favor of TEDs since several
modulesareplacedabitapartfromeachotheronthesurfacetobecontrolled;avalueof
110 W.m2 isquotedas typical by Chapter & Johnsen (1973)[24]. On the other side, the
laminar singlephase flow of water through a tube of cylindrical cross section, with an
internal diameter of 5 x 103 m could remove 30 x 103 W.m2 from the tube wall whose
temperatureisassumedtobe50Kabovethatofwater,andthisfigurecanbeimproved
withturbulentsinglephaseflowandwithtwophaseflow.
2. A useful measure of the efficiency ofaTED moduleis the coefficient of performance,
ratio of the heat pumped to the power that is supplied to operate the module. This
48
ECSSEHB3101Part13A
5December2011
coefficientisoftheorderofunityforatypicalTED(Scott(1974)[161],p.221),whereasit
couldbeaslargeas400forasinglephasefluidloop(seeFigure135inthisPart).
3. ThecomparisonissomewhatmorefavorableforTEDwhenthepenaltiesassociatedwith
massandvolumearetakenintoaccount.Werthetal.(1966)[189]performedatradeoff
analysis in which they considered the cooling, either by a moving fluid or by a
thermoelectricdevice,ofacubicalcomponent,atuniformtemperature,havingasideof
0,3 m. The heat flux was of the order of 700 W.m2 for a surface temperature of 310 K.
Massandvolumeofthepowersystem,supportingstructure,bracketry,fittings,...were
neglectedinbothcases.ThestudyshowedthattheratioofTEDcontrolledtofluidloop
controlledrequiredmasswas1,5,andtheratioofrequiredvolumes2,7.
StudiestoreduceTEDmassandtoformulatelargescalemanufacturingprocedurestopermitdirect
applicationofTEDstospacecraftthermalcontrolhavebeensponsoredbyNASA.Thesestudieswere
concentratedonthedevelopmentofthinfilmcoatingsthatutilizevacuumdepositedsemiconductor
materials.
TEDsontheirpartpresentthefollowingadvantages:
1. Thermoelectric devices operate on electronic principles. They require no pump, valves,
piping,coolantsandthefixturesassociatedwithfluidloops.
2. Thermoelectricdevicesprovidevariableheatpumpingbyvaryingtheinputcurrent.This
allowsforprogrammableinternaltemperaturesthroughcurrentmodulation.
Needlesstosaythattheheattransferratecanbeincreasedinafluidloopbyincreasing
thefluidmassflowrate,butthisrequireseitheravariablespeedpumporanelectrically
controlledthrottle.
3. Heating or cooling can be provided by reversing the current. The fluid loop does not
exhibitasimilarfeature.
4. A given heat transfer rate can be pumped no matter how small the source to sink
temperaturedifferencecouldbe,whereasinsinglephasefluidloopsitisnotfeasibleto
reduce the source to sink temperature difference, for a given heat removal rate, by
increasingthecoolantmassflowratebeyondacertainvalue.Twophasefluidloopsdo
notpresentthislimitationtothesameextent.
Summarizing, under present state of development TEDs are only suited for heat conditioning small
componentswhichrequireveryclosetemperaturecontrol.
4.2.3 Phase change materials (pcm)

Although PCM systems are normally used to absorb the abnormal heat dissipation peaks of an
equipmentwhichtransferssomehowtheheatthatitdissipatestoasink,theyareusedsometimesas
thesoleheatsinkforachievingthethermalcontrolofsolitaryequipment.SeeECSSEHB3101Part
10.
Werthetal.(1966)[189]showedthattherequiredmassofthePCMcell(radiatorexcluded)isofthe
order of twice the mass of the cooling jacket of the) fluid circulation cooling system evacuating the
same heat flow rate. This indicates that the PCM system is competitive, on a mass basis, when it is
used to protect components which are so small that the mass of the cooling jacket becomes of the
orderofthemassofthepump,valvesandpipingassociatedwiththissystem.
Substantial gains in system mass, through radiator area savings can be achieved by incorporating
thermal capacitors in the loop. This occurs in systems which are subject to alternant heating and
49
ECSSEHB3101Part13A
5December2011
cooling regimes or when the radiator area grows to cope with s degradation (Lehtinen & Sadunas
(1985)[114].
4.2.4 Heat pipes

HeatpipesareconsideredinECSSEHB3101Part8.
Theconstantconductanceheatpipeisaverypowerfulheattransferdevice,whichisbasicallysimple
althoughcomplextomanufactureandof)limitedreliability.
Comparedwithconstantconductanceheatpipes,whichnormallydonotrequireauxiliarypower,and
which are best suited for long mission on relatively simple satellites, the fluid loops presents the
followingadvantages:
1. The heat flow rate can be controlled through the fluid flow rate, by use of appropriate
temperaturecontrolledvalvesashasbeenalreadysaid.
2. The loop can be adapted to fairly complicated and relatively large configurations,
although this advantage may be at the expense of increased power requirements. The
idea of matching fluid loops and heat pipe radiators to decrease loop tubing and
associated power requirements has been often set forth. However, heat pipe radiators
require an additional heat exchanger to produce a temperature jump across the heat
exchangerradiatorinterface,resultinginlargersystemmass.
3. Different temperature levels, suited to each individual component, can be achieved by
using regenerative heat exchangers. These are devices in which the heat exchanging
surfacescomeincontactsuccessivelywithfluidmediaatdifferenttemperatures.
The noncondensable gas generated in a constant conductance heat pipe represents a problem; in a
variableconductanceheatpipe(VCHP),however,itispurposefullyintroduced(seeECSSEHB3101
Part 8, clause 7) to provide a variable blockage of the condenser region and then to varythe
conductanceasafunctionoftheheatload.
As the required heat transfer rates increase and remote and (or) deployable radiators are being
considered, the limits of current heat pipes are exceeded. Pumpassisted heat pipes are being
consideredatpresent.
4.2.5 Short-term discharge systems

Opencycledischargesystemsarenormallyusedtoabsorbshorttermpeakthermalloads.Although
thesedeviceslooksimple,afairlycomplexshieldingsystemisrequiredtoavoiddirectimpingingof
thecoolantonthecomponent.Whencryogenicboilingliquidsareused,thisdirectimpingingresults
inveryextremelocalcoolingeffects.
In opencycle highpressure gas systems, highpressure gas provides the necessary cooling. When a
realgasisexpandedatconstantenthalpy(JouleThomsonorthrottlingprocess)thetotaltemperature
of the gas decreases when the pressure decreases, provided that gas temperature is below a certain
inversion temperature. This effect results in cooling. Notice that under normal pressure and
temperature conditions the gas behaves as a perfect gas and there is no total temperature change
during throttling, thence no cooling effect would appear. Details concerning JouleThomson open
refrigerators,aswellasotherdischargesystems,canbefoundinECSSEHB3101Part14,clause5.3.
Discharge systems are uneconomic under orbital operation. In addition, the propulsive effect of the
ejectedfluidonthespacecraftflighttrajectoryshouldnotbeoverlooked.Forexample,Nast,Barnes&
Wedel(1976)[131]reportthetemporaryupsettingoftheNimbusFspacecraftbecauseoftheventing
50
ECSSEHB3101Part13A
5December2011
ofcoolingmethaneandthesubsequentimpingementoftheexhaustplumeonthespacecraft.Inspite
ofthesedrawbacks,dischargesystemscouldbeusefulinsomecases,forinstance,thermalprotection
duringreentryofreusablespacevehiclescanbeimprovedbymeansofexpendablecoolantsystems
(Helenbrook,Anthony&Fisher(1971)[84].
51
ECSSEHB3101Part13A
5December2011
5
Analysis of a fluid loop
5.1 General
Thepurposeofthisclauseistodescribethewayinwhichafluidloopisanalyzedandtoshowhow
thedatapresentedinthefollowingclausesmaybeusedtothisend.
Averysimplefluidloopisconsideredforconvenience.Itconsistsofacircuitformedbyfluidcarrying
tubeswhichconnectaheatsinktoaheatsource.Acirculationpumpmovesthefluidthroughtheloop
so that heat is transported from source to sink by forced convection. Figure 51 is a schematic
representationofthefluidloop.
Figure51:Schematicrepresentationofthefluidloop.
The heat source could be an energy dissipating component such as electronic equipment, an
experimental module involving exothermic reactions or any other heatproducing device. In some
applications the heat source element is mounted on and thermally linked to a could plate, and the
coolantflowsthroughthelatter,inwhichcasethecoldplatemaybeconsideredastheheatsource.In
the following analysis the heat source is simply identified by its heat production rate, Q, and no
considerationisgiventothedetailsoftheheattransfermechanismfromsourcetocoolantfluid.
Theheatsinkcanbeaheatexchanger,inwhichcaseheatistransferredtoacoldfluidcirculatingin
anotherloop(dualloopconfiguration),oritmaybearadiatorwhichrejectstheheatdirectlytodeep
space. In the present example transfer to a second loop through an interface heat exchanger is
considered.
The procedure presented in this clause to analyze the fluid loop in Figure 51 may be easily
generalizedtoothersystems.
Thefollowingdataareassumedtobeknown:
1. General layout of the fluid loop, including heat exchanger surface geometry and
dimensions.
2. Heatflowratetobeevacuated,Q.
52
ECSSEHB3101Part13A
5December2011
3. Massflowrateineachloop:mh,mc.
4. Inlettemperatureofthecoldfluid,Tci.
From these data, the temperatures throughout the loop and the pumping power requirements are
computed. The mass flow rate through the cold fluid loop is often determined by means of some
optimizationprocedure(seeClause13).Inthepresentanalysis,however,onlytheperformanceofthe
fluidloopforagivenmassflowrateisconsidered.
5.2 Thermal performance

Assuming,1)thermallyinsulatedtubesforthefluidtransportbetweenheatsourceandheatsink,2)
notemperatureincreaseinthecirculationpump,and3)notemperatureincreaseduetofriction,the
temperatureattheentranceofthesink,Thi,equalsthetemperatureattheexitofthesource,andthe
temperatureattheexitofthesink,Tho,equalsthetemperatureattheentranceofthesource(seeFigure
51).
To calculate the temperature Thi, Tho, and Tco, the following equations for the heat flow rate may be
used:
Q sC1 Thi Tci [51]
Q C h Thi Tho [52]
Q C c Tco Tci [53]
whereC1isthesmallestofChandCc.
Eq.[51]describestheheattransferratefromhotfluidtocoldfluidintheheatexchanger,intermsof
theheattransfereffectiveness, (seeEq.[54]ofclause11.2.2whereaslightlydifferentnomenclature
hasbeenusedforconvenience).Eqs.[52]and[53]aresimplyanenergybalanceforthehotandcold
fluid,respectively.FromEqs.[51]to[53],thetemperaturethroughtheloopareobtained.
Q
Tco Tci [54]
Cc
Q
Thi Tci [55]
sc1
1 1
Tho Tci Q [56]
sc1 ch
Iftheheatsourceisacouldplate,thereisanadditionalequationtodescribetheheattransferprocess
attheplate,namely
53
ECSSEHB3101Part13A
5December2011
Q scp ch TS Tho [57]
wherecpistheeffectivenessoftheheattransfersurfaceattheheatsource,andTSisthetemperatureof
thesource.Eq.[57]determinesthetemperatureatthesource.
Summarizing, to compute the temperatures in the loop, the heat transfer effectiveness of the heat
exchangermustbecalculated.Thefollowingstepsareneededtothisend:
1. Compute heat transfer areas, free flow areas and frontal areas, for the hot side and the
coldsideoftheexchanger.Todescribeaheattransfersurfacethefollowingparameters
areoftenused:
A, Totalheattransferareaononeside.[m2].
AFL, Freeflowareaononeside.[m2].
AFR, Frontalareaononeside.[m2].
DE, Hydraulicdiameter.[m].
L, Flowlengthononeside.[m].
, Ratiooftotalheattransferareaononesideofaclosed
typeexchangerorrecuperatortototalexchanger
volume.[m2.m3].Thisratioiscalledareadensity.
, Ratiooffreeflowareatofrontalareaofonesideof
exchanger.Thisratioiscalledporosity.
The following equations relate surface factors with core factors, for each side of the
exchanger,andmaybeusedtocalculateheattransferareasandfreeflowareas,
AFL
DE 4 L [58]
A
AFL DE
[59]
AFR 4
2. Compute fluid properties. The fluid properties are evaluated at an arithmetic mean
temperatureasfollows,
Tci Tco Thi Tho

Tc , Th [510]
2 2
54
ECSSEHB3101Part13A
5December2011
whereTcoandThiaregivenbyEqs.[54]and[55]respectively,andThomaybecomputed
from Eq. [56] assuming a heat transfer effectiveness = 0,8. This estimation has to be
verifiedaposteriori.
Seeclause6.2.4forfurtherdetailsontheevaluationoftheaveragetemperature.
The following fluid properties are needed: density, , specific heat, cp, thermal
conductivity,k,dynamicviscosity,,andPrandtlnumber,Pr=cp/k.Table101givesthe
valuesofthesepropertiesforrelevantliquidcoolants.
3. ComputeReynoldsnumber.TheReynoldsnumberisgivenby
mDE
Re [511]
AFL
Itiscomputedforeachsideoftheexchanger.
4. ComputetheNusseltnumberNuandtheconvectiveheattransfercoefficient,h.
ThethermalanalysisdevelopedinClause6presentsdatatocomputetheNusseltnumber
as a function of the Reynolds number, Prandtl number and geometry (Figure 61 to
Figure638),formostbasicconfigurations.Thesefiguresmaybealsousedtoestimatethe
Nusseltnumberinmorecomplicatedgeometries.
If the heat transfer surface is one of those presented in clause11.3, the Nusselt number
may be obtained from figures given by Kays& London (1964) [102], which relate the
Stantonnumber.
Nu
St [512]
PrRe
totheReynoldsnumber,formanydifferentheatexchangersurfaces.
Clause 9 presents data to estimate the Nusselt number, when turbulence promoters or
othertypeofheattransferenhancementdevicesareused.
Theconvectiveheattransfercoefficient,h,isgivenby
kNu
h [513]
DE
or
C
h St [514]
AFL
5. Compute the temperature effectiveness of the surface due to fins, . The temperature
effectivenessisgivenbyEq.[53]ofclause11.2.2,wherefineffectivenesss, f,fortypical
finsaregiveninFigure117andFigure118.
6. Computetheoverallthermalconductance,U,foreachsideoftheexchanger.
55
ECSSEHB3101Part13A
5December2011
ThiscoefficientisgivenbyEq.[52]andofclause11.2.2,intermsofboththehotandthe
coldsideoftheexchanger.Whenfoulinghastobeconsidered,theaboveequationshould
bereplacedbytheequationwhichappearsinclause11.6.2.
7. Computethenumberofheattransferunits,Nt,asdefinedinEq.[55]ofclause11.2.2.
8. Computetheexchangereffectiveness,.
Theheattransfereffectiveness, ,oftheexchangerispresentedinClause11asafunction
ofthenumberofheattransferunits,Ntu,capacityrateratio,R,andflowarrangement.The
results are given in graphical form, for many different arrangements, in Figure 119 to
Figure1128.Thesefiguresmaybealsousedtoestimatetheheattransfereffectivenessfor
manyflowarrangementsnotexplicitlyconsideredinClause11.
Insomeinstances,particularlyforheatexchangersofveryhighheattransfereffectiveness,theeffects
discussedinclause11.4shouldbetakenintoaccount.
If the resulting effectiveness is very different from the one used in point (2) to compute the fluid
properties,thestepsjustdescribedshouldberepeated.
Once the heat transfer effectiveness is known, Eqs. [54] to [56] yield the temperature through the
loop.
5.3 Power requirements

Thepower,P,whichissuppliedtothecirculationpumpcanbeexpressedas
1 m
P p [515]
p
where pistheoverallefficiencyofthepump, pthepressurelossthroughtheloop,mthefluidmass

flowrateand thefluiddensityatthelocationofthepump.m/isthenthefluidvolumeflowrate
throughthepump.EquationsandgraphicstoevaluatethepressurelossaregiveninClause7.Values
ofpfortypicalspacebornepumpscanbefoundinClauses12and9.Generally,thepressurelossmay
besplitintothefollowingterms:
1. pressurelossinthesupplyplumbing;
2. pressurelossinthereturnplumbing;
3. pressurelossintheheatsource,and
4. pressurelossintheheatsink.
Tocomputetheselosses,thefollowingstepsareneeded.
1. ComputetheReynoldsnumberinthesupplyandreturnplumbing,
4 m
Re [516]
DE
56
ECSSEHB3101Part13A
5December2011
where DE is the diameter of the tube and the fluid viscosity. Since the viscosity is
temperaturedependenttheReynoldsnumberisdifferentoneachlegoftheplumbing.
2. Compute the friction factor for the straight parts of the ducts, from Figure 72 and
Figure 73 in Clause 7. The friction factor is a function of the Reynolds number and
shouldbecomputedforthesupplyandforthereturnplumbing.
3. Computepressurelossduetofrictionalongthetubewalls,
L 1
p V 2 [517]
DE 2
4. Computepressurelossforeachoneofthebendsofthecircuit,
1
p C k V 2 [518]
2
whereCkcanbeobtainedbyuseofFigure74toFigure714inClause7.
5. Computepressurelossinvalves,manifolds,entrances,etc.byusingFigure715toFigure
727 or the data submitted by the manufacturer of the particular component. Data for
entranceeffectscanbefoundinclause7.2.5.
6. Computepressurelossinheatsourceandheatsink.Thispressurelossmaybeestimated
fromFigure72toFigure730inClause7.Iftheheatsourceorsinkisaheatexchanger
with a coresurface geometry like the ones presented in Clause 11.3, the friction factor
maybeobtainedfromfiguresgivenbyKays&London(1964)[102].
57
ECSSEHB3101Part13A
5December2011
6
Thermal analysis
6.1 General
Dataconcerningtheconvectiveheattransferbetweenafluidandthetubewallsurfacearepresented
inthisclause.Thesedatacanbeused,togetherwiththeheatexchangertheorypresentedinclause11.2
fortheanalysisofheatexchangers.
The data are divided into two main parts. The first part concerns internal flows, i.e., configurations
where the fluid flows inside the domain bounded by the heat transfer surface. In the second part
external flows are considered, i.e., configurations where the fluid flows outside of the domain
boundedbythesurface.
Dataforbothlaminarflowandturbulentflowarepresented.
Concerning the heat flux distribution, two cases are considered, constant wall temperature and
constantheattransferrateperunitlength.Thecaseofconstantheattransferratecanbeeasilydealt
with theoretically, nevertheless most experiments are performed under nearly constant wall
temperature.Inpracticalinstancestheconstantwalltemperatureboundaryconditionisapplicablein
such heat exchangers as evaporators, condensers and in any heat exchanger where one fluid has a
verymuchhighercapacityratethattheother.Theconstantheattransferrateboundaryconditionis
applicable in counter flow heat exchangers, when the fluid capacity rates are similar. These two
boundaryconditionscovertheusualextremesmetinheatexchangerdesign.
The data forconstant heat flux may be applied directly to other boundary conditions provided that
the real variation of the heat transfer rate, Q(x), is smooth enough or, more precisely, that the
followinginequality
dQ Q
[61]
dx D
issatisfied.Additionalcommentonthispointaregiveninclause9.1.4.
6.2 Analytical background
6.2.1 Heat transfer coefficient

Theconvectivelocalheattransferbetweenthetubewallandafluidmovingrelativetothewallcanbe
calculated in term of the temperature difference, TwTb, and a local heat transfer coefficient, hx, by
meansofthefollowingexpression:
58
ECSSEHB3101Part13A
5December2011
q x hx Tw Tb x [62]
Usuallythewalltemperature,Tw,andthelocalheattransfercoefficientareuniformateachsection,x,
ofthetube,whilethefluidbulktemperatureisgivenbytheexpression:
Qx
Tb Ti [63]
mc p
Thedefinitionofthebulktemperatureintermsofthelocalpropertiesisgiveninclause6.2.4.
Sufficiently far from the tube entrance, the fluid velocity and temperature profiles for conditions of
constant wall temperature and of constant heat rate become selfsimilar. The non dimensional
temperatureprofilesareinvariantwiththeaxialdistancex,andthereforetheheattransfercoefficient
hxbecomesindependentofx.
Theheattransferuptothestationxisevaluatedbyusingthemeanheattransfercoefficienthm
Q x hm Tm xD [64]
wherehmisdefinedbythefollowingaveragingprocess:
x
hm Tm x hx Tw Tb x dx [65]
0
Three terms of the coefficient hm may be used, according to the following three definitions of
temperaturedifference,Tm:
1. Inletmeantemperaturedifference:
Tim Tw Tb i [66]
2. Arithmeticmeantemperaturedifference:
Tw Tb i Tw Tb o
Tam [67]
2
3. Logarithmicmeantemperaturedifference:
59
ECSSEHB3101Part13A
5December2011
Tw Tb i Tw Tb o
Tlm
Tw Tb i
[68]
ln
Tw Tb o
6.2.2 Dimensionless groups

Itiscustomarytointroducethefollowingdimensionlessparameters,whicharedefinedeitherlocally
orthroughsomeaveragingprocess.
1. Nusseltnumber
hDE
Nu [69]
k
The Nusselt number indicates what multiple of mere thermal conduction is transferred
byconvection.Itisusuallydefinedintermsoftheequivalentorhydraulicdiameter,DE.
2. Prandtlnumber
c p
Pr [610]
k
whichmeasurestheratiooffluidviscoustothermaldiffusivities.
3. Reynoldsnumber
VDE
Re [611]

givingtheratioofinertiatoviscousforces.
SometimestheheattransferdependsontheproductPrReandtheGraetznumberisused
topresenttheavailabledata.
DE
Gz PrRe [612]
4 L
FromtheNusselt,PrandtlandReynoldsnumberstheStantonnumberisdeduced.
Nu
St [613]
PrRe
60
ECSSEHB3101Part13A
5December2011
4. Grashofnumber
gDE3 2 Tw Tb
Gr [614]
2
TheGrashofnumberexpressestheratioofbuoyancytoviscousforces.
A dimensionalanalysis indicates that the heat transfer coefficient may be related to the
fluidpropertiesandflowparametersinthefollowingway.
Nu f Pr , Re, Gr , x / DE [615]
6.2.3 Simplifying assumptions

Thefollowingsimplifyingassumptionsareintroduced:
1. Buoyancycanbeneglected,especiallyduringorbitalflight,everforverylowfluidmass
flowrates;Gr=0.Nevertheless,buoyancycanbeintroducedintentionally,byswirlingof
theflow,inordertoenhancetheconvectiveheattransfer(seeclause9.1.1andFigure98
toFigure911).
2. Fluid properties are evaluated at some reference temperature. The resulting values are
assumedtoprevailthroughthefluidbulk.Thisassumptionisdiscussedinclause6.2.4.
3. The effect of surface roughness on heat transfer is neglected. Sometimes surface
roughness is created intentionally to enhance the heat transfer (see clause 9.1.1 and
Figure91toFigure99).
4. Thefluidconfigurationissteady.
6.2.4 Temperature-dependence of fluid properties

Theheattransferdatapresentedinthisclausearebasedontheassumptionthatthefluidproperties,
whichareevaluatedatareferencetemperature,aretemperatureindependent.Inapplicationswhere
temperaturedifferencesbetweenthefluidandthesurfacearelarge,thevariationsoffluidviscosity,
thermalconductivityanddensityoughttobetakenintoaccount.
The variation of fluid properties differs substantially for gases and liquids. For gases, thermal
conductivity,viscosity,anddensityallvarywithtemperaturetoaconsiderabledegree.Forliquids,on
theotherhand,theonlypropertythatishighlytemperaturedependentisviscosity,whichvarieswith
temperaturemuchmorethananyofthegasproperties.
Temperaturedependenceeffectsataparticularflowcrosssectioncanusuallybecompensatedforby
evaluatingcertainorallofthepropertiesatthefluidbulktemperature,Tb,forthiscrosssection,
uTdA FL
[616]
Tb
AFL

VAFL
61
ECSSEHB3101Part13A
5December2011
whereuandTarethelocalvelocityandlocaltemperature,AFListhesurfaceareaofthecrosssection,
andVthemeanfluidvelocity.Tbissometimestermedthemixingcuptemperature.
In many cases the mean film temperature, the arithmetic average of Tw and Tb, is used as reference
temperature.
Therearecasesinwhichtheuseofareferencetemperatureisinadequate,forexamplewhenthewall
tobulk temperature ratio differs greatly from one. The problem is then approached by introducing
into the transfer law a loading factor of the form (Tw/TR)m or (w/R)m, where TR is the reference
temperatureindependentlyofthewayinwhichithasbeendefined.
Itisnottobeexpectedthatthecorrectionhasthesameformforgasesandliquids,norforfluidsofthe
same phase but widely differing Prandtl numbers, nor for different transferred entities (heat or
momentum),norforlaminarandturbulentflow.
Forgasesatwalltobulktemperatureratiosintherange0,5<Tw/Tb<2,theeffectontheheattransfer
coefficientofradialpropertyvariationsdoesnotexceed10percent(ESDU68006(1968)[48]).Values
oftheexponentmforawiderrangeofTw/Tb,forseveralphysicalsituations,andforbothheattransfer
andfriction,canbefoundinKays&London(1964)[102],p.88.
Inthecaseofliquids,theviscosityisnormallyastronglydecreasingfunctionoftemperature.When
the walltobulk temperature difference is large, the fluid velocity near the wall departs from that
calculatedassuminguniformviscosity.Thisdistortionofthevelocityprofilechangestheheattransfer
coefficient.Experimentaldataindicate(ESDU68006(1968)[48])thatthiseffectcanbeaccountedfor
bymeansofthefollowingformula
0 ,14
Nu b
[617]
Nub w
in the range 0,004 < b/w < 20. This expression, extensively quoted, was suggested by Deissler for
laminarflowincircularducts.Exponentsforturbulentflow,deducedafterexperimentsbyDeissler,
aregiveninKays&London(1964)[102],p.90.InallthesecasesthereferencetemperatureisTb.
Inmanyapplications,temperaturevariationsalongthetubearemorepronouncedthanthoseoverthe
cross section, and a mean temperature with respect to the tube length ought to be used to evaluate
fluidproperties.ThisreferencetemperatureisidentifiedasTbm.
TheevaluationofTbmisnotaneasytask,andthedesignershouldexercisehisjudgmentsforavoiding
fairlytedioustrialanderrorprocedures.Tbmdependsonthevariationsofheatfluxandtemperature
differentialalongthetube.Severaltypicalcasesareconsidered:
1. Forconstantheatflux,thewalltobulktemperaturedifferenceremainsnearlyconstant,
and since the wall temperature usually varies linearly, the mean fluid temperature is
merelythearithmeticaverageofinflowandoutflowtemperatures.
Tbi Tbo
Tbm Tbam [618]
2
62
ECSSEHB3101Part13A
5December2011
2. Constantwalltemperature.Thefluidtemperaturevariationalongthetubewillbeclose
to exponential, provided that the heat transfer coefficient does not vary markedly with
tubelength.
Tbo Tbi
Tbm Tw
4 L / De St [619]
whereStisestimatedbeforehand.
3. When the heat transfer coefficient decreases with tube length, as occurs with laminar
flow, the true mean temperature with respect to tube length is close to the fluid exit
temperature, since the large heat transfer rates upstream are responsible for the
comparativelysharpchangesinthefluidbulktemperaturenearthetubeentrance.
6.2.5 Laminar versus turbulent fluid flow

ThecharacteristicsofthefluidflowinapipedependontheReynoldsnumber.
ThemotionislaminarforsufficientlylowvaluesoftheReynoldsnumber,becomingturbulentwhen
sometransitionvalueisexceeded.
ForfluidsflowingthroughpipesthetransitionReynoldsnumberisnormallyoftheorderof2400.In
therange2000<Re<2800thefluidflowischaracterizedbyanintermittentturbulencewhichtakesthe
form of alternating slugs of laminar or turbulent fluid moving down the pipe. For Re 4000 the
turbulentregimeisfullydeveloped.
Thedisorderassociatedwiththeturbulenceimprovestheheattransportprocessatthecostofalarger
energydissipation.
Severalconsiderationsconcerninglaminarversusturbulentfluidflowshouldbemade.
1. Normally laminar flow should be preferred because the power requirements are lower
than in the turbulent case. Nevertheless, the low velocities required for laminar flow,
together with low thermal conductivity of most fluids (particularly gases) result in low
heattransferratesperunitsurfacearea.Thus,largesurfaceareasarerequired.
2. Lowvelocityflowaremoresensitivetocontrolthanthehighvelocityones.
3. Transitionalflowwiththeassociatedintermittencyshouldbe,asageneralrule,avoided.
Incompactheatexchangerapplications,however,theReynoldsnumberrangeofinterest
isusuallyabout500to15000;thusthetransitionregioncanhardlybeavoided.
4. IftheReynoldsnumbermustbedecreasedpreservingtheflowrate,theinnerdiameterof
thetubemustbeincreased,andconversely.
6.2.6 Heat transfer to internal flows

Data on convective heat transfer to internal flows are presented in Clause 6.3 (Figure 61 to Figure
627).ThedataaregivenintermsoftheNusseltnumber,Nu,eitherlocal,averagedorboth.Theheat
transfer coefficient, h, is deduced from the value of Nu, the fluid thermal conductivity, k, and the
hydraulicdiameteroftheduct,DE.
63
ECSSEHB3101Part13A
5December2011
Thesequenceoffiguresisarrangedasfollows:
1. Figure61toFigure612concernlaminarflow(clause6.3.1.1),andFigure614toFigure
627turbulentflow(clause6.3.1.3).ThepredictionofNusseltnumbersinthetransitional
flow regimen (2000 < Re < 4000) is difficult and thus the available information is fairly
limitedinscope(clause6.3.1.2).Figure613,whichshowsatypicalrangeoflocalNusselt
numbers for transitional gas flow, is intended to indicate that the Nusselt numbers are
somewherebetweenthelaminarandturbulentvalues.
2. Withineachflowregimethefirstfewfiguresapplytotheregionfarremovedfromthe
entrancetothetube,wherebothafullydevelopedvelocityprofileandafullydeveloped
temperatureprofileexist.ThisisthecaseofFigure61toFigure65forlaminarflow,and
Figure614toFigure622forturbulentflow.
3. Thenextgroupoffiguresineitherthelaminarorturbulentregimeapplytothecasein
which the velocity profile is fully developed and remains fixed, while the temperature
profiledevelops.Intheseconfigurationsthefluidtemperatureupstreamofsomesection
isassumedtobeuniformandequaltothewallsurfacetemperature,therebeingnoheat
transferinthisregion.Followingthementionedsectionheattransfertakesplace.These
thermalentrylength solutions only apply rigorously when a hydrodynamic starting
length is provided so that the velocity profile is fully developed before heat transfer
starts, a condition rarely encountered in practice. However, when the temperature
changeisconcentratednearthewall(Pr>>1),thetemperaturegradientatthewallreacts
quickly to changing conditions, and these solutions are excellent approximations to the
realcases.
4. Finally, Figure 610 to Figure 612 apply to the case of combined hydrodynamic and
thermal entry length; thats, when both the velocity and the temperature profile are
uniformattheentranceofthetube.Therelevantvariableforthesethermalentrylength
solutionsisthenondimensionalaxialdistance
x/r
x [620]
PrRe 2Gz
whichisrelatedtotheinverseoftheGraetznumber,Gz.
Mostofthedatarefertocylindricaltubesofcircularcrosssection.Forturbulentflowthesedatacanbe
appliedtoothercrosssectionalshapesprovidedthattheequivalentorhydraulicdiameter,DE,isused
ascharacteristiclength.
Exceptionstothisruleconcerntheheattransfertoafluidflowingthroughacylinderofannularcross
section or between two parallel plates (which is a limiting case of annulus). These are the simplest
geometrical configurations allowing different heating rates from different walls of the same cross
sectionoftheduct.Onlytwofundamentalsolutionssufficetoobtainthesolutionforanycombination
of heat fluxes through these walls. The mentioned fundamental solutions are: a) inner wall heated
withouterinsulated(subscript11),b)outerwallheatedwithinnerinsulated(subscript22).Theheat
transfercoefficientsontheinnerandoutertubes,h1andh2,aredefinedasfollows:
q1 h1 T1 Tb , q2 h2 T2 Tb [621]
64
ECSSEHB3101Part13A
5December2011
andtheNusseltnumbersas:
h1 DE h2 DE
Nu1 , Nu 2 [622]
k k
These Nusselt numbers can be calculated from the fundamental solutions for any heat flux ratio,
throughtherelationships
Nu11 Nu 22
Nu1 , Nu 2
1 q2 / q1 Z1 1 q1 / q2 Z 2
[623]
where Z1 and Z2 are the influence coefficients which may be evaluated from the fundamental
solutions. The values of Nu11, Nu22, Z1, Z2 are presented in graphical form in Figure 63, Figure 64,
Figure 68 and Figure 618 to Figure 621 for annuli. Similar data, for parallel plates, are given in
Figure66,Figure67,Figure612,Figure616,Figure617andFigure627.
Curved ducts are widely used in heat transfer equipment. In these ducts the heat transfer (and the
friction) isaugmented by secondary flows which appear because of the fact that the radial pressure
gradient imposed by thehigh velocity core is not balanced by the local centripetalacceleration. The
effectofthesecondaryflowisaccountedforthroughanadditionaldimensionlessparameter,theDean
number,K,whichisdefinedas:
r
K Re [624]
R
whereristheradiusofthetubecrosssectionandRtheradiusofthecoil.
Itshouldbesaidthattheonsetofturbulenceincoiledpipesisdelayedowingtothesuppressionof
transversemotionsoffluidelementsbycentrifugalforces.Kubair&Kuloor(1966)[111]proposethe
following formula to predict the critical Reynolds number for transition from laminar to turbulent
flow,
2
r
Re 1,273 10 4 [625]
R
Figure65,Figure69andFigure622presentdataontheheattransfertofluidmovingthroughtube
coils.
6.2.7 Heat transfer to external flows

In many configurations of practical interest the heated or cooled surface is submerged in the fluid
insteadofboundingit.
65
ECSSEHB3101Part13A
5December2011
The convective heat transfer rate between the external surface of a cylinder and a fluid moving
relativetoitcanbecalculatedintermsofthetemperaturedifference,TwTb,andameanheattransfer
coefficient,h,byusingthefollowingexpression:
Q hDLTwm Tb [626]
whereDisacharacteristiclengthofthecylindercrosssection,Listhecylinderlength,Tbisthebulk
temperature corresponding to the undisturbed fluid, and Twm is a mean temperature of the outer
surfaceofthecylinder.AperimetermeanvalueofTwisnormallyused,althoughdifferentdefinitions
of the average value may be considered without substantially influencing the experimental data on
Nusseltnumbers.
Adimensionalanalysisindicatesthattheheattransfercoefficientcanberelatedtothefluidproperties,
cylindergeometryandflowparametersinthefollowingway:
Nu f Pr , Re, Gr , L / D, [627]
beingtheinclinationofthecylinderaxistotheincomingflow.
TheinfluenceofGrashofnumbercanbeneglectedasabove,whereasendeffectsareinsignificantfor
L/Dgreaterthanabout3.
Figure628andFigure629(clause6.3.1.2)givedataforaconstanttemperaturecircularcylinder.
The problem is far more complicated in the case of tube banks. The convective heat transfer rate
between a fluid and a tube bank constituted by n circular cylinders of length L and diameter D is
givenby
Q nh DLTwm Tb [628]
where h isthemeanheattransfercoefficientfortheparticulartubebankunderconsideration.
Adimensionalanalysisshowsthat,whenfreeconvectioncanbeignored,themeanNusseltnumber,
Nu ,ofatubebankisrelatedtotheflowparameters,fluidpropertiesandgeometrybytheexpression
Nu f Reb , Prb , Prw , st / D, sl / D, N , L / D, [629]
whereinthisinstance
bVmax D
Re b [630]
b
and Vmax is based on the minimum free area available for fluid flow, regardless of whether this
minimumareaoccursinthetransverseordiagonalopenings.
66
ECSSEHB3101Part13A
5December2011
Itiscustomary,ESDU73031(1973)[57],tobaseheattransferdataonaNusseltnumber,Nur,which
correspondstothefollowingreferenceconditions:
1. Constantfluidproperties.
2. Numberofrowsinthetubebank,N=10withatleast6tubesineachrow.
3. Cylinderaxisnormaltoflow,=90.
4. L/D>5.
Reference Nusselt numbers, Nur, vs. Reynolds number, Reb, are shown in clause 6.3.2.2. Figure 631
andFigure 632 areforinline tube banks,andFigure 633forstaggered tubebanks. Thefluid bulk
Prandtlnumber,Prb,isassumedtobeequaltooneinallthesethreefigures.CorrectionsforPrandtl
numbereffectscanbededucedfromFigure634.
OnceNurisknown,theMeanNusseltnumber, Nu ,isexpressedinthesimplifiedform,
Nu Nu r F1 F2 F3 [631]
wherethefactorF1takesintoaccountthevariationofthefluidproperties,F2isacorrectionfactorfora
bankhavingN10,andF3correctsfortheyawangleofthebank.Thesethreecorrectionfactorsare
giveninFigure635,Figure636andFigure637,respectively.
Tocalculatetheheattransferforaparticularrow(thenth)anewfactor,F4,relatingtheperformanceof
thatrowtotheaverageforthe10rowsbankisused(Figure638).
Nu n Nu r F1 F3 F4 [632]
Heat transfer data for other exchanger configurations can be found in Kays & London (1964) [102].
GeometricaldescriptionsofthemanyconfigurationstestedbytheseauthorsaregiveninClause11.3.
Each group of figures giving the data for both internal and external flows is preceded by tables
allowing a quick assessment of the content of each figure. The table in the next page is for internal
flows.Thetableforexternalflowsisplacedimmediatelybeforethecorrespondingfigures.
67
ECSSEHB3101Part13A
5December2011
6.3 Thermal performance data
6.3.1 Heat transfer to internal flow

Thefollowingsummarizesthedatapresentedinthefollowingpages.
Thedataarecategorizedaccordingto:flowcharacteristics,heatfluxdistribution,andductgeometry.
Thelabelsusedtodistinguishthesecategoriesarethefollowing:
Typeofflow:L,laminar.T,turbulent.IT,transitional.
Degree of flow development: FL, fully developed flow. TEL, thermal entry length. HTEL,
combinedhydrodynamicalandthermalentrylength.
Heatfluxdistribution:q,constantheattransferrate.T,constantwalltemperature.

Type Flow Heatflux Geometry InputVariables Output Fig. References
of devel. distribution
flow
L FL q&T Several Geometry Nuq,NuT Figure Kays&
61 London(1964)
[102]
Straighttubeof a/b Nuq,NuT Figure
rectangularcross 62
section
q straighttubeof r1/r2 Nu11,Nu22 Figure
annularcross 63
section
Z1,Z2 Figure
64
Coil K,Pr=0,1,0,2,0,7,1,4, Nu Figure Morietal.
10 65 (1965)[128]
TEL T Circularcylinder, x+ Nux,Nu1m Figure Kays(1966)
parallelplates, 66 [101]
straighttubeof
rectangularcross
section
q Circularcylinder, Nux,Z Figure
parallelplates 67
Straighttubeof x+,r1/r2=0,1,0,5 Nu11,Nu12, Figure
annularcross Z1,Z2 68
section
T Coil x+,r/R=0,025,0,1 Nux Figure Kubairetal.
69 (1966)[111]
HTEL Circularcylinder x+ Nux,Nu1m Figure ESDU68006
610 (1968)[48]
q x+,Pr=0,01,0,7,10 Nux Figure Heatonetal.
611 (1964)[82]
68
ECSSEHB3101Part13A
5December2011
Type Flow Heatflux Geometry InputVariables Output Fig. References

of devel. distribution
flow
Parallelplates Nux,Nuz Figure
612
IT FL T Circularcylinder Re,Pr=0,7 Nux Figure ESDU68006
613 (1968)[48]
T FL q Re,Pr=0,7,1,3,10,50, Nu Figure Petukhov&
102,103,104,105 614 Roizen
(1975)[143]
ratio Re,Pr=0,0,005,0,01, Nuq/NuT Figure Sleicher&
Nuq/NuT 0,02,0,03,0,04,0,10, 615 Tribus(1957)
0,718 [167]
q Parallelplates Re,Pr=0,01,0,03,0,7, Nu Figure Kays(1966)
3,10,30,102,103 616 [101]
Z Figure
617
Straighttubeof Re,Pr=0,01,0,7,10, Nu11,Z1 Figure
annularcross r1/r2=0,2 618
section
Nu22,Z2 Figure
619
Re,Pr=0,01,0,7,10, Nu11,Z1 Figure
r1/r2=0,5 620
Nu22,Z2 Figure
621
Coil Re,r/R=0,05,0,1 NuPr0,4 Figure Sebanetal.
622 (1963)[162]
TEL T Circularcylinder x/D,Re=0,5x105,105, Nux,Nu1m Figure Kays(1966)
2x105,Pr=0,01 623 [101]
x/D,Re=0,5x105,105, Figure
2x105,Pr=0,7 624
q Nux/Nu Figure625
x/D,Pr=0,01,0,7,10, Figure626

Re=105
Parallelplates x/DE,Pr=0,01,0,1,10 Nux/Nu,Z Figure
Re=7096,73612 627
69
ECSSEHB3101Part13A
5December2011
6.3.1.1 Laminar flow
SHAPE(L/DE>100) Nuq NuT
Circle 4,364 3,66
EquilatorTriangle 3,00 2,35
Square 3,63 2,89
Parallellines 8,235 7,54
Parallellines,oneofthem 5,385 4,86
insulated
Figure61:Nusseltnumbers,Nu,forfullydevelopedlaminarflowthrough
straightpipesofseveralcrosssectionalshapes.NuqistheNusseltnumberfor
constantheattransferratealongtheduct,andNuTthatforconstantwall
temperaturealongtheduct.FromKays&London(1964)
[102].
Figure62:Nusseltnumbers,Nu,vs.ratio,a/b,ofshortsidetolongsideforfully
developedlaminarflowthroughstraightpipesofrectangularcrosssection.From
Kays&London(1964)[102].
Figure63:Nusseltnumbers,Nu,vs.ratioofinnertoouterdiameter,r1/r2,forfully
developedlaminarflowinconcentriccirculartubeannuli.Constantheattransfer
rate.FromKays&London(1964)[102].
70
ECSSEHB3101Part13A
5December2011
Figure64:Influenceofcoefficients,Z,vs.ratioofinnertoouterdiameter,r1/r2,for
fullydevelopedlaminarflowinconcentriccirculartubeannuli.Constantheat
transferrate.FromKays&London(1964)[102].
Figure65:Nusseltnumber,Nu,vs.Deannumber,K,forfullydevelopedlaminar
flowincurvedpipeofcircularcrosssection.Constantheattransferrate.Results
areshownfordifferentPrandtlnumbers,Pr.CalculatedbythecompilerafterMori
&Nakayama(1965)[128].
71
ECSSEHB3101Part13A
5December2011
Figure66:ThermalentrylengthNusseltnumbers,Nu,vs.nondimensionalaxial
distance,x+,forlaminarflowthroughstraightpipes.Constantwalltemperature.
CalculatedbythecompilerafterKays(1966)[101].
Figure67:ThermalentrylengthNusseltnumber,Nux,vs.nondimensionalaxial
distance,x+,forlaminarflowthroughstraightpipes.Constantheattransferrate.
Alsoshowntheinfluencecoefficient,Z,forlaminarflowbetweenparallelplates
withonesideinsulated.CalculatedbythecompilerafterKays(1966)[101].
72
ECSSEHB3101Part13A
5December2011
Figure68:ThermalentrylengthNusseltnumbers,Nux,andinfluencecoefficients,
Z,vs.dimensionlessaxialdistance,x+,forlaminarflowinconcentriccirculartube
annuli.Constantheattransferrate.CalculatedbythecompilerafterKays(1966)
[101].
Figure69:ThermalentrylengthNusseltnumber,Nux,vs.nondimensional
distancealongthecoilcenterline,x+,forlaminarflowthroughacoil.Theresults
aregivenfortwovaluesoftheratio,r/R,betweenthecrosssectionalradiusand
thecoilradius.Constantwalltemperature.CalculatedbythecompilerafterKubair
&Kuloor(1966)[111].
73
ECSSEHB3101Part13A
5December2011

Figure610:Nusseltnumbers,Nu,vs.nondimensionalaxialdistance,x+,forthe
combinedhydrodynamicalandthermalentrylength.Laminarflowthrough
straightpipesofcircularcrosssection.Constantwalltemperature.Pr=0.7.
ReplottedbythecompilerafterESDU68006(1968)[48].
Figure611:LocalNusseltnumber,Nux,vs.nondimensionalaxialdistance,x+,for
thecombinedhydrodynamicalandthermalentrylength.Laminarflowthrough
straightpipesofcircularcrosssection.Constantheattransferrate.Resultsare
shownfordifferentPrandtlnumbers,Pr.CalculatedbythecompilerafterHeaton
etal.(1964)[82].
74
ECSSEHB3101Part13A
5December2011
Figure612:LocalNusseltnumber,Nux,andinfluencecoefficient,Z,vs.
dimensionlessaxialdistance,x+,forthecombinedhydrodynamicalandthermal
entrylength.Laminarflowbetweenparallelplates,oneoftheminsulated.
Constantheattransferrate.ResultsareshownfordifferentPrandtlnumbers,Pr.
CalculatedbythecompilerafterHeatonetal.(1964)[82].
6.3.1.2 Transitional flow
Figure613:LocalNusseltnumber,Nux,vs.Reynoldsnumber,Re,forfully
developedtransitionalflowthroughcylindricalductsofcircularcrosssection.
Constantwalltemperature.GasFlow(Pr0.7).FromESDU68006(1968)[48].
75
ECSSEHB3101Part13A
5December2011
6.3.1.3 Turbulent flow
Figure614:Nusseltnumber,Nu,vs.Reynoldsnumber,Re,forfullydeveloped
turbulentflowthroughcylindricalducts.Constantheattransferrate.Resultsare
shownfordifferentPrandtlnumbers,Pr.Calculatedbythecompilerafter
Petukhov&Roizen(1975)[143].
Figure615:RatioofNusseltnumberatconstantheattransferrate,Nuq,toNusselt
numberatuniformwalltemperature,NuT,vs.Reynoldsnumber,Re,forfully
developedturbulentflowthroughastraightpipeofcircularcrosssection.Results
areshownfordifferentPrandtlnumbers,Pr.FromSleicher&Tribus(1957)[167].
76
ECSSEHB3101Part13A
5December2011
Figure616:Nusseltnumber,Nu,vs.Reynoldsnumber,Re,forfullydeveloped
turbulentflowbetweenparallelplates,oneoftheminsulated.Constantheat
transferrate.ResultsareshownfordifferentPrandtlnumbers,Pr.Calculatedby
thecompilerafterKays(1966)[101].
Figure617:Influencecoefficient,Z,vs.Reynoldsnumber,Re,forfullydeveloped
turbulentflowbetweenparallelplates.Constantheattransferrate.Resultsare
shownfordifferentPrandtlnumbers,Pr.CalculatedbythecompilerafterKays
(1966)[101].
77
ECSSEHB3101Part13A
5December2011
Figure618:Nusseltnumber,Nu11,andinfluencecoefficient,Z1,vs.Reynolds
number,Re,forfullydevelopedturbulentflowinconcentriccirculartubeannuli.
r1/r2=0,2.Constantheattransferrate.ResultsareshownfordifferentPrandtl
numbers,Pr.CalculatedbythecompilerafterKays(1966)[101].
78
ECSSEHB3101Part13A
5December2011
79
ECSSEHB3101Part13A
5December2011
Figure622:NusseltnumbertimesPrandtlnumbertotheminus0.4power,NuPr
0.4,vs.Reynoldsnumber,Re,forfullydevelopedturbulentflowinhelicallycoiled
tubes.Theresultsaregivenfortwovaluesoftheratio,r/R,betweenthecross
sectionalradiusandthecoilradius.Constantheattransferrate.Calculatedbythe
compilerafteranexperimentalcorrelationobtainedbySeban&McLaughlin
(1963)[162]fromdataforwater.
distance,x/D,forfullydevelopedturbulentflowthroughastraightpipeofcircular
crosssection.Constantwalltemperature.Pr=0.01.Resultsareshownfordifferent
Reynoldsnumbers,Re.CalculatedbythecompilerafterKays(1966)[101].
80
ECSSEHB3101Part13A
5December2011
distance,x/D,forfullydevelopedturbulentflowthroughastraightpipeofcircular
crosssection.Constantwalltemperature.Pr=0.7.Resultsareshownfordifferent
Reynoldsnumbers,Re.CalculatedbythecompilerafterKays(1966)[101].
Figure625:RatioofthermalentrylengthNusseltnumber,Nux,toNusseltnumber
forfullydevelopedturbulentflow,Nu,vs.nondimensionalaxialdistance,x/D.
Straightpipeofcircularcrosssection.Constantheattransferrate.Pr=0.01.Results
areshownfordifferentReynoldsnumbers,Re.Calculatedbythecompilerafter
Kays(1966)[101].
81
ECSSEHB3101Part13A
5December2011
forfullydevelopedturbulentflow,Nu,vs.nondimensionalaxialdistance,x/D.
Straightpipeofcircularcrosssection.Constantheattransferrate.Re=105.Results
areshownfordifferentPrandtlnumbers,Pr.CalculatedbythecompilerafterKays
(1966)[101].
forfullydevelopedturbulentflow,Nu,vs.nondimensionalaxialdistance,x/DE.
Parallelplatesatdistance2DE,oneoftheminsulated.Constantheattransferrate.
Alsoshowntheinfluencecoefficient,Z.Resultsareshownforthreedifferent
Prandtlnumbers,Pr,andtwoReynoldsnumbers,Re.Calculatedbythecompiler
afterKays(1966)[101].
82
ECSSEHB3101Part13A
5December2011
6.3.2 Heat transfer to external flows

Thetablebelowsummarizesthedatapresentedinthefollowingpages.Allthesedatacorrespondto
external flows, i.e. configurations where the fluid flows outside the domain bounded by the heat
transfersurface.
Geometry Input Output Fig. References

Variables
CircularCylinderNormaltoFlow. Re,Pr=0,7,5,3,5 Nub Figure ESDU69004[50],

Fluidwithconstantpropertyvalues. 628 Fig.6.
CircularCylinderNormaltoFlow. 0,061<wm/b< Nu/Nub Figure ESDU69004[50],

Liquidswithvariablefluid 0,936 629a Fig.7.
properties.
CircularCylinderNormaltoFlow. 0,20<Twm/Tb<4,47 Nu/Nub Figure ESDU69004[50],

Gaseswithvariablefluidproperties. 629b Fig.8.
CircularCylinderYawedtoFlow. 40<<90 Nu/Nu90 Figure ESDU69004[50],

629c Fig.9.
InLineTubeBank. Figure ESDU73031[57],

630 Fig.5.
InLineTubeBank. Reb,Prb=1 Nur Figure ESDU73031[57],

Fluidswithconstantproperty 631 Fig.6.
values.
InLineTubeBank. Reb,Prb=1 Nur Figure ESDU73031[57],

Fluidwithconstantpropertyvalues. 632 Fig.6.
StaggeredTubeBank. Reb,Prb=1 Nur Figure ESDU73031[57],

values.
InLineorStaggeredTubeBank. Prb Nu r Figure ESDU73031[57],

634 Fig.8.
Fluidswithconstantproperty Nur Prb 1
values.
InLineorStaggeredTubeBank. Prb/Prw F1 Figure ESDU73031[57],

values.
InLineorStaggeredTubeBank. N F2 Figure ESDU73031[57],

636 Fig.10.
InLineorStaggeredTubeBank 2090 F3 Figure ESDU73031[57],

YawedtoFlow. 637 Fig.11.
83
ECSSEHB3101Part13A
5December2011
Geometry Input Output Fig. References

Variables
InLineorStaggeredTubeBank. 1n20 F4 Figure ESDU73031[57],

Characteristicsofthenthrow. 638 Fig.12.
6.3.2.1 Cylindrical bodies
Figure628:Nusseltnumber,Nu,vs.Reynoldsnumber,Re.Flowofafluidhaving
constantphysicalpropertiesoveraconstanttemperaturecircularcylinderwhose
axisisnormaltotheincomingflow.FromESDU69004(1969)[50].
84
ECSSEHB3101Part13A
5December2011
Figure629:Effectofvariablefluidproperties,(a)and(b),andofinclinationangle,
(c),ontheNusseltnumbercorrespondingtotheflowofafluidoveraconstant
temperaturecylinder.Nub(Nu90)canbededucedfromFigure628.FromESDU
69004(1969)[50].
85
ECSSEHB3101Part13A
5December2011
6.3.2.2 Tube banks
Figure630:GuidefortheselectionofthecurvesgiveninFigure631andFigure
632concerninginlinetubebanksofdifferentrelativepitches.FromESDU73031
(1973)[57].
86
ECSSEHB3101Part13A
5December2011
Figure631:ReferenceNusseltnumber,Nur,forPrb=1,asafunctionofReynolds
number,Re.Inlinetubebanks.SeeFigure630forthemeaningofthenumbers
whichappearonthecurves.FromESDU73031(1973)[57].
87
ECSSEHB3101Part13A
5December2011
number.Re.Inlinetubebanks.SeeFigure630forthemeaningofthenumbers
whichappearonthecurves.FromESDU73031(1973)[57].
88
ECSSEHB3101Part13A
5December2011

number.Re.Inlinetubebanks.Staggeredtubebanks.FromESDU73031(1973)
[57].
Figure634:EffectofthePrandtlnumber,Prb,onthereferenceNusseltnumber,
Nur,forbothinlineandstaggeredtubebanks.FromESDU73031(1973)[57].
89
ECSSEHB3101Part13A
5December2011
Figure635:ThefactorF1toaccountforvariablefluidproperties.FromESDU
73031(1973)[57].
Figure636:ThefactorF2accountingforabnormalnumberofrowsvs.that
number,N.FromESDU73031(1973)[57].
Figure637:ThefactorF3accountingfortheeffectofyawvs.theinclinationangle,
.FromESDU73031(1973)[57].
90
ECSSEHB3101Part13A
5December2011
Figure638:ThefactorF4forestimatingtheNusseltnumberofthenthrow.From
ESDU73031(1973)[57].
91
ECSSEHB3101Part13A
5December2011
7
Frictional analysis
7.1 General
Oncethethermalanalysisofthefluidloophasbeenperformed,thepressuredropthroughtheloop
willbeestimated.Theaimofthisestimationistwofold:
1. Tofixthepressureleveloftheworkingfluid.Thislevelisofconcerninmanycases.For
instance,vaporizationisnormallyavoidedwhenliquidcoolantsareused,becauseofthe
associatedcoolingeffectivenessdegradation.Thisisnotalwaysthecasesinceadvantage
is taken of boiling (and subsequent condensation) in several heat exchanger devices to
storetheheatinthefluidduringitstransferfromheatsourcetoheatsink.However,only
fluidsthatremainintheirliquidstatethroughouttheloopareconsideredhere.
2. Toselecttheappropriatepump.Thepower, pp,requiredtoforcethefluidthroughthe
ductisdeducedfromthepressuredrop, p,themassflowrate,m,andthefluiddensity,
,byuseofthefollowingexpression:
m
p p p [71]

This clause is aimed at estimating the pressure drop, p, of a loop having fairly general geometric
configuration, in terms of the pressure loss at each individual component and the appropriate
interferencefactors,whentheyareavailableandrelevant.
Ifthepowerrequirementscalculatedbymeansofthepertinentdataaretoostringent,acarefullookto
thepipeworkcouldberequired.Reductionsinthemassflowrateand,thence,inthepressureloss
wouldimplyhigheroperatingtemperaturedifferentialsandlargerheatrejectingsystems.
Whenthemassflowrateisreducedthethermalanalysisoftheloopshouldbeundertakenagain.
7.2 Analytical background
7.2.1 Introduction
Theflowofarealfluidthroughapipeischaracterizedbythesignificanceoffrictionforces.Associated
withtheseforcesthereisapressuredrop,p.
92
ECSSEHB3101Part13A
5December2011
Adimensionalanalysisindicatesthat,forapipesystemofgivengeometryrunningfull,thepressure
drop between two sections may be related to flow properties, fluid properties and geometrical
characteristicsofthepipeinthefollowingway:
p e
f , Re, M , , geometry
1
V 2 D [72]
2
Intheapplicationswhichwillbeconsidered,compressibilityeffectswhichintheaboveexpression
aremeasuredbytheMachnumber,M,andtheratioofspecificheats,willbeneglected.
7.2.2 Fully developed flow in straight pipes

Thepressuredropduetofrictioninacylindricalpipeofanycrosssectionrunningfullisaccountedas
L 1
p V 2 [73]
DE 2
Manyauthors,especiallyfromtheUnitedStates,usetheFanningfrictionfactor,f= /4,insteadof .
WewilluseindiscriminatelybothandfthroughoutthisclauseandClauses8,9,10,11and13.
For laminar flow the friction factor does not depend on the relative roughness of the tube wall. In
particular,whenthecrosssectionofthepipeiscircular:
64 / Re [74]
which is the well known HagenPoiseuille formula, valid for Re < 2400 and fully developed flow.
Values of (of f) for fully developed laminar flow and other crosssectional shapes can be found in
Figure73,inESDU66027(1966)[46]andinKays&London(1964)[102],p.103.
In the transition between laminarandturbulent flows, the value of the direction factor isuncertain.
However, it is surely bounded below and above by the extrapolation of the laminar line and the
turbulentlinecorrespondingtotherelativeroughnessofthepipewall,respectively.
In the case of fully developed turbulent flow the friction factor, , strongly depends on the relative
roughnessofthewall,butisfairlyinsensitivetotheshapeofthecrosssection.Referenceismadeto
Schlichting(1960)[157]fordetailsconcerningtheturbulentflowthroughpipesofnoncircularcross
section.
Thelawoffrictionforroughpipesdependsonthesize(andshape)oftheroughnessascomparedto
thethicknessofthelaminarsublayeroffluidclosetothewall.
1. Whentheroughnesssizeissosmallthatallprotrusionsarecontainedwithinthelaminar
sublayer, the pipe is hydraulically smooth. For these pipes at very large Reynolds
numberthefollowinglawoffrictionapplies:
93
ECSSEHB3101Part13A
5December2011
1

2,035 log Re 0,91 [75]
Thisexpressionisoftenwrittenas:
2
f

2,5 ln Re f 0,85 [76]
whichisusuallyknownasKarmanNikuradsefrictionlaw.
Manyauthors(Schlichting(1960)[157])usetheslightlydifferentexpression
1

2 log Re 0,8 [77]
knownasPrandtlsuniversallawoffrictionforsmoothpipes.Thislawhasbeenverified
by Nikuradses experiments up to a Reynolds number of 3,4 x 106, and may be
extrapolatedtoarbitrarilylargeReynoldsnumbersprovidedthatthetubeissmooth.
2. Inthecaseofroughpipesthelawoffrictiondeviatesfromthatcorrespondingtosmooth
pipe,oncetheReynoldsnumberexceedsadefinitevaluewhosemagnitudeincreasesas
the relative roughness, e/DE, decreases. Above some even higher value of the Reynolds
numberthefrictionfactorbecomesconstantanddependsontherelativeroughnessonly
(seeFigure72).
In this regime, called the completely rough regime, all protrusions reach outside the
laminarsublayer,whosethicknessdecreaseswhentheReynoldsnumberincreases.
ThelawoffrictionforthecompletelyroughregimewasfirstderivedbyvonKrmnin
theform:
2 D
2,5 ln E 4,75 [78]
f 2e
Intermsofthefrictionfactor,
1 DE
2,035 log 1,679 [79]
2e
AcomparisonwithNikuradsesexperimentalresultsshowsthatcloseragreementcanbe
obtainedbyuseofthefollowingexpression(Schlichting(1960)[157]):
1 DE
2 log 1,74 [710]
2e
94
ECSSEHB3101Part13A
5December2011
3. Intheintermediateregion,protrusionsextendpartlyoutsidethelaminarsublayer.
AfrictionsimilaritylawwhichisvalidoverawiderangeofroughnesssandReynolds
numbershasbeendevelopedbyNikuradseforsandgraintyperoughness.
2
f
D

2,5 ln E ue e 3,75
2e
[711]
The roughness function, ue+(e+), for sandgraintype roughness, is shown in Figure 91b. For large
valuesofthedimensionlessprotrusionheight,e+,theroughnessfunctiontendstoanasymptoticvalue
closetoue+=8,50,yieldingthepreviouslyquotedexpressionforthecompletelyroughregime.
Expressions similar to that developed by Nikuradse are useful reducing the experimental work
involvedinthedeterminationoff.Sincetheconstants2,5and3,75areuniversal,itisonlynecessaryto
determinethefunctionue+,whichnormallydependsone+andonthetypeofroughness.ue+forseveral
discreteroughnesssisshowninFigure92bandFigure93b.
For commercially rough pipes the Moody diagrams or the ColebrookWhite formula can be used.
Moodys diagramis based onan equivalentsand roughness, tobe determined, for each commercial
pipe,byuseofanauxiliaryequivalencechart(seeSchlichting(1960)[157],pp.528529).
Table71givestheequivalentroughnessheight,e,foranumberofcommonsurfaces.
Table71:Effectiveroughnessheightofanumberofcommonsurfaces.
TheColebrook&Whiteformula(Schlichting(1960)[157])correlatesthewholetransitionregionfrom
hydraulicallysmoothtocompletelyroughflow,
1 2e 18,7
1,74 2 log
[712]
DE Re
95
ECSSEHB3101Part13A
5December2011
Whene0thisequationyieldsthatforthehydraulicallysmoothregime.ForRe,ittransforms
into the equation for the completely rough regime. In the transition region it gives results which
resemblethoseobtainedbyuseoftheMoodydiagram.
Inpracticethesurfaceroughnesscanbedistributedineitherarandomorasystematicway,orcanbe
concentratedatintervalsalongthesystemasoccurswithweldingseamsinpipesorinternalsupports.
Althoughtheseimperfectionsmaybecomeinsignificantwithcarefulmanufacturing,andtheireffect
ontheflowissmallexceptatveryhighReynoldsnumber,specifictestsarerequiredtodeterminethe
pressure loss induced in each particular case. Some general trends of the friction law for several
roughness types can be deduced from the analysis of Figure 71 in order to separate the several
curves, they have been plotted for different values of the equivalent roughness, e, otherwise their
asymptoticvalue(forRe)wouldbethesame.
Figure71:Frictioncharacteristicsassociatedwithfourtypesofroughness
geometry.Noticethattheequivalentroughnessisdifferentineverycase.From
Reynolds(1974).
Type A: Sand grains and other closely spaced elements. As the Reynolds number rises all peaks
emerge at once from the laminar sublayer. Once the peaks emerged the stress is transmitted by the
dragofthepeaks,thencebecomingindependentoftheReynoldsnumber.
TypeB:Protrusionsofvarioussizes,asproducedbymanymanufacturingprocessesorbyaging.The
emergenceisgradualandtheelementsprogressivelystarttotransmitstressbydrag.
Type C: Isolated protrusions on an otherwise fairly smooth wall. The drag coefficient of the large
elements becomes Reynolds number independent, while the friction coefficient for the surrounding
wall continues to fall as the Reynolds number increases. This behavior may be expected when less
thanhalfofthesurfaceareaisoccupiedbytheroughnesselements.
Type D: Isolated depressions or grooves in an otherwise fairy smooth wall. The contribution of the
depressionrisesslowlyastheviscouslayerthins.Thisissobecausethistypeofroughnesshaslittle
effectuntilthefreestreamturbulencepenetratesintothecavity.
There are surfaces, however, which do not fit into these categories. This happens to be the case, for
example,whenlargeinsertsareusedtoenhancetheheattransfer(seeclause9.1.1).
96
ECSSEHB3101Part13A
5December2011
7.2.2.1 Power-law approximations for the hydraulically smooth regime
Numericalworkcanbereducedbyrepresentingthefrictionfactorbyapowerlawapproximation
f C1 C2 Re n [713]
withC1,C2andnsochosenastofittheabovelogarithmicformulae.
ThefollowingexpressionshavebeenborrowedfromReynolds(1974)[149].
Table72:ConstantsforPowerLawApproximation.HydraulicallySmooth
Regime.
C1 C2 n ReRange Accuracy Comments
0 0,079 0,25 <105 2%oftheexperimental Blasiusfrictionformula.

values Asaccurateasthe
logarithmicformula
0 0,046 0,20 105106 Notgiven
0,0014 0,125 0,32 <107 3%ofthelogformula

forRe<106.
7%ofthelogformula
forRe<107
7.2.3 Temperature-dependence of fluid properties

Thefrictiondatapresentedinthisclausecorrespondtoisothermalconditions.
The simplest approach to account for property variation is to evaluate all fluid properties at the
referencetemperature(seeclause6.2.4)andtointroduceatemperatureratiocorrectionfactorofthe
form(Tw/TR)nor(w/R)n.
The correction is inherently small for turbulent flow, whereas the available experimental evidence
suggeststhatitismuchlargerinthelaminarflowcase.
Table73belowpresentsrelevantdataforfullydevelopedflowincylindricaltubesofcircularcross
section.
97
ECSSEHB3101Part13A
5December2011
Table73:LoadingFactorsAccountingforTemperatureDependenceofDiabatic
Friction
TypeofFlow Fluid LoadingFactor n Comment
/ b
Fluid Fluid
Heating Cooling
LaminarFully Gas (Tw/Tb)n 0,45 0,30 Intherange

DevelopedTubesof 1/3<Tw/Tb<3
Circular
CrossSection Liquid (w/b)n 0,58 0,50 More
experimental
confirmationis
required
Turbulent Gas (Tw/Tb)n 0,10 0 Intherange

1/3<Tw/Tb<3
FullyDeveloped Liquid (w/b)n 0,03 0,05

Pr=10
TubesofCircular Liquid 0,04 0,01

Pr=102
CrossSection Liquid 0,12 0,02

Pr=103
NOTE FromKays&London(1964)[101]
7.2.4 Several definitions of pressure loss coefficient

Thetotalpressure,pt,inagivensectionofaductisdefinedas
1
pt p V 2 gz [714]
2
1. Acoefficientoftotalpressurelossbetweentwostations1and2alongapipeisdefinedas
follows:
pt1 pt 2
cK t
1 [715]
V 2
Whenthecrosssectionalareas1and2aredifferent,thelosscoefficientmaybeexpressed
in terms of the value of the dynamic pressure, V2/2, at either section. Care should be
98
ECSSEHB3101Part13A
5December2011
exercisedtoensurethatthecorrectreferencedynamicpressureisusedwheninterpreting
thepressurelosscoefficientdata.
ALL THE DATA PRESENTED HERE, WITH THE ONLY EXCEPTION OF THE
SUDDEN CONTRACTION, ARE BASED ON THE CONDITIONS EXISTING
UPSTREAMOFTHECOMPONENTINVOLVED.
2. Inrealflowsthetotalpressurevariesoverthecrosssection,whilethestaticpressureis
uniformprovidedthatthestreamlinesarealmostparallel.Thence,inpractice,itisoften
simpler to define the mean static pressure, rather than the mean total pressure at a
section.
Since the contribution of the geometrical head term, gz, depends on the orientation of
the pipework, the quoted data correspond to static pressure differences between two
points in the same horizontal plane. The contribution of gz, when significant, is
accountedforindependently.
Thestaticpressurelosscoefficientbetweentwostations1and2alongthepipeisdefined
asfollows:
p1 p2
cK s
1 [716]
V 2
2
Noticethatwhenthecrosssectionalareaincreasesinthedirectionoftheflowthestatic
pressurecanincrease,inwhichcasethepressurelossisnegative.
Inthecaseofincompressibleflowthroughconstantcrosssectionductsthetotalpressure
loss, cKt, equals the static pressure loss, cKs. In that case the subscripts t or s have been
omitted.
3. The pressure loss due to a component, let say a sudden area change or a bend, can be
consideredtobemadeupoftwoterms.Oneisthepressurelossthatwouldhaveexisted
intheabsenceofthecomponenti.e.thatpressurelosscorrespondingtofullydeveloped
flows in lengths equal to the straight portions upstream and downstream of the
componentandtheotheristhepressurelossdirectlyattributabletothecomponent.
Thegrosspressureloss,pG,isdefinedas:

pG p1 p2 p x1 p x2 [717]
where px1 and px2 are the static pressure losses due to friction for fully developed
flowsinlengthsofstraightpipeequaltox1andx2respectively.
Thenetpressureloss,pN,isdefinedas:
p N pG p L [718]
wherethestaticpressurelossduetofrictionforfullydevelopedflowthroughthelength,
L,ofthecomponenthasbeensubtracted.
99
ECSSEHB3101Part13A
5December2011
Theusefulnessofthesedefinitionsisthat,providedthestaticpressurelossisreferredtostations1and
2intheregionsoffullydevelopedflow,thecorrespondingcoefficientsareindependentofx1andx2.
The use of the gross static pressure loss should be preferred to that of the net static pressure loss
because:
1. Componentswithunequalentryandexitareascanbedealtwithconveniently.
2. ThedefinitionofLisoftenarbitraryasoccursinthecaseofverylocalizedsingularities
suchas:suddenareachanges,orifices,valves,...
ForsimplemitrebendsLisequaltozero,sothatthereisnodifferencebetweengrossandnetpressure
losses.
ALLTHEDATAPRESENTEDAREBASEDONTHEGROSSPRESSURELOSS
7.2.5 Entrance effects

Thefullydevelopedflowdoesnottakeplacejustfromthetubeentrance,ratherandentrancelength
is required for the velocity profiles to relax from the initial practically uniform distribution to the
parabolicdistributionwhichischaracteristicoflaminarflows.
Theentrancelength,Lent,canbededucedfromthefollowingapproximateformula(Nekrasov(1969)
[132]).
Lent / D 0,029 Re [719]
The pressure loss in the entrance length is greater than in subsequent sections because of the larger
velocitygradientsattheductwallassociatedwiththenearuniformvelocityprofiles.
Thefrictionfactorcorrespondingtolaminarflowthroughacylindricaltubeofcircularcrosssection,
whoselength,L,iscommensurablewiththeentrancelengthisgiven(Nekrasov(1969)[132])by:
D 64
1,165 [720]
L Re
Noticethatanadditiveterm,D/L,incorrectlyneglectedbyNekrasov(whogives0,165D/Linsteadof
1,165D/L) appears in this expression. This term accounts for the change in dynamic pressure, V2/2,
betweentheentranceandthegivensection.
Theabovevalue(k=1,165)wasdeducedanalyticallybySchiller.Experimentallyobtainedvaluesare
somewhat larger (see, for example, Lundgren, Sparrow & Starr (1964) [121]). The analytical value
obtainedbytheseauthors(k=4/3)isveryclosetothatresultingfromSchillersexperiments.
ValuesofkforseveralcrosssectionalshapescanbefoundinLundgrenetal.(1964)[121],whoapplied
their theory to circular, elliptical, rectangular, isosceles triangular and annular ducts. The case of
circularsegmentcrosssectionshasbeenconsideredbySparrow&HajiSheikh(1966)[169].
The inlet length in turbulent flow is considerably shorter that in laminar flow. According to
experimentsperformedbyNikuradse(quotedbySchlichting(1960)[157])thefullydevelopedvelocity
profileexistsafteraninletlengthof25to40diameters.
100
ECSSEHB3101Part13A
5December2011
7.2.6 Interferences and networks

ThepressurelosscoefficientsofthedifferentcomponentscanbededucedfromFigure72toFigure
730orfromthedatasubmittedbythemanufacturerofthecomponent,whenthesedataareavailable.
Allthesecoefficientswillbereferredtothenominalcrosssectionoftheloopbyusingtheequationof
masspreservation.
Theinteractionbetweenconsecutivecomponentsoftheloopcanonlybetakenintoaccountinspecial
cases. When data are available for particular interactions, the components concerned should be
consideredassingleentities.
Thepressurelosscoefficientofelementsplacedinseriescanbeadded,oncetheyarereferredtothe
nominalcrosssection.
Whennelementsareinparallelanequivalentpressurelosscoefficientiscalculatedonthebasisofthe
followingtwofacts:
(a) The mass flow approaching any junction or branch must equal the mass flow
leavingthem.
(b) The overall pressure loss through several routes in parallel is the same for all of
them.
From these considerations the following alternative expressions of the equivalent pressure loss
coefficientcanbededuced:
(a) Forlaminarflows,wherethepressurelossisproportionaltothemeanvelocity:
n
1 1
[721]
cK equiv i 1 cKi
(b) For turbulent flows, where the pressure lossmean velocity relationship is
quadratic:
n
1 1
[722]
cK equiv i 1 cK i
Thepressureloss, p,intheloopisrelatedtothetotalcoefficientofpressureloss,cK,bymeansofthe
equation
1
p cK V 2 [723]
2
whereVisthemeanflowvelocityinthenominalcrosssection.
101
ECSSEHB3101Part13A
5December2011
7.2.7 Flow chart

The flow chart is intended to guide in the selection of the appropriate pressure loss data. This
guidance is achieved by means of a series of questions concerning the parameters involved. The
figuresquotedintheseveralboxesarethoseenclosedinClause7.3.
Inadditiontothepressurelossdatawhicharecollectedinclause7.3,thefollowingsourcesofdataare
worthyofbeingmentioned.
Two updated data items on pressure losses in curved ducts have been issued by ESDU. The item
ESDU 77008 (1977) [58] deals with single bends, whereas ESDU 77009 (1977) [59] gives interaction
factorsfortwobendsinseries.
Flowfriction design data for many heat exchanger configurations can be found in Kays & London
(1964)[102].Thegeometriesoftheconfigurationsconsideredbytheseauthorsaredescribedinclause
11.3ofthisPart.
Additionalinformationonpressurelossdataforbranchingflowsthroughplanarthreelegjunctionsis
given in ESDU 73022 (1973) [55]. Data for Yjunctions, angled junctions, general dividing junctions,
andtaperedlegjunctionsarepresentedinthementioneddataitem.Similarinformation,butnowfor
confluentflows,isgiveninESDU73023(1973)[56].
Finally, the book by Idelcik (1969) [97] should be mentioned. This book presents data on straight
pipes (for both fully developed flow and entrance effects), area changes, bends, dividing and
combiningflows,orifices,valvesandotherconfigurationsoftechnicalinterest.
FlowchartfortheSelectionofPressureLossData
102
ECSSEHB3101Part13A
5December2011
103
ECSSEHB3101Part13A
5December2011
104
ECSSEHB3101Part13A
5December2011
7.3 Pressure loss data
7.3.1 Straight pipes
Figure72:Frictionfactor,c,asafunctionofReynoldsnumber,Re,fordifferent
valuesoftherelativeroughness,e/D:Cylindricaltubesofcircularcrosssection.
FromIdelcik(1969)[97].
Figure73:Correctionfactor,K,tobeusedwhenthecrosssectionoftheductisnot
circular.Laminarflow.K=1forturbulentflowthroughhydraulicallysmooth
ducts.FromESDU66027(1966)[46].
105
ECSSEHB3101Part13A
5December2011
7.3.2 Bends
Figure74:Boundarybetweenshortandlongcirculararcbends.FromESDU67040
(1967)[47].
Figure75:Boundariesbetweenlaminar,transitionalandturbulentflowsinlong
circulararcbends.FromESDU67040(1967)[47].
106
ECSSEHB3101Part13A
5December2011
Figure76:Pressurelosscoefficientperunitbendangle,cK/,asafunctionofthe
dimensionlessradiusofcurvatureofbendcenterline,R/D,fordifferentvaluesof
Reynoldsnumber,Re.Eithercircularorsquarecrosssection.FromESDU67040
(1967)[47].
107
ECSSEHB3101Part13A
5December2011
Figure77:Pressurelosscoefficient,cK,asafunctionofthedimensionlessradius
ofbendcenterline,R/D,fordifferentvaluesofReynoldsnumber,Re.Laminar
flowthroughshortcirculararcbends.FromESDU67040(1967)[47].
108
ECSSEHB3101Part13A
5December2011
Figure78:Pressurelosscoefficient,cK,asafunctionofthedimensionlessradius
ofbendcenterline,R/D,fordifferentvaluesofbendangle,.Turbulentflow
throughshortcirculararcbends.Eithercircularorsquarecrosssection.From
ESDU67040(1967)[47].
109
ECSSEHB3101Part13A
5December2011
Figure79:Pressurelosscoefficient,cK,forshortcirculararcbends,havingashort
downstreamtangentoflength,Ld,asafunctionofLd/D,fordifferentvaluesofthe
dimensionlessradiusofbendcenterline,R/D.Turbulentflow.Eithercircularor
squarecrosssection.FromESDU67040(1967)[47].
Figure710:Thefactor1toaccountfortheaspectratioofthebendcrosssection.
FromESDU67040(1967)[47].
110
ECSSEHB3101Part13A
5December2011
Figure711:Thefactor2toaccountforthebendangle.FromESDU67040(1967)
[47].
Figure712:Pressurelosscoefficient,cK,forsinglemitrebends,asafunctionof
bendangle,,fordifferentvaluesofthedimensionlesslength,Ld/D,ofthe
downstreamtube.Turbulentflow.Eithercircularorsquarecrosssection.From
ESDU67040(1967)[47].
111
ECSSEHB3101Part13A
5December2011
Figure713:Factor,whichaccountfortheinteractionbetweentwo90circular
arcbends,asafunctionofthedimensionlessdistancebetweenbothbends,La/D.
FromESDU68035(1968)[49].
112
ECSSEHB3101Part13A
5December2011
Figure714:Factor,whichaccountfortheinteractionbetweentwomitrebends,
asafunctionofthedimensionlessdistancebetweenbothbends,La/D.FromESDU
68035(1968)[49].
7.3.3 Sudden changes of area
Figure715:Totalpressurelosscoefficient,cKt,asafunctionofReynoldsnumber,
ReD1,fordifferentvaluesofthearearatio,.Enlargementwithaductdownstream
4D2long.UniformincomingflowatlowReynoldsnumber.FromESDU72011
(1972)[54].
113
ECSSEHB3101Part13A
5December2011
Curve Umax/U Typicalof: Curve Umax/U Typicalof:
1 1,0 Uniformflow 4 1,4 Exitfromgrid,highporosity

2 1,24 Fullydevelopedturbulent 6 2,0 Exitfromgrid,lowporosity
3 1,44 flow 7 2,9 Exitfromsharpradiusbend,
5 1,87 Exitfromconicaldiffuser2 radiusratio<1,5,andexit
halfangle fromplanediffuser815
Exitfromconicaldiffuser4 halfangle
halfangle
Figure716:Differentvelocityprofilesupstreamofasuddenenlargement.From
ESDU72011(1972)[54].
Figure717:Totalpressurelosscoefficient,cKt,asafunctionofarearatio,.
Enlargementwithaductdownstream4D2long.Numbersoncurvesindicatethe
velocityprofileinFigure722forwhichthecurveapplies.FromESDU72011(1972)
[54].
114
ECSSEHB3101Part13A
5December2011
Figure718:Staticpressurelosscoefficient,cKs,asafunctionofarearatio,.
Enlargementwithaductdownstream4D2long.Numbersoncurvesindicatethe
velocityprofileinFigure722forwhichthecurveapplies.FromESDU(1972)[54].
Figure719:Totalpressurelosscoefficient,cKt,asafunctionofReynoldsnumber,
ReD2,fordifferentvaluesofthearearatio,.Thepressurelosscoefficientis
expressedintermsofthedynamicpressureatclause6.FromIdelcik(1969)[97].
115
ECSSEHB3101Part13A
5December2011
7.3.4 Orifices and diaphragms
Figure720:Referencevaluesofthepressurelosscoefficient,cK,asafunctionof
theratio,,oftheareaavailableforfluidflowtothetotalareaoftheductcross
section.Perforatedplatesandorifices.FromESDU72010(1972)[53].
116
ECSSEHB3101Part13A
5December2011
Figure721:Thefactor3toaccountfortheeffectofplatethicknesswhent/d<0,8.
cKoisgiveninFigure719.FromESDU72010(1972)[53].
117
ECSSEHB3101Part13A
5December2011
Figure722:Thefactor4toaccountfortheeffectofplatethicknesswhent/d0,8.
cK0,8isgiveninFigure719.FromESDU72010(1972)[53].
118
ECSSEHB3101Part13A
5December2011
Figure723:Comparisonbetweenthepressurelosscoefficients,cK,inthe
intermediateregioncalculatedbyassumingeitherofthetwoextremecases,fully
separatedorreattachedorificeflow.FromESDU72010(1972)[53].
7.3.5 Screens
Figure724:Referencepressurelosscoefficient,cKr,asafunctionofporosity,.
Roundwiregauzes.FromESDU72009(1972)[52].
119
ECSSEHB3101Part13A
5December2011
Figure725:Factor5toaccountforlowReynoldsnumbereffectsinroundwire
gauzes.Reynoldsnumberbasedonthewirediameter.FromESDU72009(1972)
[52].
7.3.6 Valves
Figure726:Referencepressurelosscoefficient,cKr,asafunctionofReynolds
number,Re,fordiaphragmandbutterflyvalvesfullyopen.Preparedbythe
compilerafterESDU69022(1969)[51].
120
ECSSEHB3101Part13A
5December2011
Figure727:Factor6,whichaccountsforthepartialopeningofthevalve,asa
functionofthedegreeofvalveopening,.isdefinedastheratioofvalvecontrol
travelfromclosedpositiontototalvalvecontroltravel.FromESDU69022(1969)
[51].
7.3.7 Tube banks
To Ti
cK ARe N 2
m
[724]
Ti
N:numberoftransversalrows.
1. Whenstsl,thenA=ab
whereaisgiveninFigure728a,bisgiveninFigure728band(Re)m,for(std)/(sld)=1is
giveninFigure728c
2. Whenst>sl,thenA=bc
where c is given inFigure 728a, b isgiven inFigure 728b and(Re)m is given in Figure
728c
121
ECSSEHB3101Part13A
5December2011
Figure728:Graphicsforestimatingthepressurelosscoefficient,cK,forinline
tubebanksofseveralrelativepitches,sl,st,andyawangles,.Theinfluenceofthe
heatexchangeonthepressurelossistakenintoaccountthroughthetubebank
inletandexittemperatures,TiandTo,respectively.FromIdelcik[97].
cK ARe
0 , 27
N 1 To Ti [725]
Ti
N:numberoftransversalrows.
1. Whenst/d<2and0,14(std)/(sld)<1,7,thenA=ab+3,2
whereaisgiveninFigure729a,bisgiveninFigure729band(Re)0,27,isgiveninFigure
729c
2. Whenst/d2and0,14(std)/(sld)1,7,thenA=3,2
3. Whenst/d1and1,7(std)/(sld)5,2,thenA=c
wherecisgiveninFigure729a
122
ECSSEHB3101Part13A
5December2011
Figure729:Graphicsforestimatingthepressurelosscoefficient,cK,forstaggered
tubebanksofseveralrelativepitches,sl,st,andyawangles,.Theinfluenceofthe
heatexchangeonthepressurelossistakenintoaccountthroughthetubebank
inletandexittemperatures,TiandTo,respectively.FromIdelcik(1969)[97].
123
ECSSEHB3101Part13A
5December2011
7.3.8 Branching of tubes

SKETCH CURBENUMBER
Allcrosssections
CONFLUENCE BRANCHING CONFLUENCE BRANCHING
havethesame
diameter,D p32 p13 p12 p12
ck32 ck13 ck12 ck12
1 1 1 1
V22 V12 V22 V12
2 2 2 2
1 1 2 2
r/D=0,13 r/D=0,15
r/D=0,24 r/D=0,26
m4/m3=27 10
m4/m3=18
m4/m3=0,59

Figure730:Pressurelosscoefficient,cK,asafunctionoftheratiooflateraltototal
massflowratesinbranchingtubes.Themixedconfluencebranchingcaseisnot
considered.FromIdelcik(1969)[97].
124
ECSSEHB3101Part13A
5December2011
8
Combined thermal and frictional analysis
8.1 General
Theaimofthisclauseistoprovideinformationrelatingheattransferandfluidfrictioninturbulent
flowthroughsmoothpipes.
Theinterplaybetweenheattransferandfluidfrictioncomesfromtwodifferentcauses:
1. Thetransportofanyfluidentityinaturbulentflowdependsonthemeanandfluctuating
velocity fields. Thus the analysis of a turbulent transfer process requires some
understanding of momentum transfer and the associated velocity variation, as will be
showninthisclause.
2. In many instances, there is a dependence of the velocity field on the heat transfer. The
temperaturevariationsassociatedwithrapidheattransfer(orwithveryhighvelocities)
give rise to changes in the molecular transport properties of the fluid, with the
consequent modifications in the structure of the fluid layers close to the wall where
moleculartransportstilldominates.Thispointhasbeenalreadydiscussedinclause6.2.4
andclause7.2.3.
Ithasbeenmentionedabovethatthisclauseexclusivelyconcernsflowthroughsmoothpipes.Dataon
convectiveheattransferandfluidfrictioninroughpipesaregiveninclause9.2.
8.2 Analogies between momentum and heat transfer

Itcanbededucedfromthefirstconsiderationinclause8.1thatanalogiesrelatingratesofheattransfer
tomomentumtransferexist.Theseanalogiescouldbeusefultotheengineerwhendataonturbulent
heattransferarenotavailablewhereasdataonfluidfrictionare.
Fromthemanyanalogiesexistingintheliteratureonlythosemostcommonlyusedwillbeintroduced
here.ReferenceismadetothebookbyA.J.Reynolds(Reynolds(1974)[149])forbothanaccountofthe
motivation and justification of these analogies and for a survey of those exhibiting simple
mathematicalformulations.
8.2.1 The Reynolds analogy

In 1874 Osborne Reynolds postulated that the laws governing turbulent transfer of momentum and
heatwerethesame.
Turbulenttransferisinducedbytherandomfluctuationsinvelocitywhicharesuperimposedtothe
meanflow.Becauseofthesefluctuationssmalllumpsoffluidmovebackwardsandforwardsacross
125
ECSSEHB3101Part13A
5December2011
the stream. This movement must involve the transfer of momentum (or heat) when there is a
transversevelocity(ortemperature)gradient.
If the fluid particles move on the average from a region where the mean velocity is V and the bulk
temperatureisTb,tothewallsurfacewherethemeanvelocityiszeroandthetemperatureisTw,the
Reynoldsanalogywouldimply,
Rate of heat transfer q c p Tw Tb

[81]
Rate of momentum transfer w V
whereqistheheatflowrateperunitwallsurfaceareaandwisthewallfriction.
Anexpressionfortheheattransfercoefficient,h=q/(TwTb),orfortheStantonnumber,St=q/Vcp(Tw
Tb),maybeobtainedfromEq.[81]
cp
h w [82]
V
f
St [83]
2
wheref=2w/V2istheFanningfrictionfactor(f=/4).
In order to check the validity of the Reynolds analogy, experimental values of the dimensionless
parameter2St/f(whichaccordingtoEq.[83]shouldbeequaltounity)havebeenplottedinFigure81
asafunctionofReynoldsnumber,Re,fordifferentvaluesofthePrandtlnumber,Pr.Theexperimental
curveshavebeendeducedbycombinationofdataforStfromEagle&Ferguson,andfrictionaldata
for f from Stanton and Pannell (both quoted by Goldstein (1950) [73]). The experiments by Eagle &
FergusonextendedoverarangeofvaluesofPrfrom3to10,andarangeofReynoldsnumbersfrom5
x103to2x105.Thefollowingempiricalcorrelationhasbeenusedtoplotthesecurves:
2St 1

Pr 1 Pr 12
[84]
f
where,andaregivenasfunctionsofRe,inTable65,p.659,byGoldstein.ThecurveforPr=1has
beenobtainedbyuseofthecorrelation(4)outsideitsrangeofprovenvalidity(3Pr10).
126
ECSSEHB3101Part13A
5December2011
Figure81:Theratio2St/f,forturbulentflowinconstantwalltemperature
cylindricaltubes,ascalculatedbyuseofseveralexpressions,vs.theReynolds
number,Re.E:Correlationofexperimentalresults.FromGoldstein(1950)[73].R:
ReynoldsAnalogy.P:PrandtlAnalogy.K:vonKrmnAnalogy.Calculatedbythe
compiler.
ItcanbededucedfromFigure81thattheReynoldsanalogyfailsundertwocircumstances:
1. ForfluidsforwhichPrsubstantiallydiffersfromunity.Thissobecauseoftheexistenceof
alayernearthewall(thelaminarsublayer)wherethetransferofheatandmomentumby
thermal conductivity and viscosity are of importance. These laminar transfer processes
are strictly similar when Pr = 1. For common gases, however, Pr = 0,7 and Reynolds
analogycanbeapplied.
2. When Re is not large, Eq. [83] cannot be applied when the flow is laminar. Its use for
transitionalflowshasnotbeenvalidatedbytheexperimentseither.Nevertheless,ithas
been indicated (Reynolds (1974) [149], p. 244) that the analogy gives fairly accurate
results in several instances oflaminar pipe flow.Letus consider,for example, the fully
developedlaminarflowthroughstraightpipesofcircularcrosssection.Forconstantheat
transferrateNu=4,364(seeclause7.2.1,Figure61),whereasf= /4=16/Re(seeclause
7.2.2).IntermsofStandf
2St 6
[85]
f 11Pr
Theanalogybeingmostaccurateforcommongases(Pr0,7).
127
ECSSEHB3101Part13A
5December2011
8.2.2 The Prandtl analogy

AccordingtoPrandtl,theReynoldsanalogyisstrictlyvalidonlyforthecentralcoreoffullydeveloped
turbulentflow,outsidethelaminarsublayer.
Inthatcase,Eq.[81]becomes
q c p T Tb
[86]
w V u
wheresubscriptreferstoconditionsattheedgeofthelaminarsublayer.
Inthelaminarsublayer.whosethicknessis ,wherethetemperatureandvelocityprofilesarelinear
functionsofthedistancetothewall,
k Tw T u
[87]
q w
EliminationofTbetween[85]and[86]yields:
f /2
St
1 Pr 1u / V
[88]
which reduces to Eq. [83] when Pr = 1. Eq. [87] takes partly into account the velocity distribution
acrossthetubethroughtheratiou/V.
InordertoapplyEq.[87]toparticularcasesitisnecessarytomakeasuitableassumptionaboutthe
ratioofthevelocity,attheouteredgeofthelaminarsublayertothevelocityfarfromthewall.Inthe
caseofturbulentflowinasmoothpipe(Schlichting(1960)[157]),
u f
5 [89]
V 2
withthisapproximationPradntlsequationbecomes
f /2
St
1 5 f / 2 Pr 1
[810]
CurveslabeledPinFigure81havebeendeducedfromEq.[88]withthevaluesoff(Re)tabulatedby
Goldstein(1950)[73].Alternativeexpressionsoffforsmoothtubescouldbealsoused(seeclause7.2.2
andTable72ofclause7.2.2.1).Itcanbeseenthattheanalyticalresultsarenowinbetteragreement
withtheexperimentaldatathantheywerewhentheReynoldsanalogywasused.
128
ECSSEHB3101Part13A
5December2011
8.2.3 The Von Karman analogy

Von Krmn improved further the Prandtl analogy by introducing an intermediate buffer layer in
which molecular and turbulent diffusion were active. He maintained the same boundaries for the
several layers of the velocity and temperature variations, and assumed that the coefficients for
turbulenttransportofheatandmomentumwerethesame.
TheresultingexpressionfortheStantonnumberis
f /2
St
1 5 f / 2 Pr 1 ln1 5 / 6 Pr 1
[811]
ThereaderinterestedinthedetailsleadingtothisresultcouldconsultthebooksbyGoldstein(1950)
[73],Hinze(1959)[89]orSchlichting(1960)[157],amongothers.
Curves deduced from Eq. [89], with the values of f(Re) tabulated by Goldstein (1950) [73], are
represented in Figure 81. It can be seen that the agreement between these curves and the
experimentalonesisexcellentexceptforthelowerpartoftheReynoldsnumberrange.
8.2.4 Other analogies

Manyotherattemptshavebeenmadetorelateheattransfertofluidfrictioninturbulentpipeflows.
Nevertheless, since the algebra of the more advanced models becomes very complicated, only two
examples which will be used in other clauses of this Part will be presented here. A short survey of
analogiescanbefoundinKnudsen&Katz(1958)[107],pp.417455andinReynolds(1974)[149]pp.
27273.
1. MartinelliretainedvonKrmnsdistributionforturbulenttransportcoefficientsofheat
and momentum, but he allowed these coefficients to be different in the viscous core,
assuminginsteadthesamelineardistributionfortherateofheattransferacrossthetube
asexistsfortheshear.
AdetaileddescriptionofthisanalogycanbefoundinKnudsen&Katz(1958)[107].
2. Colburn analogy. This simple and very useful analogy, which is entirely based on
empiricaldata,iswrittenas:
f
j [812]
2
fbeingthefanningfrictionfactorandjtheColburnfactorforheattransferj=StPr2/3.See
Knudsen&Katz(1958)[107]fordetailsonthisanalogy.
129
ECSSEHB3101Part13A
5December2011
9
Heat transfer enhancement
9.1 General
Theperformanceofconventionalheatexchangerscanbesubstantiallyimprovedandequipmentsize
reducedbyanumberofaugmentativetechniques.Existingsystemscanoftenbeupgradedbyusing
anaugmentativemethod,whileinthespaceborneheatexchangersanaugmentativeschememaybe
mandatoryinorderforthesystemtofunctionproperlyandmeetthesize(and/orpower)limitations
imposed. This may be particularly true for systems transferring heat between a liquid which is
normallyaneffectiveheatcarrierandagaswhichisnotsoeffective.
An introduction to the literature on the subject has been given by Bergles (1969) [8]. References to
morerecentstudiescanbefoundinBergles(1973)[7].
Table 91, from Bergles (1973) [7], classifies both techniques and modes of heat transfer. The
techniques are grouped according to passive methods, which require no external power, and active
schemes, which require external power. A final category, compound augmentation techniques,
enclosesthosesituationswheretwoormoreofthesebasictechniquesareutilizedsimultaneously.
AXinTable91meansthatthetechniquehasbeenused,accordingtotheliterature,toenhancethe
heattransfermode.
Table91:HeatTransferEnhancementTechniques
ModeofHeat Single Single Pool Flow Condensation
Transfer Phase Phase Boiling Boiling
Free Forced
Technique
Convection Convection
PASSIVE
TreatedSurfaces X X X
RoughSurfaces X X X
ExtendedSurfaces X X X X X
DisplacedEnhan.Devices X X
SwirlFlowDevices X X
SurfacesTensionDevices X
AdditivesforLiquids X X X
130
ECSSEHB3101Part13A
5December2011
ModeofHeat Single Single Pool Flow Condensation

Transfer Phase Phase Boiling Boiling
Free Forced
Technique
Convection Convection
AdditivesforGases X
ACTIVE
MechanicalAids X X
SurfacesVibration X X X X X
FluidVibration X X X X
ElectrostaticFields X X X X
InjectionorSuction X X X
COMPOUND X
NOTE FromBergles(1973)[7]
This clause presents data concerning the singlesphase forced convection mode of heat transfer
(second row in table below). Techniques involving complicated or massive devices, such as surface
vibrationorelectrostaticfields,whosereliableandsageuseinspaceisopentoquestion,havenotbeen
considered.
9.1.1 Basic augmentation mechanisms

Themostusualpassivetechniquesforimprovingsinglephaseforcedconvectionarebasedonone,or
more, of the following mechanisms: 1) Turbulence promotion; 2) Extension of the heat transfer
surface,and3)Swirlingoftheflow.
1. Turbulence promotion. Transport processes are more effective under turbulent than
underlaminarconditions.Nevertheless,therateofheattransfertoasurface,nomatter
theturbulenceoftheflow,iscontrolledbyapurelymolecularpropertyofthefluid,its
thermalconductivity.Thisissobecauseofthelaminarsublayerwhichappearscloseto
thesurface.
The resistance to heat transfer of the laminar sublayer is proportional to its thickness,
thenceanyreductionofthethicknesswillresultinanincreaseintheheattransfer.
One means of reducing the thickness of the laminar sublayer consists in increasing the
level of turbulence of the flowing stream. This can be achieved by use of turbulence
promoters,eitherattachedtoordetachedfromtheheattransfersurface.
2. Extension of the heat transfer surface. It is clear that, even in the absence of turbulence
promotion, the heat transfer coefficient referred to the nominal area increases by the
additionoffins.
Inordertogetthedesiredincreaseintheheattransfercoefficient,especiallywhenliquid
heatcarriersareused,goodthermalcontactbetweenthefinandthetubewalliscrucial.
131
ECSSEHB3101Part13A
5December2011
3. Swirlingoftheflow.Swirlingimprovestheheattransferduetoincreasedvelocitynear
thetubewall,andtoacentrifugalconvectioneffectwhichispresentwhenthedirection
ofheatflowisfromlargertosmallerradii.
Swirlingcanbeeasilyinducedbymeansoftwistedmetallicstripsinsertedintheflowduct.
Inmanycasestheaugmentationachievedwithagiventechniquemaybeattributedtovariousofthe
mentionedmechanisms,renderingitdifficulttheabsoluteclassificationofthetechnique.Forexample,
spiralfinscouldsimultaneouslyprofitfromthethreebasicmechanisms.
Althoughseveralactivetechniquesaimatscrapingawayorsuckingoffthelaminarsublayer,the
augmenting effect is obtained in many other systems by means of periodic flow oscillations
superposedtothesteadyflow.
The effect of low frequency pulsations, which probably have the most potential for practical
application,canbeeasilyunderstoodonthebasisofaquasisteadyapproach.Atverylowfrequencies
onecanassumethattheinstantaneousheattransferrateisequaltothatcorrespondingtothesteady
flowattheinstantaneousvelocity.Thefollowingconclusionsresult:
1. Slowfrequencyoscillationsdonotinfluencethefullydevelopedlaminarflow(provided
that no transition is induced) since the steady flow heat transfer rate, h, is velocity
independent.
2. Whentheoscillationsaresuperposedtoaturbulentflowadecreaseintheoscillationisso
largethatreverseflowappearsduringpartofthecycle.
3. Whentheoscillationsaresuperposedtoaturbulentflowadecreaseintheheattransfer
results,unlesstheamplitudeoftheoscillationissolargethatreverseflowappearsduring
partofthecycle.Thereasonisthat,beingtheheattransferrateproportionaltoapower
ofthevelocitycloseto0,5,thegainachievedbyincreasingtheinstantaneousvelocityina
givenamountislessthanthelossresultingfromdecreasingitinthesameamount.
Fluid oscillations may be generated either vibrating the duct, pulsating the mass flow rate, or
acousticallydisturbingthefluidflow.
Thefirstprocedure,althoughusedtocoolsmallcomponentssubmergedinacoolantbath,isdifficult
to apply and, as a general rule, leads to mechanical failure of the equipment in a short time.
Modulation of the mass flow rate can be induced by means of a pistontype pump. Acoustic fields
applied to air flows are generally impractical because of the required sound intensities well above
humantolerance.Withliquidsthereisconsiderabledifficultyindisturbingsystemsofnottoosmall
sizes.
9.1.2 Criterion for the evaluation of the several techniques

Toestablishagenerallyapplicableselectioncriterionforaugmentativetechniquesisdifficult.Evenif
costconsiderations,andreliabilityandsafetyfactorsaredisregardedwhichshouldnottheproblem
arisesthattheaugmentativetechniqueincreasesatconstantvelocityboththeheattransferandthe
friction.
It appears that among the many available ways to relate heat transfer and friction for evaluation
purposes, the most favorable one consists in representing the ratio of augmented, ha, to non
augmented,ho,convectiveheattransfercoefficientforconstantpower,P.Therefore,thisratiohasbeen
usedinallcasesthroughoutthepresentitem.
The comparison on the basis of equal pumping power is natural when one aims ate upgrading the
heattransfercapabilityofanexistingheatexchangerorwhen,ashappenstobethecaseinspacecraft,
theavailablepowerislimited.
132
ECSSEHB3101Part13A
5December2011
When (ha/ho)P is larger than unity, the heat transfer rate for a constant temperature difference, TwTb,
canbeincreased,orforthesameheatflux,q,thetemperaturedifferencecanbedecreased,atnoextra
pumping power consumption. Even when (ha/ho)P < 1, a gain in the performance of the loop could
result, since the decreased mass flow rate reduces the pressure loss in the remainder of the system,
reducing therequired overall pumping rate. In order to ease theestimation of this reduction by the
user, data on both heat transfer and wall friction under augmentative conditions are presented by
meansofempiricalcorrelations.
Regarding active techniques, P will still be the pumping power but the required external power
shouldbetakenintoaccountintheevaluationofthetechnique.
9.1.3 Index of the compiled data.

Thedatawhicharepresentedinclause9.2areindexedintheTableshownbelow.Twopointsshould
betakenintoaccountinconnectionwiththeclassificationoftheseveraltechniques.
1. Somedifficultycouldarisewhenidentifyingagivensystemeitherassurfaceroughness
orasextendedsurface.Insertsforwhichnoparticularcarehasbeenexercisedtooptimize
solidheattransfertotheheatedsurfacehavebeenclassedeithersurfaceroughnesswhen
theycontactthesurfaceorextendedsurfaceswhentheydonot.Twistedtapeswhichdo
contact the surface are the single exception to this rule. Nevertheless, no matter how
conductivethiscontactmightbe,thesetapesswirltheflowandaregroupedinaseparate
class.
2. Anotherdifficultyappearswhentryingtodistinguishbetweenmachinedroughnessand
othersmallinternalprotrusions.Providedthattheseprotrusionsareparalleltothetube
axisorslightlyspiralled,theydonotincreasethelevelofturbulenceandareclassedas
fins,otherwisetheywouldbeincludedundersurfaceroughness.
9.1.4 Validity of the empirical correlations

The validity of the empirical correlations given in clause 9.2 rests upon the fulfillment of several
conditions,whicharelistedsidebysidewiththecorrelation.Theseconditionsarebasedeitheronthe
particularexperimentdevisedtoobtainthecorrelation,onthetheoreticalbackgroundonwhichitis
based,oronboth.Sometimestheoreticalandexperimentalconditionsconflict.
The user interested in applying the correlations under slightly different conditions should take into
accountthefollowingconsiderations.
1. Lengthtodiameterratio,L/D,requiredtoachievefullydevelopedconditions.
In most cases the value L/D listed under Conditions is that of the experimental set up.
The correlation can be used for other values of L/D provided that they are larger than
both the thermal and frictional entrance lengths. See clause 6.3.1 Figure 66 to Figure
612), clause 6.3.1 (Figure 623 to Figure 627) and clause 7.2.5 for estimates of these
lengths.
2. Temperaturedependenceofthefluidproperties.
The convective local heat transfer coefficient, hx, is calculated in terms of the local
temperaturedifference,(TwTb)x,asisindicatedinclause6.2.1.
The fluid properties are evaluatedat the bulk temperature, Tbx,of the tube section, x, as
definedinclause6.2.4.
133
ECSSEHB3101Part13A
5December2011
Tb
AFL
uTdAFL
[91]
VAFL
Inseveralinstancesthefluidpropertiesareevaluatedattheaveragetemperature,Tavx,of
thesection,x.
Tav
AFL
TdAFL
[92]
AFL
The difference between both temperatures is insignificant, particularly for turbulent

flows,andwhenthevariationofTalongthetubeissmall(Reynolds(1974)[149],pp.174
175).
Both procedures are, however, inadequate when large wall to fluid temperature
differentials exist. In these instances a loading factor, of the form (Tb/Tw)m or (b/w)m,
mustbeintroduced(seeclause6.2.4andclause7.2.3).
Inmanypracticalinstancestheheattransfercoefficientisdeducedbyuseofanaveraging
experiment.Thishappenstobethecase,forexample,whentheheattransferisrelatedto
thepowerrequiredtoholdconstantthetemperatureofahighlyconductivetube.Since
the bulk temperature do change along the duct, the average value of the heat transfer
coefficientarebasedonanaveragebulktemperature.
3. 3)Heatfluxdistribution.
The case of constant heat fluxis fairly easily dealt with theoretically, nevertheless most
experimentsareperformedundernearlyconstanttemperaturedifferentials.Normally,a
heat transfer coefficient based on the local temperature differential will be lower for a
constant wall temperature than for constant flux, since the large differentials upstream
leavethetemperaturenearthewallhigherthanitwouldbeforaconstantdifferential.
TheTable92belowindicatesthatforturbulentflow,thedifferenceintheheattransfercoefficientis
virtually negligible, compared with other sources of uncertainty, except at very small Prandtl
numbers.
Table92:Ratioofheattransfercoefficientsforconstantwalltemperature,NuT,
andconstantheatflux,Nuq,forturbulentpipeflow.FromReynolds(1974)[149].
PrRe 106 105 104 103 102 10
NuT/Nuq 0,96 0,95 0,90 0,83 0,75 0,73
ForlaminarflowstheeffectislargerascanbeseeninFigure61,clause6.3.1.1.HoweverNuT/Nuqis
largerthan0,8inmostcases.
134
ECSSEHB3101Part13A
5December2011
IndexofCompiledData
SinglePhaseForcedConvection
Technique Main Description Duct Fluid EnclosedData Fig.

Operating
(ha/ho)p orf NuorSt
Mode
PASSIVE
Rough Inducing CommercialRoughness Tube

Surface Transition
toTurbulence SandGrainRoughness Water X X X Figure
91
MachinedRoughness
SmallRingTypeInserts Several X X X Figure

93
LargeRingTypeInserts
WireCoilRoughness
MeshInserts
ProtuberancesorGroves Annulus Water X X X Figure

96
Extended Increasing StraightInternalFins Tube Air

Surfaces Transfer
Area Water X X X Figure
99
LowSpiralInternalFins X X X Figure
910
HighSpiralInternalFins X X X Figure
911
ExternallyFinnedTubes Annulus
Displaced Inducing RingandDiscs Tube

Enhancement Transition
Devices toTurbulence Discs Water
StreamlinedBodies X Figure
912
Rings Annulus
SwirlFlow Increasing TwistedTapeInserts Tube Liquid X X X Figure

Devices FluidPath 913
Length
Air X X X Figure
914
Water X X X Figure
916
135
ECSSEHB3101Part13A
5December2011
X X X Figure
917
BrushTypeInserts
PropellerTypeInserts
InletVortexGenerators
ACTIVE
Surface Removing RotatingBlade Wallin

Scrapers thelaminar Channel
Sublayer

Fluid Increasing AcousticVibrations Tube

Vibration Local
Convection MechanicallyInduced
Vibrations
9.2 Single-phase forced convection data
SANDGRAINTYPEROUGHNESS
TURBULENTFLOWINTUBES
WATER
Figure91:Constantpowerheattransferratio,(haho)P,vs.Reynoldsnumberbased
onnonaugmentativeconditions,Reo.FromBergles(1969)[8].
136
ECSSEHB3101Part13A
5December2011
Explanation
Key A1 B1 C1 D1 A2 B2 C2 D2 A3 B3 C3 D3
Investigator Dipprey&Sabersky(1963)[35]
Dx103[m] 10
L/D(Diabatic) 38
(e/D)x104a 24 138 488
Ti[K] 416,5b 338 312,5 300 416,5b 338 312,5 300 416,5b 338 312,5 300
(TwTb)[K]c 0
Pr 1,20 2,79 4,38 5,94 1,20 2,79 4,38 5,94 1,20 2,79 4,38 5,94
Rea 1,4x1045,2x105
Heating/Cooling H
a Theroughnessisdefinedbymeansofthehydraulicequivalentsandgrandsize.Thissizeisdeterminedby
comparingthefrictionfactorofthepresentpipeswiththatpredictedbythevonKrmnNikuradselawfor
thecompletelyroughrefine(see7.2.2).
b Thewaterispumpedbynitrogengaspressure.Thepressurelevelisnotquotedinthesource.
c Theempiricalcorrelationsdeducefromtheseexperimentaldataarebasedontheconstantfluidproperties
assumption.Testsatthreedifferentvaluesoftheheatfluxwereperformedforeachofthenominal
combinationsofReynoldsnumbersandPrandtlnumber.Toobtainthereportedisothermalheattransfer,
devoidedoftheeffectsofradialtemperaturegradients,thedatafromeachtestwerefirstadjustedto
correspondtothenominallocalReynoldsandPrandtlnumberconditions.Theadjustedheattransfer
coefficientswereplottedagainstwalltofluidtemperaturedifference.Anextrapolationofthestraightlinebest
approximatingthethreepointsprovidedthezerotemperaturedifferencevalue.Thisvalueistakento
representtheisothermalheattransfer.
Thethreeroughtubesusedfortheexperimentswereproducedbyelectroplatingnickelovermandrels
coated with closely graded sand grains. The mandrels were subsequently dissolved with chemicals,
leaving a pure nickel shell which served as the test tube. The smooth reference tube was also
producedbythesamemethodusingasmoothmandrel.
137
ECSSEHB3101Part13A
5December2011
EmpiricalCorrelations Conditions Comments
DISPREY&SABERSKY(1963)[35] Straighttubeofcircularcross ue+(e+)forNikuradsesandgrain

section.Fluidproperties roughnessisgiveninFigure
2 D

2,5 ln ue e 3,75 evaluatedatTb.RebasedonDM. 92b
f 2e Re>2x103,L/D>2540. Whene+>67,ue+=8,48,thence
oritsequivalent Statisticallyuniformroughness thevonKrmnNikuradselaw
pattern,geometricallysimilar offrictionforthecompletely
1 D
2,035 log fromtubetotubewithonlyone roughregimeresults.(See7.2.2).
2e scalefactor.
0,354ue e 1,323

f /2 Straighttubeofcircularcross Itisassumedthatthetime
St c

1 f / 2 g e , Pr 8,48 section.Fluidproperties meanflowinandaboutthe
evaluatedatTb,andassumedto cavitiesofdeptheconstituting
beconstant. theroughwallisinducedbya
RebasedonDM.Re>2x103,e+> patternofstandingvortices
67,L/D>2540.Statistically withaxesnormaltothemean
uniformroughnesspattern, velocity.Thencethefunctiong
geometricallysimilarfromtube isexpressedas:
totubewithonlyonescale c p ue Tw Te
factor. g Ste1
q
Constantheatfluxatthewall.
Theanalysisoftheexperiments
TheexpressionforStisbased
leadsto:
onthefollowingfouradditional
g=5,19(e+)0,2Pr0,44
assumptions:
Theprovenrangeofvalidityof
1)Closedtothewallthevelocity thisexpressionis:
distributiondependsexclusively e+>65,1,2<Pr<6.
onthelocalconditions(lawof
thewall).
2)Thelawofthewallappliesto
thetemperaturedefectprofile,
TwT,aswellastothevelocity
profile,u.
3)TheReynoldsanalogyisvalid
inthefullyturbulentcore.
4)Velocityandtemperature
reachtheircrosssectionalmean
valuesatthesamedistancefrom
thewall.
OWEN&THOMSON(1963)[138] Sameasabove Thetimemeanflowisinduced

f /2 byhorseshoevorticeswhich
St c

wrapthemselvesaroundthe

1 f / 2 g e , Pr 17,8 f / 2
individualexcrescencesand
traildownstream.Themain
differencewiththeabovemodel
appearsinthevelocitynearthe
138
ECSSEHB3101Part13A
5December2011
edgeoftheroughnesselements
whichhereisassumedtobe
smallcomparedtou*.
g(Ste)1=0,52(e+)0,45Pr0,8
undertheconditions
e+>67,0,7<Pr<7,5.
CommentsconcerningthedatapresentedinFigure91a.
ThegainobtainedbyrougheningincreaseswiththeroughnessandwiththePrandtlnumber.
ForeachvalueofeandPrthereisavalueofRewhichoptimizestheratio(haho)P.Theseoptimumconditions
occurinthetransitionregionratherthaninthecompletelyroughregion.
Figure92:Roughnessfunctionue+(e+)forNikuradsessandroughness.(1)Hydraulically
smooth.(2)ue+=8,48,completelyrough.FromSchlichting(1960)[157].
SMALLRINGTYPEINSERTS
SEVERALFLUIDS
onnonaugmentativeconditions,Reo.CurvesAtoDarefromBergles(1969)[8],
curvesEandFhavebeencalculatedbythecompilerafterWebb,Eckert&
Goldstein(1971)[186].
139
ECSSEHB3101Part13A
5December2011
Explanation
Key A B1 B2 C D1D2 D3 D4 E1 E2 E3 F1 F2 F3
Investigator Nunner(1956)[134] Webb,Eckert&Goldstein(1971)[186]
Dx103[m] 50 36,8
L/D(Diabatic) 20 41
Roughnessa
Shape
(e/D)x104 400 800 200
s/e 1,25 2 7,5 2 0,52
b/e 20,420,4 81,7 2 5,1 10,2 20,4 10 20
Ti[K] 300320
(TwTb)[K] 2080
Fluid Air Air Water 1Butanol Air Water 1Butanol
Pr 0,71 0,71 5,10 21,7 0,71 5,10 21,7
Reod 5x1027x104 6x103105
Heating/Cooling H H
a Theseconfigurations,usefulforlaboratorystudies,aredifficulttoadapttocommercialpractice.
b
c Notgiven.Intheliquidtests,thepowerinputwasselectedtogivePr/Prw1,1.
d NunnerdefinestheReynoldsnumberonthebasisofDM,andofWebb,Eckert&GoldsteinonthebasisofD.
140
ECSSEHB3101Part13A
5December2011
WEBB,ECLERT&GOLDSTEIN(1971) Straighttubeof ue+(e+,b/e)forrepeatedribroughnessis

[186] circularcrosssection. giveninFigure94b.
Fluidproperties Whene+30,ue+=0,95(b/e)0,53.
2 D

2,5 ln ue e , b / e 3,75 evaluatedatTb. Thefollowingcorrelationhasbeen
f 2e RebasedonD.Re>2x usedbythecompilerfordata
oritsequivalent 103 reduction
0 , 53 0,95 10b / D
Repeatedrib 0 , 05 e
1 D ue b / e

2,035 log
exp 0,0343e
roughness.

2e
0,354ue e , b / e 1,323

for e / D 0.1

f /2 Straighttubeof Theanalysisoftheexperimentsleads
Stb
1
e
f / 2 g e , b / e, Pr u e, b / e
circularcrosssection. to:
Fluidproperties g=4,50(e+)0,28Pr0,57
evaluatedatTb,and theeffectofb/ebeinginsignificant.
assumedtobe Theprovenrangeofvalidityofthis
constant. expressionis:
RebasedonD.Re>2x e+>25,0,71<Pr<38
103,L/D>2540. Thefollowingexpressioncanbeused
Repeatedrib toaccountfortheeffectofradial
roughness.Heat temperaturegradients.
transfercoefficientin m
St Prb
termsofthetubearea.
TheexpressionforSt Stb Prw
isbasedonthe m=0,15forsmoothtubes
followingadditional m=0,25forroughtubes.
assumptions:
1)Closedtothewall
thevelocity
distributiondepends
exclusivelyonthe
localconditions(law
ofthewall).
2)Thelawofthewall
appliestothe
temperaturedefect
profile,TwT,aswellas
tothevelocityprofile,
u.
3)TheReynolds
analogyisvalidinthe
fullyturbulentcore.
4)Velocityand
temperaturereach
theircrosssectional
meanvaluesatthe
samedistancefrom
thewall.
141
ECSSEHB3101Part13A
5December2011
GOMELAURI(1974)[74] Straighttubeof (b/e)=[(b/e)opt/(b/e)]

Nu=0,021Re0,8Pr0,43(Prb/Prw)0,25r circular,annularor whenb/e(b/e)opt
rectangularcross (b/e)=[(b/e)/(b/e)opt]
r=1,04Pr0,04exp[0,85(b/e)]
section.Fluid whenb/e(b/e)opt
propertiesevaluated (b/e)opt=12to14
atTbam.
NuandRebasedon
DE.,104<Re<5x105,
0.7<Pr<250.
AnempiricalcorrelationforSthasbeenalsogivenbyNunner(1956)[134].Althoughthiscorrelationfits
thedataforringinsertsatPr=0,7,itdeviatesforothertypesofroughnessandhigherPrandtlnumbers.
SeecriticismsbyBergles(1969)[8],Webbetal.(1971)[186]andGomelauri(1974)[74].
Isothermalfrictiondatahavebeenusedinthecomputationsleadingto(haho)PforthecurvesfromWebb
etal.
Theoptimumvalueofb/eisabout1020forthecasesconsidered.Theexistenceofthisoptimumisrelated
totheflowpatternnearthewall(Figure94c).Separationoccursattherib,andreattachment68rib
heightsdownstreamfromtheseparation.Thelocalheattransferismaximuminthereattachmentzone.
Reattachmentdoesnotoccurforb/e<8.Nevertheless,becauseoftherecirculation,thelocalheattransfer
intheseparatedflowregionislargerthanthatcorrespondingtoanattachedboundarylayer.
ThegainachievedbyrougheningincreaseswiththePrandtlnumber.
Figure94:Roughnessfunctionue+(e+,b/e)forrepeatedribroughness.FromWebbet
al.(1971)[186].
Figure95:Flowpatternnearthewallfordifferentvaluesofb/e.
142
ECSSEHB3101Part13A
5December2011
DISCRETEROUGHNESS
TURBULENTFLOWINANNULI
AIRORWATER
onnonaugmentativeconditions,Reo.CurvesA,B,CarefromBergles(1969)[8],
curvesDtoGhavebeencalculatedbythecompilerafterSheriff&Gumley(1966)
[166].
Explanation
Key A B C D E F G
Investigator Kemeny&Cyphers(1961) Sheriff&Gumley(1966)[166]
[103]
D1x103[m] 12,7 26,2
D2x103[m]a 17,5 79,5
L/(D2D1)(Diabatic) 82 48
RoughnessShapeb
[e/(D2D1)]x104c 276 244 281 23,8 95,2 135,7 190,5
s/e 2 1
b/e 10 10
Ti[K] 300320 310
(TwTb)[K] 10 50
Fluid Water Air
Pr 5,5 0,71
Rea 4x1038x104 1042x105
Heating/Coolingd H H
143
ECSSEHB3101Part13A
5December2011
a Sheriff&Gumleybasetheeffectivediameterandheattransferareaonanannuluswiththeinnersurface
takenasthecylinderenvelopingtheroughnesspeaks.Thediametergivenherecorrespondstothebase
surfaceoftheroughnesselements.
b Helicalelementwhosecrosssectionissketched.
c Roughnesselementsaddedtotheinnertube.Theoutertuberemainssmooth.
d Heatistransferredfromtheinnertube.Theoutertubeisadiabatic.
144
ECSSEHB3101Part13A
5December2011

SHERIFF&GUMLEY(1966)[166] Straighttubeofcircular u (e ,b/e)forwirecoil
e+ +

2 D crosssection. roughness,andb/e=10,is
2,5 ln H ue e , b / e 3,75 Fluidproperties giveninFigure97b.
fH 2e
evaluatedatTb Thefollowingcorrelationhas
oritsequivalent ReHbasedonDH.ReH>2 beenusedbythecompilerfor
1
2,035 log
DH

0,354u e e , b / e 1,323
x103, datareduction
H 2e L/DH>2540 ue+=4,93,6/e+fore+>1.
Wirecoilroughness. Thevalueofue+forthefully
subscriptHreferstotheHalltransformpassage.
Frictionfactorbasedon roughregimeissomewhat
Thetransformationisdescribedinthenext.
thecylindrical largerthanthatquotedinthe
surfaceenvelopingthe source,whichis4,65.Thisis
roughnesspeaks. duetothefactthatthedata
pointsfore=0,01in.,which
lookdefective,havebeen
excluded.
fH fH g e , b / e, Pr Straighttubeofcircular Theanalysisofthe
1
ue e , b / e m
crosssection.Fluid experimentsleadsto:
2St H 2 propertiesevaluatedat g=5,387(35+e+35)0,198
Tb,andassumedtobe Theprovenrangeofvalidity
constant. ofthisexpressionis:
ReHbasedonDH.ReH>2 e+>10,b/e=10,Pr=0,71
x103, Thevalueofmdependson
L/DH>2540 thevelocityandtemperature
Wirecoilroughness. distributions,whichinturn
Constantheatfluxatthe dependontheratioofthe
wall.Heattransfer eddydiffusivities,M/H.
coefficientbasedonthe Valuesofmaretobefound
cylindricalsurface fromthemeasuredvelocity
envelopingthe andtemperatureprofiles.A
roughnesspeaks.The meanvaluem=1,8is
expressionforStisbased quotedinthesource.
onthefollowingfour
additionalassumptions:
1)Closetothewallthe
velocitydistribution
dependsexclusivelyon
thelocalconditions(law
ofthewall).
2)Thelawofthewall
appliestothe
temperaturedefect
profile,TwT,aswellas
tothevelocityprofile,u.
3)Inthefullyturbulent
corethedimensionless
temperatureprofile,T+
T+maxandthe
dimensionlessvelocity
profile,u+maxu+,are
145
ECSSEHB3101Part13A
5December2011

similar.
4)Thesameprofiles
referredtotheir
respectivemeanvalues,
areslightlydifferent
becauseoftheaveraging
procedureusedto
calculateTb.The
difference(temperature
minusvelocityis
expressedasm.
Althoughgroovesarecheapertoproduceandarelesssubjecttoerosiondamagethanprotrusions,theyhave
littleeffectuntilthefreestreamturbulencepenetratesintothecavity.
Attached protrusions are less effective than integrally machined protrusions because of the thermal
resistance between wire and wall. This resistance plays a more important role the higher the overall heat
fluxbecomes.AtthelowestReynoldsnumbers,however,theslightlygreaterheightofthewireresultsinan
improvedheattransferovertheintegralprotrusion.
TheinfluenceoftheroughnesscanbeseenincurvesDtoG.Thegainobtainedbyrougheningincreaseswith
the roughness. For each value of e there is a value of Re which optimizes the ratio (haho)P. This optimum
valueisassociatedtotheminimumexhibitedbygate+=35.
ThetubeusedforcomparisonbySheriff&Gumley(1966)[166]isnotsmooth.ByuseoftheMoodydiagram
forcommerciallyroughpipes(Slichting(1960)[157])anequivalentsandroughnessofe/DH=4,2x104results
fromfrictiondataintherelevantReynoldsnumbersrange.Heattransferdataforthistubecanbecorrelated
bymeansofthefollowingformula:
Nu=0,0125Re0,845Pr0.4,(Pr=0,71)
The Reexponent (larger than 0,8) and the constant factor (smaller than 0,023) support the evidence of
roughness.
BothKemeny&CyphersandSheriffandGumleyusedapitchtoheightratiob/e=10..DatafromBrauer,
quotedbyBergles(1969)[8],indicatethattheoptimumb/efortheannulargeometryappearstotheabout3,
whichislowerthanthevalue,10,quotedfortubesofcircularcrosssection.
Figure97:Roughnessfunction,ue+(e+,b/e),forwirecoilroughness.Plottedbythe
compilerafterSheriff&Gumley(1966)[166].
146
ECSSEHB3101Part13A
5December2011
Hallstransformationofheattransferdataforannularducts
This transformation (Hall(1962)[78]) reduces the dataforannular passageswith only the core tube
heatedandroughenedtothoseforcircularcrosssectiontubeshavingthesamerelativeroughness.
Thevelocityandtemperaturedistributionsacrosstheannularpassageareassumedtobeknown.
1. Fromthevelocitydistributiontheradiusofmaximumvelocityisdetermined.Thisradius
definesacylindricalsurfaceatwhichthereisnoshearstress.Thisisnotstrictlytruefor
turbulent flow, as shown by Hanjalic & Launder (1972) [80], who used a channel with
parallelplates,oneofthemroughened,orbyRehme(1975)[148],usingconcentricannuli
withsmoothwalls.Thisfeature,however,isdisregardedhere.
2. By applying a force balance on the region of the annular passage between the inner
radius,D1/2,andtheradiusofnoshear,Dm/2,thefrictionfactorontheroughsurface,fH,
is related to the experimental pressure loss, dp/dx, to the kinetic energy of the flow
through the passage (per unit volume), (V2)H/2, and to the hydraulic diameter of the
transformedpassage,DH=(Dm2D12)/D1.
V 2 H H V 2
dp 1 4f 1 4f
[93]
dx 2 DH 2 DE
This equation relates also the average friction factor, f, for the whole of the annular
passagetothefrictionfactor,fH,forthetransformedpassage.
3. TheassumptionismadethattheresultingfrictionfactorReynoldsnumberrelationisthe
same as that which would exist for a circular tube having diameter DH and identical
relativeroughness.
4. Thetemperaturedistributionacrossthepassageisadjustedsothatthenoshearsurface
becomes adiabatic too (Figure 98c). The transformation is based on the following
features:
Figure98:VelocityandTemperaturedistributionsacrosstheannulus.
(a) Thelocalvalueofthecoefficientofthermaldiffusivityisassumedtobeunaffected
bytheconditionofnoheatfluxatDm.
(b) Theheatfluxattheroughenedsurfaceinthetransformedsystemisassumedtobe
thesameasintheexperimentalsystem.
147
ECSSEHB3101Part13A
5December2011
Applicationofathermalbalanceontheregionoftheannularpassagebetweenthe
inner radius, D1/2, and a given radius, r, for both the experimental and the
transformed system yields, after some manipulation, the following ratio of
transformedtoexperimentalradialtemperaturegradients:
T / r H mD1 , D2 mr , Dm

T / r mD1 , Dm mr , D2 [94]
b
with m(a, b) 2 urdr
a
Forgases,sincetheHalltransformationdoesnotchangethepressurefield, T1.
Forliquidsdisappearsaltogetherfromtheequation.
Equation above is Hintegrated to yield TH(r). The integration constant may be
chosensothateitherthewalltemperatureorthefluidbulktemperaturebecomes
invariant under the transformation. Although the second option looks
inconvenientfromthecomputationalpointofview,itsimplifiesthetransformation
ofthedatasincethefluidpropertiesaretheninvariantunderthetransformation.
(c) c.Theheattransfercoefficientsareintheinverseratiotothevaluesofthewallto
bulk temperature differences. The ratio of transformed to experimental Stanton
numbersisthen:
St H T T V
w b
St Tw Tb H V H [95]
5. AccordingtoSheriff&Gumley(1966)[166]thevelocityprofilesaregivenbymeansofa
logarithmic law, valid in the logarithmic layer near the rough wall, and a velocity
defect law, valid up to Dm. The experiments indicate that the defect law is universal,
withintheroughnessrangeconsidered(e/DE0.02),ashappenstobethecasefortubesof
circularcrosssection.Thence,DmisindependentofbothReande/DE,anditsvaluecanbe
deduced from a single experiment. In this particular case (DmD1)/(D2D1) = 0,476. This
value is very close to that which would be obtained if the flow were laminar, which
resultstobe0,454.Thisstrickingcoincidencehasbeenalreadynotedforsmoothannuli
(Knudsen&Katz(1958)[107]).
ThevalueofDmforlaminarflowisgivenby:
D22 D12
Dm
2 ln D2 / D1
[96]
Thefollowingtablegivesratiosoftransformedtoexperimentalvaluesofseveralvariablesasdeduced
fromoneoftheexperimentsbySheriff&Gumley[166].Alltheseratios,excepthH/h,canbeassumed
to be equal to unity. Values of hH/h for different Re and e are given by the authors in their Table 6.
AlthoughaslightReynoldsnumberdependencecanbenoted,thequotedvaluesdeviatefrom1,08by
lessthan2%.
148
ECSSEHB3101Part13A
5December2011
H/
Tb 1,013
h 1,080
V 1,006
V2 1,028
V/ 1,013
STRAIGHTINTERNALFINS
WATER
Figure99:Constantpowerheattransferratio,(ha/ho)P,vs.Reynoldsnumberbased
onnonaugmentativeconditions,Reo.FromCarnavos(1974)[19].
Explanation
Key
Investigator Carnavos(1974)[19]
Shape

NumberofFins 0a 6 10 12 14 20
Dx103[m] 26,0 11,5 14,2 11,6 18,0 25,3
DEx103[m] 26,0 7,87 8,53 6,99 10,8 14,3
149
ECSSEHB3101Part13A
5December2011
Key
e/D 0 0,148 0,111 0,094 0,099 0,073
b/DE 0 0,494 0,348 0,312 0,287 0,213
AFLx103[m2] 0,529 0,099 0,143 0,100 0,237 0,478
L/Db 94 211 171 210 135 96
NominalWt. 0,1995 0,0884 0,1088 0,0889 0,1381 0,1944

Area[m2]
UnfinnedWt. 0,1995 0,0684 0,0645 0,055 0,0872 0,1118

Area[m2]
FinnedWt. 0 0,0604 0,0986 0,0856 0,1318 0,2140

Area[m2]
TotalWt. 1,00 1,46 1,50 1,58 1,59 1,68

Area/Nom.Wt.
Area
a Smoothtubetestedforcomparison.
b L=2,435m.

CARNAVOS(1974)[19] FluidpropertiesevaluatedatTb.fa Fortheexperimentsreportedby
fa=0,406Re0,39(b/DEa)0,16 andRebasedonDEa. Carnavos(1974)[19]thestandard
5x103<Re<7,5x104 deviationis10,8%.
e/DEa<0,21 Nodetailsaregiveninthesource
0,21<b/DEa<0,50 ontheapparatusandprocedure
L/D100 usedintheseexperiments.
Checkedwithwater.
Nuab=0,212Re0,6Pr1/3(b/DEa)0,34 FluidpropertiesevaluatedatTb. Fortheexperimentsreportedby

NuabasedonDEaandtotalwetted Carnavos(1974)[19]thestandard
area.RebasedonDEa. deviationis10%.
Nua/Nub=(b/w)0,14 5x103<Re<x105 Nodataonthetemperaturelevel
e/DEa<0,21 prevailingintheexperimentsis
0,21<b/DEa<0,50 giveninthesource.Thefactor
L/D100 (b/w)0,14,accountingforthe
Checkedwithwater. temperaturedependentviscosityis
validintherange0,004<b/w<20.
(Seeclause6.2.4)
CommentsconcerningthedatapresentedinFigure99.
TheperformanceofstraightfinnedtubesapproachesthatofasmoothtubeastheReynoldsnumber
increases.
InthelowReynoldsnumberrange(Re104)straightfinnedtubesgivelowerperformancethanspiral
finnedtubes,comparedatainFigure99withthoseinFigure910andFigure911.
InthehigherReynoldsnumberrange(Re105)theperformanceofstraightfinnedtubesisnearly
150
ECSSEHB3101Part13A
5December2011
equaltothatofspiralfinnedtubes.
Foreaualtotalwettedareastheimprovementachievedbytheuseofinternalfinsincreaseswiththe
ratiooftheaveragedistancebetweenfins,b,tothehydraulicdiameter,DE.Thisiswhatmightbe
expecteduponinspectionoftheempiricalcorrelationswhichrevealthattheNusseltnumber
increasesmorerapidlywithb/DEathandoesthefrictionfactor.
Muchworkshouldbedonetodrawgeneraconclusionsconcerningtheoptimumfinconfiguration
andnumberoffins.Attemptshavebeenmadebyseveralauthors(Hu&Chang(1973)[92]),Masliyah
&Nandakumar(1976)[123])tooptimizeinternallyfinnedtubesforheattransfer.Theirtheoretical
approaches,however,whichdealonlywithfullydeveloped(bothfrictionalandthermal)laminar
flowandconstantheatfluxalongthewall,arefairlylimitedinscope.
LOWSPIRALINTERNALFINS
WATER
Figure910:Constantpowerheattransferratio,(ha/ho)P,vs.Reynoldsnumber
basedonnonaugmentativeconditions,Reo.FromCarnavos(1974)[19].
Explanation
Key
Shape

Numberof 0a 30 30 32 32 50
Fins
Dx103[m] 26,0 20,0 25,7 14,7 24,8 31,5
DEx103[m] 26,0 13,6 15,7 8,76 14,1 21,4
151
ECSSEHB3101Part13A
5December2011
Key
e/D 0 0,037 0,037 0,047 0,042 0,026
b/DE 0 0,123 0,129 0,116 0,120 0,072
y 0 3,81 2,97 10,4 3,46 3,12
AFLx103[m2] 0,529 0,299 0,517 0,161 0,479 0,761
L/Db 94 122 95 166 98 77
NominalWt. 0,1995 0,1531 0,1965 0,1127 0,1899 0,2413

Area[m2]
Unfinned 0,1995 0,0794 0,1224 0,0471 0,0874 0,0841

Wt.Area
[m2]
FinnedWt. 0 0,1556 0,2211 0,1330 0,2648 0,2880

Area[m2]
TotalWt. 1,00 1,54 1,75 1,598 1,855 1,54

Area/Nom.
Wt.Area
b L=2,435m.
CARNAVOS(1974)[19] Fluidpropertiesevaluatedat Fortheexperimentsreported

fa=0,535Re y
0,39 0,2 Tb.faandRebasedonDEa. byCarnavos(1974)[19]the
5x103<Re<7,5x104 standarddeviationis12,3%.
e/DEa<0,34 Nodetailsaregiveninthe
0,07<b/DEa<0,52 sourceontheapparatusand
3<y<17 procedureusedinthese
L/D100 experiments.
Checkedwithwater.
Nua=0,306Re0,63Pr1/3(b/DEa)0,21y0,27 Fluidpropertiesevaluatedat Fortheexperimentsreported

Tb.NuabasedonDEaand byCarnavos(1974)[19]the
totalwettedarea.Rebased standarddeviationis12,8%.
Nua/Nub=(b/w)0,14 onDEa. Nodataonthetemperature
5x103<Re<x105 levelprevailinginthe
e/DEa<0,34 experimentsisgiveninthe
0,07<b/DEa<0,52 source.Thefactor(b/w)0,14,
3<y<17 accountingforthe
L/D100 temperaturedependent
Checkedwithwater. viscosityisvalidintherange
0,004<b/w<20.
(Seeclause6.2.4)
152
ECSSEHB3101Part13A
5December2011
Spirallyfinnedtubesprovideconvectiveheattransferaugmentationbyextensionoftheheat
transfersurfaceandbyinductionofspiralflow.Hence,thesetubesgiveaslightlyhigher
performancethanstraightfintubes.Thiscouldalsobeexpecteduponinspectionoftheempirical
correlationswhichrevealthattheNusseltnumberincreasesmorerapidlywithspirallingthan
doesthefrictionfactor.Bycomparisonwithdataforstraightfinnedtubes(Figure99),itis
deducedthattheheattransferaugmentationduetospirallingmotionislargeratlowReynolds
numbers(Re104).
Thetopperformertubehas50triangularfinswithtwisty=3,12.Itisdifficult,however,todraw
generalconclusionsontheoptimumconfigurationfromthesedatabecauseofthewidevarietyof
geometriestested.Additionalworkshouldbedonewithtrapezoidalandtriangularshapedfinsto
betterunderstandtheirinfluenceonperformance.
Oneofthetubes(square)exhibitsasuddendecreaseinheattransferforaReynoldsnumbervery
nearlycoincidentwiththatcorrespondingtoanincreaseinfrictionfactor.Thereasonforthis
drasticperformancechangeisbynomeansobvious.
HIGHSPIRALINTERNALFINS
WATER
basedonnonaugmentativeconditions,Reo.FromCarnavos(1974)[19].
153
ECSSEHB3101Part13A
5December2011
Explanation
Key
Shape

Numberof 0a 6 8 16 16 16 16 16
Fins
Dx103[m] 26,0 11,8 30,4 10,7 14,0 17,2 20,4 25,4
DEx103[m] 26,0 7,54 17,93 4,47 6,60 8,59 11,30 14,96
e/D 0 0,158 0,149 0,144 0,119 0,105 0,097 0,088
b/DE 0 0,516 0,433 0,244 0,265 0,240 0,270 0,243
y 0 7,26 3,34 16,66 10,00 6,48 4,98 3,00
AFLx103[m2] 0,529 0,100 0,675 0,077 0,139 0,218 0,310 0,474
L/Db 94 206 80 228 174 142 119 96
Nominal 0,1995 0,0905 0,2327 0,0817 0,1070 0,1314 0,1562 0,1944

Wt.Area
[m2]
Unfinned 0,1995 0,0684 0,1518 0,0465 0,0580 0,0748 0,0969 0,1115

Wt.Area
[m2]
FinnedWt. 0 0,0598 0,2267 0,1282 0,1568 0,1750 0,1782 0,2081

Area[m2]
TotalWt. 1,00 1,417 1,626 2,139 2,00 1,90 1,76 1,64

Area/Nom.
Wt.Area
b L=2,435m.

154
ECSSEHB3101Part13A
5December2011
CARNAVOS(1974)[19] Fluidpropertiesevaluatedat Fortheexperimentsreportedby

fa=0,535Re0,39y0,2 Tb.faandRebasedonDEa. Carnavos(1974)[19]thestandard
5x103<Re<7,5x104 deviationis12,3%.
e/DEa<0,34 Nodetailsaregiveninthesourceon
0,07<b/DEa<0,52 theapparatusandprocedureusedin
3<y<17 theseexperiments.
L/D100
Checkedwithwater.
Nua=0,306Re0,63Pr1/3(b/DEa)0,21y Fluidpropertiesevaluatedat Fortheexperimentsreportedby

0,27 Tb.NuabasedonDEaand Carnavos(1974)[19]thestandard
totalwettedarea.Rebased deviationis12,8%.
onDEa. Nodataonthetemperaturelevel
Nua/Nub=(b/w)0,14 5x103<Re<x105 prevailingintheexperimentsisgiven
e/DEa<0,34 inthesource.Thefactor(b/w)0,14,
0,07<b/DEa<0,52 accountingforthetemperature
3<y<17 dependentviscosityisvalidinthe
L/D100 range0,004<b/w<20.
Checkedwithwater. (Seeclause6.2.4)
Spirallyfinnedtubesprovideconvectiveheattransferaugmentationbyextensionoftheheattransfer
surfaceandbyinductionofspiralflow.Hence,thesetubesgiveaslightlyhigherperformancethan
straightfintubes.Thiscouldalsobeexpecteduponinspectionoftheempiricalcorrelationswhich
revealthattheNusseltnumberincreasesmorerapidlywithspirallingthandoesthefrictionfactor.By
comparisonwithdataforstraightfinnedtubes(Figure99),itisdeducedthattheheattransfer
augmentationduetospirallingmotionislargeratlowReynoldsnumbers(Re104).
Thetopperformertubehas8trapezoidalfinswithtwisty=3,34.Itisdifficult,however,todraw
generalconclusionsontheoptimumconfigurationfromthesedatabecauseofthewidevarietyof
geometriestested.Additionalworkshouldbedonewithtrapezoidalandtriangularshapedfinsto
betterunderstandtheirinfluenceonperformance.
Oneofthetubes(square)exhibitsasuddendecreaseinheattransferforaReynoldsnumbervery
nearlycoincidentwiththatcorrespondingtoanincreaseinfrictionfactor.Thereasonforthisdrastic
performancechangeisbynomeansobvious.
Avalue(ha/ho)P=2,19forRe104,whichisnotenclosedinFigure911,isalsogiveninthesourcein
connectionwiththistube.
155
ECSSEHB3101Part13A
5December2011

STREAMLINEDAXIALBODIES
WATER
basedonnonaugmentativeconditions,Reo.FromBergles(1969)[8].
Explanation
Key A B C D E
Investigator Evans&Churchill(1963)[61]
Dx103[m] 25,53
L/D(Diabatic) 64
Dbx103[m]a 15,88 19,05 22,23
Db/D 0,622 0,746 0,871
Lb/Da 4 8 12
DTx103[m]a 3,17
Ti[K] 300
(TwTb)[K] 1050b
Rea 5x1035x104
Heating/Cooling H
NOTE Thestreamlinedbodiesarecomposedofaconeandahemisphere,machinedfromnylon.Theyare
fittedonacentralrodthroughaholedrilledintheaxisofthebody.Thebodiesareheldinplaceby
smallscrews,andcenteredbymeansofthreealuminiumpins,installedatthesectionofmaximum
diameterandjusttouchingthetubewall.
156
ECSSEHB3101Part13A
5December2011
a
b Estimatedbythecompilerfromreportedvaluesoftheheattransfercoefficientandpowerspentinheating
electricallythewall.
TWISTEDTAPEINSERTS
LAMINARFLOWINTUBES
LIQUIDS
basedonnonaugmentativeconditions,Reo.CalculatedbythecompilerafterHong
&Bergles(1976)[91].
157
ECSSEHB3101Part13A
5December2011
Explanation
Key A B C D
Investigator Hong&Bergles(1976)[91]
Liquid Water Ethylene Glycol
Dx103[m] 10,2
L/D(Diabatic) 120
y 2,45 5,08 2,45 5,08
tx103[m] 0,46
Tapefita 5
mx103[kg.s1] 5,59,5 1,833
qx103[W.m2] 0,538,5 5,227,6
Pr 3,07,0 84192
Rea 832460 13390
Heating/Cooling H
a Thetapefitsarecategorizedbymeansofnumbersasfollows:
NOTE 1.Specialcaretakentoincreasethethermaljointconductance.
2.Tighttapefit.Nomeasurablegapbetweenthetapeandthetubewall.
3.Snugtapefit.Gapoflessthan0,25x103m.
4.Loosetapefit.Gapestimatedtobegreaterthan0,25x103m.
5.Specialcaretakentoreducethethermaljointconductance.Inthepresentcasethetapeswere
coveredatbothedges2,5x103widestripsofblackNo.33Scotchelectricaltape,andbackedina
furnaceat530Kforaboutoneminute.
158
ECSSEHB3101Part13A
5December2011

HONG&BERGLES(1976)[91] Straighttubeof Datafortubeswithtwisted
2 semicircularcross tapeinsertsfromHong&
63,06
2 2 dt D
a
Re 1,463 L
section.Diameter,D. Bergles(1976)[91]correlate
4 d
t
a Fluidproperties withtheanalyticalresultsof
withthisexpressionfora: evaluatedatTb. Sparrow&HajiSheikh(1966)
L1 RebasedonD.Re<2,4 [169]forcirculartubeswitha
P a V 2 x103 diametraldividingwall.The
D2 Sparrow&HajiSheikh additivetermaccountingfor
(1966)[169]. entranceeffectsshouldbe
takenintoaccountforRe103.
Thereductionsinpressure
dropbecauseofheatingwere,
intheseexperiments,too
smalltobeaccurately
assessed.
64 D Straighttubeofcircular Seeclause7.2.5.
o 1,333 crosssection. Theentrancelength
Reo L
Fluidproperties coefficient,K=1,333,isfrom
HagenPoiseuilleformula
evaluatedatTb. Sparrow&HajiSheikh(1966)
RebasedonD.Re<2,4 [169].
x103
5/ 4 Fluidproperties Alinearvariationinbulk
Re
Nu a 5,172 1 5,484 10 3 Pr 0, 7 a evaluatedatTb. temperaturefromtheinletto
5 Constantwall theexitoftheheatedlengthis
temperature. assumed.
NuandRebasedonD. Wheny,Nua5,172.
10<Re<2x103 Thisvaluecorrespondstoa
3<Pr<200 straightcirculartubewitha
2,5<y<5 diametraladiabaticdividing
L/D=120 wall.
Nuo=3,656 Straighttubeofcircular Thelimitationimposedon

crosssection. GrPrshouldbefulfilledto
Fluidproperties keepthebuoyancyeffectsata
evaluatedatTb. negligiblelevel.Moreprecise
Constantwall dataregardingthislimitation
temperature. canbefoundinESDU69004
NubasedonD. (1969)[50].
GrPr<101forRe>1
(seecomments)
L/D>0,125PrRe.
Knudsen&Katz(1958).
159
ECSSEHB3101Part13A
5December2011
Theincreaseinpressuredropwithtapegeneratedswirlflowislessthantheincreaseinheattransfer
coefficient.Providedthatbuoyancyeffectscanbeneglected,asituationwhichisofthegreatestinterest
forspaceapplication,themaximumincreaseintheheattransfercoefficientisoftheorderoftentimes
theemptytubeconstantpropertyvalue,whilethecorrespondingpressuredropisoftheorderofthree
timestheemptytubepressuredrop.Thisisincontrasttoturbulentflow,wherecomparableincreasesin
heattransfercoefficientresultinpressuredropswhichareseveralordersofmagnitudelargerthatthe
emptytubevalues.
SincetheNusseltnumberfortheemptytubeflowisindependentofReynoldsnumber,theusual
coupling
betweenaugmentedandemptytubesdoesnotappearhere.Hence,Nua/Nuovs.Reaisthesamefora
constraintoffixedflowrate,pressuredroporpumpingpower.Forconsistencywithotherdatainthis
itemitis,however,convenienttoplotha/hoasafunctionoftheReynoldsnumber,Reo,basedonnon
augmentativeconditions.TorelatedbothaugmentativeandnonaugmentativeReynoldsnumbers,the
requiredofequalpumpingpoweriswrittendownas:
(Re3)o=(Re3)a.
IncorporatingfrictionfactorsandReynoldsnumbers,allofthembasedontheinnertubediameter,and
theperimeterratio,a/o=(+22t/D)/,theaboveexpressionyields:Reo=2,186Rea.
TWISTEDTAPEINSERTS
TURBULENTFLOWINSMOOTHTUBES
AIR
basedonnonaugmentativeconditions,Reo.CurvesAtoIarefromBergles(1969)
[8],curvesJtoMhavebeencalculatedbythecompilerafterThorsen&Landis
(1968)[178].
160
ECSSEHB3101Part13A
5December2011
Explanation
Key A B C D E F G H I J K L M
Investigator Colburn&King(1931)[27] Evans&Sarjant(1951) Smithberget Thorsen&Landis
[61] al.(1964)[168] (1968)[178]
Dx103[m] 66,67 63,5 35,1 25,35 25,37
L/D(Diabatic) 13,7 33,6 29 36
y 0,57a 1,14a 2,67 2,80 3,60 4,80 5,60 1,81 11 1,58 4,00 1,58 4,00
tx103[m] b 2,38 0,56 b
Tapefitc 4 4 3 3
Ti[K] 590 390170 380660 <755 300 300 <519
TwTb 160 60270 220280 105125 <50 <350 200
Tw/Tb 0,670,730,550,84 0,520,82 0,790 1,2 1,9 0,6
Rea 1,9x1035,6x103 3,3x1031,5x104d 1,5x1049x104 5x103105
Heating/Cooling C C H H C
a Thetapeiswrappedaroundasmalldiameteraxialrod.
b Notgiven,itispresumablynegligiblecomparedtoD.
c Thenumberscategorizethetapefitasfollows.
3. Snugtapefit.Gapoflessthan
4. Loosetapefit.Gapestimatedtobegreaterthan9,25x103m.
d BasedonD.
161
ECSSEHB3101Part13A
5December2011
THORSEN&LANDIS(1968)[178] Fluid ReTbasedontheaveragetotalvelocityandDE.Re

fab/f= properties basedonaverageaxialvelocityandDE.
evaluatedatTb. ReT/Re=(2/3)([1+(/2y)2]3/21)/(/2y)2
2
f
2,5 ln ReT
f 0,85 fabandfbased ForReT>104,f=0,046ReT0,2.
onReTandDE. Kistheflowpathcurvaturenearthewall.
oritsequivalent 5x103<Re<105 KD/2=(/2y)2/[1+(/2y)2]
1

2,035 log ReT 0,91 1,58<y<4
K>10m1
L/D=36
=1,126+0,0094K
0 ,1 Sameasabove.
f T
w 0,6<Tw/Tb<1,9
f ab Tb
Nuab=F(ReT,Pr,) Thevalidityof Nuab/vs.ReT,forseveralvaluesofPr,isgivenin

Correspondstoisothermalflowthroughthe theanalytical Figure915b.Nuab/ispracticallyindependentof
spiralchannel. expression .Theerrorincurredinneglectingthis
relatingNuabto dependencyisoftheorderof0,6%forPr=0,71,
ReT,Prand andincreaseswithPr.ForPr=50thaterrorisof
hasbeen theorderof9,5%.
experimentally Thorsen&Landis(1968)[178]suggesttheuseof
checkedforair theDittusBoelterformula,modifiedas:
(Pr0,71) ThecomputationsleadingtoFigure915bdonot
under validatethisalternative,particularlyregarding
theabove thePrdependence,exceptwhenPr1.
stated
conditions.
Theanalysis
requiresthat
heatfluxand
shearstressat
thewallvary
inthesame
way.
Heating Fluid Forgases=Tf1.
Aw D 2t 1 / 2 y L
0 , 32
2DH / D Tw Tb properties
Nu a Tw
2
1 evaluatedat
Nu ab Tb 2 y 4 Tbamexcept, Aterm,Qf,accountingfortheconductionheat
Givestheradialconvectioneffect. evaluatedat transferthroughthetapemustbeaddedfor
NuandRe evaluatingthetotalenergytransferredtoorfrom
basedonDE. thefluid.Thistermisgivenby:
5x103<Re<105
DLTw Tb / Q f sD / 2k f t 1 / 2 y 1 / h
2
1<Tw/Tb<1,9
Constantheat whereistheradialgapbetweenstripand
fluxatthewall. wall,andthefineffectiveness.
hbasedonAw.
162
ECSSEHB3101Part13A
5December2011
Cooling Sameasabove Intheexperimentswhosedataarebeing

except correlated,coolingwasachievedbycounter
2DH / D Tb Tw
0 ,1
Nua T
1,07 w 1 0,6<Tw/Tb<1 flowingwaterthroughanouterconcentrictube.
Nuab Tb 2 y 4 Tbisdefinedas Tbisdefinedbycombinationofthearithmetic
Givestheradialconvectioneffect. indicated meanwatertemperature,Tc,andthelogarithmic
under meanfluidtowatertemperaturedifference.
Comments. Tci Tco Tbi Tco Tbo Tci
Tb
2 T Tco
ln bi

Tbo Tci
LOPINA&BERGLES(1969)[120] Thevalidityofthiscorrelation,setforthforwaterandmoderate
SeeTableofFigure916. temperaturedifferences,hasbeencheckedforheliumandfor
nitrogenbyBollaetal.(1973)[15]intherange104<Re<3x105.The
effectonNuaoftheradialthermalgradientisaccountedforbythe
factor(Tw/Tb)0,575.Althoughtheseauthorsdonotreportthe
temperatureratio,Tw/Tb,theyquoteavalueofthethermalfluxof
45x104W.m2,morethanthreetimeslargerthanthehighestvalue
quotedbyThorsen&Landis(1968).
OnlyThorsen&Landis(1968)[178]reporttheappropriatediabaticfrictionfactordata.
Thecoolingdataareslightlylowerthantheheatingdata.Someimprovementisexpectedinanycaseduetothe
increasedvelocitybecauseoftheswirlinducedbythetape.
Thetightesttwistratioisnotnecessarilythebest.Atwisty=2,5isquotedasoptimumforair.
Figure915:IsothermalNusseltnumber,Nuab.dividedbytheratiooffrictionfactors,,vs.
theReynoldsnumber,ReT,fordifferentvaluesofthePrandtlnumber,Pr.Calculatedbythe
compilerafterThorsen&Landis(1968)[178].
163
ECSSEHB3101Part13A
5December2011
TWISTEDTAPEINSERTS
TURBULENTFLOWINSMOOTHTUBES
WATER
Figure916:Constantpowerheattransferratio,(he/ho)P,vs.Reynoldsnumber
basedonnonaugmentativeconditions,Reo.FromBergles(1969)[8].
Explanation
Key A B C D E F G H I J
Investigator Greene(1960)[76] Ibragimovet Smithberg& Lopina&Bergles(1969)

al.(1961)[96] Landis(1964) [120]
[168]
Dx103[m] 22,6 12,0 35,1 4,91
L/D(Diabatic) 57 29 6976
(e/D)x104 1 <3 0,57(tube)

0,61(tape)
y 0,28 0,56 1,12 2,10 4,56 1,81 11 2,48 3,15 5,26 9,20
tx103[m] 1 0,56 0,343
Tapefita 2 3 1
Ti[K] 288 283366
(TwTb)[K] <50 678
px103[Pa] 210690
Rea 1045x104 8x1036x104 8x1031,3x105
Heating/Cooling H H H H
164
ECSSEHB3101Part13A
5December2011
NOTE 1.Specialcaretakentoincreasethethermaljointconductance;f.e.,byredrawingofthetubeoverthe
twistedtape.
3.Snugtapefit.Gapoflessthan0,25x103m.
4.Loosetapefit.Gapestimatedtobegreaterthan0,25x103m.
5.Specialcaretakentoreducethethermaljointconductance;f.e.,insulatingthetapebypaint.
LOPINA&BERGLES(1969) FluidpropertiesevaluatedatTb. Forsmoothtubeintherange5x103<

[120] fabasedonDEa.foandReobasedon Re<105,fo=0,046Reo0,2(seeTable
fab/fob=2,75y0,406 D. 72).ForawiderReynoldsnumber
9x103<Re<1,2x105 rangeseeclause7.2.2,where=4f.
e/D<3x104 Forthefrictionfactortobecome
1,8<y<11 constantL/D>0,623Re1/4.
Checkedwithbothwaterandair.
fa/fab=(w/b)0,35(DEa/D) FluidpropertiesevaluatedatTbam Thedifferencesbetweenisothermal

fabasedonDEa.RebasedonD. andheatedfrictionfactorsareless
9x103<Re<1,2x105 thanfortheemptytube.Thisisso
e/D<3x104 becausethetapeisnearlyadiabatic
2,48<y<9,20 and,thence,thereductioninthe
CheckedwithwateratTb<360K liquidviscositybecauseofthe
increasedtemperatureismuch
smallernearthetapethannearthe
tubewall.
Thiseffectisaccountedforbythe
factorDea/Dintheexponent.
Nua=Nusc+Nucc+Nuf Liquidorgasesundermoderatetransversetemperaturegradients.
Nusc=0,023Re0,8Prn Straighttubesofcircularcross Nuscispredictedfromconventional

DittusBoelterformula. section. straighttubecorrelations,
FluidpropertiesevaluatedatTbam. accountingfortheincreased
Correspondstoflowthrough
NuandRebasedonD.Re>104. velocityneartubewallcausedby
thespiralchannel. 0,7>Pr>100 thetapeinsertionandspiralling
n=0.4forheating,n=0,3for flowpath.ForthistheReynolds
cooling. numbermustbedefinedas:
L/D>60 Re=VDE/,where2=1+(/2y)2.
Knudsen&Katz(1968)[107] Theconstantfactor,0,023,canbe
Thefactthatnisslightlydifferent modifiedasfollows:0,0237forRe<
forheatingthanforcoolingisdisre 1,5x104,and0,025forRe<1,5x104.
gardedhere.n=0,4inanycase.
Nucc=0,114(GrPr)1/3 Horizontalheatedplate,widthL, TheGrashofnumber,Gr,isbased

FishendenSaujndersformula facingupward. onthecentrifugalacceleration.
Givesthecentrifugal FluidpropertiesevalautedatT0,5, Gr/Re2=2(/2y)2(DEa/D)(TwTb).
convectioneffectwhenthe GrbasedonL.2x10 7<GrPr<3x1010. THISTERMWILLNOTBE
heat McAdams(1954)[125].Afactor PRESENTWHENCOOLING.
flowstowardthetubeaxis. 0,14insteadof0,114isquotedin
165
ECSSEHB3101Part13A
5December2011

this,aswellasinothersources.
Nuf=(F1)(Nusc+Nucc) ThetablebelowgivesvaluesofFin Thistermisdeterminedbythe

Givestheheattransferbythe typicalinstances. designerforeachparticularsystem
tapeactingasafin. ofinterest.
KLACZAK(1973)[106] FluidpropertiesevaluatedatTbm.
Nua=1,46Re0,44Pr0,36y0,33 NuandRebasedonD.
1,7x103<Re<2x104
2,5<Pr<9
1,63<y<3,79
L/D=35.
Thecurvesexhibitratherwidescatter.Forsimilarvaluesoftapetwistdifferencesareduetovariationin
thefineffectandcentrifugalconvectioneffect.
Thetightesttwistratioisnotnecessarilythebest.Extremelytighttwistratios(lowvaluesofy),suchas
thoseinGreene(1960)[76],havebeenaccomplishedbywrappingthestriparoundasmalldiameterrod.
Allthereferences,exceptLopina&Bergles(1969)[120],onlypresentisothermalfrictionfactors.
AnattempthasbeenmadetoincludeinthefiguredatafromtheextensivetestsbyKreith&Margolis
(1958).Theresultsforwater,althoughshowingthecorrecttrendregardingthevariationwiththe
Reynoldsnumber,leadtofartoolargevaluesof(ha/ho)P,whichatRe=5x104areoftheorderof2,5fory
=2,58and1,7fory=5.Theseresultsseemtobeaffectedbyexperimentalerrors(seeSmithberg&Landis
(1964)[168]).

TypicalvaluesoftheFinEffectMultiplier,F

Fita Fluid F Reference
1 Water 1,137forRe<3x104 Lopina&Bergles(1969)

1,10forRe>3x104 [120]
3 Water 1,10 Smithberg&Landis

(1961)[168]
5 He,N 1,00 Bollaetal.(1973)[15]

twistedtape.
3.Snugtapefit.Gapoflessthan0,25x103.
5.Specialcaretakentoreducethethermaljointconductance;f.e.,insulatingthetapebypaint.
166
ECSSEHB3101Part13A
5December2011
TWISTEDTAPEINSERTS
TURBULENTFLOWINROUGHTUBES
WATER
basedonnonaugmentativeconditions,Reo.CurvesA,B,CfromBergles(1969)[8],
curvesD,E,FfromBergles,Lee&Mikic(1969)[9].
Explanation
Key A B C D E F
Investigator Gambill,Bundy&Wansbrough Bergles,Lee&Mikic(1969)[9]

(1961)[69]
Dx103[m] 6,32 6,35 6,30 6,35
L/D(Diabatic) 45 5479 5985 5479
(e/D)x104 0,84(tube)a 12(tube) 1,21(tube) 12(tube)

0,24(tape) 0,605(tape) 0,60(tape)
y 2,30 5,00 8,00 No 2,42 2,55

Tape
tx103[m] 0,381 0,472
Tapefitb 2 1
Ti[K] 290300 283366
(TwTb)[K] 80c 678
px103[Pa] 2503500 210690
Rea 5x1034,3x105 3x1038x104
Heating/Cooling H H
167
ECSSEHB3101Part13A
5December2011
a Thisvalueisexceedinglylow.FromtheswirlflowfrictiondatabyGambilletal.(1961)[69]intheirFig.10,
equivalentsandroughnessesoftheorderofe/D=20x104result.Lopina&Bergles(1969)[120],intheirFig.
04,ascribetothesetubesthevaluee/D=17x104.
b Thenumberscategorizethetapefitasfollows:
twistedtape.
c EstimatedfromTwTb=(D/L)[(TboTbi)/4St].
Empirical Conditions Comments

Correlations
BERGLES,LEE& FluidpropertiesevaluatedatTb. Expressionstocalculatethe

MIKIC(1969)[9] fabasedonDEa.foandReobased isothermalemptytubefriction
fab/fob=2,75y0,406 onD. factorsforeithersmoothorrough
3x103<Re<8x104 tubesaregiveninclause7.2.2.
e/D<12x104
2,55<y<6,16
fa/fab=(w/b)0,35(DEa/D) L/D=6080 Thedifferencesbetweenisothermal
Thecompensationofviscosity andheatedfrictionfactorsareless
variationeffectshasbeenchecked thanfortheemptytube.Thisisso
withwateratTb<360K becausethetapeisnearlyadiabatic
and,thence,thereductioninthe
liquidviscositybecauseofthe
increasedtemperatureismuch
smallernearthetapethannearthe
tubewall.
Thiseffectisaccountedforbythe
factorDEa/Dintheexponent.
Nua=Nusc+Nucc+Nuf Liquidorgasesundermoderatetransversetemperaturegradients.
Nusc=0,0088Re0,915Pr0,4 FluidpropertiesevaluatedatTbam Nuscispredictedfromconventional

Correspondstoflow NuandRebasedonDE. straighttubecorrelations,
throughthespiral 3x103<Re<8x104 accountingfortheincreased
channel. e/D<12x104 velocityneartubewallcausedby
2,55<y<6,16 thetapeinsertionandspiralling
L/D=6080 flowpath.ForthistheReynolds
numbermustbedefinedas:
Re=VDE/,where2=1+(/2y)2.
Fortheeffectofroughnessand
Prandtlnumberonboththe
constantfactorandReexponentsee
ECSSEHB3101Part3.
168
ECSSEHB3101Part13A
5December2011
Empirical Conditions Comments

Correlations
Nucc=0,114(GrPr)1/3 Horizontalheatedplate,widthL, TheGrashofnumber,Gr,isbased

FishendenSaujnders facingupward. onthecentrifugalacceleration.
formula FluidpropertiesevaluatedatT0,5, Gr/Re2=2(/2y)2(DEa/D)(TwTb).
Givesthecentrifugal GrbasedonL.2x107<GrPr<3x1010. THISTERMWILLNOTBE
convectioneffectwhen McAdams(1954)[125].Afactor PRESENTWHENCOOLING.
theheatflowstoward 0,14insteadof0,114isquotedin
thetubeaxis. this,aswellasinothersources.
Nuf=(F1)(Nusc+Nucc) RepresentativevaluesofFare Thistermisdeterminedbythe

Givestheheattransfer givenintheTableofFigure916. designerforeachparticularsystem
bythetapeactingasa ofinterest.
fin.
Dataonsmoothswirltubesandonrough(noswirl)tubeshavebeenincludedforcomparison
purposes.
TheimprovementobtainedbyinsertingtwistedtapesinroughtubesstronglydependsonReoor
thepumpingpowerlevel.AtlowvaluesofReotheclearchoicewouldbeasmoothswirltube;
butathighervaluesofReotheroughtubesoutperformthesmoothswirltubes.Forthesevalues
ofReotheroughswirltubeisclearlythetopperformer.
169
ECSSEHB3101Part13A
5December2011
10
Working fluids
10.1 General
Thisclauseisintendedasanaidintheselectionofcoolantfluids.
Coolantfluidsforsatellitetemperaturescanbegaseousorliquids.Moltenmetalsareusedathigher
temperatures.
Theselectionofthecoolantismadeonthebasisofthethermalandpowerrequirementsforachieving
agivencoolingduty.Inthecaseofnonexpandablecoolantsthebasicvariablesare:density,specific
heat, thermal conductivity and viscosity, as well as temperature limits which should be compatible
with the desired application. The most important factors to be taken into account when selecting
expandablecoolantsare:specificheatandusefultemperaturerange.
Theabovementionedrequirements,however,donotgivethecompletepictureoftheproblem.Many
fluids fulfilling the appropriate thermal and frictional requirements can be aggressive to metallic or
elastomericmaterials.Severalcoolantsaretoxicandtheiruseareruledoutwhenmannedoperations
areenvisaged;severalothersareflammable,....
Additionalrequirementscanbedefinedinparticularcases.Forinstance,goodlubricityisdesirableto
performhydraulicfunctionsaswellascooling;dielectricpropertiesbecomeimportantwhensingle
walled jackets are used and the fluid is in contact with electrical equipment. Concern with vapor
pressure,storagetemperature,safetyandhandlingappearsinthedetaileddesignofthesystem.
Accordingtotheaboveconsiderations,thisclauseisdividedintotwodistinctclauses.
The first part deals with the selection of the fluids which meet the cooling and pumping power
requirements. This selection is based on the fluid thermal and frictional properties. A drastically
simplifiedcoolingloopisassumedinordertosubstantiatethecomparisonbetweenthefluids.Eight
fluidcoolantsareconsideredbywayofexample,fourofthembeingliquidsandtheotherfourgases
under normal operating conditions. The method used for ranking these coolants can be applied to
other fluids provided that their relevant physical properties are known. From the worked cases the
conclusion is drawn that liquids are far superior to gases, although the useful temperature range of
liquidsisrestrainedbyboiling,highvaporpressureandthermaldegradation.
Inthesecondpartoftheclause,thephysicalandenvironmentalpropertiesofthirtyliquidcoolantsare
compiled.Theclauseendswithaclausegivingthermalpropertiesofdryairinthetemperaturerange
270Kto310K,atapressureof105Pa.
10.2 Cooling effectiveness of a fluid

The cooling effectiveness of a fluid is measured by the ratio, F, of the heat transfer rate, Q, to the
mechanicalpower,p(m/),requiredtodrivethefluidthroughtheloop,ftheheattransferrate
170
ECSSEHB3101Part13A
5December2011
Q
F
m [101]
p

A preliminary estimate of the fluid coolant merit should be based on simplified fluid loops. In our
particular case the loop is constituted by a cylindrical smooth tube, of circular cross section, inner
diameter D, length LE, and diabatic length L (L LE). More complex loops can be reduced to this
simple equivalent configuration once the pressure losses through them, for given flow conditions,
havebeenevaluated.
The merit of several gaseous and liquid coolants are compared on the basis of the ratio F, for both
laminar and turbulent flow, in Figure 61. The calculations have been made for particular, yet
representative,valuesoftheheatfluxq(q=Q/DL),innerdiameterofthetube,D,anddiabaticlength,L.
Figure101:Productofcoolingeffectiveness,F,ofseveralfluidstimesthe
equivalentlengthoftheloop,LE,asfunctionsofthedifferencebetweentheheat
sourceandtheinlettemperature,TSTi,forthefollowingreferencevalues:Inner
diameteroftheduct,D=102m.Diabaticlengthoftheduct,L=1m.Heatflux,q=
250W.m2forAir,CarbonDioxide,CarbonTetrachloride,HydrogenandNitrogen,
q=1000W.m2forEthyleneGlycol,FlutecPP50andWater.Calculatedbythe
compiler.
171
ECSSEHB3101Part13A
5December2011
ItcanbededucedfromFigure101thatliquidaresuperiortogases.Gases,however,canbeusedover
a much wider temperature range since common liquid are restrained by freezing and/or high
viscosities atlow temperatures, and by boiling, highvapor pressure or thermal degradationat high
temperatures. It is also seen that the ranking of the coolants considered is relatively unaffected by
laminarorturbulentflowconditions.
Detailsconcerningthecalculationsarebrieflypresentedinthefollowingclauses.
10.2.1 Simplified fluid loop configuration

ThefluidispumpedthroughthecircularcylindertubeoflengthLEanddiameterD.Themassflow
rate, m, as well as the heat transfer rate, Q, and the inlet fluid temperature, Ti, are assumed to be
constant.
Someregulatingdeviceisrequiredtoenforceinpracticalcasestheseconstraints.Hereitisassumed
thataheatexchangerplacedbetweenthepumpandthetubeinletcontrolstheinlettemperature,Ti,
withoutchangingthefluidmassflowrate,m.
The layout of the simplified configuration is sketched in Figure 102. The controller picks the wall
temperatureandcontrolsthemassflowratethroughtheheatexchanger.
Figure102:Schematicrepresentationofthefluidloopconsideredforestimating
thefluidcoolingeffectiveness.
10.2.2 Thermal performance of the simplified loop

Assumingthatthefluidisheatedorcooledoverthediabaticlength,L,theheatflowrate,Q,willbe
givenby
q ham Tam LD mc p Tbo Ti [102]
seeclause6.2.1.
The source temperature, TS, is assumed to be equalto the wall temperature, Tw, of the diabatic
cylindricaltube.Thiscouldbethecasewhenthistubeisthefluidductofacoldplateonwhichthe
heatsourceismounted.
Taking into account the definitions of the Nusselt, Prandtl and Reynolds numbers given in clause
6.2.2,Qbecomes
172
ECSSEHB3101Part13A
5December2011
k
Nu
Q D DLTS Ti [103]
Nu L
1 2
PrRe D
TheNusseltnumberisestimatedwiththeaidofthefollowingcorrelations:
FullyDevelopedLaminarFlow(Re<2400,Gz<9),Nu=3,66.SeeFigure61.
ThermalEntryLengthLaminarFlow(Re<2400,Gz9),Nu=1,75Gz1/3.SeeESDU68006
(1968)[48],p.10,oralternatively,Figure66ofthisPart.
FullyDevelopedTurbulentFlow(Re>4000),Nu=0,023(Re2Pr)0,4.Seeclause9.2,Dittus
Boelterformula.
In the transitional region both the laminar and turbulent values should be calculated. These values
willcorrespondtothelowerandupperlimitsoftheNusseltnumber.
10.2.3 Power requirements of the simplified loop

Thepressurelossduetofrictioninthecylindricalpipeis:
LE 1
p V 2 [104]
D 2
seeclause7.2.2.Inthisequationisgivenby:
FullyDevelopedLaminarFlow(Re<2400), =64/Re.Seeclause7.2.2,HagenPoiseuille
formula.
Fully Developed Turbulent Flow (Re > 4000), = 0,316Re0,25. See Table 72, Blasius
formula.
Again the transitional friction factor, , is supposed to be bounded by the values obtained by
extrapolationofthelaminarandturbulentformulae.
10.2.4 Several examples

Values of the cooling effectiveness, F, of several fluids were plotted as functions of the temperature
difference,TSTi,inFigure101.Thecomputationsweremadeforreferencevaluesoftheheatflux,q,
theinnerdiameterofthetube,D,andthediabaticlength,L.
The same computations have been replotted for four different fluids, namely: Air, Ethylene Glycol,
FlutecPP50andWaterinfourseparatefigures(heavylinesinFigure104toFigure107).
InordertoestimatehowsensitivetheplottedvaluesofFLEaretochangesintheassumedvaluesofq
andDasimplegraphicalmethodcanbeused.ThismethodisintroducedbyresortingtoFigure103.
173
ECSSEHB3101Part13A
5December2011
Figure103:Graphicalmethodallowingforvaluesofheatflux,q,andinner
diameteroftheduct,D,differentfromthoseusedinFigure101.
InordertomovefrompointBrtopointB:
1. AsegmentwillbedrawnfromBr,paralleltoDaxis,whoselengthcorrespondstothenew
valueoftheductdiametermeasuredintheappropriatescale.
2. From the resulting point, B1, a segment parallel to q axis and of the appropriate length
willbedrawn,givingthepointB.
BothqandDscalesarelogarithmic.
PointArwillbemovedinthesamemannerforpinpointingA,althoughdifferentobliquecoordinates
qandDmustbeusedasshowninFigure104toFigure107.
174
ECSSEHB3101Part13A
5December2011
Figure104:Graphicforestimatingtheproductofthefluidcoolingeffectiveness,
F,timestheequivalentlengthoftheloop,LE,asafunctionofthedifference
betweentheheatsourceandtheinletfluidtemperature,TSTi.Fluid:Air.
Referencevalues:D=102m,L=1m,q=250W.m2.ValuesFLEfordifferentDand
q,yetL=1,canbecalculatedgraphicallyasisindicatedinthetext.Preparedbythe
compiler.
175
ECSSEHB3101Part13A
5December2011
betweentheheatsourceandtheinletfluidtemperature,TSTi.Fluid:Ethylene
Glycol.Referencevalues:D=102m,L=1m,q=1000W.m2.ValuesFLEfor
differentDandq,yetL=1,canbecalculatedgraphicallyasisindicatedinthetext.
Preparedbythecompiler.
176
ECSSEHB3101Part13A
5December2011
betweentheheatsourceandtheinletfluidtemperature,TSTi.Fluid:FlutecPP50.
compiler.
177
ECSSEHB3101Part13A
5December2011
betweentheheatsourceandtheinletfluidtemperature,TSTi.Fluid:Water.
compiler.
Inordertodrawthenewlaminarflow(orturbulentflow)curverelatingthecoolingeffectivenessper
unitequivalentductlength,F,withthetemperaturedifference,TSTi,forthechosenvaluesoftheheat
flux,q,andtheinnerdiameteroftheduct,D,astraightlineisdrawfromBparalleltothatdrawnfor
Br.
This graphical method is only approximate. The reference curves are not strictly straight lines nor
shouldbetranslatedinparallell.howevertheerrorresultingfromtheuseoftheseoversimplifications
hasbeenoftheorderof5%inallthecaseswhichhavebeenconsidered.
10.3 Properties of liquid coolants

Relevantpropertiesofthirtyselectedliquidcoolantshavebeencompiledinthisclause.
Thedataarearrangedasfollows:Basicphysicalpropertiestobeusedinthepreliminaryselectionof
the coolant have been assembled in Table 101. These properties are: operating temperature limits,
178
ECSSEHB3101Part13A
5December2011
flammability data, density, specific heat, thermal conductivity and dynamic viscosity. When the
availabledatacannotbepresentedinthesmallspaceallottedtothemintheTable,additionalfigures
and tables are used. As a general rule the additional figures precede the additional tables, but are
placedaftertheopeningTable101andTable102.
Environmental properties such as, odor, toxicity and materials compatibility,together with some
reference to previous applications of the several coolants are presented in Table 102, plus the
additionalTable104toTable109.
179
ECSSEHB3101Part13A
5December2011
Table101:PhysicalPropertiesofTypicalLiquidCoolants
LIQUID CHEMICAL FREEZING BOILING FLASH VAPOR DENSITY SPECIFIC TERMAL DYNAMIC
FORMULA POINT POINTat POINT PRESSURE HEAT CONDUC. VISCOSITY
[K] 105Pa [K] psat [kg.m3] cp k
[K] [Pa] [J.kg1.K1] [W.m1K1] [Pa.s]
Water H2O 273 373 None Figure108 Figure109 Figure1010 Figure1011 Figure1012
Carbon CCl4 250a 350a Noneb Figure1013 Figure1014 Figure1015 Figure1016 Figure1017
Tetrachloride
Coolanol15c 177 422 350 Figure1018 Figure1019 Figure1020 Figure1021 Figure1022

Coolanol25c,d 189 372 436
Coolanol35c 189 483 450
Coolanol45c,e 208 461
DC200 Dimethyl Table103 Table103 Figure1023

Siloxane
Polymer
EthyleneGlycol C2H6O2 Figure 470f Figure1025 Figure1026 Figure1027 Figure1028 Figure1029

Glycol/Water 1024
FlutecPP2g 223h,i 349i Figure1030 Figure1031 Figure1033 Figure1034

FlutecPP9 g 193 j 433 i 1080 j
FlutecPP50g C5F12 133j 302j Figure1032

Freon11k CFCl3 162 297 None Figure1035 Figure1036 Figure1037 Figure1038 Figure1039
Freon12k CF2Cl2 115 243 None
Freon13k CF3Cl 92 192
Freon21k CHFCl2 138 282 None
Freon22k CHF2Cl 113 232
Freon113k C2F3Cl3 238 321
180
ECSSEHB3101Part13A
5December2011
LIQUID CHEMICAL FREEZING BOILING FLASH VAPOR DENSITY SPECIFIC TERMAL DYNAMIC
FORMULA POINT POINTat POINT PRESSURE HEAT CONDUC. VISCOSITY
[K] 105Pa [K] psat [kg.m3] cp k
[K] [Pa] [J.kg1.K1] [W.m1K1] [Pa.s]
Freon114k C2F4Cl2 179 277 None
Freon142k C2H3F2Cl 142a 263a
FreonE1k (C2F3O)C3HF8 119 314 None Figure1040 Figure1041 Figure1042 Figure1043 Figure1044
FreonE2 k (C2F3O)2C3HF8 150 377
FreonE3k (C2F3O)3C3HF8 166 425
FreonE4 k (C2F3O)4C3HF8 179 467
FreonE5k (C2F3O)5C3HF8 189 497
FC75l 160 372 None Figure1045 Figure1046 Figure1047 Figure1048 Figure1049

Methanol/Water 231j 355j 294m 900j 3430j Figure1050 Figure1051
60/40
Methanol/Water 170m 343m 287m 850m 2880m

80/20
Monsantoc 208 644 461 Figure1052 Figure1053 Figure1054 Figure1055 Figure1056

OS59
Oroniten 200 422 383 5066 891 1985 0,0986 2,5x103

FloCool100
NOTE Alldatainthistable,unlessotherwisestated,arefromFilippi&Guerra(1977).
a FromVargaftik(1975)[183].
b FromWeat(1966)[185].
c SislicateestermanufacturedbyMonsantoChemicalCo.
d AlsoknownasFFP103andMonsantoOS139.
e AlsoknownasOS45
f FromRaznjevic(1970)[147].
181
ECSSEHB3101Part13A
5December2011
g ManufacturedbyImperialSmeltingCo.Ltd.
h Meltingpoint.
I FromDunn&Reay(1976)[40].
j FromKissner(1973)[105].Propertiesaregivenat298K.
k FreonisaregisteredtrademarkofE.I.DuPontdeNemours&Co.,Inc.
l FC75isafluorochemical.
m FromIgnations&Mitchell(1974)[98].Thetemperaturescorrespondingtothermodynamicandtransportpropertiesarenotreported.Theapparentrageseemstobefrom
283Kto298K.
n OroniteisasilicateestermanufacturedbyOroniteChemicalCo.Propertiesaegivenat298Kexceptvaporpressureandspecificheatwhicharegivenat422Kand298K
respectively.
182
ECSSEHB3101Part13A
5December2011
Table102:EnvironmentalPropertiesofTypicalLiquidCoolants
LIQUID ODOR TOXICITY MATERIALSCOMPATIBILITY APPLICATIONS
(Mac)a
Metals Elastomers
Water None Nontoxic CorrosivetoAlalloys Noswell Skylab,Shuttle
CarbonTetrachloride Noxiusb (315mg.m3)b CorrosivetoCu Unusablec
alloysc
Coolanol15 Slight ToxicatT>422K Table104 Moderateswelld GeminiSpaceCapsule
Coolanol25 Lowtoxicity
(1000ppm)d
Coolanol35 Compatiblewith Testfluid
BunaNd
Coolanol45
DC200 Nonee Nontoxice Chemicallyinerte Liquiddielectric
andcoolant
EthyleneGlycol (27,7mg.m3)f Compatiblewith Highvicosity.gNormally
Aluminiumg mixedwithwater
Glycol/Water60/40 Slight 114mg.m3 Corrosive Noswell Antifreezehmixture.
Apollo
FlutecPP2
FlutecPP9 Breathingvaporshould Extremelygood Coolantforcomputerand
beavoided electronicequipment,can
beusedashydraulicand
FlutecPP50
dielectricfluidh
Freon11 Table105 Compatiblewithsteel, Table106
castiron,brass,copper,
Freon12
tin,lead,aluminium
Freon13 andotherstructural
metallicmaterials
Freon21 VeryNoxius Skylab,Shuttleg
183
ECSSEHB3101Part13A
5December2011
LIQUID ODOR TOXICITY MATERIALSCOMPATIBILITY APPLICATIONS

(Mac)a
Metals Elastomers
Freon22 undernormal
conditionsofuse.
Freon113 MgalloysandAlalloys
Freon114 containingmorethan
2%Mgarenot
Freon142 recommendedforuse
insystemscontaining
Freonwerewatermay
bepresent.Zincisnot
recommendedforuse
withFreon113d
FreonE1 None Extremelylow.Toxicat Slightswell
533Kd
FreonE2 None Table107 Heattransfermedium,
dielectriccoolant
FreonE3 None Table108i
FreonE4
FreonE5
FC75 Lowtoxicity Table107 Coolantforelectronic
(1000ppm)d equipmenth
Methanol/Water60/40 Noxius 26mg.m3 Corrosive Noswell
Methanol/Water80/20
MonsantoOS59 (1000ppm)d
OroniteFloCool100 Table109
NOTE Alldatainthistable,unlessotherwisestated,arefromIgnationis&Mitchell(1974)[98].
184
ECSSEHB3101Part13A
5December2011
a Maximumallowableconcentration
b FromWeast(1966)[185].ThevalueunderToxicitycorrespondstotheIndustrialLimitwhichisnormallytentimesthatallowedinSpacecraftfor90d.
c FromRabald(1968)[146].
d FromFilippi&Guerra(1977)[64].
e FromDOWCORNING(1963,1972)[37]&[38].
f FromRittenhouse&Singletary(1969).ThisvaluecorrespondtoSpacecraft1hEmergencywhichisnormallyfivetimesthatallowedinspacecraftfor90d.
g FromKissner(1973)[105].
h FromScott(1974)[161].
I Alsoincludescompatibilitywithplasticsandwirecoatings.
185
ECSSEHB3101Part13A
5December2011
Figure108:Vaporpressure,psat,ofWatervs.temperature,T.FromVargaftik(1975)
[183].
Figure109:Density,,ofWatervs.temperature,T.FromVargaftik(1975)[183].
Figure1010:Specificheat,cp,ofWatervs.temperature,T.FromVargaftik(1975)
[183].
186
ECSSEHB3101Part13A
5December2011
Figure1011:Thermalconductivity,k,ofWatervs.temperature,T.FromVargaftik
(1975)[183].
Figure1012:Dynamicviscosity,,ofWatervs.temperature,T.FromVargaftik
(1975)[183].
Figure1013:Vaporpressure,psat,ofCarbonTetrachloridevs.temperature,T.From
Vargaftik(1975)[183].
187
ECSSEHB3101Part13A
5December2011
Figure1014:Density,,ofCarbonTetrachloridevs.temperature,T.From
Figure1015:Specificheat,cp,ofCarbonTetrachloridevs.temperature,T.From
Figure1016:Thermalconductivity,k,ofCarbonTetrachloridevs.temperature,T.
FromVargaftik(1975)[183].
188
ECSSEHB3101Part13A
5December2011
Figure1017:Dynamicviscosity,,ofCarbonTetrachloridevs.temperature,T.
FromVargaftik(1975)[183].
Figure1018:Vaporpressure,psat,ofCoolanol15,25,35and45vs.temperature,T.
FromFilippi&Guerra(1977)[64].
Figure1019:Density,,ofCoolanol15,25,35and45vs.temperature,T.From
Filippi&Guerra(1977)[64].
189
ECSSEHB3101Part13A
5December2011
Figure1020:Specificheat,cp,ofCoolanol15,25,35and45vs.temperature,T.
Figure1021:Thermalconductivity,k,ofCoolanol15,25,35and45vs.
temperature,T.FromFilippi&Guerra(1977)[64].
Figure1022:Dynamicviscosity,,ofCoolanol15,25,35and45vs.temperature,T.
190
ECSSEHB3101Part13A
5December2011
Figure1023:Kinematicviscosity,,ofDC200vs.temperatureT.Numberson
curvesindicatethestandardviscosityincs.FromDOWCORNING(1972)[38].
Figure1024:Freezingpoint,T,ofWater/GlycolSolutionsvs.Glycolmassfraction,
s.FromFilippi&Guerra(1977)[64].
191
ECSSEHB3101Part13A
5December2011
Figure1025:Vaporpressure,psat,ofWater/GlycolSolutionsvs.temperature,T.
NumbersoncurvesindicateGlycolmassfraction,c.FromFilippi&Guerra(1977)
[64].
Figure1026:Density,,ofWater/GlycolSolutionsvs.temperature,T.Numberson
curvesindicateGlycolmassfraction,c.FromFilippi&Guerra(1977)[64].
192
ECSSEHB3101Part13A
5December2011
Figure1027:Specificheat,cp,ofWater/GlycolSolutionsvs.temperature,T.
[64].
Figure1028:Thermalconductivity,k,ofWater/GlycolSolutionsvs.temperature,
T.NumbersoncurvesindicateGlycolmassfraction,c.FromFilippi&Guerra
(1977)[64].
193
ECSSEHB3101Part13A
5December2011
Figure1029:Dynamicviscosity,,ofWater/GlycolSolutionsvs.temperature,T.
[64].
Figure1030:Vaporpressure,psat,ofFlutecPP2,PP9andPP50vs.temperature,T.
DataarefromDunn&Reay(1976)[40]exceptthosecorrespondingtoFlutecPP50
whicharefromWynRoberts(1974)[193].
194
ECSSEHB3101Part13A
5December2011
Figure1031:Density,,ofFlutecPP2,PP9andPP50vs.temperature,T.Dataare
fromDunn&Reay(1976)[40]exceptthosecorrespondingtoFlutecPP50which
arefromWynRoberts(1974)[193].
Figure1032:Specificheat,cp,ofFlutecPP50vs.temperature,T.FromWyn
Roberts(1974)[193].
195
ECSSEHB3101Part13A
5December2011
Figure1033:Thermalconductivity,k,ofFlutecPP2,PP9andPP50vs.
temperature,T.DataarefromDunn&Reay(1976)[40]exceptthosecorresponding
toFlutecPP50whicharefromWynRoberts(1974)[193].
Figure1034:Dynamicviscosity,,ofFlutecPP2,PP9andPP50vs.temperature,
T.DataarefromDunn&Reay(1976)[40]exceptthosecorrespondingtoFlutecPP
50whicharefromWynRoberts(1974)[193].
196
ECSSEHB3101Part13A
5December2011
Figure1035:Vaporpressure,psat,ofFreon11,12,13,21,22,113,114and142vs.
temperature,T.FromVargaftik(1975)[183].
Figure1036:Density,,ofFreon11,12,13,21,22,113,114and142vs.temperature,
T.FromVargaftik(1975)[183].
Figure1037:Specificheat,cp,ofFreon11,12,13,21,22,113,114and142vs.
197
ECSSEHB3101Part13A
5December2011
Figure1038:Thermalconductivity,k,ofFreon11,12,13,21,22,113,114and142vs.
temperature,T.FromVargaftik(1975)[183].
Figure1039:Dynamicviscosity,,ofFreon11,12,13,21,22,113,114and142vs.
temperature,T.DataarefromVargaftik(1975)[183]exceptthosecorrespondingto
Freon13whicharefromFilippi&Guerra(1977)[64].
198
ECSSEHB3101Part13A
5December2011
Figure1040:Vaporpressure,psat,ofFreonE1,E2,E3,E4andE5vs.temperature,T.
Figure1041:Density,,ofFreonE1,E2,E3,E4andE5vs.temperature,T.From
Figure1042:Specificheat,cp,ofFreonE1,E2,E3,E4andE5vs.temperature,T.
199
ECSSEHB3101Part13A
5December2011
Figure1043:Thermalconductivity,k,ofFreonE1,E2,E3,E4andE5vs.
Figure1044:Dynamicviscosity,,ofFreonE1,E2,E3,E4andE5vs.temperature,
T.FromFilippi&Guerra(1977)[64].
Figure1045:Vaporpressure,psat,ofFC75vs.temperature,T.FromFilippi&
Guerra(1977)[64].
200
ECSSEHB3101Part13A
5December2011
Figure1046:Density,,ofFC75vs.temperature,T.FromFilippi&Guerra(1977)
[64].
Figure1047:Specificheat,cp,ofFC75vs.temperature,T.FromFilippi&Guerra
(1977)[64].
Figure1048:Thermalconductivity,k,ofFC75vs.temperature,T.FromFilippi&
Guerra(1977)[64].
201
ECSSEHB3101Part13A
5December2011
Figure1049:Dynamicviscosity,,ofFC75vs.temperature,T.FromFilippi&
Guerra(1977)[64].
Figure1050:Thermalconductivity,k,ofMethanol/WaterSolutionsvs.
temperature,T.NumbersoncurvesindicateMethanolmassfraction,c.From
202
ECSSEHB3101Part13A
5December2011
Figure1051:Dynamicviscosity,,ofMethanol/WaterSolutionsvs.temperature,
T.NumbersoncurvesindicateMethanolmassfraction,c.FromVargaftik(1975)
[183].
Figure1052:Vaporpressure,psat,ofMonsantoOS59vs.temperature,T.From
203
ECSSEHB3101Part13A
5December2011
Figure1053:Density,,ofMonsantoOS59vs.temperature,T.FromFilippi&
Guerra(1977)[64].
Figure1054:Specificheat,cp,ofMonsantoOS59vs.temperature,T.FromFilippi
&Guerra(1977)[64].
Figure1055:Thermalconductivity,k,ofMonsantoOS59vs.temperature,T.From
204
ECSSEHB3101Part13A
5December2011
Figure1056:Dynamicviscosity,,ofMonsantoOS59vs.temperature,T.From
205
ECSSEHB3101Part13A
5December2011
Table103:PropertiesofDowCorning200Fluids(DimethylSiloxanePolymers)
Kinematica Freezing Boiling Flash Density SpecificHeat,cp Thermal
Viscosityat Point Point Point at298K, [J.kg1.K1] Conductivityat
298K, [K] [K] [K] 298K,
313K 373K 473K
x106 [kg.m3] k
[m2.s1] [W.m1.K1]
0,65 205 373c 272 759 0,100
1 187 425c 311 816
1,50 197 465c 336 851 0,105
2 189 343373d 352 870 0,109
5 208 393433d 408 917 0,117
10 173 436 931 0,134
20 189b 505 946 1457 1478 1516 0,142
50 203b 558 957 0,151
100 208b 588 0,155
200 967
350 1403 1444 1478 0,159
500 223b 968
1,000 594
12,500 227b 972
30,000 230b
60,000 232b 973
100,000 240b 974
206
ECSSEHB3101Part13A
5December2011
a Numeralsinthiscolumnofthetablecorrespondtostandardviscositygradesincs(1cs=106m2.s1).ThesenumeralsareusedtoidentifythevariousliquidsintheDow
Corning200series.
b Pourpoint(ASTMD9739,Sec.57).Duetosupercoolingeffectsthistestmethodyieldsvaluesbelowthetemperaturesatwhichthesefluidssolidifywhenheldatagiven
temperatureforlongeritems.
c At105Pa.
d At67Pa.
NOTE FromDowCorning(1963,1972)[37]&[38].
207
ECSSEHB3101Part13A
5December2011
Table104:CorrosionandOxidationTextDataforCoolanolLiquids
Liquid Coolanol Coolanol Coolanol Coolanol
15 25 35 45
Testa Ab Bc A B A B A B
MetalWeightChange[g.m2]
Aluminium +0,4 +0,1 +0,5 0,2 +0,1 0,1 +0,1
Magnesium +0,1 +0,2 0,2 0,1
Copper +0,6 0 24. 0,2 0,3 +0,5 0
Silver +0,3 0 0
Steel +0,4 +0,2 +0,2 0,1 0,2 +0,2 0
CadmiumPlated +0,8 0 0,1 +0,1

Steel
Acidity[gofKOHperkg]
Initial 0,11 0,11 0,03 0,03 0,01 0,14
Final 0,06 1,00 0,03 3,00 0,01 0,81
KinematicViscosityat328K
x106 Initial 3,19 3,12 4,46 4,21 8,44 8,51

[m2.s1]
Final 3,27 5,71 4,54 8,39 8,63 9,05
Increase[percent] 1,9 83,0 1,8 99,0 2,3 5,8
TotalFluidWeightLoss[percent] 1,8 13,5 2,8 13,0 <1 <10

a TestweremadeasindicatedinthefollowingTable(AandBrefertocodeinTable104).
Code ExposureTime Temperature AirFlowRate Specification

[h] [K] [m3.s1]
A 168 394 1,389 MILH5606
B 72 478 MILH8446
b Coolanol15isthermallystableat394K,andmaybeusedintypicalcirculatingsystems.However,inthistest,
whenexposedtoapurgeofairatthistemperature,itcanvaporize,oxidizeandperhapsgel.
c Coolanol15isnotrecommendedforcontinuoususeat478K.
NOTE FromFilippi&Guerra(1977)[64].
208
ECSSEHB3101Part13A
5December2011
Table105:ToxicityofSeveralFreonLiquids
Liquid Maximum Group ToxicityRatea
Concentration Definition
Freon11 5a 6.Gasesorvaporswhichinconcentrationsuptoatleast
about20%byvolumeanddurationsofexposureatthe
Freon12 500mg.m3 6 orderof2hdonotappeartoproduceinjury.
Freon13 6 5a.GasesorvaporsmuchlesstoxicthanGroup4butmore
toxicthanGroup6.
Freon21 420mg.m3 4&5
4&5.MuchlesstoxicthanGroup4butsomewhatmore
Freon22 5a toxicthanGroup5.
4.Gasesorvaporswhichinconcentrationsoftheorderof
Freon113 4&5
2to2,5%anddurationsofexposureoftheorderof2hare
Freon114 500mg.m3 6 lethalorproduceseriousinjury.
a ThecriterionforevaluatingthetoxicityisthatemployedbyUnderwritersLaboratoriesofChicago.According
tothiscriterionthefluidsaredividedinto6categoriesrangingfromGroup1(mosttoxicfluids)toGroup6
(leasttoxicfluids).ThecharacterizationofthegroupsinvolvedisgivenintheabovetableunderDefinition.
Table106:SwellingofElastomersinSeveralFreonLiquidsa
Liquid Freon Freon Freon Freon Freon Freon Freon Freon
Elastomer 11 12 13 21 22 113 114 142
BunaN(butadieneacrylonitrile) 6 2 1 48 26 1 0 3
BunaS(GRS)(butadienestyrene) 21 3 7 49 4 9 2 4
Butyl(GRI) 41 6 0 24 1 21 2 3
Hypalon40b 3 1 1 24 3 1 1 4
(chlorosulphonatedpolyethylene)
NaturalRubber 23 6 1 34 6 17 2 5
NeopreneGNb 17 0 0 28 2 3 0 3
NeopreneW 9 1 3 11 0 1 0 0
PolyvinylAlcohol 0 8 1 9 6 1 2 3
ThiokolFA(polysulfide) 2 1 0 28 4 1 0 1
VitonA(PVnF)b 4 10 3 26 37 25
VitonB(PVnF)b 6 9 4 22 28 7 9 29
SiliconeRubber 38 20 34
UrethaneRubber 20 5 28 7
a NumbersintheTableindicatepercentincreaseinlengthatroomtemperature.
b ManufacturedbyE.I.DuPontdeNemours&Co.,Inc.
209
ECSSEHB3101Part13A
5December2011
Table107:CompatibilityofFreonE2andFC75withElastomersa
Elastomers Liquid Swelling TensileStrength,x104[Pa]
Change
Before Afterexposure Change
[Percent]
exposure [Percent]
BunaN FreonE2 +2,3 11,92 13,17 +10,5
FC75 +0,7 12,19 +2,6
VitonAb FreonE2 0,3 10,48 9,72 7,3
FC75 0,7 9,84 6,1
a NumbersintheTableindicateswellingandtensilestrengthchangesafter100hexposure.
b ManufacturedbyE.I.DuPontdeNemours&Co.,Inc.
Table108:CompatibilityofFreonE3withElastomers,PlasticsandWireCoatingsa
Elastomers Testb Plastics Testb Wire Testb
Coatings
327 425 327 425 377 425
K K K K K K
AdipreneC(Urethane 1 5 Alathom7050c 1 5 Acrylex 6

Rubber) (PolyethyleneResin) (Acrylic)
AdipreneL(Urethane 1 3 Alathom9140c 1 1 Alkanex 6

Rubber) (PolyethyleneResin) (Terephthalate
Polyester)
BunaN 1 5 Delrin 1 3
(AcetalResin)
BunaS 1 4 EthylCellulose 2 5 Anavar 6

(Isocyanate
Buthyl 2 2 G103675 1 3 Modified
(EpoxyResin) Polyvinyl
Formal)
Hypalon40c 2 5 Kralasticd 1 5 Ensolex 6

(Chlorosulfon.Polyethyl.) (ABSPolymer) (Acrylic)
NaturalRubber 1 2 Lexane 1 2 Epoxy 6

(Polycarbonate
Resin)
NeopreneW 1 2 Lucitec 1 1 Formvarf 7 8

(AcrylicResin) (Polyvinyl
Formal)
Nordel(Hydrocarbon 1 3 PolyvinylAlcohol 2 3 Nylclad 7 8

Rubber) (Nylon
Coated
NSRX5602(Nitrile 1 5 PolyvinylChloride 1 4
210
ECSSEHB3101Part13A
5December2011
Elastomers Testb Plastics Testb Wire Testb

Coatings
327 425 327 425 377 425
K K K K K K
SiliconeRubber) (Unplasticized) Polyvinyl
Formal)
Silastic50(Silicone 1 2 Styron457 1 5 PlainEnamel 7 8

Rubber) (Polystyrene) (Oleoresinous)
ThiokolFA(Polysulfide) 1 5 Teflonc 1 2
(Tetrafluoroethylene)
VitonAb 1 2 Zitel101(Nylon 1 3
(Fluoroelastomer) Resin)
1 Suitableforuse 5 Materialdisintegratesor 7 Veryslightcrazing,

dissolvesintheliquid stresscracking,bucling
2 Probablysuitableforuse and/ortemporary
softeningofmaterial
3 Probablynotsuitableforuse 6 Nocrazing,stress 8 Severecrazing,stress

cracking,bucklingand/or cracking,buckling
4 Notsuitableforuse temporarysofteningof and/ortemporary
material softeningofmaterial
a NumeralsinthisTableindicatetestratingsaccordingtothescoreinthesecondtable.
b Testwereperformedatquotedtemperaturesfor100h.
c ManufacturedbyE.I.DuPontdeNemours&Co.,Inc.
d AcrylonitriteButadieneStyrene.ManufacturedbyUnitedStatesRubberCo.,NaugatuckChemicalDiv.
e ManufacturedbyGeneralElectricCompany,ChemicalMaterialsDept.
f ManufacturedbyShawiniganResinsCorp.
211
ECSSEHB3101Part13A
5December2011
Table109:CompatibilityofOroniteFloCool100withElastomers
Elastomer Swelling Weight Tensilestrength,x104[Pa] Compression Cross
Change Change Set Sectional
Before After Change
[Percent] [Percent] [Percent] Area
exposure exposure [Percent]
Change
[Percent]
BunaN 0,2a 11,92 9,95a 16,5 9,7b
ButhylRubberB +5,3 +31. 10,31 8,16 20,8 19,8 10.

3187
Ethylene +0,4 +27,9 11,40 12,04 +5,6 8.

PropyleneE515
8
FluorosiliconeL +3,7 5,12 5,88 +14,8 12,6 2,9

3088
FluorosiliconeL +4,6 8,25 5,86 29. 6,2 2.

6086
LCS +0,02 12,84 12,82 0,02 10,6 0,1

Fluorocarbon(per
MILR83248)
NeopreneC294 1,2 16,28 9,60 41,1 12,7 0

6
NeopreneC526 +1. 18,37 7,32 60,2 8,1 0,5

7
Silicone54186 +13,2 +117,4 5,44 3,53 35,1 45,7
Silicone56047 +26,4 +65,6 6,76 6,06 10,3 27.
VitonA +1,1a 10,48 9,98a 4,8 63,6c

NOTE AlldatainthisTable,unlessotherwisestated,havebeenobtainedafter30dexposure.
a After7dexposure.
b After14dexposure.
c After19dexposure.
212
ECSSEHB3101Part13A
5December2011
10.4 Properties of dry air
Figure1057:Density,,ofAiratapressureof105Pavs.temperature,T.From
Figure1058:Specificheat,cp,ofAirvs.temperature,T.FromVargaftik(1975)
[183].
Figure1059:Thermalconductivity,k,ofAirvs.temperature,T.FromVargaftik
(1975)[183].
213
ECSSEHB3101Part13A
5December2011
Figure1060:Dynamicviscosity,,ofAirvs.temperature,T.FromVargaftik(1975)
[183].
214
ECSSEHB3101Part13A
5December2011
11
Heat exchangers
11.1 General
Heat exchangers are devices whose primary purpose is the transfer of energy between two fluids.
Among the different types of heat exchangers, regenerators and recuperators are those most
commonlyusedintheaerospaceindustry.
PeriodicFlowTypeExchangersorRegeneratorsareexchangersinwhichthehotandcoldfluidsflow
alternativelythroughthesamematrixsurfacewithaslittlephysicalmixingbetweenthetwostreams
aspossible.Theamountofenergytransferdependsonthepropertiesofthefluidstreams,aswellas
onthegeometryandthermalpropertiesofthematrixsurface.Twotypesofregeneratorsareshownin
Figure111.
Figure111:Typicalregenerators.a)Rotarytype.b)Valvedtype.FromKays&
London(1964)[102].
ClosedTypeExchangersorRecuperatorsareexchangersinwhichhotandcoldfluidareseparatedby
atubewallorsomekindofsurface.Energyexchangeisthusaccomplishedfromonefluidtoasurface
byconvection,throughthewallorplatebyconductionandthenbyconvectionfromthesurfacetothe
secondfluid(Figure112).
215
ECSSEHB3101Part13A
5December2011
Figure112:Typicalrecuperators.a)Counterflowheatexchanger.b)Crossflow
heatexchanger.FromWelty,Wicks&Wilson(1969)[188].
Recuperators are of primary importance in practice and therefore, the present review is mainly
concernedwiththestudyofthistypeofexchangers.
Recuperatorsareusuallyclassifiedaccordingtotheirconfigurationandthenumberofpassesmadeby
each stream as it traverses the heat exchangers. Singlepass heat exchangers are those in which the
fluidflowsthroughtheexchangeronlyonce.Dependingontherelativedirectionsofthetwostreams
the singlepass heat exchangers are divided in parallel flow, counterflow and crossflow heat
exchangers.Amultipassheatexchangerisoneinwhichthefluidflowsthroughtheexchangermore
thanonce.Multipassingallowsalargevarietyofarrangements.
Formostheattransfersurfacesusedinheatexchangers,theheattransferrateperunitofsurfacearea
increaseswiththefluidflowvelocity,andthisratevariesassomethinglessthanthefirstpowerofthe
velocity. The friction power expenditure also increases with fluidflow velocity, but in this case the
exponentofthevelocitydependencevariesbetweentwoandthree.
Pumping power limitations generallyforce the designer of the heat exchanger to use lowvelocities;
therefore the heat transfer rate tends to be low, and large heat transfer surfaces are required. To
accomplishasmuchheattransferinaslittlespaceaspossible,itisnecessarytoconstructexchangers
with large surface area densities (surface area density is the ratio of total heat transfer area to total
exchanger volume). Exchangers with large surface area densities are known as compact heat
exchangers.
Several typical compact heat transfer surface arrangements are shown in Figure 113. It is observed
that an effective way to increase surface area density is to use fins in one or both fluid sides of the
surface as is shown in Figure 113b. To accomplish large area density on both fluid sides of the
surface, the platefin arrangement shown in Figure 113d may be used. In this arrangement the
exchangerisbuiltupasasandwichofflatplatesbondedtointerconnectingfins.Thetwofluidsare
carried between alternate pairs of plates and can be arranged in either counterflow or crossflow,
which provides an added degree of flexibility. The stripfin surfaces shown in Figure 113e exhibit
veryhighheattransferratesbecauseoftheshortflowlengthassociatedtoeachfin,thuspreventing
thethickeningoftheviscouslayer.
216
ECSSEHB3101Part13A
5December2011
Figure113:Sometypicalexamplesofcompactheatexchangersurfaces.FromKays
&London(1964)[102].a)Circulartubebundle.b)Finnedcirculartubesurface.c)
Finnedtubesurface,flattubes,continuousfins.d)Platefinarrangement.e)Strip
finsurface.f)Regeneratorcompactmatrix.
OtherprocedurestoenhanceheattransferarediscussedinClause9.
Toaccomplishasmuchenergytransferinaslittlespaceaspossible,itisdesirabletoutilisemultiple
passesofoneorbothfluids.Amultipassingconfiguration,whichiscommonlyusedforcommercial
and general process applications, is the shellandtube exchanger shown in Figure 114. In this
particularcase,theshellfluidgoesthroughtwoshellpassesandthetubefluidflowsthroughsixtube
passes.Goodmixingoftheshellsidefluidisaccomplishedwiththebafflesshown,whichpreventthe
formationofstagnantregions.
217
ECSSEHB3101Part13A
5December2011
Figure114:a)Shellandtubeexchangerwithtwoshellpassesandfourtube
passes.b)Schematicrepresentationoftheexchangerwhichwillbeusedinclause
11.2.3.
Sometimes several exchangers are connected together in series, in parallel or in seriesparallel, to

achieveheattransferrate.Anotherusefularrangementistheliquidcoupledindirectheatexchanger,
which consists essentially of two recuperators coupled with a pumped heat transfer medium as
shown in Figure 115. With this arrangement the circulating transfers thermal energy from the hot
fluidtothecoldfluid.Thecouplingliquidcapacityrate(massflowtimesspecificheat),whichisan
independentvariable,canbechosentooptimisesystemperformance.
Figure115:Liquidcoupledindirecttransfertypeofheatexchanger.FromKays&
London(1964)[102].
11.2 Basic analysis
11.2.1 Introduction
Thepurposeofthisclauseistosummariseexchangerheattransferperformancetheory.Thedesigner
usesthistheory,inconjunctionwiththebasicdatareportedinClauses6to9,toapproximatelysize
theheatexchangerforaspecifiedheattransferrateandpressuredrop.
Theproblemofoptimisingtheheatexchangerdesigninafluidloop,isacomplexprobleminwhich
many qualitative judgements are introduced. Although no attempt is made to outline hear an
218
ECSSEHB3101Part13A
5December2011
optimisationanalysis,thepresenttheoryprovidesbasicrelationshipsthatmaybeusedinoptimizing
thedesignonthebasisofanycriterionpreviouslyselected(seeClause13).
Theequationswhichappearinthefollowingclausesconcernclosedtypeexchangersorrecuperators
of either the conventional directtransfer type or of the liquidcoupled indirecttransfer type. The
methodology used can, however, be extended with only minor changes to periodicflow type
exchangersorregenerators.
11.2.2 Analytical background

Theheatflux,q,perunittransferareaatacertainsectionintheexchangerisexpressedintermsofan
overallthermalconductance,U,asfollows:
q U Th Tc [111]
U includes the convective and conductive mechanisms which are responsible for the heat transfer
fromthehottothecoldfluid.ThereciprocalofUisanoverallthermalresistancewhichpresentsthe
followingseriescomponents:
1. Ahotsidefilmconvectioncomponent,whichincludesthetemperatureineffectivenessof
theextendedsurfaceorfinareaonthisside.
2. Awallconductioncomponent.
3. A coldside filmconvection component, which includes the temperature ineffectiveness
oftheextendedsurfaceorfinareaonthisside.
AccordingtothisideaUcanbeexpressedeitheras:
1 1 a 1

U h h hh Aw / Ah k Ac / Ah c hc
[112]
whenUisbasedonaunitofhotsidetotalarea(includingtheextendedarea),oras:
1 1 a 1

U c c hc Aw / Ac k Ah / Ac h hc
[113]
whenitisbasedonaunitofcoldsidetotalarea.ObviouslyUhAh=UcAc.
Intheseequations:
Awistheaveragewallarea.IfnoextendedsurfacesareusedthenAw(Ah+Ac)/2.
aisthewallthickness.
kisthewallthermalconductivity.ValuesofkforseveralcommonmaterialsaregiveninFigure116.
219
ECSSEHB3101Part13A
5December2011
Figure116:Thermalconductivity,k,ofseveralmetalsvs.temperature,T.From
Kays&London(1964)[102].
h and c denote the temperature effectiveness of the total area Ah and Ac, respectively due to the
presence of fins. The temperature along the fin is not equal to that existing at the surface, thus
reducingtheeffectivenessofthesurface.Intheabsenceofextendedsurfaces isunity,otherwiseitis
aweightedaverageofthe100%effectivenessofthesurfaceandtheeffectivenessofthefinsurface.
1
Af
1
f
[114]
A
Theparameter fisthefineffectivenessdefinedastheratiooftheamountofheatactuallyevacuated
throughthefintothatwhichwouldbeevacuatedifthefinsurfacewereheldatthetemperatureofthe
fin base.Values of f versus thedimensionless filmlengtharegiven inFigure 117andFigure118.
When the fin extends from wall to wall, the effective length, L,is taken as half the spacing between
walls.
220
ECSSEHB3101Part13A
5December2011
Figure117:Heattransfereffectiveness,f,oftrapezoidalfins,vs.dimensionless
finlength,L[2h/k(t+b)]1/2.CalculatedbythecompilerafterJakob(1958)[99].
Figure118:Heattransfereffectiveness,f,ofcircularfins,vs.dimensionlessfin
length,(rori)(h/k)1/2.CalculatedbythecompilerafterJakob(1958)[99].
hh,hcareconvectiveheattransfercoefficients.Clause6presentsdatatocomputetheirvaluesformost
practicalconfigurations.
Theheattransferrateequation(Eq.[111])iscombinedwithanenergyequationequatingthelossof
enthalpyofthehotfluidtothegainofenthalpyofthecoldfluid,soastorelatethedifferentvariables
oftheexchanger.Toobtainthesolutiontothisequationinaconvenientform,itisusefultointroduce
thefollowingnondimensionalvariables:
1. Thecapacityrateratio,R,
C1
R [115]
C2
221
ECSSEHB3101Part13A
5December2011
whereCi(i=1,2)isthemassflowratetimesspecificheatforthestreami.Subscript1will
bealwaysassignedtothestreamwithlowerC.
2. Theheattransfereffectiveness,,
Q T T T T
1i 10 2i 20
Qmax T1i T2i RT2i T1i
[116]
istheratiooftheactualheattransferratetothethermodynamicallylimitedmaximum
rate, obtained only in a counterflow exchanger of infinite transfer area. Thus, is the
exchangerthermodynamiceffectiveness.
3. Thenumberofexchangerheattransferunit,Ntu,
A
N tu AU / C1 UdA / C1 [117]
o
where A is the transfer area used in the definition of U. Ntu is a nondimensional

expression of the heat transfer size of the exchanger. For small Ntu, the exchanger
effectiveness is low. When Ntu increases, the effectiveness, , approaches asymptotically
thelimitimposedbyflowarrangementandthermodynamicconsiderations.
Ingeneralthesolutiontotheenergyequationyields
t N tu , R, flow arrangement [118]
11.2.3 Exchanger performance

ThisclausepresentsdatatosolveEq.[118]forseveralflowarrangementsusedinpractice.Thesedata
areintendedasillustrativecases.Theusershouldresorttoadditionalreferenceswhentheparticular
arrangementofitsinterestisnotincludedhere.
Mostofthefigureswhichappearinthefollowingpagesrelate toNtu.Asageneralrule increases
withNtuforagivencapacityrateratio.Thedatapresented,however,showtheasymptoticcharacterof
this increaseand, thence, the design engineer is confronted with the costs ofattaininga large Ntuin
terms of capitalisation, size, mass and power requirements. Since this clause deals only with the
thermal aspect of the problem, no data on which the estimate of these costs could be based are
reportedhere.
Figure 119 shows the Ntu relationship for a counterflow configuration. The effectiveness, ,
approachesunityforlargevaluesofNtu,nomatterthevalueofthecapacityrateratio,R.Notethatthe
smallerthecapacityrateratio,thehighertheeffectivenessforagivenNtu.
222
ECSSEHB3101Part13A
5December2011
Figure119:Heattransfereffectiveness,,vs.numberofheattransferunits,Ntu,in
acounterflowexchanger.FromKays&London(1964)[102].
TheparallelflowcaseisconsideredinFigure1110.Nowtheasymptoticvalueforeffectivenessisless
than unity, except for R = 0. Obviously the maximum effectiveness is reached when the exit
temperaturesofbothflowstreamsbecomeidentical.ForR=1,themaximumachievableeffectiveness
inparallelflowisonly0,5.
aparallelflowexchanger.FromKays&London(1964)[102].
Figure1111andFigure1112areforcrossflowarrangements.Nowtheresultsdependonwhetheror
notcrossmixingisallowed.InthecaseconsideredinFigure1111itisassumedthateachoneofthe
two fluid streams has been broken up into a large number of separate flow tubes with no cross
mixing.Theresultingeffectivenessisbasedonthemixedmeantemperaturesoftheoutletfluids,i.e.,
thetemperaturewhichthefluid,passingthecrosssectionalexitareaofanytubeperunittime,would
have if collected and perfectly mixed with those fluids leaving the other tubes in the same time
interval.
223
ECSSEHB3101Part13A
5December2011
acrossflowexchangerwithfluidsunmixed.Calculatedbythecompilerafter
Mason(1954)[124].
acrossflowexchangerwithonefluidmixed.FromKays&London(1964)[102].
In Figure 1112 it is assumed that one stream is perfectly cross mixed, so that its temperature is
uniformonsectionsnormaltoflowdirection.
ThecasewhenbothfluidsaremixedcanbefoundinKays&London(1964)[102].
ItistobenotedthatforthelimitingvalueR=0,theresultsinallthecasesmentionedarethesame,
indicating that the flow arrangement is irrelevant when one capacity rate is much larger than the
other.
Itisusualtocharacterisethedifferentcrossflowarrangementsbymeansofsketchessimilartothose
shownbelow.
224
ECSSEHB3101Part13A
5December2011
Multipass crossflow exchangers are extensively used in the aerospace industry. Nearly all of the
exchangersusingthesocalledcompactheatexchangersurfacesareofthistype.
Stevens, Fernndez & Woolf (1957) [170] have analysed these exchangers for one, two and three
passes,witheithercountercurrentorcurrentflow.
Ten different arrangements are possible for both twopass and threepass exchangers. Analytical
expressionsrelatingtheheattransfereffectiveness, ,tothe(total)numberofheattransferunits,Ntu,
can be found in five of the ten cases. Data for the cases that require numerical computation are
presentedbytheabovementionedauthors.Thesedataaregivenintheformofcorrectionfactor,which
are the ratio of the exchanger effectiveness to the effectiveness of a pure counter or parallel flow
exchangerforthesamevaluesofRandNtu.
Figure1113givestheeffectiveness, ,versusthenumberofheattransferunits,Ntu,forone,twoand
threepass countercrossflow exchangers, when fluid A is unmixed throughout and fluid B mixed
throughout, and with passes connected in reverse order. Streams with the same capacity rate have
beenassumed.Reverseordermeansthatthemixedfluidsurroundsatubeaftertheotherinacertain
orderwhenflowingthroughoneofthepasses,anditsurroundsothersectionsofthesametubesinthe
reverseorderwhenflowingthroughthenextpass.
Figure1113:Heattransfereffectiveness,,vs.numberofheattransferunits,Ntu,
forthecaseofnpasscountercrossflowexchangers,whenfluidAisunmixed
throughoutandfluidBmixedthroughout,andwithpassesconnectedinreverse
order.CalculatedbythecompilerafterStevens,Fernandez&Woolf(1957)[170].
Figure1114givessimilardatabutnowthepassesareconnectedinidenticalorder.
225
ECSSEHB3101Part13A
5December2011
forthecaseofnpasscountercrossflowexchangers,whenfluidAisunmixed
throughoutandfluidBmixedthroughout,andwithpassesconnectedinidentical
order.CalculatedbythecompilerafterStevens,Fernandez&Woolf(1957)[170].
AdditionalresultscanbefoundinKays&London(1964)[102].Thesedataareforcountercrossflow
exchangers,withonetofourpasses,inthefollowingthreeparticularcases.
1. Bothfluidsmixedbetweenpasses,butunmixedineachpass.
2. Oneofthefluidsmixedbetweenpasses,unmixedineachpass.Theotherfluidunmixed
throughout.Invertedorder.
3. Bothfluidsunmixedthroughout,bothinvertedorder.
ComparisonofdatainFigure1113andFigure1114withthoseinFigure119demonstratesthatthe
larger the number of passes, the more nearly the performance approaches pure counterflow
performance, but a large number of passes is often not warranted. Tradeoff studies performed by
Trusch & Nason (1975) [180], in connection with compact heat exchangers for the Space Shuttle,
indicatedthat,althoughthereisasignificantreductioninexchangermasswhenthenumberofpasses
increases from 2 to 4, the decrease in mass when the number of passes is further increased to 6 is
negligible.
Figure1115toFigure1120givetheexchangereffectivenessforparallelcounterflowconfigurations
whenoneofthefluidsismixed.Thisisthecaseoftheshellandtubeheatexchanger,whichisoneof
themostcommonconfigurationsforgeneralprocessapplications.Simpleversionsoftheseexchangers
havebeenusedinSpacecraft(Ebersole(1970)[41]).Nevertheless,lightweightcompactexchangersare
ofthecountercrossflowtype,ashasbeensaidalready,andthisissobecauseofreasonswhichwillbe
madeapparentverysoon.
226
ECSSEHB3101Part13A
5December2011
amultipassexchangerwith1shellpassand2ormoretubepasses.FromKays&
London(1964)[102].
amultipassexchangerwith2shellpassesand4,8,12,....tubepasses.FromKays&
London(1964)[102].
227
ECSSEHB3101Part13A
5December2011
amultipassexchangerwith3shellpassesand6,12,18,....tubepasses.FromKays
&London(1964)[102].
amultipassexchangerwith4shellpassesand8,16,24,...tubepasses.FromKays
&London(1964)[102].
228
ECSSEHB3101Part13A
5December2011
amultipassexchangerwith5shellpassesand10,15,20,...tubepasses.FromKays
&London(1964)[102].
fordifferentnumberofshellpasses,inamultipassexchangerwithR=1.Thecase
R=0isalsoshownforcomparison.CalculatedbythecompilerafterKays&
London(1964)[102].
The mathematical expression in Figure 1115 is for one shell pass and two tube passes. Numerical
resultsforfour,six,eight,...tubepassesareveryclosetothoseshowninthefigure.
Figure 1116 to Figure 1118 present the effect of multipassing the basic parallel counterflow
arrangement.
SomeoftheresultsfromthepreviousfiguresaresummarisedinFigure1120.Leavingasidethecase
R=0,whoseeffectiveness, ,doesnotdependonthenumber,n,ofshellpasses,onecandeduce,from
considerationsofthecasesR=1,thatlittleisgainedbyincreasingthenumberofshellpassesoverfour
orfive.
Acomparisonamongthedatapresenteduptothemomentindicatesthat,foragivenNtu,thelargest
effectiveness is achieved with pure counterflow configurations (Figure 119), whereas the smallest
229
ECSSEHB3101Part13A
5December2011
effectiveness corresponds to parallel flow configurations (Figure 1110). Crossflowfluids unmixed
configurations(Figure1111)yieldresultsslightlybelowthepurecounterflowcase.Allowingoneof
the fluids to be mixed lowers a bit the performance of the cross flow (Figure 1112), but these
performancescanbeincreasedbymultipassingasinthecountercrossflowconfigurationsconsidered
in Figure 1113 and Figure 1114. On the other side, parallel counterflow arrangements of the shell
andtube type (Figure 1120) yield performances substantially above those for the parallel flow but
stillbelowthoseforcrossflow.
Anotherwaytowardthepurecounterflowperformanceistheliquidcoupledindirectheatexchanger.
The couplingliquid capacity rate is an independent variable in such system, and this additional
freedomcanbeusedtomaximisetheoveralleffectivenessforgivenheattransferareas,givenhotand
coldfluidcapacities,andgivenoverallthermalconductances.
Figure 1121 to Figure 1124 analyse the performance of a liquidcoupled indirect heat exchanger.
Subscripts 1 and 2 refer, respectively,to properties of the exchanger with smaller and larger flow
streamcapacityrate.SubscriptLreferstopropertiesoftheliquidcouplingstream.
Figure1121:Optimumliquidflowcapacityrate,C1/CLopt,tomaximizetheheat
transfereffectivenessvs.ratioofthenumberofheattransferunitsNtu1/Ntu2,of
exchanger1toexchanger2.CalculatedbythecompilerafterHolmberg(1975)[90].
230
ECSSEHB3101Part13A
5December2011
Figure1122:Heattransfereffectiveness,,vs.liquidflowcapacityrate,CL/CLopt,
fordifferentvaluesoftheoverallnumberofheattransferunits,Notu.Ntu1/Ntu2=1.
CalculatedbythecompilerafterHolmberg(1975)[90].
Figure1123:Heattransfereffectiveness,,vs.liquidflowrate,CL/CLopt,for
differentvaluesoftheratiobetweenthenumberofheattransferunitsof
exchanger1toexchanger2,Ntu1/Ntu2.Notu=2.Calculatedbythecompilerafter
Holmberg(1975)[90].
231
ECSSEHB3101Part13A
5December2011
Figure1124:Overallheattransfereffectiveness,,vs.arearatiobetween
exchanger1and2,(A1/A2)/(A1/A2)opt,forthecaseofoptimumliquidflowcapacity
rateandR=1.CalculatedbythecompilerafterHolmberg(1975)[90].
Theoptimumliquidflowcapacityrate,C1/CLoptt,isgiveninFigure1121versustheratioNtu1/Ntu2,for
differentvaluesoftheratio,R,ofcoldfluidtohotfluidcapacityrates.Optimisationisperformedby
choosingthecouplingliquidflowcapacityrate,CL,sothattheoveralleffectivenessismaximisedfor
givenNtuvaluesofbothexchangers.
The relationship between the optimum value of the overall effectiveness and the overall number of
heat transfer units for the liquidcoupled system, Notu, is the same as that for pure counterflow
(Holmber(1975)[90]).Thence,Figure119canbeusedtocalculatetheoveralleffectivenessinthecase
ofoptimumcouplingliquidflowcapacityrate,onceNotuisdeducedfrom
1
1 1
N
o
[119]
N tu N tu
tu
1 2
NonoptimumperformancecanbediscussedbymeansofFigure1122andFigure1123.Theoverall
effectiveness, ,isrepresentedinthesefiguresversusCL/CLoptforsomevaluesoftheparametersR,Notu
andNtu1/Ntu2.Itcanbeseenthattheachievementofoptimumperformancebecomescriticalonlyfor
largevaluesofNotu1andthatitbecomeslesscrucialwhenRgoestozero.
One could further optimise the area ratio A1/A2 for given total heat transfer area, A1 + A2, assuming
thatthecouplingliquidcapacityrateisoptimum.Itisshown(Holmberg(1975)[90])that
A1 U2
[1110]
A2 opt U1
whereU1andU2aretheoverallthermalconductancesforbothstreams.
232
ECSSEHB3101Part13A
5December2011
Figure 1124 shows how a nonoptimum distribution of total heat transfer area affects the overall
effectiveness.Onecandeducefromthisfigurethattheoptimumisveryflatforallrealisticvaluesof
theparameters.
Heat exchangers can be combined to achieve a given heat transfer duty. One resorts to these
combinations either because existing exchangers are used, or because the optimum configuration is
impracticalduetomanufacturinglimitations.
Theexchangerscanbecombinedinseries,parallel,orseriesparallel.Theyareoftenofnonidentical
type (or size), and a multiplicity of associations are possible. Typical associations of exchangers are
shownbelow.
TheresultsgiveninFigure1125toFigure1128,whichhavebeenborrowedfromDomingos(1969)
[36],arebasedonthefollowingassumptions:
Figure1125:Overallheattransfereffectiveness,t,ofanassemblyofnidentical
exchangersinparallel,vs.effectiveness,,ofasingleexchanger.Calculatedbythe
compilerafterDomingos(1969)[36].
233
ECSSEHB3101Part13A
5December2011
exchangersincounterflow,vs.effectiveness,,ofasingleexchanger.Calculated
bythecompilerafterDomingos(1969)[36].
exchangersinparallelinthestreamoflowercapacityrate,vs.effectiveness,,ofa
singleexchanger.CalculatedbythecompilerafterDomingos(1969)[36].
234
ECSSEHB3101Part13A
5December2011
exchangersinparallelinthestreamofhighercapacityrate,vs.effectiveness,,ofa
singleexchanger.CalculatedbythecompilerafterDomingos(1969)[36].
1. Theoverallthermalconductance,U,isaconstantforeachexchanger.
2. Fluidcapacityrateratio,R,isconstant.
3. Eachfluidiscompletelymixedattheinletandoutletofeachexchanger.
4. InthecasesconsideredinFigure1125toFigure1128alltheexchangersareidentical.
Analytical expressions for nonidentical exchangerscan be found in Domingos (1970)[36]where, in
additiontotheassembliessketchedabove,mixedassembliesareconsidered.Insuchassembliessome
elementsareinoverallcounterflowandsomeothersinoverallparallelflow.
DatainFigure1125toFigure1128relatetheoveralleffectiveness,t,oftheassociationtothoseofthe
components,,irrespectiveofthetypeofexchangersinvolved.
Figure1125showsthedrawbacksassociatedtoassemblinginparallelanevennumberofexchangers.
The maximum achievable effectiveness of the association equals that for a parallel flow exchanger
(Figure1110).Toincreasetheeffectivenessoftheindividualcomponentwouldgivequitedefective
results;theoveralleffectivenesswoulddecrease!Asalimitingcase,noticethatwhentwofluidswith
thesamecapacityrate(R=1),flowthroughtwoidenticalcounterflowexchangers,witheffectiveness
= 1, associated in parallel, the exit conditions for the second exchanger are exactly equal to those
prevailingattheentranceofthefirstone.
Thesituationforcounterflowassembliesisquitedifferent(Figure1126).Theoveralleffectivenessof
the assembly is greater than that for each individual component, and the gain increases with the
numberofcomponents.
Figure 1127 and Figure 1128 give data for assemblies in parallel in one of the streams. In these
assemblies one of the streams is equally divided between n exchangers, each with the same
effectiveness.InFigure1127thestreamoflowercapacityrateistheonedivided,whereasinFigure
1128thestreamofhighercapacityrateistheoneflowinginparallel.
Onecandeducefromthesefiguresthat,whenanassemblyinparallelinoneofthestreamisused,the
streamwithlargercapacityshouldbethatsubdividedbetweentheexchangers.Buttheadvantagesof
thisconfigurationcanbefullyexploitedwhenthenumberoexchangers,n,is,atmost,equalto1/R.
235
ECSSEHB3101Part13A
5December2011
TheTablebelowsummarizesthedataonexchangerperformancewhicharepresentedinthefollowing
pages.
Arrangement Description Sketch Input Output Fig. References

Variables
Basic Counterflow Ntu Figure Kays&

arrangements R=0 119 London
0,25 (1964)[102]
Parallelflow 0,50 Figure
0,75 1110
1
Crossflow Fluids Figure Mason
Exchangers unmixed 1111 (1954)[124]
Onefluid Figure Kays&

mixed 1112 London
(1964)[102]
Countercross Reverse Ntu Figure Stevens,

flow order R=1 1113 Fernndez
Exchangers. n=1 &Woolf
Onefluid Identical 2 Figure (1957)[170]
unmixed order 3 1114
Parallel Oneshell Ntu Figure Kays&

counterflow pass R=0 1115 London
exchangers. 0,25 (1964)[102]
Onefluid Twoshell 0,50 Figure
passes 1116
unmixed 0,75
(Shelland 1
Threeshell Figure
tube
passes 1117
exchangers)
Fourshell Figure
passes 1118
Fiveshell Figure
passes 1119
Onetosix Ntu,R=1 Figure

shellpasses n=1,2,3,4,5,6 1120
Indirect Liquid Ntu1/Ntu2 C1/CLopt Figure Holmberg

Transfer coupled R= 1121 (1975)[90]
Heat heat 0,0,25,0,50,0,
exchangers exchangers 75,1

CL/CLopt,Ntu Figure

=0,5,1,2,10 1122
Ntu1/Ntu2=1,
R=0,1
236
ECSSEHB3101Part13A
5December2011
Arrangement Description Sketch Input Output Fig. References

Variables
CL/CLopt,Notu Figure
=2 1123
Ntu1/Ntu2=
1,5,R=0,1
(A1/A2)/(A1/ Figure
A2)opt+U1/U2= 1124
0,5,1,2
R=1,CL=CLopt
+U1(A1+A2)/
C1=5,10
Assemblyof Parallel t Figure Domingos

identicalheat assembly R=0 1125 (1969)[36]
exchangers 0,5
Counterflow Figure
1
assembly 1126
n=2
4
Inparallelin Figure
10
thestreamof 1127
lower
capacityrate
Inparallelin Figure
thestreamof 1128
higher
capacityrate
11.3 Exchanging surface geometries

The purpose of this clause is the geometrical description of the compact heat transfer surfaces for
whichthermalandfrictionaldataaregiveninKays&London(1964)[102],Clause14.
Besidesthesketch,designationanddimensionsofthesurfaces,allofthemreproducedinthisclause,
thefollowingdataareenclosedintheoriginalsource.
1. Coregeometrydatawhichrequiresomeelaborationsuchas:hydraulicdiameter,DE;free
flowarearatio,=AFL/AFR;areadensitiesor,whicheversignificant,andsoforth.
2. Finmaterialusedforthetests.Thiscouldbesignificantbecauseofthereasonsindicated
inclause11.5.3.
3. Heattransferandflowfrictiondata.
Heat transfer data are presented in terms of the Colburn factor, j = StPr2/3. Actually the tests were
made heating air either by means of a steady state steamtoair system or by means of a transient,
heatedwall,procedure.
FrictiondataaregivenintermsoftheFanningfrictionfactor,f.
237
ECSSEHB3101Part13A
5December2011
TheabsciseareinbothcasestheReynoldsnumber,Re.TodefineRethemeanvelocityisbasedonthe
minimumfreeflowarea.Inthecaseofmatrixsurfacesthefreeflowareaisdeducedfromtheporosity,
, and the frontal area of the exchanger, AFR. The characteristic length for calculating Re is the
hydraulicdiameterofthecore.
TheReynoldsnumber,Re,variesapproximatelyintherange200Re104.
ThefollowingTablepresentsseveralsymbolsusedbyKays&London(1964)[102]inconnectionwith
theabovedata.TheequivalencetothesymbolsusedinthisPartisalsoindicated.
Kays&London thisPart Definition

(1964)[102]
G V Flowstreammassvelocity
NPr Pr Prandtlnumber
NR Re Reynoldsnumber(NR=4rhG/)
NSt St Stauntonnumber(NSt=h/Gcp)
NStNPr2/3 j Colburnfactor
f f Fanningfrictionfactor
p Porosity
4rh DE Hydraulicdiameter
Areadensity.(Heattransferarea/Totalvolume)
Areadensity.(Heattransferarea/volumebetween
plates)
AFL Freeflowarea
11.3.1 Tubular surfaces
Sketch FundamentalTypesofsurfaces Table
Flowinsidecircularandflattenedcirculartubes Table
111
Flownormaltobanksofbaretubes Table
112
238
ECSSEHB3101Part13A
5December2011
Table111:FlowInsideCircularandFlattenedCircularTubes
Sketch Surface Dimensions Forheattransferand
Designationa frictiondataseethe
followingtablesand
figuresinKays&
London(1964)[102]
Table Fig.
ST1 SeeSketch 101 101

(Dimensionsare
giveninmm)

FT1 101 102
FTD1 101 103

a DesignationisthesameasinKays&London(1964)[102].STindicatesflowinsidestraighttubes,whileFT
referstoflattenedtubesandFTDtoflattenedanddimpledtubes.
239
ECSSEHB3101Part13A
5December2011
Table112:FlowNormaltoBanksofBareTubes
Sketch Surface Dimensionsb[m] ForHeatTransfer
Designation andFrictiondata
a seethefollowing
tablesandfigures
inKays&London
(1964)[102]
ax103 bx103 cx103 Table Fig.
S1,501,25(s) 6,4 9,5 7,9 102 104
S1501,25 9,5 14,3 11,9 105
S1,251,25 11,9 11,9 106

S1,501,00 14,3 9,5 107
S1,501,50 14,3 14,3 108
S2,001,00 19,1 9,5 109
S2,500,75 23,8 7,2 1010
I1,501,25(s) 6,4 9,5 7,9 102 1012
I1,501,25 9,5 14,3 11,9 1013
I1,251,25 11,9 1014

FT2 SeeSketch 102 1017

(Dimensionsaregiveninmm)
FTD2 102 1018
a DesignationisthesameasinKays&London(1964)[102].ThecapitalsSorIindicatewhetherthecircular
tubesarestaggeredorinline.Thenumeralswhichfollowgiverespectivelythetransverseandthe
longitudinalpitchratios.FTreferstoflattenedtubesandFTDtoflattenedanddimpledtubes.
b AdditionaldetailsonsurfacegeometryaregiveninKays&London(1964)[102],Table112.
240
ECSSEHB3101Part13A
5December2011
11.3.2 Plate-fin surfaces
PlateFinsurfaces,plainfins Table
113
PlateFinsurfaces,louveredfins Table
114

PlateFinsurfaces,stripfins Table
114
PlateFinsurfaces,wavyfins Table
116
PlateFinSurfaces,perforatedfins Table
117
PlateFinSurfaces,pinfins Table
118
Table113:PlateFinsurfaces,plainfins
a seethefollowing
tablesandfigures
inKays&
London(1964)
[102]
ax103 bx103 cx103 dx103 Table Fig.
2,0 305 19,1 13,6 10,2 103 1019
3,01 9,5 5,7 1020
3,97 6,4 4,8 1021
241
ECSSEHB3101Part13A
5December2011
Table113:PlateFinsurfaces,plainfins(continued)
a seethefollowing
tablesandfigures
inKays&London
(1964)[102]
4,00 SeeSketch 103 1022


6,2 103 1024
1,11(a) 103 1027
5,3 63,5 11,9 9,6 103 1023
11,1 6,4 4,6 1026

14,77 8,4 3,4 1028
19,86 65,6 6,4 2,6 1030
9,03 302 20,9 5,6 103 1025
15,08 174 10,6 3,4 1029
10,27T 63,5 13,8 4,9 103 1031
11,94T 127 6,3 4,3 1032
12,00T 63,5 6,4 4,2 1033
46,45T 66,8 2,5 1,1 1037
16,96T 127 6,5 3 103 1034
25,97T 63,5 5,2 2 1035
30,33T 8,8 1,7 1036

DesignationisthesameasinKays&London(1964)[102].Thenumeralsindicatefindensityinfinsperinch.
a
AdditionaldetailsonsurfacegeometryaregiveninKays&London(1964)[102],Table133.
b
242
ECSSEHB3101Part13A
5December2011
Table114:PlateFinsurfaces,louveredfins
a seethefollowing
tablesand
figuresinKays&
London(1964)
[102]
ax103 bx103 cx103 dx103 ex103 Table Fig.
3/86,06 9,5 6,4 8,4 2,8 1,4 104 1028
1/26,06 12,7 1040

3/87,8 9,5 6,4 5,8 1,5 1,4 1042
3/1611,1 4,8 6,4 4,6 0,9 1,4 1044
1/411,1 6,4 1045
3/811,1 9,5 1047
1/211,1 12,7 1049
3/8(a)6,06 9,5 6,4 8,4 0,9 3,3 104 1039
1/2(a)6,06 12,7 1041
3/8(a)8,7 9,5 6,4 5,8 0,9 2 1043
1/4(b)11,1 6,5 6,4 4,6 0,9 1,4 104 1046
3/8(b)11,1 9,5 6,4 4,6 0,9 1,4 104 1048
3/411,1 19,1 6,4 4,6 1,3 1 104 1050
3/4(b)11,1 10,1 6,4 4,6 1,3 1 104 1051

a DesignationisthesameasinKays&London(1964)[102].Thefirstnumeralsindicatefinlength,a,ininches,
thesecondindicatefindensityinfinsperinch.
243
ECSSEHB3101Part13A
5December2011
Table115:PlateFinsurfaces,stripfins
a seethefollowing
tablesandfigures
inKays&London
(1964)[102]
1/4(S)11,1 SeeSketch 105 1052

3/3212,22 2,4 12,3 2,1 105 1053
1/85,2 3,2 10,5 1,7 105 1054
1/813,95 9,5 1,8 1055
1/211,94(D) 12,7 6 2,1 105 1056
1/415,4(D) 6,4 5,2 1,7 1057
1/612,18(D) 4,2 9 2,1 1058

1/715,75(D) 3,6 7,7 1,6 1059
1/820,06(D) 3,2 5,1 1,3 1060
1/819,82(D) 5,2 1,3 1061
1/816,12(D) 5,2 1,6 1062
1/816,00(D) 6,5 1,6 1063
1/816,12(T) 3,2 8 1,6 105 1064

a DesignationisthesameasinKays&London(1964)[102].Thefirstnumeralsindicatefinlength,a,ininches,
thesecondindicatefindensityinfinsperinch.
244
ECSSEHB3101Part13A
5December2011
Table116:PlateFinsurfaces,wavyfins
Designation andFrictiondatasee
a thefollowingtables
andfiguresinKays
&London(1964)
[102]
ax103 bx103 cx103 dx103 Table Fig.
11,443/8W 9,5 10,5 2,2 2 106 1065
11,53/8W 9,5 2,2 2 1066
17,83/8W 10,5 1,4 2 1067

a DesignationisthesameasinKays&London(1964)[102].Thefirstnumberindicatesfindensityinfinsper
inchandthesecondfinwavelength,a,ininches.Wisforwavefinsurfaces.
Table117:PlateFinsurfaces,perforatedfins
Designationa andFrictiondatasee
thefollowingtables
andfiguresinKays
&London(1964)
[102]
Table Fig.
13,95(P) SeeSketch 107 1074

(Dimensionsaregivenin
mm)

a DesignationisthesameasinKays&London(1964)[102].Thefirstnumeralsindicatefindensityinfinsper
inch.Pisforperforated.
245
ECSSEHB3101Part13A
5December2011
Table118:PlateFinsurfaces,pinfins
a seethefollowing
tablesandfigures
inKays&London
(1964)[102]
Table Fig.
AP1 SeeSketch 107 1068

(Dimensionsaregiven
inmm)
AP2 107 1069
PF3 107 1070
PF4(F) 107 1071
PF9(F) 107 1072
PF10(F) 107 1073
a DesignationisthesameasinKays&London(1964)[102].Thisdesignationisnotdescriptiveofthesurface.
246
ECSSEHB3101Part13A
5December2011
11.3.3 Finned tubes

Finnedtubes,circulartubes,circularfins Table119
Finnedtubes,circulartubes,continuousfins Table1110
Finnedtubes,flattubes,continuousfins Table1111

Table119:Finnedtubes,circulartubes,circularfins
Sketch Surface Dimensionsb[m] ForHeatTransferand
Designation Frictiondataseethe
a followingtablesand
figuresinKays&
London(1964)[102]
ax103 bx103 cx103 dx103 ex103 fx103 Table Fig.
CF7,34 24,8 20,3 9,7 23,4 3,5 0,5 108 1075
CF8,72 2,9 1076
CF8,72(C) 10,7 21,9 2,9 1077
CF11,46 9,7 23,4 2,2 0,4 1078
CF7,05/8J 31,3 34,3 16,4 28,5 3,6 0,3 1079
CF8,75/8JA 31,3 34,3 16,4 28,5 2,9 0,3 1080
CF8,75/8JB 46,9
CF9,053/4J 39,5 44,5 19,7 37,2 2,8 0,3 1081

A
CF9,053/4J 50,3
B
CF9,053/4J 69,2
C
CF9,053/4J 69,2 20,3

D
CF9,053/4J 50,3 34,9

E
CF8,81,0JA 49,8 52,4 26 44,1 2,9 0,3 1082
CF8,81,0JB 78,2
a
DesignationisthesameasinKays&London(1964)[102].CFreferstocircularfins.Thefirstnumeralsreferto
thenumberoffinsperinch,thesecondrefertothenominaloutsidediameterofthetube,d,ininches.
247
ECSSEHB3101Part13A
5December2011
Table1110:Finnedtubes,circulartubes,continuousfins
a seethefollowing
tablesandfigures
inKays&London
(1964)[102]
ax103 bx103 cx103 dx103 ex103 Table Fig.
8,03/8T 25,4 22 10,2 3,2 0,3 1083
7,755/8T 38,1 44,5 17,2 3,3 0,4 1084

a DesignationisthesameasinKays&London(1964)[102].Thefirstnumeralsrefertothenumberoffinsper
inch,thesecondrefertothenominaloutsidediameterofthetube,c,ininches.
Table1111:Finnedtubes,flattubes,continuousfins
Designation andFrictiondatasee
a thefollowingtables
andfiguresinKays
&London(1964)
[102]
9,680,87 0 Otherdimensionsare 108 1085

given,inmm,inthe
9,680,87R 0,6 6,4 sketch 1087
9,10,737S 0 2,8 Other 108 1086

dimensions
9,290,737SR 0,6 6,4 2,7 aregiven,in 1088
mm,inthe
11,320,737SR 2,2 1089
sketch

a DesignationisthesameasinKays&London(1964)[102].Thefirstnumeralsrefertothenumberoffinsper
inch,thesecondrefertothetubedimensionintheflowdirection.Sindicatesstaggeredarrangement.
248
ECSSEHB3101Part13A
5December2011
11.3.4 Matrix surfaces
Table1112:CrossedRod,wovenscreenandspherematrices
Sketch Matrix Surface ForHeatTransfer
Surface Descriptionb andFrictiondata
Designationa seethefollowing
tablesandfigures
inKays&London
(1964)[102]
Porosity Transverse Table Fig.

Pitch,st
CrossedRod, 0,832 4,675 109 1090

Inline
Stacking 0,766 3,356
0,675 2,417
0,602 1,974
0,500 1,571
CrossedRod, 0,832 4,675 109 1091

Staggered
0,675 2,417

0,602 1,974
0,500 1,571
CrossedRod, 0,832 4,675 109 1092

Random
0,766 3,356
0,725 2,856
0,675 2,417
0,602 1,974
0,500 1,571
249
ECSSEHB3101Part13A
5December2011
Sketch Matrix Surface ForHeatTransfer

Surface Descriptionb andFrictiondata
Designationa seethefollowing
tablesandfigures
inKays&London
(1964)[102]
Porosity Transverse Table Fig.

Pitch,st
Wovenscreen 0,832 78
79
0,817
0,766

0,725
0,675
0,602
Spherematrix 0,37to0,39 1010 710
a DesignationisthesameasinKays&London(1964)[102].
b AdditionaldetailsoncrossedrodmatrixgeometryaregiveninKays&London(1964)[102],Table135.
11.4 Deviations from basic analysis
11.4.1 Introduction
The exchanger theory presented in Clause 11.2 is based on several idealisations. It is assumed, for
example,thatthermalconductionintheflowdirectioncanbeneglected,andthatthefluidflowdoes
not detach from the heat exchanger surfaces and is evenly distributed among the several flow
passages.
The conditions for the validity of these idealisations are usually met. Nevertheless, we can confront
configurationswhereoneoranotherofthesimplifyingassumptionsfails.
As an example of situation where the axial heat conduction along the exchanger core can hardly be
neglected, let us consider the heat exchangers used in typical cryogenic refrigeration cycles. These
exchangers are characterised by its very high effectiveness and limited pressure drop, and these
requirementsarefulfilledbyusinglargeheattransfersurfacesandshortflowpassages,thenceaxial
heatconductionbecomessignificant.
Simplifyingassumptionsconcerningthefluidflowpatterncannotbeeasilyintroducedinallcases.For
example, the flow in the shell side of shellandtube heat exchangers is so complicated that some
allowanceforthedepartureofactualflowfromthatassumedinthetheoreticalpredictionsshouldbe
takenintoaccount.Thisdeparturefromtheidealflowpatternsisusuallycalledmaldistribution.
250
ECSSEHB3101Part13A
5December2011
Flowmaldistributionappearsinexchangerswithparallelchannelsasanimbalanceofthefluidmass
flowrateamongthepassages.Thisimbalanceismainlyduetotolerancesinmanufacturing,pressure
drop in the manifolds, and deposition of impurities on the exchanging surfaces. The extent of the
problemincreasesrapidlyasthedesiredefficiencyisraised,andisalsomoreseverewithincreasing
the number of heat transfer units. Although the magnitude of the deviations from the nominal
performanceswillvarywiththecareusedindesigningandfabricatingtheheatexchanger,thereare
someupperlimitstothemanufacturingaccuracywhichcannotbeexceededatareasonablecost,thus
sometechniqueshouldbedevisedtocompensatetheimbalancewheneveritissignificant.
11.4.2 Longitudinal heat conduction

Ishasbeenassumeduptonowthatthereisnoheatconductionintheflowdirection,eitherthrough
thefluidorthroughthesolidwall.
The effect of axial heat conduction through the fluid is usually neglected in conventional heat
exchangerdesign,sincethethermalconductivityoffluidsissmall(liquidmetalsexcepted).Theorder
ofmagnitudeofthelongitudinalfluidconductioncomparedwithlongitudinalfluidconvectionequals
the inverse of the Peclet number, Pe = PrRe, so that the effect of longitudinal conduction can be
ignored in comparison with longitudinal convection effects for Peclet numbers exceeding 100. The
analysis by Schmidt & Zeldin (1970) [159] could be used to estimate the heat transfer rate in the
entranceregionofductstakingintoaccounttheeffectoflongitudinalconduction.
Wellconductioneffectsmaybeimportantinheatexchangersdesignedforhigheffectiveness(>0,9).
The effect of wall conduction is to reduce the temperature gradient along the fluid, thus decreasing
theheatexchangereffectiveness.
Aroughestimateoftheeffectofheatconductionthroughthewallmaybeobtainedbyequatingthe
ratiobetweenthedecreaseineffectivenessduetowallconduction,andtheeffectivenessinabsenceof
wall conduction, to the ratio between heat conducted through the wall and heat convected by the
flow.
k / L Ak
[1111]
C1
wherekisthewallthermalconductivity,Ltheflowlengthoftheduct,andAkthecrosssectionalarea
of the wall for longitudinal conduction. is the conduction parameter, defined as shown. For high
effectivenessexchangers(1),theexchangerineffectiveness,1,isapproximatelyequalto.
To obtain a better estimate of the effect of wall conduction it is necessary to solve the differential
equationsdescribingtheheatexchangerprocesswiththeappropriateboundaryconditions.Kroeger
(1967) [110] carried out such an analysis, for the case of a counterflow exchanger, both for balanced
flow,R=1,andforunbalancedflowR<1,underthebasicassumptionofconstantmaterialproperties.
Figure1129toFigure1131showtheineffectiveness,1 ,asafunctionoftheoverallnumberofheat
transferunits,
1 1 1
[1112]
N tu N tu1 N tu2
forthreedifferentvaluesofR.
251
ECSSEHB3101Part13A
5December2011
Figure1129:Heatexchangerineffectiveness,1,vs.numberofheattransfer
units,Ntu,foracounterflowheatexchanger.ResultsareshownforR=1and
differentvaluesofthewallconductionparameter,.FromKroeger(1967)[110].
units,Ntu,foracounterflowheatexchanger.ResultsareshownforR=0,95and
252
ECSSEHB3101Part13A
5December2011
units,Ntu,foracounterflowheatexchanger.ResultsareshownforR=0,90and
Typical values of Ntu range widely. For the Spacelab Freon to Water Interloop Heat Exchanger (see
clause 11.7) Ntu varies from about 5 to about 20, depending on the operating conditions (Owen,
Sessions & Walker (1976) [139], pp. 5455). On the other side, Cowans (1974) [28] describes a
counterflowheatexchanger,foracryogeniclifesupportsystem,whichwasdesignedtooperateata
valueofNtuaslargeas200.
Figure1132givestheineffectiveness,1 ,asafunctionoftheconductionparameter, ,forNtu=50
andseveralvaluesofthecapacityrateratio,R.
Figure1132:Heatexchangerineffectiveness,1,vs.dimensionlesswall
conduction,,foracounterflowheatexchanger.ResultsareshownforNtu=50and
differentvaluesofthecapacityrateratio,R.Calculatedbythecompilerafter
Kroeger(1967)[110].
Curveslabelled =0inthesefiguresrelatetheineffectiveness,1 ,to thenumberofheattransfer

units,Ntu,whenwallconductioneffectsareabsent,thencetheincreaseinineffectivenessforagiven
253
ECSSEHB3101Part13A
5December2011
nonzero value of can be estimated from the figures. It is seen that the wall conduction effects
increasewiththenumberofheattransferunits.ForlargevaluesofNtutheineffectivenessapproaches
anasymptoticminimumwhich,accordingtoKroeger(1967)[110],isgivenby:

lim 1 , for R 1 [1113]
N tu 1 2
or
1 R
lim 1 1 R
, for R 1
N tu R 1 R [1114]
e R
3R 1
It should be indicated that while the expression for R = 1 is exact, within the validity of the
assumptionsusedthroughouttheanalysis,theexpressionforR<1involvesfurtherassumptions.
TheresultspresentedinFigure1129toFigure1132alsoindicatethatthecapacityrateratio,R,hasa
strong influence on the ineffectiveness, the most severe deterioration arising when both fluids have
thesameflowcapacityrate(R=1).
11.4.3 Flow maldistribution

Flowmaldistributionappearswhenthefluidflowisnotuniformlydistributedovertheheattransfer
surfaceor,stillworse,whenportionsofthestreambypassthesurfaceentirely.
The problem of predicting performance as a function of flow maldistribution cannot be completely
solved because of the infinite number of ways the flow can distribute in a heat exchanger. The
purposeofthisclauseisthentodiscusstheseveralwaysinwhichmaldistributioncantakeplace;to
predict performance degradation in some very simple cases; to indicate some cures to performance
degradation,andtointroducetherelevantliteratureonthesubject.Noattemptismadeheretorelate
flowmaldistributiontomanufacturingtolerances,poorheaderdesignorinservicedegradationofthe
exchangingsurfaces,ratherthelevelofmaldistributionisdefinedbymeansofappropriateparameters
and the effect of these parameters on the performance is assessed. Manufacture and in service
degradation will be considered in Clause 11.5. The effect of header design on flow distribution has
beendiscussedbyCichelly&Boucher(1956)[26]andbyFreas&Ozisik(1965)[67],amongothers.
11.4.3.1 Simple analyses

In order to introduce the subject let us discuss several simple examples where performance
degradationcanbeestimatedwithoutdifficulty.
First we will consider the case of surface bypassing. The fluid velocity and, hence, the heat transfer
coefficient, h, are decreased by departure of a portion of the total outside stream from the available
area through the active surface. Thus the bypass stream reaches the exchanger outlet uncooled (for
unheated) in the worst conceivable case, and only there mixes with the active stream. The resulting
outlettemperaturewillbe:

To Ti 1 To1 To1 Ti T01 [1115]
254
ECSSEHB3101Part13A
5December2011
where isthemassfractionbypassed,andTo1istheoutlettemperatureoftheactivestream,whose
mass fraction is 1 . If the desired heat duty is to be accomplished, the activeoutlet to inlet
temperaturedifferenceisincreasedaccordingto:
1
To1 Ti To Ti [1116]
1
Thisformofmaldistributionappearsintheshellsideofshellandtubeheatexchangers,andwillbe
discussedwithsomedetailinclause11.4.3.2.
Flow maldistribution is not so drastic in the case of parallel channels although performance
degradationcanbemoreimportantifhighefficienciesarerequired.
To illustrate the effect of flow maldistribution on the performance of heat exchangers formed by
parallelchannels,wewillconsiderthefollowingsimpleexample,borrowedfromWeimer&Hartzog
(1973)[187].Letusstartwiththeproblemofcountercurrentflowbetweentwofluidstreamswiththe
samecapacityrateratio(Ch=Cc=C).Then,foraperfectlybalancedexchanger(seeclause11.2.2),
Q C Thi Tho C Tco Tci AU T [1117]
and
AU
N tu [1118]
C
Ifthetwofluidsarenotdividedevenlyamongtheseveralparallelsectionsoftheexchanger,thenfor
oneofthesesections,n,withtheinlettemperaturesfixed,
Qn C h ,n Thi Tho ,n C c ,n Tco,n Tci AU n Tlm ,n [1119]
wherethelogarithmicmeantemperaturedifference(seeclause6.2.1)isusedsincethehotfluidtocold
fluidtemperaturedifferenceisnolongerconstant,alongthepassagesbecauseoftheflowimbalance
(Ch,nCc,n).
Thi Tco,n Tho ,n Tci

Tlm ,n
Thi Tco,n [1120]
ln
Tho ,n Tci
Combinationoftheaboveequationsyields:
255
ECSSEHB3101Part13A
5December2011
1 Qn C
1
ln
N tu Q Cc ,n
N tu
AU n C

C
[1121]
1 Qn C AU C
1 h ,n C c ,n
N tu Q C h ,n
Illustrative results are given in Figure 1133 and Figure 1134. To calculate the data shown in these
figures the heat exchanger is assumed to be divided into two equal parts, each having surface An =
A/2, U n = U andCc,n=C/2,buthavingthehotfluidmaldistributedsuchthatinoneoftheExchangers
Ch,1=C(1+ )/2,whereasintheotherCh,2=C(1 )/2, beingthemaldistributionparameter.Itisalso
assumedthatthefluidspecificheatisindependentoftemperature.
Figure1133:Relativecapacity,Qn/Q,ofasimpletwofluidheatexchangervs.the
maldistributionparameter,,forseveralvaluesofthenominalnumberofheat
transferunits,Ntu.FromWeimer&Hartzon(1973)[187].
Figure1134:Relativesurfacerequirements,An/A,ofasimpletwofluidexchanger
forfixedtotaldutyvs.themaldistributionparameter,.Resultsareshownfor
severalvaluesofthenominalnumberofheattransferunits,Ntu.FromWeimer&
Hartzog(1973)[187].
256
ECSSEHB3101Part13A
5December2011
Figure 1133 shows the reduction in duty of the heat exchangeras afunction of the maldistribution
parameter,,forseveralvaluesofNtu.
Itcanbeseenthattheeffectofagivendegreeofmaldistributionisstronglyaffectedbythenominal
numberofheattransferunitsintheexchanger.Itshouldbesaid,however,thatlargevaluesof are
highly unrealistic; for a countercurrent platefin exchanger, maldistribution parameters of 0,1 are
normallytolerated.
Figure1134indicateshowmuchsurfaceshouldbeaddedtotheexchangertoachievethesameheat
transferdutyforamaldistributedcaseasforaperfectlydistributedcase.
NoticethatforeachdesignvalueofNtuthereisalimitingdegreeofmaldistributionbeyondwhichno
furtherincreaseinexchangersurfacecanrestoretheoriginalduty.
A slightly more general approach to performance prediction in terms of maldistribution has been
takenupbyFleming(1967)[65],whoconsideredasplitcounterflowheatexchanger,Figure1135.One
sideofthisheatexchanger,theuniformside,hasflowdistributeduniformlyamongallitschannels.
The other side, the nonuniform side,which carries the same totalflow rate as the uniform side,is
divided into two parts, the first one carries a lowerthanaverage flow in each channel uniformly
distribution among the fraction FLow of the total channels, whereas the second part carries a higher
thanaverage flow in each channel uniformly distributed among the fraction 1 FLow of the total
channels.
Figure1135:Flowdistributionmodelforthenonuniformside.FromFleming
(1967)[65].
ThefirstproblemconsideredbyFlemingcorrespondstothecaseofpairedchannels.Pairedchannels
meansthateachchannelwithhotfluidisinthermalcontactwithasinglechannelwithcoldfluidfor
the whole length of the exchanger. A purely counterflow shellandtube exchanger or a platefin
exchangeralmostfitsintothiscategory,althoughthechannelsarenotpairedstrictlyonaontoone
basis,butrather,eachchannelmayexchangeheatwithmorethanonechannelontheotherside.
Representative results are given in Figure 1136. Data on Figure 1136a have been computed for a
nominalnumberofheattransferunitsNtu=10,andthoseinFigure1136bforNtu=100.
257
ECSSEHB3101Part13A
5December2011
Figure1136:Overalleffectiveness,,andeffectivenumberofheattransferunits,
Ntueff,vs.thefraction,FLow,ofchannelswhichcarrylowerthanaverageflowonthe
nonuniformsideofapairedchannelsheatexchanger.Resultsareshownfor
differentvaluesoftheratioofthecapacityrate,CLow,ofasinglechannelwith
lowerthanaverageflowtothecapacityrate,CHigh,ofasinglechannelwithhigher
thanaverageflow.(a)isforanominalnumberofheattransferunitsNtu=10and
(b)forNtu=100.FromFleming(1967)[65].
ItcanbededucedfromFigure1136thattheeffectofflowmaldistributionincreaseswithincreasing
thenominalnumberofheattransferunitsandcanbeverysevereforhighNtuvalues,thenceastrong
emphasisshouldbeplacedongoodheaderdesignandonchanneluniformity.Asademonstrationof
the effect of flow maldistribution, suppose an exchanger designed with Ntu = 100 for an overall
effectivenessof0,99.If,however,halfthetubesononesideeachcarries0,8oftheflowcarriedbyeach
tubeoftheotherhalf(FLow=0,5,CLow/CHigh=0,8inFigure1136b)theactualoveralleffectivenesswillbe
only=0,944andtheeffectivenumberofheattransferunits,Ntueff16,5.
Thesituationismuchmorefavourableifitisassumedthattheuniformflowsideiscompletelyand
continuously mixed throughout the heat exchanger. An example ofexchanger which comes close to
thiscasewouldbethepurelycounterflowshellandtubeexchanger,ifthetubesideisconsideredto
bethenonuniformside.Asaconsequenceofthecompletemixingassumptionthefluidtemperature
intheuniformsideisuniformatanycrosssection.
ResultsfromcalculationsperformedbyFleming(1967)[65]aregiveninFigure1137.Itcanbeseen
that even for large non uniformities (CLow/CHigh = 0,5) there is little effect on the exchanger thermal
performance,andthatthiseffectismorepronouncedforsmallervaluesofNtu.
Figure1137:Overalleffectiveness,,vs.thefraction,FLow,ofchannelswhichcarry
lowerthanaverageflowonthenonuniformsideofaheatexchangerwithuniform
258
ECSSEHB3101Part13A
5December2011
sidemixed.Resultsareshownfordifferentvaluesofthenominalnumberofheat
transferunitsNtu,andoftheratioCLow/CHigh.FromFleming(1967)[65].
Complete mixing at any cross section is an ideal limit which cannot be realised in any real heat
exchanger.Astepinthedirectionofcontinuousmixingwouldbethecollectingandmixingofoneor
bothstreamsatintervalsalongtheexchangerlength.ThiscasehasbeenalsoconsideredbyFleming
(1967).Inordertomovebeyondtheconsiderationspresentedabove,itisnecessarytoparticularisethe
studytoconfigurationsofactualheatexchangersandtodevelopmathematicalmodelsforsimulating
actualexchangerperformance.Resultsforhelicalcoilwoundexchangersandforplatefinexchangers,
bothtwofluidandfourfluidexchangers,havebeengivenbyWeimer&Hartzog(1973)[187].
11.4.3.2 Maldistribution compensating techniques in shell-and-tube heat

exchangers
The ideal flow distribution in the shell side of a baffled exchanger is shownschematicallyin Figure
1138.
Figure1138:Idealflowdistributionintheshellsideofshellandtubeheat
exchangers.
The actualsituationis farfrom ideal, however, unless especial provisions are taken to minimise the
adverseeffects.TheseveralfluidpathwayswhichappearinpracticalcasesareshowninTable1113.
Although the performance degradation associated with each fluid pathway can be estimated (see a
critical survey of the available design methods by Taborek (1974) [171]) engineering practices for
reducingmaldistributioninthesellsideofshellandtubeexchangersarewellknown.Thesepractices
aresummarisedinTable1114.
259
ECSSEHB3101Part13A
5December2011
Table1113:FluidPathsintheShellSideofShellandTubeExchangersa
Stream Definition Sketch SourceofMaldistribution Comments
B Truecrossflowstream Basisforcomparisonof
bothheattransferand
pressuredrop.
A Flowthroughthegap Standardbaffletubeholesare Fullyeffectiveforheat

betweenthetubeandthe drilled0,79x103m(1/32)over transfer.Decreasesthe
baffletubehole,whichis thetubeouterdiameterfor overallpressuredrop.
inducedbythepressure unsupportedtubelengths

differentialonthetwosides below0,914m.Forlarger
ofthebaffle lengthsorforpulsating
conditionstubeholesmaybe
drilledsmallerthanstandard
(TEMA(1968)[176]).
C Flowbypassingthebundle Bundletoshellclearance. Partiallyeffectivefor

Blockingdevicesarebeused heattransferasthe
whentheseclearancesare fluidcontactstheheat
largerthanapproximatelyone transferringsurfaces

tubelayoutpitch. ononesideofthe
tubularfieldonly
D Flowthroughopenpassages Openpassagesintubelayout Moreeffectiveforheat

intubelayoutpartitions partitions. transferthanCstream.
whentheyareinthe Useofblockingdevicesis
directionofthemain recommended.
crossflowstream
260
ECSSEHB3101Part13A
5December2011
Stream Definition Sketch SourceofMaldistribution Comments
E Leakageflowbetweenbaffle Shelltobaffleclearance.For Veryineffectivefor

andshell. heatexchangersofgeneral heattransfer.Itgreatly
purposeapplicationsthis distortsthe
clearanceisequalto2,54x103m temperatureprofiles.

fornominalshellinside
diametersintherange0,152m
to0,330m(TEMA(1968)[176]).
a ArrangedbythecompilerafterTaborek(1974)[171].CapitalletterslabelingthedifferentpathwaysarethoseusedbyTaborek.
Table1114:EngineeringPracticesforReducingMaldistributionintheShellSideofShellandTubeExchangersa
Configurationandcorrectionpractice Sketch Comments
Basicconfiguration. Theflowintheshellsideisessentiallyparalleltothe
Conventional4passlayout.Nobafflesareused. tubes,andbecauseoftheopenspaceandthefrictional
Thefluidinthetubesideflowsperpendicularlytothe resistanceofthetubes,theactualvelocityinthevicinity
planeofthefigure. ofthetubesisconsiderablylessthanthevelocitythrough
Intwoofthebundlesthefluidflowtowardtheobserver, theopenspaces.Calculationsoftheheattransfer
andintheothertwofromtheobservertowardthefigure, coefficientbasedonthemeanvelocitythroughtheshell

thusthefluidinthetubesidetraversesfourtimestheheat sidewillbeinerror.
exchangerbeforeleavingit.
Additionofsegmentedbaffles. Evenassumingthatclearancesbetweeninsidediameter
Abafflecutofapproximately20%isprovidedasindicated oftheshellandbafflediameterarekepttoaminimum
inthesketch. andthatotherbypasspathsareminimized,mereaddition
ofsegmentalbafflesdoesnotguaranteeagreement
betweentheoreticalpredictionandactualperformance.
Theflowispartlyparallelandpartlynormaltothetubes

and,inaddition,bypasslanesstillexist.
261
ECSSEHB3101Part13A
5December2011
Configurationandcorrectionpractice Sketch Comments
Additionoftierodsandconcentricspacersmaybeso Tierodsandconcentricspacersareusedtoholdbafflesin
chosenastoblockthebypassareabetweentheoutermost positionagainstmechanicalforcesduringbundle
tubesandtheinsideoftheshell. assemblyorremoval,andagainstfluidforcesduring
operation.
Insertionofrodsordummytubes. Theaimoftheserodsistoblockthebypassareintube
layoutpartitions.Thenumberandsizeofthedummy
tubesvarywiththesizeofthetubebundleandthewidth
ofthepasslane.Itisconsideredgoodpracticetousea
dummytubeateveryfifthrow.
a ArrangedbythecompilerafterGilmour(1965)[72].
262
ECSSEHB3101Part13A
5December2011
11.4.3.3 Maldistribution compensating techniques in parallel counterflow heat

exchangers
Awaytosolvetheimbalanceprobleminparallelcounterflowheatexchangershasbeensuggestedby
Cowans (1974) [28]. The device, which can only be used with gases, is based on the effects of
maldistributionuponthefluidtemperatureprofilesalongtheheatexchangers,Figure1139.
Figure1139:Schematicofthetemperaturedistributionalongtheheatexchanger.
(a)Balancedcase.(b)Imbalancedcase.FromCowans(1974)[28].
Figure1139ashowsthetemperatureprofileforabalancedheatexchanger.Theresultingcurvesare
parallelstraightlines.Figure1139b,initsturn,correspondstoacaseofflowimbalance.Itisrealised
from these figures that when the mass flow rate of the hot flow becomes higherthanaverage the
temperatureprofilesshiftup,particularlyinthemiddleoftheheatexchanger,andthattheywillshift
down when the mass flow rate of the hot flow becomes below average. This shift in temperature
inducesashiftinpressuredrop.
Forlaminarflow,
dp
~ [1122]
dx
sincedp/dxm/(seeclause7.2.2).
Forturbulentflow,ontheotherhand,
dp 0, 2
~ [1123]
dx
sincedp/dx m2/and Re0,2 to 0,25(seeclause7.2.2andTable72inclause7.2.2.1).Inaddition,for

mostgasesT0,6to0,8andT1.Thus,dp/dxincreaseswhenthetemperatureincreasesforbothlaminar
andturbulentflowofgases.
Theshiftintemperaturedoesnotmitigatetheflowimbalanceinatypicalheatexchangersinceboth
sides of the exchanger are affected in the same manner because the temperature profiles are quite
similar.Cowans(1974)[28]conceivedtheideaofshapingthetwosidesoftheheatexchangerpassages
differently(Figure1140)sothatthetemperatureshiftinthemiddleoftheexchangerstronglyaffects
the pressure drop of the hot flow only. According to Cowans the power of this feedback is quite
substantial. He reports that the performance of a highefficiency ( > 0,9925) balancedflow
263
ECSSEHB3101Part13A
5December2011
countercurrentheatexchangerdesignedforoperationatNtu=200,wassodegradedbyflowimbalance
thattheactualeffectivenessoftheprototypewas =0,95andNtu20.Afterimprovementofdesign
and manufacturing techniques the performance improved to = 0,97, Ntu = 33. No further
improvements in manufacturing techniques could be attained. After the feedback compensation
technique was conceived another unit was designed and built identical to the last unit but with the
passagestaperedfrom0,7x103mto0,3x103minonedimension,asindicatedinFigure1140.All
other parameters remained unchanged. The efficiency of the modified unit improved to = 0,994
givingNtu=167.
Figure1140:Flowimbalancecompensatingtechniqueforgastogasheat
exchangers.FromCowans(1974)[28].
Thisfeedbackcompensationtechniquecannotbeusedwithliquidssinceinthiscase ispractically
temperatureindependentwhereasdecreaseswhenthetemperatureincreases.Thence,anincreasein
temperatureproducesadecreaseinpressuredropandconversely.
Noinformationhasbeenfoundonflowimbalancecompensatingtechniquesforliquidtoliquidheat
exchangers.
11.5 Manufacturing defects
11.5.1 Introduction
The advantage which present heat exchangers employing high area density surfaces can be fully
exploitedwhenmanufacturingtolerancesarekeptunderclosecontrol.Nevertheless,therearesome
upperlimitstotheachievableaccuracy.Forinstance,itisdifficulttofabricateheattransfersurfacesto
adimensionaltolerancebetterthan5%.Thisfigurewouldbetypicalofverygoodpractice.
The aim of this clause is to relate manufacturing imperfections to performance degradation in real
heat exchangers. In several instances analytical models exist which provide useful guidance in the
definitionofthemanufacturingtolerances.Inothercases.however,onemustresorttothetestingof
representative configurations. The influence of manifold irregularities is difficult to assess since any
effect due to this factor can be hidden in the overall design by increasing the total pressure loss
throughtheexchangercore.AccordingtoCowans(1974)[28],a5%variationinthepressurelossfor
agivenmassflowrateduetothiscausecouldbeconsideredexcellentpractice.
11.5.2 Variations of the flow passages

A theoretical analysis of the performance degradation in laminar flow regenerators (clause 11.1)
because of small variations in spacing between corrugations was made by London (1970) [118]. The
nonuniformitiesconsideredareeitheroftheplatespacingtype,Figure1141a,orofthefinspacing
type,Figure1141b.ThegeometricalcharacteristicsofthesurfacesaresummarisedinTable1115.
264
ECSSEHB3101Part13A
5December2011
Figure1141:Platespacing,(a),andfinspacing(b),typeflowpassagenon
uniformities.FromLondon(1970)[118].
Table1115:NominalDimensionsforNumericalResults
Glassceramicisoscelestriangular Stainlesssteeldeppfoldconfiguration
passagegeometry
Cellbasecx103[m]:0,643 Passagewidthcx103[m]:0,360
Cellheightdx103[m]:0,457 Passageheightbx103[m]:2,88
Cellheighttowidthratio:0,710 Passageaspectratiob/c:8,00
Apexangle[angulardegrees]:109,2 Foilthicknessax106[m]:53,3
PassagecountNx106[Passages.m2]:3,4 Fractionoffoilsurfacecoveragec,y:0,125
Porositya:0,650 PassagecountNx106[Passages.m2]:0,809
Areadensityb[m2.m3]:7900 Porositya:0,840
FlowareaperpassageAFLx106[m2]:0,190 Areadensityb[m2.m3]:5250
HydraulicdiameterDEx103[m]:0,330 FlowareaperpassageAFLx106[m2]:1,037
HydraulicdiameterDEx103[m]:0,640
a Porosity,istheratioofflowaretofrontalareaofmatrixcore.
b Areadensity,,istheratioofheattransfersurfacearea(twosides)tocorevolume.
c Fractionoffoiledsurfacecoverage,y,isthefractionoffoilstocksurfacethatcoversitselfforthedeepfold
surface.Forexactlyrectangularfolding:y=(c+2a)/(2c+b+2a).
NOTE FromLondon(1970)[118].
TheassumptionsonwhichtheanalysisisbasedarelistedinTable1116.Theapproachcanbeapplied
toothergeometricalconfigurations,othertypesofnonuniformitiesandotherflowconditions.
265
ECSSEHB3101Part13A
5December2011
Table1116:Basicassumptionsforthetheoreticalanalysis
1.Heattransfercoefficientandfrictionfactorareconstantwithpassageflowlength.
2.Theflowisassumedtobelaminar(thepassageReynoldsnumber,Re,islessthan500).Although
inthecaseoftriangularpassagegeometry,thecellsareisoscelestrianglesofapex109deg,datafor
equilateraltrianglesareusedintheanalysis.
3.Endlossesaresmallenoughrelativetotubefrictionsothattheirinfluenceisnegligible.
4.Fluidproperties,,andPrarethesameforallpassagesandcanbetreatedasconstant.
5.Todealwithflowpassagecrosssectionnonuniformities,thefollowingassumptionsaremade.
5.1.Walltemperatureincreaseslinearlyalongeachpassagereachingacommonmaximumvalueat
theexit.
5.2.Thefluidbulktemperatureatthepassageentranceisthesameinanycase.
5.3.Walltofluidtemperaturedifferenceremainsconstantalongthetube.
Asaconsequenceoftheseassumptionsthefollowingsimplerelationsbetweenpassageheat
transfereffectiveness,,andpassagenumberofheattransferunits,Ntu,holds:
N tu ,n
n , n 1,2
1 N tu ,n
whereandNtuaredefinedasfollows:
Tbo Tbi

Two Tbi
hDE L Q
N tu
mc p mc p Tw Tb
Ntueffforthecombination12passagesystemcanbeexpressedas
av
N tueff
1 av
whereavrepresentstheweightedaverageofthemixedflowoutletstreamtemperatures,
m1 1 m2 2
av
m1 m2
5.4.Themassflowratesm1andm2throughpassages1and2respectively,aredistributedsoasto
ensurethatthepressuredropisthesameforbothpassages.
NOTE ArrangedbythecompilerafterLondon(1970)[118].
TheresultsarepresentedinFigure1142asaheattransferperformancepenaltyandapressureloss
reduction (a gain) relative to the behaviour of the exchanger with uniform geometry, versus the
geometrical nonuniformity, which is measured in terms of the relative change of a characteristic
lengthofthepassagecrosssection.
266
ECSSEHB3101Part13A
5December2011
Figure1142:Degradationintheheatexchangerthermalperformance,measured
byCostNtu,andrelativegaininpressureloss,1(p1/pn),asfunctionsofthe
deviationinchannelsize,1(DE1/DEn).Resultsin(a)applytoanycylindrical
passageprovidedthatthenonuniformitiesaregeometricallysimilar.Resultsin
(b)arefornonuniformitiesofthefinspacingtype.FromLondon(1970)[118].
The degradation in the heat exchanger thermal performances is expressed by means of the cost
fraction
N tueff
Cost N tu 1 [1124]
N tu
Ntucorrespondingtothenominalconfiguration.
Manufacturingtolerancescanbetentativelyspecifiedonthebasesoftheseresults.Itisnotyetclear
whatstatisticalmethodologyshouldbeusedtorelatemanufacturedcores,withadistributedrangeof
sizes,totheequivalenttwosizepassagemodelusedintheanalysis.Asatentativeproposal,basedon
Figure1142aandb,thefollowingmanufacturingtolerancetargetsarerecommended.
1. Geometries where the distortions tend to preserve geometrical similarity with the
nominalconfiguration
D
2

1 E 20 percent
D [1125]
En av
and95percentofthepassageswithin40percent.Thiscorrespondsto1(DE1/DEn)=0,10
inFigure1142a.
267
ECSSEHB3101Part13A
5December2011
2. Deepfoldplatefinsurfacesforfinspacingnonuniformities
c
1 10 percent [1126]
cn av
and95percentofthepassageswithin20percent.
11.5.3 Fin leading edge imperfections

Thiseffectisparticularlynoteworthywhenstripfinsurfaces,asthoseshowninFigure113e,areused
incompactheatexchangers.Finsofthistypearegenerallyconstructedbyamachinecuttingprocess
thatinevitablyleavesaslightlybentandscarfedfinedge,whosegeometricalcharacteristicsdepend
upon the fin material and upon the cutting tool. Since a few microns of scarfing can have a
considerable effect on the friction performance of the surfaces, it is difficult to either dimensionally
describethiseffectortoexactlyreproducethetestsurfaces.
Asoftmaterial,suchasaluminium,maybeexpectedtoyieldfinedgeswhicharesomewhatdifferent
that those obtained from a hard material, such as stainless steel. Thence, some experimental
uncertainty should be accounted for when performance data obtained with aluminium test surfaces
areappliedtosurfacesofstainlesssteelconstruction.Inthesesituations,someconservatismindesign
isadvisable(Kays&London(1964)[102]).
11.5.4 Brazing
Brazing is used in compact exchanger surfaces for structural stability purposes and for insuring
adequatethermalcontactbetweenfinandbaseplate.
The influence of brazing on the heat transfer characteristics is small. Brazing results in coverage of
surface with consequent reduction of the total heat transfer area. This is detrimental in the case of
regeneratorswherebrazingisusedonlyforstructuralpurposes.
On the other hand, the influence of brazing on flow friction is quite significant because of the
roughnessitintroduces.Thisroughnessisoftwotypes:1)afinegrainedsurfaceroughnessand2)that
produced by the flow of excess braze material during the fusion process which partially blocks the
flow passages. The influence of brazing roughness decreases as the core becomes less compact,
because then both the dimensionless roughness, e/DE, and the blocking of the flow passages are
reduced.
The influence of brazing on the performances of compact heat exchangers has been studied
experimentally by Shah & London (1970) [164]. Although these authors tested three different heat
transfersurfaces,onlytwoofthemwillbeconsideredherebecausetheirgeometriesareverysimilar
andbecausetheirperformancesweremeasuredbyusedofthesameexperimentalmethod.
Theconfigurationschosenforthetestareoffsetrectangularplatefinsurfacesoftheplatefinstripfin
variety.ThegeneralfeaturesofthefinsaredepictedinFigure1143,whereasTable1117providesa
comparisonofthetestedsurfaces.
268
ECSSEHB3101Part13A
5December2011
Figure1143:Fincenterofoffsetrectangularplatfinsurface.FromShah&London
(1970)[164].
Table1117:CharacteristicsofOffsetRectangularPlateFinSurfaces
CoreNumbera 501 501MOD
Platespacing,bx103[m] 0,673 0,665
Finpitch,cx103[m] 0,688 0,711
Finoffsetlength,lx103[m] 1,27 1,27
Platethickness,ax103[m] 0,154 0,025
Finthickness,x103[m] 0,025 0,025
Porosityb, 0,744 0,896
Areadensityc,[m2.m3] 4610 6240
Areadensityd,[m2.m3] 5650 6483
Hydraulicdiameter,DEx103[m] 0,645 0,574
Dimensionlesscorelengthinflowdirection,L/DE 101,2 113,2
Dimensionlessfinoffsetlength,l/DE 1,97 2,21
Aspectratio,*=(b)/(c) 0,977 0,933
Brazedconstruction yes no
a Authorsdesignation.
b Porosityistheratiooffreeflowairononesideoftheheatexchangercoretofrontalareaofthatside.
c Areadensityaistheratiooftotalheattransferareaofonesideoftheheatexchangertototalexchanger
volume.
d Areadensitybistheratiooftotalheattransferareaononesideofaplatefinheatexchangertothevolume
betweentheplatesonthatside.
NOTE FromShah&London(1970)[164].
269
ECSSEHB3101Part13A
5December2011
Bothsurfaceswerebuiltintocoresfortheheattransfertestingwitha82,5x103by82,5x103mtest
crosssectionanda65,3x103mairflowlength.InCore501,thefinswereseparatedbyplatesof0,152
x103mthicknessandthecorewasbrazed.InCore501MODthefinsandtheplatesbothwere0,025x
103mthickandwerestackedalternativelywithoutbrazing.Materialfortheplatesandfinsforboth
coreswasstainlesssteel(AISI347type).
Thevaluesof,,andDEinTable1117arecalculatedonthebasisofsomegeometricalidealisation
ofthecellsconstitutingthecore.Figure1144showsthemodelsusedtoderivetheseproperties.
Figure1144:Assumedcoregeometries.FromShah&London(1970)[164].
Model (b) represents fairly realistically the unbrazed surface, which did not exhibit as complete
mechanicalcontactbetweenfinsandplateascouldbewronglydeducedfrommodel(c).Thevaluesof
the relevant parameters as deduced from models (b) and (c) are compared in Table 1118. It can be
seen that model (b) yields the higher area density, and consequently the lower DE, by about 14
percent.
Table1118:Core501MODGeometriesDerivedfromDifferentModels
Table1117 RectangularModel ModifiedModel
Figure1144c Figure1144b
Cellheight,dx103[m] 0,691 SAME
Cellwidth,cx103[m] 0,711
Platethickness,ax103[m] 0,025
Finthickness,x103[m] 0,025
Porosity, 0,896a 0,893 0,902
Areadensity,[m2.m3] 6240 5381 6240
Hydraulicdiameter,DEx103[m] 0,575 0,667 0,574
a Porositydeterminedgravimetrically.
NOTE FromShah&Londono(1970)[164].
270
ECSSEHB3101Part13A
5December2011
Heattransfertestsweremadebythesingleblowtransienttestingtechnique.Asummaryofthetheory
underlyingthistechnique,adescriptionofanexperimentalfacility,andcommentsontheapplicability
ofthetestresultscanbefoundinPucci,Howard&Piersall(1967)[145].Pressurelosswasmeasured
bytheconventionalsteadyflowprocedure.
The dimensionless heat transfer and fluid friction characteristics are presented in Figure 1145 as
functions of the Reynolds number, Re, based on the hydraulic diameter, DE. Heat transfer data are
presentedintermsoftheColburnfactor,j(seeEq.[1112]inclause8.2.4).Fluidfrictiondataaregiven
in terms of the Fanningfriction factor, f (see clause 7.2.2). Estimates of the experimental uncertainty
areasfollows:j13%,f3%andRe2%.
It can be seen from Figure 1145 that both surfaces exhibit quite closely the same heat transfer
characteristics.AtlowReynoldsnumberthegrossblockageduetobrazingwouldreducejslightly.At
high Reynolds number, however, brazing roughness induces an earlier transition to turbulence
improvingtheheattransfer.Thedifferentslopeofthetwocurvescouldbealsoduetoa10percent
differencein1/DE,asindicatedinTable1117.
Figure1145:Heattransfer,j,andfriction,f,characteristicsasfunctionsof
Reynoldsnumber,Re,forsurfaces501and501MOD.FromShah&London(1970)
[164].
Thefrictionalcharacteristicsare,ontheotherhand,stronglyinfluencedbybrazing.Thesurfacewith
the greatest relative roughness (Core 501) has 3550 percent higher friction factor relative to the
smoothsurface(501MOD).
Figure 1146 shows the flow area goodness factor, measured by the ratio j/f, as a function of the
Reynoldsnumber,Re.
271
ECSSEHB3101Part13A
5December2011
Figure1146:Flowareagoodnessfactor,j/f,asafunctionofReynoldsnumber,Re,
forsurfaces501and501MOD.FromShah&London(1970)[164].
11.6 In service degradation
11.6.1 Introduction
Performancedegradationofrealheatexchangerscouldbeduetoincompatibilityoftheworkingfluid
withthematerialsurfacesandtoimpuritydeposits.Dataonmaterialcompatibilityoftypicalliquid
coolantswithbothmetalsandelastomersaregiveninclause10.3,Table102toTable109.Theeffect
ofimpuritydepositscanvaryenormously.Ifthestreamispureandthetimebetweenoverhaulsshort,
theeffectcanbenil.Becauseoftherelativelyshortorbitaltimeofpresentmannedspacecraft,which
aretheonlysatellitesinwhichtheuseofrelativelysophisticatedfluidloopsisenvisaged,theeffectof
impuritydepositsisnegligible.Nevertheless,theproblemwillbecomemoreacuteinthefuturewith
theadventofreusablespacevehicles.
11.6.2 Fouling
The term fouling is used in the literature to describe any undesirable deposit on heat exchanger
surfaces which increases resistance to heat transfer. There are a great variety of fouling forms
(polymerisation, chemical reaction deposits, crystallisation, sedimentation, corrosion, etc.). This fact,
preventsasystematictreatmentoftheproblem.
In general, the effect of fouling is described in terms of a fouling resistance, rf (often referred to as
foulingfactor),whichisaddedtotheinversethermalconductance(seeEqs.[112]inclause11.2.2).
TheexpressionforUh,thusbecomes
1
1 a 1 A
Uh r f h r fc [1127]
h hh Aw / Ah k Ac / Ah c hc Ac
272
ECSSEHB3101Part13A
5December2011
wherethevalueofrfineachparticularcasecanbefoundinsomeexperiencebasedtable.Themost
comprehensivesourceoffoulingdatahasbeencompiledbytheUSTubularExchangerManufacturers
Association(TEMAStandards).SomeusefulvaluesaregiveninTable1119.
Table1119:FoulingResistanceofSeveralCommonMaterialsa
Material rfx104
[K.m2.W1]
GasesandVapors
Steam 0,88
RefrigerantVapors 3,5
CompressedAir 3,5
IndustrialOrganicHeatTransferMedia 1,8
Liquids
DistilledWater 0,88
RefrigerantLiquids 1,8
HydraulicFluids 1,8
IndustrialOrganicHeatTransferMedia 1,8
MoltenHeatTransferSalts 0,88
a FromTEMAStandars(1968)[176].
While the fouling process is obviously a time function, a constant value of the fouling resistance is
generallyusedfordesign.Thisiftheninterpretedasavaluetobereachedinsomereasonabletimeat
whichtheuseroftheequipmentiswillingtocleanit.However,thefactthatduringshorttimesthe
equipment is clean and, therefore, will operate under frequently drastically different conditions is
rarely examined. Designing with a large fouling resistance, which is thought to be reached after a
certaintime,oftenleadstoamorerapiddeteriorationoftheexchanger(Taborek,Aoki,Ritter,Palen&
Knudsen(1972a)[173]).
Inpractice,largefoulingresistancesoccuroftenasaconsequenceofpoordesign.Toavoidthis,care
shouldbetakentopreventstagnantregions,bypasses,flowmaldistributions,inadequatelocationof
bafflesandrods,etc.(Gilmour(1965)[72]).
For many typical heat transfer processes, an improvement in the ability to predict the heat transfer
coefficient will be useless, unless accompanied by a correspondingly increased understanding of
fouling.
Yet,theeffortspresentlydevotedtofoulingarealmostnegligible.AccordingtoTaboreketal.(1972a)
[172],themainreasonsforthisstateofaffairsare:
1. Foulingisillsuitedforuniversitybasedinvestigations.
2. Itisdifficulttocontroltherelevantparametersinalaboratory.
3. Acontinuoussupplyofflow,alwaysunderthesameconditions,isrequired.
4. Longtimeisneededtoobtainmeaningfuldata.
5. Therearemanydifferenttypesoffouling.
6. Foulingisinfluencedbymanydifferentfactors.
Predictive methods for fouling behaviour are scarce. A good review of the models proposed in the
pasthasbeenrecentlycompletedbyTaboreketal.(1972b)[173].
273
ECSSEHB3101Part13A
5December2011
Thedifferentformsoffouling,andtheeffectofthedifferentparametersforeachfoulingprocessare
discussedinthefollowingtables,especialemphasisisplacedonthoseprocessesrelevantinthermal
control technology. Table 1120 describes the various fouling mechanisms. The effects on fouling of
differentphysicalparametersaresummarisedinTable1121.
Table1120:FoulingMechanismsa
Mechanism Description
Crystallizationof Oneofthemostcommoncausesoffouling.
inversesolubilitysalts.b
Crystallization fouling exhibits different behavior depending on
whetherthesolutioncontainsasinglesaltoramixtureofsalts.
1.Whenthesolutioncontainspurely(orprimarily)asinglesalt,the
crystallineformationsshowfreatadherencetothesurfaceandbetween
themselves.Becauseofthisstrongbond,theremovalmechanismis
ineffectiveornegligible,sothatfoulingisinsensitivetoflowvelocityand
thicknessofthedepositlayer.Thence:
Foulingtimecurveisgrowingfunctionwithnoasymptoticvalue.
Generally, an induction period exists during which no fouling
deposit occurs, until suddenly fouling starts and keeps growing
thereafter.Theinductionperiodisafunctionofsurfacematerialand
roughness.
2.Whenthesolutionisamixtureofsalts,crystallineclustersbuildupin
irregularpatterns,formingcavitiesbetweenthemwhichpermitdeposition
ofsuspendedparticles,thusdecreasingthecrystallinecohesion.Inthiscase
theremovalmechanismisafunctionoftheflowvelocity,andofthe
thicknessofthedepositlayer.Thence:
Foulingtimecurveapproachesanasymptoticvalue.
The variety of crystalline forms, makes fouling initiation less
sensisitivetothesizeanddistributionofnucleationcavities.
Sedimentation Occursinmanypracticalcasesundernormalgravityconditions.
Some of the sedimented particles can act as catalysts or enter into
complexreactions.
Therearemanytypesoffoulingbysedimentationandthereforeitis
difficulttostateconclusions.
Verysensitive,ingeneral,toflowvelocity.
ChemicalReaction Inmanycasescombinedwithcrystallization.
andPolymerization
Acriticalvariableisthesurfacetemperature,sincethereactionrate
willbeverysensitivetochangesinsurfacetempearture.
Normally the foulingtime curves approach an asymptotic value,
indicating the existence of flow velocitydependent removal
mechanisms.
OrganicMaterial Commontountreatedwater.
274
ECSSEHB3101Part13A
5December2011
Mechanism Description
Growth
In most cases combined with crystallization or sedimentation
fouling,itcan,however,existasthepredominantform.
Existing studies are largely confined to specification of proper
treatmentmaterial.
Corrosion Duetotwodifferentmechanism.
1.Creationofacrusty,heatresistanceproducinglayer,suchasrust.
2.Rougheningofthesurface,invariablyconnectedwithcorrosion,will
producenucleationsitesforcyrstallizationandsedimenetation.
Systematic studies on this form of fouling are rare since it is often
difficulttoseparatecorrosionfoulingfromotherfoulingprocesses.
a ArrangedbythecompilerafterTabored,Aoki,Ritter,Palen&Knudsen(1972a)[172]
b Inversesolubilitysaltsarethoseexhibitingdecreasingsaturationconcentrationbeyondacertaintemperature.
Asaconsequencedepositsstarttoformwhensuchsolutionscontactheatedsurfaces.Watercouldcontain
manysaltsoftheinversesolubilitytype,especiallyCaCO3andCaSO4.
Table1121:EffectsofDifferentPhysicalParametersonFoulinga
Parameter Effect
Flowvelocity Verystrongeffectonmostfoulingprocesses
Wallsurfacetemperature Affectsmanyfoulingprocessesparticularlycrystallizationand
chemicalreaction
Fluidbulktemperature Affectsreactionandcrystallizationrates
Material Importanteffectonfoulinginitiation.
Possiblecatalyticactionandcorrosion
Surfacefinish Importanteffectonfoulinginitiation.Roughness,sizeand
densityofcavitieswillaffectcrystallinenucleation,
sedimentationandadherencetendencyofdeposits.
Equipmentdesign,geometry Considerabledifferencesexistdependingonflowpatterns,
andflowpattern velocityandwalltemperaturedistribution,turbulencelevel,etc.
a ArrangedbythecompilerafterTaborek,Aoki,Ritter,Palen&Knudsen(1972a)[172].
275
ECSSEHB3101Part13A
5December2011
11.7 Existing systems

FREONTOWATERINTERCHANGER
1. Developer: Hamilton Standard, Division of United Aircraft Corporation (now United
TechnologiesCorporation),BradleyFieldRoad,WindsorLocks,Conn.06096.
2. Function:TransfersheatfromSpaceShuttleAtmosphereRevitalisationSubsystem(ARS)
waterlooptotheFreonCoolantLoop(FCL).Normallyheatistransferredfromoneofthe
tworedundantwaterloopstobothFreonloops.Inanabortsituation,onlyoneofthetwo
Freonloopsmaybeoperational.
3. Sketch:
4. Description: Platefin configuration, Counterflow arrangement except for the tent top
areaatbothendsofthecore.

Fins Part. Area CoreSize Dry

Sheets Density Mass
[kg]
Height Thickness Finsper Thickness, [m2.m3] Hot Clod No
hx103 ax103 meter ax103 Flow Flow Flow
[m] [m] [m1] [m] Length, Length, Length,
Lh[m] Lc[m] Lnf[m]
ARS 0,51 0,051 1260 0,127 6300 0,102 0,610 0,048 15,01
Water
FCL 0,51 0,051 1260 6300

Freon
21
Material:Fins,partingsheets,headers,bosses,etc.Stainlesssteel.

276
ECSSEHB3101Part13A
5December2011
5. Performances:

Heat Temperatures FluidMass Effectiveness Pressuredroppx

transfer Flowrate,m 103[Pa]
rate[W] Ti[K] To[K] [kg.s1]
ARS 14140 321,5a 280,5 0,0580,129 0,93b 20,7form=0,120kg.s

Water 1
FCL 277,5 318a 0,567 18,6form=0,567kg.s

Freon21 (2loops) 1
a DatafromHamiltonStandard,CEINo.MC25000010040.ItemNo.7.7.Postlandingconditions.
b Fromthepreviouslyquotedtemperatures.
6. References:ThedataarefromTrusch&Nason(1975)[180],unlessotherwisestated.

GSEHEATEXCHANGER
2. Function: Rejects heat from the Space Shuttle Freon Coolant Loop (FCL) to the Ground
SupportEquipment(GSE)coolantloops,duringprelaunchandgroundcheckoutmission
modes. Normally heat is transferred from one of the two redundant GSE loops to both
FCLloops.
3. Sketch:
277
ECSSEHB3101Part13A
5December2011
4. Description: Platefin configuration. Counterflow arrangement except for the tent top
areaatbothendsofthecore.TheFCLloopflowusesthewayofminimumpressuredrop
to minimise in flight Freon pumping power, whereas the GSE loop uses the wayside
ports,whichleadtohigherpressuredrop,eventhoughthemassflowrateisgreaterthan
thatfortheFCL.ThisisnobecausepumpingpoweroftheGSEisoflessconcern.

Fins Area
Part. CoreSize Dry
Density
Sheets Mass
[kg]
Height Thickness Finsper Thickness, [m2.m3] Hot Clod NoFlow
hx103 ax103 meter ax10
3 Flow Flow Length,
[m] [m] [m ]
1 [m] Length, Length, Lnf[m]
Lh[m] Lc[m]
FCL 0,51 0,051 1260 0,127 6300 0,063 0,198 0,086 6,58
Freon
21
GSE 0,51 0,051 1260 6300

Freon
21
Material:Fins,partingsheets,headers,basses,etc.Stainlesssteel.

5. Performances:

Heat Temperatures FluidMass Effectiveness Pressuredroppx

rate[W] Ti[K] To[K] [kg.s ]
1
FCL 31,194 355,5a 275 0,693 88b 12,1form=0,693kg.s1

Freon21 (2loops)
GSE 264 286a 1,260 103,4form=1,260kg.s1

Freon21
278
ECSSEHB3101Part13A
5December2011
a DatafromHamiltonStandard,CEINo.MC25000010025.ItemNo.7.5.Theheatexchangerdescribedinthis
documentfromHS,whichresultedfromaprecontractawardstudy,slightlydiffersfromthatdescribedhere.
Nevertheless,inputandoutputtemperaturesarepresumablythesameforbothheatexchangers.Thequoted
temperaturescorrespondtoprelaunchconditions.
b Fromthepreviouslyquotedtemperatures.

6. References:ThedataarefromTrusch&Nason(1975)[180],unlessotherwisestated.
AVIONICSHEATEXCHANGER
2. Function:CoolsrecirculatedairinthreeclosedcompartmentsoftheSpaceShuttlewhich
contain aircooled electronic equipment. It is a component of the Atmosphere
RevitalisationSubsystem.
3. Sketch:
4. Description: Platefin configuration. The core is composed of alternate layers of air,

primarycoolant,andsecondarycoolantfinsseparatedbypartingsheets.
Air is circulated by one of two redundant fans which are mounted directly to the split
inletheader.Aircoolingisprovidedbyoneoftworedundantwatercoolantloops,both
ofwhichpassthroughtheunitinafourpasscountercrossflowmanner.
279
ECSSEHB3101Part13A
5December2011

Fins Part.
Area CoreSize Dry
Sheets
Density Mass
[kg]
hx103 ax103 meter ax10
3 Flow Flow Length,
[m] [m] [m ]
1 [m] Length, Length, Lnf[m]
Lh[m] Lc[m]
ARS 5,08 0,051 945 0,102 2350 0,066 0,284 0,122 6,31
Air
ARS 1,27 0,051 1102 3900

Water
Material:Fins,partingsheets,headers,basses,etc.Stainlesssteel.

5. Performances:

Heat Temperatures FluidMass Effectiveness Pressuredroppx10

ARS 1541 331,5a 311 0,110 0,65b 0,164form=0,110kg.s1

Air
ARS 300 313,5a 0,040 9,997form=0,044kg.s1

Water
a Datadeducedfromthevaluesof,mandcpforAirandWater.SeeEq.[1116]inclause11.2.2WithR=0,664.
b Additionaldataonthepredictedperformanceofthisheatexchangerareplottedbelow.
Effectiveness, ,oftheAvionicsHeatEschangervs.watermassflowrate,mH2O,forseveralvaluesof
airmassflowrate.
280
ECSSEHB3101Part13A
5December2011
6. References:ThedataarefromTrusch&Nason(1975)[180],unlessotherwisestated.The
lastsketchisfromHamiltonStandard,CEINo:MC62100080002.ItemNo.2,30.

FREONTOWATERINTERLOOPHEATEXCHANGER
1. Developer: Microtecnica, Via Madama Cristina 147 Casella postale 210/Ferr., 10100
Torino(Italy).
2. Function:TransfersheatfromaFreonloop,coolingthepalletmountedequipment,tothe
Spacelabwaterloop.
A heat exchanger based on an identical core is also provided for additional equipment
cooling.ThisheatexchangeriscalledExperimentDedicatedHeatExchanger(EDHX).
3. Sketch:
4. Description:Platefinconfiguration.Counterflowarrangementoncefolded.Thestripfin
platefinsurfacedesigned1/819,82(D)isdescribedinTable115,clause11.3.1.

Fins Part. Area CoreSize Dry

Sheets Density Mass
[kg]
hx103 ax103 meter ax103 Flow Flow Length,
[m] [m] [m1] [m] Length, Length, Lnf[m]
Lh[m] Lc[m]
Payload 5,21 0,102 780 0,152 2231 10,2a

Freon21
Spacelab 5,21 0,102 780 2231

Water
a DatafromBellomo(1975)[6].
Material:Finsnickel.Partingsheets,corestrips,brakets,headingsandbossesStainless
Steel(Bellomo(1975)[6]).

281
ECSSEHB3101Part13A
5December2011
5. Performances:Datacorrespondingtotwotypicalcasesarepresented.

Heat Temperatures FluidMassFlow Effectiveness Pressuredropapx

transfer rate,m[kg.s1] 103[Pa]
rate[W] Ti[K] To[K]
Payload 4000 341 280 0,063 0,999 0,98form=0,063kg.s1

Freon21
Spacelab 280 295,5 0,063 1,88form=0,063kg.s1

Water
Heat Temperatures FluidMass Effectiveness Pressuredropapx

Payload 4000 296,5 286,5 0,378 0,938 8,31form=0,378kg.s1

Freon21
Spacelab 280 295,5 0,063 1,88form=0,063kg.s1

Water
a DatafromBellomo(1975)[6].
6. References: The data are from Owen, Sessions & Walker (1976) [139], unless otherwise
stated.

COLDPLATE
1. Developer: Microtecnica, Via Madama Cristina 147 Casella postale 210/Ferr., 10100
Torino(Italy).
2. Function:CoolingofavionicsequipmentonboardSpacelab.
3. Sketch:
282
ECSSEHB3101Part13A
5December2011
4. Description:Threedifferenttypeshavebeendeveloped.Theirdimensionsare:
(a) 0,762mx0,508mx5,08x103m.
(b) 0,381mx0,508mx5,08x103m.
(c) 0,305mx0,508mx5,08x103m.
Material:Stainlesssteel.
5. Performances:Whenwaterisusedascoolantliquid.
Overallthermalconductance,U=3500W.m2.K1.
Massflowrate,m=0,063kg.s1.
Pressuredrop,p=689Pa.
WhenFreon21isusedascoolantliquid.
Overallthermalconductance,U=2000W.m2.K1.
Massflowrate,m=0,38kg.s1.
Pressuredrop,p=8620Pa.
The photograph below shows the fins internal disposition and the inner part of the
bosses.
6. References: The data are from Microtecnica (1977) [126]. Some information concerning
coldplatedesigndatacanbefoundinBaum(1969)[5].
283
ECSSEHB3101Part13A
5December2011
12
Pumps
12.1 General
Apumpisamachinewhichimpartsenergytoafluid,inmostcasesaliquid.Compressors,fans,and
blowersimpartenergytogasesratherthantoliquids.
Thenetamountofenergy,En,transferredtotheunitliquidmassequalstheratioofthepressurerise
impartedtotheliquid,p,totheliquiddensity,.
Most manufacturers still use the head, H, to measure the pressure rise to density ratio, p/. The
dimension of head is length, the physical meaningof the concept being potential energy of the unit
liquidweightatanelevationH(gravitypotential).Sincetheintroductionofgravityishighlyartificial
wewillavoid,asfaraspossible,theuseofthetermheadthroughoutthisclause.
Thepressureriseenablestheliquidtoovercomehydraulicresistances(seeClause7).
A pump consumes more power than it gives off. The ratio of the specific energy (energy per unit
liquidmass)developedbythepump,En,totheenergyperunitliquidmasssuppliedtoit,Es,isthe
overallefficience, p.Inordertorelatethepresentnotationwiththatusedinclause7.1,itisindicated
thatEn=pEs=pP/m.
The (negative) difference EL = En Es is the energy lost by the pump per unit liquid mass flowing
throughit.Theenergylossaccountsforthreetypesoflosses:
1. Hydrauliclossesduetoliquidfrictionandturbulence,
2. volumelossesduetothefactthatsomeofthefluidleaksbackthroughinternalpassages
forcedbythepressureriseproducedbythepump,and
3. mechanicallossesbecauseoffrictioninbearings,packings,etc.
Spacebornepumpsseemtovaryenormouslyindesignandprincipleofaction.Nevertheless,theycan
allbedividedintotwomaintypes:rotodynamicpumpsanddisplacementpumps.
1. A rotodynamic pump imparts energy to a liquid through the work done by a rapidly
rotating vaned impeller. Examples of rotodynamic pumps are radialflow centrifugal
pumps, mixedflow pumps and axialflow pumps or propellers. Relevant rotodynamic
pumpsareclassifiedinTable121.
284
ECSSEHB3101Part13A
5December2011
Table121:RotodynamicPumps
Class SketchofaTypicalModel
Radialflow
Mixedflow
Axialfloworpropeller

NOTE FromHydraulicInstitute(1975)[95].
2. Displacementpumps,whichincludereciprocatingandrotarypumps,impartenergybya
positive displacement action. Reciprocating pumps are shown in Table 122, and rotary
pumpsinTable123.
Table122:DisplacementPumps.Reciprocating
Class SketchofaTypicalModel
Piston
Diaphragm

285
ECSSEHB3101Part13A
5December2011
Table123:DisplacementPumps.Rotary
Class SketchesofTypicalModels
Vane
Piston
Flexiblemember
Gear
Lobe
Circumferential
piston
Screw

Severalfeaturesoftypicalpumps,whichshouldbekeptinmindintheselectionofapumpforagiven
duty,arelistedinTable124.
Table124:MainFeaturesofTypicalPumps
RotodynamicPumps PositiveDisplacementPumps
Veryefficientwhenoperatingatspeedsfrom1200to Theyoperateatverylowrotatingspeed(220to
3600rpm,withintherangeofusualalternating 500rpm).
currentelectricmotors.
Cannotberunefficientlyatlowspeedstopump Efficiencies,althoughcanbehigh,arebelow
smallquantities. thoseforrotodynamicpumps.
286
ECSSEHB3101Part13A
5December2011
RotodynamicPumps PositiveDisplacementPumps
Overallefficiencyusuallyrangesfrom0,7to0,85.
Noteasilyregulated.Regulationbythrottlingis Efficiencydecreaseswhenwearincreases
simplebutwasteful.Regulationbyrunningspeed leakage.
adjustmentmaintainsmoreorlesstheefficiencybut
requiresauxiliaryequipmentwhichisexpensive, Thedischargecharacteristicisapulsatingone.
complicatedandunreliable.Theuseofmultiple Asmootherdischargeisobtainedindoubleor
windingmotorsandinvertorstocontrolmotor trebleactingunits.Rotarypumpsexhibit
speedthroughvoltageandfrequencycontrol greateruniformityofdeliverthanreciprocating
deservesconsideration. pumps.
Thedeliveryissubstantiallyconstant,
regardlessofthepressurerisedeveloped.
Canbeusedwithveryviscousliquids.Thisis
Cannotdeliverathighpressureunlessalargeand
particularlytrueinthecaseofrotarypumps.
heavytypeisused.Thepressureriskincreaseswith
Thepressurerisedropswithincreasing
increasingtheperipheralspeed.
viscosity.
Cannothandleviscousliquids.
Selfprimingandcapableofcopingwithhigh
suctionlifts.
Notselfpriming.Thiscanbeovercomeinvarious
Relieforbypassvalvesaretobeused.Unableto
ways,seetheradialflowrotodynamicpumpin
operateagainstacloseddischarge.Evenaslight
Table121.
decreaseindeliverymaycauseasubstantial
Noreliefvalvesaretobeused.Evencomplete pressurerise.
throttlingdoesnotpresentanydangertothepump
Abletohandlelargeproportionsofvapor.
orloopasnotfurtherpressurerisedevelops.
Enoughliquidshouldbepresenttoprovidea
Limitedbycavitationandpower. liquidsealingfilmfortheclearances.Suitable
forpumpinghotliquids.
Limitedbypressureandpower.
Problemsofleakageareminimizedparticularly
Leakthroughtheshaftseal.Submergedpumpscan
withdiaphragmpumps.
beusedtopreventthisdrawback.
Muchbiggerthanrotodynamicbecauseoflow
Smallerinsizethanothertypesforequalcapacity.
rotatingspeed.
Lowcost.Rugged,reliableinoperation.
Complicatedconstruction.Inletandoutlet
valvesshouldbeused(notinthecaseofrotary
pumps).Verysensitivetowearbecause
comparativelylargesurfaceareasmoveinclose
contact.Diaphragmpumpsdonotpresent
friction,butdiaphragmmaterialsareoflimited
useatelevatedtemperatures.
NOTE Arrangedbythecompilerafter:Nekrasov(1969)[132],London(1974)[118],Pollak&Cruger(1974)
[144],Scobie(1974)[160],Settlesetal.(1977)[163].
287
ECSSEHB3101Part13A
5December2011
12.2 Specified speed

Thecharacteristicsofapumparegivenbymeansofanexperimentallygeneratedcurvewhichrelates
thepowerimportedtotheliquid,pP,withthevolumeflowrate,m/,forconstantrotatingspeed.The
volume flow rate, which is usually known as pump capacity, is denoted by Q in most textbooks.
Nevertheless,wewilldepartfromthispracticesinceQrepresentstheheattransferrateinthisPart.
TypicaldimensionalcharacteristiccurvesofacentrifugalpumpareshowninFigure121.
Figure121:Typicalcharacteristiccurvesofacentrifugalpumpforagivenrotating
speed.
ItcanbededucedfromFigure121thatanoptimumoperatingpoint(pointofmaximumefficiency)
exists at a certain rotating speedflow rate combination. The numerical values which several
dimensionlessvariablesreachatoptimumconditionsprovideaquantitativebasisfortheclassification
of all pumps. It has been found, for example, that machines exhibiting similar design features have
approximatelythesamespecificspeed,ns.Thisdimensionalparameterisdefinedas:
n(m / )1/ 2
ns [121]
H 3/ 4 opt
wherenmeansrevolutionsperminute(orsixtytimesacharacteristicfrequencyoftheimpeller),and
Histheheadinm.Subscriptoptindicatesthatn,m/andHarethevaluesattheoptimumpoint.
Formulae relating efficiency to specific speed are available for rotodynamic pumps. For example,
Nekrasov(1969)[132]takesintoaccounttheleakage,thefrictionbetweenimpellershroudandfluid,
andthemechanicalfrictioninpackingandbearings.Nevertheless,heisunabletoofferananalytical
expression for the total hydraulic losses inside the pump, since these losses depend ona number of
factors whose effects are poorly known. On the other hand, expressions similar to those just
mentioned have been critiqued on the grounds that, although based on best standard practice, they
needrevisioninthelightofmoderntechnologicaldevelopments(Settlesetal.(1977)[163]).
Theuseofspecificspeed,ns,definedasinEq.[121],presentstwomaininconveniences.Thefirstone
derives from the fact that the head, H, appears in the definition of ns. In addition, ns is not a
dimensionlessnumberandthence,comparisonofmachinesoriginatingindifferentcountries,oreven
of products from different manufacturers, becomes difficult because of the wide variety of practical
288
ECSSEHB3101Part13A
5December2011
unitsinuse.Thisiswhymodernauthors(seeforexampleCsanady(1964)[29])useadimensionless
number,,proportionaltonsanddefinedas:
(m / )1/ 2

p / 3 / 4
[122]
opt
whereismeasuredinradianspersecond,m/inm3.s1,andp/inm2.s2.Noticethatdependson
theworkingliquidthroughitsdensity, .Table125couldbehelpfulinreducingdatafromvarious
sourcestoacommonnondimensionalbasis.
Table125:ConversionFactorsintheDeductionoffromns
Whennsisgivenby Divideby
rpm(cfs)1/2(ft)3/4 129
rpm(gpm)1/2(ft)3/4a 2730
rpm(hp)1/2(ft)5/4b 42
rpm(metrichp)1/2(m)5/4b 187
rpm(m3.s1)1/2(m)3/4 53
a gpm=USgallonsperminute.
b Workingfluidiswater.
NOTE FromCsanady(1964)[29].
Rotodynamicpumpsmaybeclassifiedaccordingtotheirvaluesasfollows(Nekrasov(1969)[132]):
1. Lowspeedradialflow:0,40;D2/D1=2,23,5,
2. Normalspeedradialflow:=0,400,80;D2/D1=2,21,8,
3. Highspeedradialflow:=0,801,60;D2/D1=1,81,3,
4. Mixedflow:=1,603,20;D2/D1=1,31,1,
5. Axialfloworpropeller:3,206,40;D2/D1=1.
The impeller shapes corresponding to the five types listed above are shown in Figure 122. The
changesinshapearenecessarytoaccommodatetherelativelyhigherflowratesatlarge,andtocope
withtherelativelyhighpressurerisesatsmall .Alowerboundof foranefficientrotodynamic
pumpcouldbefixednear=0,18(Csanady(1964)[29]).
289
ECSSEHB3101Part13A
5December2011
Figure122:Rotodynamicpumpimpellers.FromNekrasov(1969)[132].
It should be said that different authors quote slightly different limiting values of for the several
classesofpumps.AcorrelationofrotordesignandspecificspeedsimilartothatshowninFigure122
isgiveninCsanady(1964)[29].
In order to select the machine(s) suitable for a given duty (given rotating speed, capacity, pressure
rise,andworkingfluid), iscalculated.Then,therotodynamicmachinebestsuitedtothatdutycan
beselectedfromFigure122.
12.3 Net suction energy

Inmanycasestheworkingliquidcontainsvaryingquantitiesofvapour,andcavitationmayappear.
This results in an intermittent stream, a characteristic noise, and reductions in the pump capacity,
pressureriseandefficiency.
Cavitationappearswhenthedifferencebetweentheabsoluteliquidpressureandthevapourpressure
of the liquid at the operating temperature becomes negative, but event at (small) positive values of
this difference cavitation can be triggered by the presence of dissolved gases which are released at
reducedpressureorincreasedtemperature.
Inordertoavoidcavitationtheabsolutepressureinthesuctionloopmustexceedacertainthreshold
value. To define this value the net positive suction head (NPSH) is traditionally used. NPSH is the
absolutetotalheadofthepumpinletabovethefluidvapourpressure(measuredinconsistentunits)
attheworkingtemperature.
Theconceptofheaddisappearswhenthenetsuctionenergyperunitmass,Ev,isintroduced:
pi psat Vi 2
Ev [123]
2
where pi and Vi are pressure and velocity at the inlet section of the pump, psat is the liquid vapour
pressure,andthe(constant)liquiddensity.
Leavingasideviscosityeffects,theperformancelawofafamilyofgeometricallysimilarpumpswith
incipientcavitationwilldependontwoindependentdimensionlessparameters.Thefirstonecouldbe
the,socalled,flowcoefficient,(m/)/D3,whereDistherotordiameter.Theotherparameterwillbe
basedonthenetsuctionenergy,Ev.Althoughseveralcombinationsarepossible,themostconvenient
choiceisthesuctionspecificspeedSofWislicenus,whichisdefinedas:
m / 1/ 2
S [124]
Ev3 / 4
290
ECSSEHB3101Part13A
5December2011
Theperformancelawthenbecomes,
m/
p f ;S [125]
D
3
Mostpresentdaypumpsaredesignedfortheavoidanceofcavitation.Withthispurpose,amaximum
value S = 2,88 is widely used for conventionallydesigned machines over the whole specific speed
range(Scobie(1974)[160]).Thiswouldnarrowthechoiceofthepump,since
3/ 4
Ev
2,88 [126]
p / opt
Whenlimitation(6)isuntenableothercurestocavitationcanbefound.Someofthemare:
1. Placingthepumpintothesystematapositionwhereitdealswithfluidsatmoderately
lowtemperatures.
2. Increasingtheworkingpressure.
3. Installingaboosterpumpattheentranceofthesuctionline,and
4. Mounting an axial wheel immediately before the impeller to increase pressure and
impartwhirltothefluid.Thisresultsinareductionoftherelative(tothevane)velocity
and improves operating conditions of the impeller. Notice that, although the auxiliary
runnercanpreventcavitationfromdevelopingattheimpeller,cavitationcandevelopon
theauxiliaryrunneritselfunlessitsrotationalspeedislessthanthatoftheimpeller.
12.4 Requirements for spaceborne pumps

Thefollowingrequirementsarenormallyimposedtopumpsforspacebornefluidloops:
1. Lowmass.
2. Highoverallefficiency.
3. Lowmasstooutputpowerratio.
4. Hermeticallysealed.
5. Quietinoperationandexhibitinghighoperationalreliability.
6. Abletoacceptmissionvibrationandshockloads.
7. Compatiblewithonboardelectricalservices.
8. Madefrommaterialssuitableforuseintheaerospaceenvironment.
9. Abletohandletypicalliquidcoolants(seeTable101).
More detailed environmental, safety, reliability and compactness requirements are defined for each
particular spacecraft. In addition, self priming under reduced gravity may be required in several
instances.
291
ECSSEHB3101Part13A
5December2011
Concerning the required performances, two general categories of spaceborne liquid pumps will be
consideredhere:
1. ClassA.Pumpsusedinlargemannedspacecraft,suchastheSpaceShuttle.Typicalrates
andpressurerisesforthesepumpsare(Owenetal.(1976)[139]):
Capacity,m/,from0,5x104m3.s1to2x104m3.s1.
Pressurerise,p,from0,5x105Pato105Pa.
Fromthesevaluesitisdeducedthattheoutputpower, pP= pm/,isintherangefrom
2,5Wto20W.
2. Class B. Small pumps for unmanned satellites (Ebersole (1970) [41]), astronaut cooling
(Carson(1972))orsimilarduties.
Capacity,m/,from105m3,s1to0,5x104m3,s1.
Pressurerise,p,from105Pato2x105Pa.
Thepowerimpartedtotheliquidisinthiscaseoftheorderof1W.
12.5 Commercially available pumps

The applicability of commercially available pumps to spaceborne fluid loops has been discussed
severaltimes(Carson(1972)[20],WynRoberts(1973)[194],Berner&Schleicher(1976)[13].
Carson (1972) [20] describes the development of a diaphragm pump whose aim is providing body
coolingforanastronautperformingextravehicularwork.Thecharacteristicsofthispumparegivenin
Table 126. Early attempts to use a commercially available centrifugal pump with the purpose of
astronaut cooling proved to be unsuccessful mainly because of too low an overall efficiency for the
requiredpower.
292
ECSSEHB3101Part13A
5December2011
Table126:CharacteristicsofSeveralCommerciallyAvailablePumps
Type Pump Size Mass Capacity Pressure ElectricalInput Comments
Ht. Wd. Lg. [kg] (m/)x104 rise W V Ph/Hz
[m] [m] [m] [m3.s1] px105
[Pa]
Rotodynamic. Cadeta 0,1230,128 0,216 4,0 Figure123 ~ 220250 1/50 Domesticcentralheatingpumps.
RadialFlow Mini 40 Workingpressure:6x105Pa.Maximum
operatingtemperature:383K.
Cadeta 0,1240,128 0,216 4,3 ~
S 70
Euramob 0,1800,095 0,175 3,5 Figure124 25 220 1/50 Domesticcentralheatingpumps.
MX32E Workingpressure:6x105Pa.Temperature
range:280Kto383K
Euramob 0,1800,130 0,173 5,1 15
XA15R

c 0,0760,050 Figure125 Submersiblepumptransferringliquid
nitrogenandforuseinclosedcyclecooling
systems.
Temperaturerange:77Kto300K
Easternd 0,1020,082 0,216 2,5 Figure126 37 115 1/5060 Magnetdrivingimpellerisfullyencapsulated
Iwaki inpolypropylene,supportedbyaTFEbearing
MD15T androtatingonastationaryceramicspindle.
Material:Glassfilledpolypropylene.
Easternd 0,1270,089 0,267 3,4
Iwaki
MDR30T
AC3CMDd0,1080,102 0,206 2,9 Figure127 175 115 1/5060 Lowcostpumpforchemicals.Magnetic
drivenimpeller.
Plasticcomponentsareglassfilled
polypropylene.Pumphermeticallysealed.
293
ECSSEHB3101Part13A
5December2011
Type Pump Size Mass Capacity Pressure ElectricalInput Comments

Ht. Wd. Lg. [kg] (m/)x104 rise W V Ph/Hz
[m] [m] [m] [m .s ] px105
3 1
[Pa]
Notfullyenclosedmotor.
Displacement e 0,0860,056 0,101 0,61 0,30f 0,36f 9,7 16,80,8 DC DevelopedbyNASAforastronautcooling.
Reciprocating Diaphragm:Dacronfabricwithbutylrubber
Diaphragm coating.
Inlettemperature:289,15K.
Displacement 41371285d 0,63 Figure128 37 12 DC Vanes:DelrinorTeflon.
Rotary. RotorsandShaft:304SS.
Flexible Bodies:PhenolicorEpoxy.
Vane Orings:BunaN,VitonorEthylenePropylene.
Forhardtohandlechemicals.
Temperaturerange:230Kto390K.
a SealedMotorConstructionCo.Ltd.BristolRoad,Bridgwater.Somerset,England.Telephone4366.
b LeMatrielTlphonique,DivisionElectroHydraulique,46,47QuaiAlphonseLeGalloBP402,92102BoulogneBillancourt,CEDEX,France.
c FromEngel&Walter(1974)[43]
d GelberPumpCo.5806NorthLincolnAvenue,Chicago,Illinois50659,USA.
e FromCarson(1972)[20].
f Optimumoperatingpoint.
294
ECSSEHB3101Part13A
5December2011
Figure123:CharacteristiccurvesofSEALEDMOTORCONSTRUCTION
CentrifugalPumpsCadetMiniandCadetSpumpingwater.FromWyn
Roberts(1973)[194].
Figure124:CharacteristiccurvesofEURAMOCentrifugalPumpsCadetMX32E
andXA15Rpumpingwater.FromEURAMOPOMPESSALMSON(1977)[60].
295
ECSSEHB3101Part13A
5December2011
Figure125:CharacteristiccurveoftheCentrifugalPumpdevisedbyEngel&
WalterCadetMiniandCadetSpumpingwater.FromWynRoberts(1973)
[194].
Figure126:CharacteristiccurvesofEasternIwakiCentrifugalPumpsMD15T
andMDR30Tpumpingwater.FromGELBER(1976)[71].
296
ECSSEHB3101Part13A
5December2011
Figure127:CharacteristiccurveofCentrifugalPumpAC3CMDpumpingwater.
FromGELBER(1976)[71].
Figure128:CharacteristiccurveofPositiveDisplacementRotaryPump41371285
pumpingwater.FromGELBER(1976)[71].
WynRoberts (1973) [194] surveyed the products from 20 European manufacturers. The imposed
requirements were 4), 5) and 8) in Clause 12.4. The selected pumps are too large for the typical
performances mentioned above, with the exception of two pumps manufactured by Sealed Motor
ConstructionCo.,whosecharacteristicsaresummarisedinTable126below.
FromthesurveymadebyBerner&Schleicher(1976)[13]twopumpsbyPompesSalmsonaresingled
out.ThecharacteristicsofthesepumpsarealsogiveninTable126.
Specialpurposepumpshavebeendescribedinseveralinstances.Acentrifugalpumpfortransferring
liquid nitrogen has been developed by Engel & Walter (1974) [43]. Data on this pump are also
includedinTable126.
297
ECSSEHB3101Part13A
5December2011
12.6 European pump manufacturers

According to WynRoberts (1973) [194] the following European firms offer pumps which look
promisingforspacecraftapplication.ThetwomanufacturersmentionedinconnectionwithTable126
havebeennotincludedinthislist.
Bertin,
AlleGabrielVoisin,
B.P.3,
78Plaisir,
France.
HermeticPumpenGmbH,
7803GundelflingenbFreiburg,
Gewerbestrasse20,
Germany.
NovameticPumpenGmbH,
Mhlenkamp59,
Hamburg392000,
Germany.
WatsonMarlowLtd.,
FalmouthTR114RU,
Cornwall,
England.
298
ECSSEHB3101Part13A
5December2011
13
System optimization
13.1 General
The engineer engaged in the design of a fluid loop to meet some specified requirements, under a
certainsetofconstraints,isfacedwithanoptimizationproblem.Heshoulddecidewhich,amongaset
ofdifferentalternativesshouldbepreferredonthebasisofsomeoptimizationcriteria.Thepurposeof
thisclauseistodiscusstheseoptimizationprocedures,andtointroducethetypesoftradeoffanalyses
which could be used by the designer. It should be pointed out, however, that it is difficult to give
general rules, and that in practice each case should be analyzed by itself, considering the
requirements,constraintsandoptimizationcriteriaofthatparticularcase.Often,thedesignerdoesnot
address the question of which of all possible solutions is the best according to the optimization
criterionchosen,butinsteadselectssomeoftheparametersandcharacteristicsoftheloop basedon
experience,availabilityandotherengineeringcriteria,andperformsanoptimizationanalysisofone
ortwoparametersleftunspecified.
Duetotheabovementionedcharacteristicsoftheproblem,noattemptwillbemadetowardaformal
description of system optimization, rather, a fairly general approach devised by Barker, Stephens &
Taylor(1967)[3]willbeusedtoindicatehowabasicanalysiscanbeperformed.Inaddition,twomore
particular,yetillustrative,exampleswillbeshown.
Oneimportantparameterwhichisusedinmostofthesestudiesisthemassneededtomeetacertain
pumpingpowerrequirement.Thisequivalentmassisobtainedthroughtheuseofapumpingpower
penaltyfactorwhichcanbetakenbetween0,5k.W1(Barkeretal.(1967)[3]and0,2kg.W1(Trusch
(1974) [179]. It should be mentioned that, as can be deduced from Table 126, several conventional
pumpsshowmasstopowerratiosnofarabove0,2kg.W1
13.2 Basic analysis

Thedesignofaliquidcoolantsystemtotransferheatfromtheenvironmentalcontrolandlifesupport
systemofasatellitetoaspaceradiatorinthesatellitehasbeendiscussedbyBarker,Stephens&Taylor
(1967)[3].Theprocedureusedistosuboptimizeindividualcomponentsofthesystemonthebasisof
power requirement and mass, at various coolant flow rates. The mass due to pumping power is
determinedthroughapowerpenaltyfactorwhichistakenas0,5kg.W1.Foreachflowrateused,the
minimumtotalmasspenalties(sumofmassoftheindividualcomponentsandequivalentmassdueto
pumpingpower)arecomputedandaddedtogethertoyieldtheminimumtotalsystemmasspenalties.
Fromtheseresultsthelowestminimumtotalsystemmassiscalculated,thusobtainingtheoptimum
coolantflowrate.
The thermal control system consists of an interface heat exchanger, a bypass valve, a controller, a
spaceradiatorandsupplyandreturnplumbing,Figure131.Heatissuppliedtothesystemthrough
theheatexchangerfromafluidloopwithwaterasworkingfluid.Thecoolantinthethermalcontrol
299
ECSSEHB3101Part13A
5December2011
system is Freon 21, which receives the heat through the exchanger and transports it to the space
radiator.Theaimofthebypassvalveandvalvecontroller,whicharelocatedupstreamoftheradiator,
istomaintainthetemperaturedownstreamofthisradiatorat274,5Kinordertoprecludefreezingof
watervapourontheexteriorsurfaceofplumbing.
Figure131:SchematicrepresentationofthefluidloopconsideredbyBarker,
Stephens&Taylor(1967)[3].
The main characteristics of the system are: heat load, Q = 15 x 103 W; water inlet temperature, Thi =
316,5K;wateroutlettemperature,Tho=277,5KandFreon21inlettemperature,Tci=274,5K.Theheat
exchangerisofcounterflow,platefinsurfaceconfiguration,andismadefromstainlesssteel.Thefin
heightish=2,53x103m,finthickness,a=0,1x103mwith787finspermeter,allthesedataonboth
sides.Partingsheetthicknessisa=0,25x103m.
Thesuboptimizationprocedureisoutlinedinthefollowingclause.
13.2.1 Interface heat exchanger

The procedure consists in finding the best heat exchanger envelope at a given coolant flow rate, to
minimizethetotalmasspenalty.Itisobservedthatforeachcoolantflowrate,thewaterflowrateand
outletFreon21temperaturecanbecomputedsincetheheatloadisknown.Essentially,thefollowing
stepsareperformediteratively.
1. total penalty is initially estimated and later specified on step 14 during subsequent
iterations.
2. Heat exchanger effectiveness, , is obtained from the imposed design conditions; fluid
specificheats,flowratesandtemperatures(seeEq.[118]inclause11.2.2).
3. The number of heat transfer units, Ntu, is computed from the effectiveness, , and the
capacityrateratio,R,fortheassumedflowarrangement(seeclause11.2.3).
4. Pressuredropsareinitiallyestimatedandlaterspecifiedfromstep14duringsubsequent
iterations.
5. A U iscomputedfromtheNtuvalueandC1(seeEq.[117]inclause11.2.2).
300
ECSSEHB3101Part13A
5December2011
6. Reynoldsnumbersareinitiallyestimatedandlatercorrectedfromtheresultsofstep11
duringsubsequentiterations.
7. Theheattransfercoefficient,h,andFanningfrictionfactor,f,arecomputedfromthedata
forthespecifiedexchangercore(seeClause11.3andKays&London(1964)[102]orthe
appropriatedatafortheparticularcore).
8. Theoverallsurfaceeffectiveness,,andoverallheattransfercoefficient,U,arecomputed
(seeclause11.2.2).
9. The heat transfer area, A,is computedfrom the overall heat transfer coefficient, U, and
theproductA U computedinstep5.
10. Frontal area of the exchanger, AFR, hydraulic diameters, DE, and flow lengths, L, are
deducedfromcoregeometriesandheattransferarea(seeClause11.3ortheappropriate
datafortheparticularcore).
11. Pressure losses are computed from Reynolds numbers, lengths, flow rates and friction
factors(seesteps6,7and10).
12. Pressures losses from steps 11 are compared with previous pressure losses from step 4
and the procedure continues at step 6 or 13. The process is finished when the error in
pressurelossremainswithinspecifiedbounds.
13. Theheatexchangervolume,frontalareas,andenvelopedimensionsarecomputed.
14. The total mass penalty is computed and compared with its previous value. This total
mass penalty includes core, headers and shell, fluid and the equivalent mass due to
pumping power (headers and shell weight account for 30 % of core weight). The
procedureiseitherterminatedorcontinuedatstep4dependingonthecomparison.New
pressurelossesareestimatedifitappearsthatthemasspenaltycanbefurtherreduced.
AsaresultofthisproceduretheoptimumshapeoftheheatexchangerisobtainedfordifferentFreon
21 flow rates. As the flow rate increases, the required power increases but the heat exchanger mass
decreases.InthecaseconsideredbyBarkeretal.(1967)[3]thesumofheatexchangermassandmass
duetopumpingreacheditsminimumvalueataFreon21flowrateof0,63kg.s1.
The envelopes for these optimum heat exchangers are long and slender, thence the actual design of
such units would involve folding the heat exchanger into a more convenient package. This folding
wouldundoubtedlyresultinsomedegradationoftheheattransferandpressuredropcharacteristics.
Thesedegradationswouldideallybetradedoffagainstthebenefitsofthemoreconvenientpackage.
13.2.2 Supply and return plumbing

These circuits are optimized by determining the tube diameter and wall thickness combination
yielding a minimum value for the sum of the mass of fluid, plumbing and equivalent mass due to
pumpingpower,foragivencoolantflowrate.Plumbingmassincludesonesetoftubes,fittingsanda
redundantsetofmanifolds.Tocalculatethepumpingpower,thepressurelossinvalves,manifolds,
fittings, ... is accounted for by doubling the pressure loss based on straight tube length and bends.
Equations are used to express these masses as functions of tube diameter and wall thickness, and
these parameters are changed over a range of standard values until the optimum combination is
found.Theresultofthisprocedureshowsthatthe minimumtotalpenaltyincreaseswithincreasing
massflowrate.
301
ECSSEHB3101Part13A
5December2011
13.2.3 Radiator
Theradiatorisoptimizedthroughaproceduresimilartotheoneusedforplumbing.Inthiscase,for
eachflowratethereare3parameterstobedefined,numberoftubes,tubediameterandwallthickens,
andtheyaredeterminedbyminimizingthesumoffluidmass,tubingmassandequivalentmassdue
topumpingpower.Thewallthicknesswastakenas2,5x103m.Ontheotherhand,itwasfoundfrom
heattransferconsiderationsthatturbulentflowshouldbepreferred,sothataReynoldsnumberof104
wasused.Thisfactprovidesarelationshipbetweentubediameterandnumberoftubesforeachflow
rate,andthereforethetubediameteristheonlyvariablewhichhastobeoptimized.
TheresultsofthesesuboptimizationproceduresareshowninFigure132,whereitmaybeobserved
that the total system penalty is minimized in the vicinity of 0,38 kg.s1 Freon 21 flow rate. This
minimumislargelyduetotherapidincreaseinheatexchangerpenaltyatreducedflowrate,whereas
boththeradiator,andsupplyandreturnplumbingpenaltiesdecreasewithdecreasingmassflowrate.
Theincreaseinheatexchangerpenaltyisduetothegreaterheatexchangertransferareasrequiredat
theseflowrates.
Figure132:Resultsoftheliquidcoolantsystemoptimizationanalysis.a)givesthe
masspenalty,M+,ofthesystemcomponentsvs.Freon21massflowrate,mc.The
masspenaltyincludestheequivalentmassduetopumpingpower.b)givesthe
systemmass,M,vs.Freon21massflowrate,mc.FromBaker,Stephens&Taylor
(1967)[3].
Thereforethroughtheprocedurejustdescribednotonlytheoptimummassflowrateisderived,but
also the optimum heat exchanger shape, tube diameter and wall thickness of supply and return
plumbing,andthenumberoftubesandtubediameteroftheradiator.
13.3 Special examples

The example in Clause 13.2 was intended to illustrate how an optimization study should be
undertaken in a very general case. No constraints other than heat load and operating temperatures
were introduced a priori in the process of calculating the coolant flow rate giving minimum system
mass,andintheprocessofdevisingtheoptimumfluidloopsystem.
302
ECSSEHB3101Part13A
5December2011
Thereareproblems,however,wheretheoptimizationprocedureisnotsoclearcut,andwheresome
tradeoffbetweenalternativeoptimizationcriteriashouldbemade.Furthermore,inseveralinstances
a subsystem is integrated into an already existing system, and the possibilities of optimizing the
subsystemareseverelycurtailedbytheconstraintsimposedbytheintegration.
Thefollowingtwoexampleswillillustratehowthedesignengineerfacestheseproblems.
13.3.1 Constraints based on source temperature

There are situations where the source temperature, TS, appears in the optimization problem, either
becauseitisspecifiedorbecauseitistobeoptimized,foragivenheatload,Q.Thisintroducesanew
featureintheproblem.
Berner & Schleicher (1976) [13] have analyzed a fluid loop with a given liquid to liquid heat
exchanger; the so called Experiment Dedicated Heat Exchanger (EDHX), which has been developed
for use in Spacelab. The mentioned fluid loop is sketched in Figure 133. Details concerning a heat
exchangersimilartotheEDHXcanbefoundinclause11.7.Theheatsource,whichisthesetupused
forperformingscientificexperimentsonboardSpacelab,ismountedonacoldplate(seeclause11.7),
thencethesourcetemperatureisassumedtobethewalltemperatureoftheflowpassagesthroughthe
coldplate.
Figure133:SchematicrepresentationofthefluidloopwithEDHX.FromBerner&
Schleicher(1976)[13].
In order to control the source temperature a percentage of the coolant flow is bypassed around the
source.
Anattemptismadetooptimizethehotfluidmassflowratesothatthemassofthesupplyandreturn
plumbing plus theequivalent mass due to pumping poweris minimized under the following set of
assumptionsandspecifications.
1. Concerningtheworkingfluids.
o Hotandcoldfluidsarewater
o Massflowrateofcoldfluid,0,0631kg.s1
o Percentageofhotfluidmassflowratebypassedaroundsourcefor
temperaturecontrolpurposes,50%
303
ECSSEHB3101Part13A
5December2011
2. Concerningoperatingtemperatures.
o InlettemperatureofcoldfluidatEDHX,Tci=293K
o OutlettemperatureofcoldfluidatEDHX,Tco313K
3. concerningsupplyandreturnplumbing.
o Lengthofforwardandreturnlegsoftheloop,2,5meach
o Numberof90bendsineachleg,5
o Ratioofbendradiustotuberadius,R/r=6
o Materialofplumbing,stainlesssteel
o Thepressuredropinvalves,manifolds,etc.,isaccountedforbydoubling
thepressuredropbasedontubelengthandbends
4. Concerningpumping.
o Pumpefficiency,p=0,1
o Powerpenaltyfactor,0,5kg.W1
Foreachmassflowrate,mh,thesumofthemassesoftubesandfluid,plustheequivalentmassdueto
pumping power is calculated for different values of the inner diameter of the tubes. The resulting
values,whicharepracticallyindependentoftheheatload,Q,indicatethat,foragivendiameter,the
total mass decreases as the flow rate decreases. however, for a given heat load, Q, the source
temperature,TS,increasesasthemassflowrate,mh,decreases.Inthosesituationsinwhichthesource
temperatureisfixedthereisaparticulartubediameteroptimizingthesystemmass,otherwiseatrade
offbetweenlowsourcetemperatureandlowsystemmassisrequired.
Figure134showsanexampleofthistypeoftradeoffanalysis.ThedatashowninthisfigureareforQ
=103W.OthervaluesofQ,intherange0,5x103Wto4x103W,havebeenconsideredbyBerner&
Schleicher(1976)[13]resultinginfairlylargechangesofthevaluesofTSandonlysmallchangesinthe
mass penalty, M+. It may be noticed from Figure 134 that at a certain low constant value of M+ a
minimumvalueofTSisobtainedforanintermediatetubediameter.Likewise,forsourcetemperatures
abovesomethresholdvalueaminimumvalueofM+couldbeobtainedbyvaryingthediameter.When
both source temperature and system mass should be optimized there is no criterion on which this
optimizationcouldbebasedandacompromisebetweenlowTSandlowM+isnecessary.Sometimes
thepumpingpowerrequiredbytheoptimumsystemissolowthatitisdoubtfulthata0,1pumping
efficiency may be reached. To illustrate this, the interrupted line in Figure 134 has been drawn
throughthepointsoneachofthecandidatesystemsforwhichthepumpingpowerrequirement, pP,
equals0,5W.ItcanbededucedfromFigure134that,forconstantpumpingpower,themasspenalty
mustincreaseifthesourcetemperatureistobedecreased,buttherearevaluesoftheinnerdiameter
ofthetubeabovewhichrelativelymodestdecreasesinsourcetemperatureimplylargemasspenalties.
Thus,atleastinthecaseconsideredinFigure134,aninnerdiameterof102mcouldrepresentagood
compromise.
304
ECSSEHB3101Part13A
5December2011
Figure134:Masspenalty,M+,ofthesupplyandreturnplumbingsystemvs.
sourcetemperature,TS.M+includesmassoftubesandfluidplustheequivalent
massduetopumping.Heatload,Q=103W.FromBerner&Schleicher(1976)[13].
Figure 135 shows the relationship between the source temperature, TS, and the coolant mass flow
rate,mh,forseveralvaluesoftheheatload,Q.Thepowerrequirementsareintherange0,5to1W.
Thus,evenifarelativelylowpumpingefficiencyisassumed,thepowersuppliedtothepumpwould
beafewwattsonly.
305
ECSSEHB3101Part13A
5December2011
Figure135:Sourcetemperature,TS,vs.coolantmassflowrate,mh,fordifferent
valuesoftheheattransferrate,Q.Theinterruptedlinesaredrawnthroughthe
pointsforwhichthepowerrequirements,foragivenDi,equals1W.FromBerner
&Schleicher(1976)[13].
An important conclusion drawn from Figure 135 is that it is not possible to reduce the difference
betweensourceandsinktemperatures,foragivenheatloadbyincreasingthecoolantmassflowrate
beyond a certain value, and that this difference can be relatively high. Figure 135 shows that, for
instance,whenQ=103Wthesourcetemperaturecannotbedecreasedbelow300K.Therefore,iflower
source temperatures are specified, some refrigeration equipment, not necessarily based on a
convectiveheattransfermechanism,shouldbeincorporated.Berner&Schleicher(1976)[13]present
anoptimizationanalysisofasysteminwhichaheatpumphasbeenincorporatedtofurtherdecrease
the source temperature. The aim of this heat pump is to extract thermal energy from the fluid
returningtowardtheheatsourceandtogiveitofftothefluidflowingfromsourcetosink..Although
a system incorporating a heat pump could operate even in the case of negative source to sink
temperature differences, the power requirement of the heat pump is much larger than that of the
circulation pump. Thus, the mass of a fluid loop system with heat pump strongly depends on the
valueofthepumpingpowerpenaltyfactor.
13.3.2 Constraints imposed by the integration

Sometimes the integration of the system onboard the satellite imposes constraints which must be
taken into account in the optimisation process. In several cases the selected optimum configuration
doesnotmeetalltheimposedrequirementsandsomesortoftradeoffshouldbemade.
AcoolingsystemwithliquidFlutec9950toairheatexchanger(ECLA)hasbeenstudiedbyBerner,
Schleicher, Oesch, Nyfeler & Steinhauser (1977) [14] for use in Spacelab. The basic ECLA system is
sketched in Figure 136. The heat source, which is the setup used for performing certain scientific
experiments is mounted on a cold plate (see clause 11.7), and its temperature is controlled by
removingheatwhichisthentransferredtotheAvionicsAirLoopoftheSpacelab.Asetofliquidtoair
306
ECSSEHB3101Part13A
5December2011
crossflowheatexchangers,combinedinseries,constitutesthethermallinkbetweencoolantloopand
airloop.
Figure136:SchematicrepresentationofbasicECLAsystem.Theaimofthe
accumulator,whichisnotmentionedinthetext,istoaccommodatechangesin
coolantdensitywithtemperature.Thecoolantthrottleisusedforadjustingthe
coolantmassflowrate,mh.FromBerner&Schleicher(1976)[13].
In order to control the source temperature, TS, part of the coolant flow is bypassed around the heat
source. This is achieved by means of a flow proportioning valve monitored through the difference
betweenmeasuredanddesiredsourcetemperature.Inthiswaythesourcetemperature,TS,islimited
from above and from below for a given heat load, Q. The upper limit of TS is set by the danger of
instability in the control loop. This instability appears because the derivative of TS with respect to
coolantmassflowratethroughtheheatsourcebecomesverylargeatlowflowrates.Thelowerlimit
for TS is reached with no bypass. Sometimes the desired TS cannot be reached within the operating
limitsofthevalve,thentheairmassflowratecanbechangedbyathrottleintheairductoftheECLA
system.Toextendtheoperatingrangetolowervaluesofthesourcetosinktemperaturedifferencesa
heat pump is incorporated in the fluid loop. The performance of the ECLA system with heat pump
hasbeenconsideredbyBerneretal.(1977)[14],butitwillnotbediscussedhere.
TherequirementsspecifiedforECLAare:
1. Heatloads,Q,from100Wto1,5x103W.
2. Sourcetemperature,TS,adjustablebetween283Kand353Kwithamarginof5K.
IntegrationofECLAintheSpacelabAvionicsAirLoopintroducesotherrequirements:
1. Themaximumallowedairmassflowrateismc=0,0908kg.s1.
2. Inletairtemperature,Tci,rangesfrom293Kto297K.
307
ECSSEHB3101Part13A
5December2011
3. Pressure drop of the air between the position where it is withdrawn from the forward
lineoftheAvionicsAirLoopandgivenbacktothereturnlineshouldnotexceed250Pa
(1inchofwater),otherwiseafanmustbeincorporatedintheairductthroughECLA.
4. Airvelocitiesabove6m.s1areundesirablefromthestandpointofnoisegeneration.
Otherengineeringcriteriawhichshouldbetakenintoaccountarethefollowing:theheatexchangeris
ofthecrossflowtypesimilartothoseusedintheradiatorofmotorcars.Inordertoreachacceptable
performances a multipass arrangement is used. This can be best implemented with a series
arrangementofseveralmodules.Eachmodulehasafixedcoldflowlength.Thesizeofthesystemis
suchthatitcanbeaccommodatedintheSpacelabmoduleracks.
The aim of the optimization is the achievement of a minimum heat exchanger mass for a selected
operating condition (basically, for given Q and TS Tci). The data used in the present case are the
following:
Heatload,Q=0,5x103W
Airmassflowrate,mc=0,0303kg.s1and0,0817kg.s1
Airinlettemperature,Tci=295K
Coolantmassflowrate,mh=0,067kg.s1
Percentageofhotfluidmassflowratebypassedaroundsourcefortemperaturecontrolpurposes,10%
Productoftotalheattransferareaandoverallthermalconductanceofthecoldplate(UA)S=500W.K1
The air free flow areaat the inlet of the first heat exchanger module, AFL, isobviouslyan important
parameter for optimization. In addition, since the cold flow length of the modules is the same
regardless of the module size, the optimization of the heat exchanger is based on AFL and on the
number,n,ofmodules.
Fromanumberofdetaildesignsofheatexchangermodules,withsquareinletairflowcrosssections
(ofdifferentsizes)adependencebetweenthemass,M,ofthemodule(includingtheliquidinit)and
theairfreeflowarea,AFL,asbeenfound,namely:
M=97AFL+0,26
whereMisinkgandAFLinm2.Thisrelationisusedinthemassoptimizationoftheexchanger.
ThenumericalresultsfortwodifferentcasesaresummarizedinTable131.
Table131:OptimizationoftheLiquidtoAirHeatExchanger
Case1)mc=0,0303kg.s1,TSTci=20K
No.of AFL M nM Cold Volume,a Formc=0,0908kg.s1b
Modules [m2] [kg] [kg] Flow vx103 Airvelocity Airpressure
Length [m3] [m.s1] drop,px103
[m] [Pa]
2 0,01740 1,948 3,896 0,150 4,26 5,6 0,050
3 0,00892 1,125 3,376 0,225 3,63 10,9 0,051
4 0,00553 0,796 3,186 0,300 3,47 17,6 0,052
5 0,00395 0,643 3,216 0,375 3,52 24,7 0,053
308
ECSSEHB3101Part13A
5December2011
Table132:OptimizationoftheLiquidtoAirHeatExchanger
Case2)mc=0,0817kg.s1,TSTci=10K
No.of AFL M nM Cold Volume,a Formc=0,0908kg.s1b
Modules [m2] [kg] [kg] Flow vx103 Airvelocity Airpressure
Length [m3] [m.s1] drop,px103
[m] [Pa]
2 0,02070 2,268 4,536 0,150 4,96 4,7 0,055
3 0,01075 1,303 3,909 0,225 4,21 9,1 0,057
4 0,00688 0,927 3,709 0,300 3,98 14,2 0,059
5 0,00495 0,740 3,701 0,375 3,97 19,7 0,061
6 0,00370 0,619 3,713 0,450 4,13 26,4 0,061
a Volumeofthecubecircumscribingtheheatexchanger.
b Maximumallowableairmassflowrate.
NOTE FromBerner,Schleicher,Oesch,Nyfeler&Steinhauser(1977)[14].
It is apparent from these results that a heat exchanger with an AFL = 0,005 m2 would have a nearly
minimummassandoccupyaboutthesmallestvolume,thusitmaybeconsideredanoptimumresult.
TheairvelocityinthecrosssectionofareaAFLandtheairpressuredropthroughtheheatexchanger
have been calculated also on the bases of maximum allowable air mass flow rate. The aim of these
calculationsistotestthefulfillmentofrequirements5)and6)above.Itisseenthatrequirement6)may
notbefulfilledwhentheoptimumexchangerisused.Thencethisoptimumconfigurationisruledout
andavalueAFL=0,0225m2hasbeenchosen.WiththisvalueofAFLtheaboveoperatingconditionscan
be satisfied with n = 2 modules. Larger values of n will be probably required when Q is increased
and/orTSisdecreasedbeforesettingupaheatpumpwithitslargepowerconsumption.
309
ECSSEHB3101Part13A
5December2011
14
Two-phase flow
14.1 General
Gasliquidflowinaductisaverycomplexphenomenonbecauseaninterfacebetweentwofluids,one
ofthemcompressible,ispresent.
Differentflowpatternsexistdependingontheconditions.Thedefinitionoftheseflowpatternsandof
theirboundarieshavebeenmuchdebated.Thereadershouldbeaddressedtobasictextsonthefield,
suchasGovier&Aziz(1972)[75],Hetsroni(1982)[85],amongothers.
Figure 141 shows a flowpattern map for vertical upward flow of air and water, which has been
preparedafterHewitt(1982)[86].Asimilarmapalsoforairandwaterbutinhorizontalflowisgiven
inFigure142.
310
ECSSEHB3101Part13A
5December2011
Figure141:Flowpatternmapforverticalupwardairwaterflow.Preparedbythe
compilerafterHewitt(1982)[86].mGandmLarethegasandliquidmassflowrates,
respectively.AFListheinternalcrosssectionalareaoftheduct.
311
ECSSEHB3101Part13A
5December2011
Figure142:Bakerflowpatternmapforhorizontalairwaterflow.FromHewitt
(1982)[86].mGandmLarethegasandliquidmassflowrates,respectively.AFLis
theinternalcrosssectionalareaoftheduct.
Thistypeofmapshasbeencriticizedveryoftenonthegroundsthatsodifferenttransitionprocesses
hardly can be controlled by the same parameters. Thus, different investigators use different maps,
withdifferentboundaries,intermediatepatternsandevendifferentnomenclatures.
Data covering a wider range of fluid properties are also available. For example, in the case of
horizontal flow, and following Baker, the physical properties of the fluids are taken into account
throughthefollowingtwoparameters
2 1/ 3
L
1/ 2
0,0728 L 10 3
B G , B [141]
3
1,205 10 0,001002 L

wherethenumericalvaluescorrespondtoairandwaterunderstandardconditions(105Pa,293K).
TheflowpatternmapisthengivenintermsofmG/BAFLvs.mLB/AFL.Obviously, Band Bareequal
tounityforstandardairwaterflow.
Informationontheflowpatternsinreducedgravityisscarce.ComparisonofFigure141andFigure
142indicatethat
1. Gravityeffectsareimportant.
2. The annular flow pattern remains in the same position in both maps: low (but not too
low)liquidvelocityandlargegasvelocity.
312
ECSSEHB3101Part13A
5December2011
14.2 Pressure loss
14.2.1 Lockhart-martinelli correlation

The flow pattern under reduced gravity is uncertain. The pressure loss would depend on this flow
pattern,buttheavailableexperienceindicatesthatthisdependenceisnotlarge.
For horizontal flows through small diameter ducts (few centimeters in diameter), without phase
change or significant acceleration, the LockhartMartinelli correlation (Lockhart & Martinelli (1949)
[116])hasbeenwidelyusedduringthelastthirtyyears.Itissimpletouse,itdoesnotdependonthe
details of the flow pattern and it is often used to calculate both the void fraction and frictional
pressure loss even when other effects are not negligible. This procedure leads to progressively
increasing errors as the frictional component of pressure loss decreases in proportion to the other
terms (acceleration or phase change). Clearly, separate correlations should be used for each flow
pattern, as has been suggested by many investigators, but when data from these partiallyvalid
correlations are compared to those from LockhartMartinelli the agreement, although poor, could
sufficeformanyengineeringpurposes.
IntheLockhartMartinellicorrelationthefrictionalpressurelossofthetwophaseflow,p/L,isrelated
toeither(p/L)SLor(p/L)SGthroughpressurelossmultiplierswhicharedefinedasfollows:
p / L p / L
2L , 2L
p / L SL p / L SG [142]
where(p/L)SListhepressurelossthatwouldoccuriftheflowrateofliquidwereflowingaloneinthe
duct,
(p/L)SGisthepressurelossthatwouldoccuriftheflowrateofgaswereflowingaloneintheduct.
Thepressuregradient,dp/dx,isveryoftenusedforthepressureloss,p/L.p/L=dp/dx.
Itisassumedthat(p/L)SLand(p/L)SGareknown.
In order to avoid having the unknown twophase pressure loss in both correlating parameters it is
convenienttointroduceanewvariable(theLockhartMartinelliparameter),
p / L SL G2
X2
p / L SG
[143]
2L
X2givesameasureofthedegreetowhichthetwophasemixturebehavesastheliquidratherthanas
thegas.
LetusintroduceaReynoldsnumber,Re,tocalculateXinthedifferentcases
mD
Re [144]
AFL G
misthetotalmassflowrate,[kg.s1].
313
ECSSEHB3101Part13A
5December2011
Disthediameteroftheduct,ortheequivalentdiameterifthecrosssectionoftheductisnotcircular,
[m].
AFListheinternalcrosssectionalareaoftheduct,[m2].
Gisthedynamicviscosityofthegasphase,[Pa.s].
IntermsoftheReynoldsnumber,Re,andthevaporquality,w,whichisdefinedas
mG
w [145]
m
theReynoldsnumbersofeachphasebecome,respectively,
G
ReL 1 w Re , ReG wRe [146]
L
Intermsofthefrictionfactors,fSGandfSG,theLockhartMartinelliparametercanbewrittenoutas(see
clause7.2fortherelationbetweenp/Landf=/4)
G 1 w f SG ReG
2
X2 [147]
L w f SL ReL
where isthedensity,[kg.m3],andfsthefrictionfactorofeachphaseifitwereflowingaloneinthe
duct.
Using for the friction factor vs. ReL and ReG either the HagenPoiseuille (clause 7.2) or the Blasius
formula(clause7.2.2.1),Xcanbeexpressedasfollows:
1 wn
1/ 2 n1
L 2
X E G [148]
L G w n3
Eisaconstantfactor,andn1,n2,n3areconstantexponents.
According to LockhartMartinelli, L and G are functions of X only although there is a slight
differencewhichdependsonthenature,laminarorturbulent,ofthetwoseparatephasealoneflows.
ThetransitionliquidandgasReynoldsnumbersareRe=1000,smallerthanthatforsinglephaseflow,
clause7.2(seeGovier&Aziz(1972)[75],p.534).ThefactorEandtheexponentsn1,n2andn3ofEq.[14
8]alsodependonthenatureofsinglephaseflowandaregiveninthefollowingTable.
314
ECSSEHB3101Part13A
5December2011
LIQUID GAS SUBSCRIPT E n1 n2 n3
Turbulent Turbulent tt 1 1/8 7/8 7/8
Laminar Turbulent vt 19,2/Re3/8 1/2 1/2 7/8
Turbulent Laminar tv Re3/8/14,2 1/8 7/8 1/2
Laminar Laminar vv 1 1/2 1/2 1/2
Chisholm (1967) [25] suggested a simple and accurate analytic representation of the multipliers
(Figure143).
Figure143:LockhartMartinellicorrelationforpressurelossmultipliers.The
figurehasbeendrawnbyuseofChisholmanalyticalrepresentation.Seetext.
C 1
2L 1 2 , G2 1 CX X 2 [149]
X X
whereCdependsonthenatureofthephasealoneflowasfollows
315
ECSSEHB3101Part13A
5December2011
Subscript tt vt tv vv
C 20 12 10 5
OtherrepresentationofLandGis(Wallis(1969)[184],p.50)

2L 1 X 2 / n
n

, G2 1 X 2 / n
n
[1410]
wherenalsodependsonthenatureofthephasealoneflow
Subscript tt vt tv vv
n 4 3,5 2,75forX<5 3,5forX>5
AcomparisonofEqs.[149]and[1410]withexperimentaldataisgiveninFigure144.
316
ECSSEHB3101Part13A
5December2011
Figure144:Gasphasepressurelossmultiplier,G,vs.LockhartMartinelli
parameter,X,asdeducedfromEqs.[149]and[1410]andfromexperimentaldata.
a)ttcase;b)vtcase;c)vvcase.ExperimentalpointsarefromWallis(1969)[184].
Thevoidfraction, ,representsthefractionofanelementofvolumewhichisoccupiedataninstant
bythegasphase(orthelighterphaseinastratifiedflow).1istheliquidfraction.
The LockhartMartinelli correlation also gives the void fraction, , in terms of X (Figure 145). The
curvecanbefairlyrepresentedbytheequation(Wallis(1969)[184],p.51)
1 X 0,8
0 , 378
[1411]
317
ECSSEHB3101Part13A
5December2011
Figure145:Liquidfraction,1,vs.LockhartMartinelliparameters,X.From
Wallis(1969)[184].
14.2.2 Improvements upon martinelli correlation

Equations [149] are general for all types of flow and can be easily adapted to specific data by
correlatingthefactorC.
ThelasttwotermsoftherighthandsideofEqs.[149]correspondtothetheoryforthelimitingcases
ofapproachingallgasorallliquidflowsrespectively.ThefactorChasbeenfoundtobeafunctionof
flowregime,vaporandliquidReynoldsnumbers,densityratio,geometry,andsurfacetension.
Martinelli & Nelson (1948) [122] noted that the LockhartMartinelli correlation does not become
asymptotic to the correct value as the critical pressure, pc, is approached. C should decrease with
increasingpressureandtheoreticallybecomezeroatthecriticalpressure.
Originally, it was intended to multiply the factor C by (1 p/pc). However, as pointed out by Palen
(1976)[140],limiteddataforcondensingnpentaneatp/pc0,5indicatethatthecorrectioncannotbe
extendedtoallfluids.
Based on Martinellis results, HTRI (Yang & Palen (1977) [195] developed a generalized curvefit
relationvalidwhenbothphasesareturbulent:
31 p / pc
1, 75
1
2
Ltt 1 0 , 55
1,11 [1412]
X tt X tt
wherepisthelocalpressureandpcthelocalaveragecriticalpressureofthegasphase.
Equation[1412], which was developed from data obtained in horizontal duct flow, can be usedfor
flowregimesotherthanturbulentturbulentduetooveralldatascatter.
TheLockhartMartinellicorrelationalsofailstotakeadequateaccountoftheinfluenceofflowquality,
w.IttendstopredicttoolowLathighqualityandtoohighatlowqualityflows,Figure146.
318
ECSSEHB3101Part13A
5December2011
Figure146:ComparisonofLockhartMartinelliwithavailableexperimentaldata.
PointswithCasperEq.[1413],withG/L=0arealsoshown.FromYang&Palen
(1977)[195].
Yang&Palen(1977)[195]suggestedthefollowingsimpleexpressionforCintheturbulentturbulent
caseforshearcontrolledflowregimes
1, 5
2
C=2,75 1 0,5 1 G [1413]
X tt L
Thisexpressiongivesasgoodafittoexperimentaldataasmanyofthemorecomplicatedformsandis
recommendedforcondensationpressuredrop.
Continuing attempts have been made to improve the correlations for frictional pressure loss, see a
fairlyrecentaccountinHewitt(1982)[86].
14.3 Annular flow

Annularflowisthemostubiquitousoftheflowregimes.Itconsistsofahomogeneousmixtureofgas
and droplets (the gas core) surrounded by a liquid layer near the wall (the liquid film). If the core
containsasignificantnumberofentraineddroplets,theflowisdescribedasannularmist,whichcan
be regarded as a transition between ideal annular flow and fully dispersed drop flow pattern. In
general, the phases may not be in equilibrium with respect to temperature and concentrations. In
319
ECSSEHB3101Part13A
5December2011
many cases, however, the interface heat and mass transfer rates are very high because of the large
interfacial area and high turbulence, thence the phases reach an equilibrium during their passage
throughtheduct.
Annular flow is the predominant flow pattern in evaporators, natural gas pipelines, steam heating
systemsand,ingeneral,reducedgravityflowsinducts.
14.3.1 Ideal annular flow model

Letusassumesteadymotionthroughaductofcircularcrosssectionwhennoliquidisentrainedin
thegascore.
14.3.1.1 Mass preservation equation for either phase

Asthepressureandtemperaturedonotchangewithinasectionoftheduct,thedensityofthegascore
doesnotchangewithinthatsection,thereforethemasspreservationequationforeachphaseis
dmG dmL
0 [1414]
dx dx
mbeingthemassflowrates.
CentralgascoreandouterliquidfilmareshowninFigure147.
Figure147:Theannularflowconfiguration.
14.3.1.2 Axial momentum equation for either phase

1. Themomentumequationfortheliquidfilmcanbeexpressedas:
dp 1 d r
[1415]
dx r dr
where is the shear stress in the liquid. It is assumed in the above equation that is
constantsothataccelerationtermsarezero.Gravityeffectsarealsoneglected.
If we integrate once the above equation with the boundary condition = i at r = (D
2)/2,beingitheshearstressattheinterface,weobtain
320
ECSSEHB3101Part13A
5December2011
D 2 dp D y y
i [1416]
D 2 y dx D 2y
Hereyisthedistancemeasuredfromthewalloftheduct:y=(D/2)r.Atthewall,y=0,
wehave,
D 2 dp D
w i [1417]
D dx D
2. Inthegascorethedifferentialmomentumequationcanbewrittenexactlyasbeforeifwe
neglecttheinertiatermcomparedwiththepressureterm.Thisisjustifiedwhenchanges
in pressure and temperature are small compared with the prevail ing pressure and
temperaturelevels,respectively.
Integratingoncewiththeboundarycondition=0atr=0weget,
dp r
[1418]
dx 2
whichattheinterface,r=(D2)/2,becomes
dp D 2
i [1419]
dx 4
Combination of Eqs. [1417]and[1419] relates the wall shear stress, w, to the pressure
lossperunitlengthoftheduct,dp/dx.
dp D
w [1420]
dx 4
14.3.1.3 Pressure loss vs. friction factors fl and fgi

Inordertocalculatewandi,weintroducethefrictionfactorsfortheliquidatthewall,fL,andforthe
gasattheinterface,fGi,asfollows:
1. Fortheliquidatthewall:
1
w LVL2 f L [1421]
2
VLbeingtheliquidmassaveragevelocity,
321
ECSSEHB3101Part13A
5December2011
mL
VL
L D
[1422]
andfL=fL(ReL),ReLbeingtheReynoldsnumberfortheliquid,
4 mL
ReL [1423]
D L
Withtheaboveexpressionofw,Eq.[1420]becomes:
dp 2mL2 f L
[1424]
dx L 2 D 2 D 2
2. Forthegasattheinterface:
1
i GVG2 f Gi [1425]
2
whereVGisthegasmassaveragevelocity,
4mG
VG
G D 2 2
[1426]
RecallingEq.[1419],
dp 32mG2 f Gi
[1427]
dx G 2 D 2 5
14.3.1.4 Laws of friction for fl and fgi

Equations[1424]and[1427]givetworelationsbetweenthepressureloss,dp/dx,thefrictionfactorat
thewall,fL,thefrictionfactorattheinterface,fGi,and /D.Thereforewewouldneedtwoadditional
relationstoclosetheproblem.
1. Iftheliquidwereflowingaloneintheduct,thepressurelosswouldbegivenby(clause
7.2withp/L=(dp/dx)SL,=4fSG,DE=D,=L,V=4mL/LD2).
322
ECSSEHB3101Part13A
5December2011
dp 32mL2 f SL
[1428]
dx SL L 2 D 5
where fSG is a function of the Reynolds number ReSG = 4mL/DL). Since ReSG = ReL, as
defined in clause 14.3.1.3, the friction factor at the wall will be the same as thatif the
liquid were flowing alone at the same flow rate, fL = fSG. The value of fL vs. ReL will be
givenby
16
fL [1429]
ReL
forlaminarflow(HagenPoiseuille,clause7.2.2),
f L 0,079 ReL0, 25 [1430]
forturbulentflow(Blasius,seeTable72)
For the transition regime, fL has been numericallycalculated by Hewitt (1982) [86], see
Figure148
Figure148:Liquidfilmfrictionfactor,fL,asafunctionofliquidfilmReynolds
number,ReL,forannulartwophaseflowincylindricalductsofcircularcross
section.FromHewitt(1982)[86].Thefulllinescorrespond,respectively,toHagen
PoiseuilleformulaandtoBlasiusformula.fLisequaltothefrictionfactorfsLwhich
correspondstothesinglephaseflowalongtheductatthesameReynoldsnumber
exceptforanintermediateReL.
2. The most widely used method for calculating fGi consists in assuming that the interface
affectsthecoreflowasaroughductwallshould,andtoseekanempiricalrelationship
betweenfGiandthedimensionlessfilmthickness, /D.Whalley&Hewitt(1978),quoted
byHewitt(1982)[86],suggesttheexpression
323
ECSSEHB3101Part13A
5December2011
1/ 3
f Gi
1 24 L [1431]
f SG G D
wherefSGisthefrictionfactorforthegasifitwereflowingaloneintheduct.
14.3.1.5 Expressions in terms of martinelli parameters

Thepressurelossesintheliquidandgasphase,givenrespectivelybyEqs.[1424]and[1427]should
beequal.Thus,

5

2 1 2
f Gi G mL D
2
D
[1432]
f L L mG2 4 2
1
D
SincefL=fSL,Eq.[1432]furnishesanalternativeexpressionoffGi/fSGintermsoftheLockhartMartinelli
parameter,X(Eq.[143]),clause14.2.1)

5 5

2 1 2 2 1 2
f Gi G mL2 f SL D D 2 D D
X [1433]
f SG L mG2 f SG 4 4
2 2
1 1
D D
whereX2isknownforgivenmL,mG,Dandthedensitiesofpureliquid,L,andpuregas,G.
EliminationoffGi/fSGfromEqs.[1431]and[1433]givestheequation
1/ 3
2 1
4 D
X 1 24 L
2
[1434]
G D D
5

1 2
D
whichrelates /Dwith L/GandX. /Dinitsturnisrelatedtothe(volumetric)liquidfraction,1 ,

throughthepurelygeometricexpression
4
1 1 [1435]
D D
324
ECSSEHB3101Part13A
5December2011
Finally,theMartinellipressuregradientmultiplier,L2,(Eq.[142],clause14.2.1)canbededucedfrom
Eqs.[1424]withfL=fSL,and[1428],whichyieldanexpressiononlydependenton/D
2
D 1

2

4
L 2
[1436]
1
D
L,G=XL,/Dand1vs.XareshowninFigure149
Figure149:Pressuregradientmultipliers,LandG,dimensionlessfilm
thickness,/D,andliquidfraction,1,accordingtoEqs.[1434],[1435]and[14
36],withL/G=1000.Calculatedbythecompiler.
A comparison between theoretical and experimental values of 1 and G vs. X can be seen
respectively,inFigure1410andFigure1411.
325
ECSSEHB3101Part13A
5December2011
Figure1410:Comparisonoftheliquidfraction,1,vs.LockhartMartinelli
parameter,X,asdeducedfromEqs.[1434]and[1435],withL/G=1000,with
experimentaldatafromWallis(1969)[184].
Figure1411:Comparisonofthegaspressuregradientmultiplier,G=xL,vs.
LockhartMartinelliparameter,X,asdeducedfromEqs.[1434]and[1436],with
L/G=1000,withexperimentaldatafromWallis(1969)[184].ttcase.
14.3.1.6 Summary
Itisconvenienttosummarizehere,forfurtherreference,thestepspursuedtoestimatethepressure
dropintheIDEALannularflowalongahorizontalcylindricalduct.
Themomentumbalanceineitherphasegivestheshearstressesatthewall,w,andattheinterface,i.
dp D
w [1437]
dx 4
dp D 2
i [1438]
dx 4
wherethevalueofdp/dxisthesameinbothcases.
326
ECSSEHB3101Part13A
5December2011
TheshearstresseswandiarerelatedtothefrictionfactorsfLandfGi,respectively
1
w LVL2 f L [1439]
2
beingVLthemeanliquidvelocity,VL=mL/L(D).
1
i GVG2 f Gi [1440]
2
withthemeangasvelocityVG=4mG/G(D2)2.
IntermsoffLandfGithepressuregradientalongtheductbecomes
dp 2mL2 f L
[1441]
dx L 2 D 2 D 2
dp 32mG2 f Gi
[1442]
dx G 2 D 2 5
ThefrictionfactorsfGandfGiarerelatedtothosecorrespondingtoeitherphaseflowingalonewiththe
same mass flow rate through the duct. The last mentioned friction factors, fSG and fSG, are known
functionsoftherespectiveReynoldsnumbers,ReLandReG.
1/ 3
fL f Gi
1 , 1 24 L [1443]
f SL f SG G D
Since the same combinations of friction factors appear in the LockhartMartinelli formulation, we
couldexpressX2,L2andG2=X2L2asfunctionsof/Dexclusively.
1/ 3
2 1
4 D
X 1 24 L
2
[1444]
G D D
5

1 2
D
2
D 1

2

4
L 2
[1445]
1
D
327
ECSSEHB3101Part13A
5December2011
14.3.1.7 Worked example

Measurementsperformedwithanairwaterflowat293Kinahorizontalductofdiameter2,5x102m
atmassflowratesmL=0,15kg.s1andmG=8,5x103kg.s1showedthat,whentheexitpressureis105
Pa,thepressure5mupstreamisfoundtobe1,11x105Pa.Comparethesedatawiththepredictionsof
1)pureannularflow,and2)LockhartMartinellicorrelation.
Thedensityandviscosityofairandwaterat293Kare,respectively:
G=1,21kg.m3,
L=998kg.m3,
G=1,85x105Pa.s1,
L=1,003x103Pa.s1.
TheReynoldsnumbersare:
4mL
ReL 7620
D L
[1446]
4mG
ReG 23400
DG
Therefore,theregimesareturbulentinbothphases.
UsingBlasiusformula(clause14.3.1.4)wehave:
fSG=0,079x(7620)0,25=0,0085,
fSG=0,079x(23400)0,25=0,0064.
TheLockhartMartinelliparameter,X,is(Eq.[143],clause14.2.1)
p / L SL
2
mL f SL
X 2
G 0,50 [1447]
p / L SG L mG f SG
1. Pureannularflowmodel.
UsingEq.[1434]withX2=0,50weobtain/D=0,044.
Equation[1436]with/D=0,044yieldsL2=35,3.
The pressure loss if the liquid were flowing alone through the duct would be given by
Eq.[1428]
dp 32mL2 f SL
63,6 Pa.m 1 [1448]
dx SL L D
2 5
andthepressurelossforthetwophaseflow,
328
ECSSEHB3101Part13A
5December2011
dp dp
2L 2250 Pa.m 1 [1449]
dx dx SL
2. LockhartMartinellicorrelation.
IftheLockhartMartinellicorrelationwereusedintheChisholmform,Eq.[149],withC
=20
L2=31,2,
anddeducedfromEq.[1411]wouldbe:
=0,081,
whereasthatdeducedforpureannularflow,Eq.[1435]with/D=0,044asabove,would
be:
=0,083.
TheresultsfromtheannularflowmodelarecomparedtothosefromLockhartMartinellicorrelation
inthefollowingTable.
Annularmodel LMcorrelation
L2 35,3 31,2
0,083 0,081
dp/dx[Pa.m1] 2250 1980
The measured pressure loss, 2200 Pa.m1, fairly agrees with that predicted by means of the annular
model.ThemuchsimplerLockhartMartinellicorrelationgivesanotsogoodresult.
14.3.2 Annular flow with entrainment model

Whenthecoregasvelocityislow,itseffectonthefilmisnegligible.However,withincreasingrelative
velocityadestabilizingeffectisnoted.Atthebeginningsmallripplestravellinginthedirectionofthe
filmappear.Highercorevelocitiesleadtoanincreaseintheamplitudeoftheripplesandsoonthree
dimensionaldisturbancesaregenerated.Whenthecorevelocityissufficientlyhigh,dragforcesonthe
topsoftherollwavespulloffdropletsofliquidtowardthegascore.
Thedataofmanyworkersconfirmthat,exceptatlowliquidrates,asignificantportionoftheliquid
maybecarriedasdropletsinthegascore.Theflowintheliquidfilminitiallyincreaseswithliquid
flowrate,buteventuallylevelsoutatavaluelargelydependentonthegasflowrate.Atthehighest
gasandliquidflowratesapproximately90percentoftheliquidiscarriedinentraineddroplets.
14.3.2.1 Mass preservation equation for either phase

Themassflowratesofgas,mG,andliquid,mL,remainconstantalongtheduct,asinclause14.3.1.2.
NowmLconsistsofthemassflowrateofliquidentrainedinthegascore,mE,plusthatintheliquid
film,mF.Themasspreservationequationsare:
329
ECSSEHB3101Part13A
5December2011
dmG d mE mF
0 [1450]
dx dx
14.3.2.2 Axial momentum equation for either phase

Ifweassumethatinertiatermsarenegligiblecomparedwiththepressuretermsandiftheamountof
liquidentrainedinthecoredoesnotchangealongtheduct,becausemasstransferduetoentrainment
bythecoreandtodepositionofdropletsinthefilmcanceltoeachother(hydrodynamicequilibrium
hypothesis),themomentumequationfortheliquidinthefilmreducestoEq.[1420],whereasthatfor
thegascorereducestoEq.[1419].Theproblemisnowtocalculatetheshearstressatthewall,w,and
theshearstressattheinterface,nowcalledi*,todistinguishitfromthatwithoutentrainment,i.
14.3.2.3 Pressure loss vs. friction factors ffand fgi

Herewewillfollowascloselyasfeasibletheapproachofclause14.3.1.3.
1. 1)Fortheliquidatthewall.
The shear stress at the wall is calculated as in the case of noentrainment, taking into
accountthatthemassflowrateofliquidflowinginthefilmisnowmLmE.Thatis:
1
w LVF2 f F [1451]
2
VFbeingtheliquidfilmmassaveragevelocity,
mL mE
VF
L D
[1452]
and fF the friction factor at the wall. fF = fF(ReF), ReF being the Reynolds number for the
massflowrateofliquidinthefilm,
4mL mE
ReF [1453]
D L
Withtheaboveexpressionofw,Eq.[1420]becomes:
2mL mE f F
2
dp
[1454]
dx L 2 D 2 D 2
Thefrictionfactoratthewall,fF,isthesameasthatiftheliquidinthefilmwereflowing
aloneatthemassflowratemLmE.
2. Forthegascoreattheinterface.
330
ECSSEHB3101Part13A
5December2011
Equation[1419]relatingdp/dxtoiremainsvalidbywritingi*inplaceofi.
Inordertorelate i*withtheinterfacialfrictionfactorwithentrainment,fGi*,weneeda
mean core density, c, and a characteristic velocity at the interface, be it the mean core
velocity,VG,orthetangentialvelocityjumpattheinterface.
Following Wallis (1969) [184], p. 363, we will envisage the core as a homogeneous
mixturewithameandensity,c,whichcanbewrittenas
mG mE m mE
c ~ G G [1455]
mG / G mE / L mG
sincemE/L<<mG/G.
Themeanvelocitiesofdropletsandgasareassumedtobethesame,thatis,thereisno
slipbetweenthem.Thenthismeanvelocity,VG,isgiveninclause14.3.1.3.VGisuniform
throughoutthecore.
The shear stress at the interface, i*, should be made dimensionless with the mean
density,c,andtheinterfacialvelocityjumpwhich,inprinciple,differsfromVG.
Whenthefilmthicknessissmall,theshearstresswithinthefilmisnearlyconstantand
thentheinterfacevelocityattheliquidsideresultstobetwicetheaverageliquidvelocity,
VF.Thisisjustified,atleastforlaminarflowinthefilm,becauseaconstantshearstressin
thefilmimpliesalinearvelocityprofile.
Withtheaboveassumptions,theinterfacialshearstresswillbe,
2
1 2V * 1
cVG2 1 F
*
i
f Gi cVG2 f Gi* [1456]
2 VG 2
wherefGi*istheinterfacialfrictionfactorwithentrainment.2VF/VGis,inmostcases,small.
RecallingEq.[1419],butwithi*insteadofi,weget
dp 32mG2 mG mE *
f Gi [1457]
dx G D 2
2 5
mG
14.3.2.4 Laws of friction for ff and fgi*

1. Thefrictionfactoratthewall,fF,canbeobtainedusingeitherPoiseuilleformula,Blasius
formula,orFigure148.WeonlywillreplaceReLbyReF.
2. ForthefrictionfactorattheinterfaceWallisassumesthattheexpression
1/ 3
f Gi*
1 24 L [1458]
G
*
fc D
331
ECSSEHB3101Part13A
5December2011
analogous to Eq. [1431], holds. Here fc* is the friction factor if the homogeneous core
mixturewereflowingaloneintheductwiththesameviscosityofthegas,G;thatis,fc*is
the same function of the Reynolds number Rec = 4(mG + mE)/DG, as those given by
Poiseuille,BlasiusformulaeorinFigure148,changinginanycaseReLbyRec.
14.3.2.5 Expressions in terms of martinelli parameters

Thepressurelossesintheliquidandgasphase,givenrespectivelybyEqs.[1454]and[1457]should
beequal.Thus,

5

2 2 1 2
f Gi* G mL mE mG D D

L mG mG mE 4 2
[1459]
fF
1
D
whichisakintoEq.[1432]fortheidealannularflowcase.
Introducing the LockhartMartinelli parameter, X, again deduced from the onlyliquid to onlygas
pressurelossratio,
2
mL f SL
X G
2
[1460]
L mG f SG
Eq.[1459]becomes:
2
f Gi* f f m mE
*
X 2 SG* F L
fc f c f SL mL

5

2 1 2 [1461]
mG D D

mG mE 4 1
2
Note that the friction factors fSG and fSG which appear in the LockhartMartinelli formulation
correspondrespectivelytothemassflowratesmLandmGofthewholeliquidandwholegasflowingin
theduct.fF,initsturn,correspondstothemassflowrate,mLmE,oftheliquidinthefilm,andfc*to
themassflowrate,mG+mE,ofthefluidinthecore.
EliminationoffGi*/fc*fromEqs.[1458]and[1461]resultsinthefollowingequation
2
f f F mL mE mG
X SG*
2
F [1462]
fc SL mL
f mG mE D
332
ECSSEHB3101Part13A
5December2011
where
1
4
1/ 3
D
2
F 1 24 L [1463]
D D G D
5
1 2
D
ispreciselythefunctionwhichappearsintherighthandsideofEq.[1434].InEq.[1462]mLandmGas
wellas Land GareknownandthenfSG,fSGandXcanbededuced.Ifthemassflowrateoftheliquid
entrainedinthegascore,mE,werealsoknownwecouldobtainfc*(whichdependsonmG*mE),andfF
(dependingonmLmE).Inthatcase,Eq.[1462]wouldgivethevalueof/D.
F(/D)vs. /Dfordifferentvaluesof L/GisgiveninFigure1412.Notethatchangesof L/Ginthe
range700<L/G<1100donotaffectthevaluesofF(/D).
Figure1412:ThefunctionF(/D)whichappearsinEq.[1462].Theeffectofthe
liquidgasdensityratioisnegligiblefortherangeofvaluesgiveninthefigure.
Comparethevaluesgiveninthiscurvewiththosegivenby2/Dvs.XinFigure
149.
Anadditionalrelationbetweenthemassflowrate,mE,oftheliquidentrainedinthegascoreandthe
liquid film thickness, /D, is required at this point. Figure1413,which refers to airwatermixtures,
relatestheMartinelliparameter,X,withanentrainmentparameter,RE,definedas
333
ECSSEHB3101Part13A
5December2011
Figure1413:Martinelliparameter,X,vs.entrainmentparameter,RE,asdeduced
fromairwaterflowexperiments.FromWicks&Duckler(1960)[190].NotethatRE
isnotdimensionless.
m WemE
RE G L
dp / dx SG
[1464]
L mG
In this equation RE has dimensions [m2.s]. The dimensionless critical Weber number, We = G(VG
VL)2D/, where VG and VL are entering air and water velocities, depends on how the liquid is
introducedintheduct.ThevalueWe=16istakenwhentheliquidisintroducedabruptly,aswitha
tee entry, and the value We = 22 when the liquid is introduced smoothly as in an annular entrance
device.
OncemEisknown,weobtain /DfromEq.[1462],dp/dxfromEq.[1454],and(dp/dx)SLfromEq.[14
28].Theliquidpressuregradientmultiplier,L,becomes:
2
m mE f F D
2
1
L
2

f SL 4
L 2
mL [1465]
1
D
ComparethisequationwithEq.[1436].
14.3.2.6 Additional data on entrainment

Informationonentrainmentforannularflowinductsisscarce.Themechanismofentrainmentisnot
wellunderstoodandtheexperimentaldifficultiesarelarge.
Figure1414showsacollectionofdata,byHutchinson&Whalley(1973)[93]forverticalannularflow
inductsofairwaterandairalcoholsystems.Thediametersoftheductsareintherange9,5x103mto
31,8x103.
334
ECSSEHB3101Part13A
5December2011
Figure1414:Concentrationofentraineddropletsinthegascore,GmE/mG,vs.
dimensionlessnumberi*/.ReplottedbythecompilerafterHutchinson&
Whalley(1973)[93].Differentsymbolsareusedtoindicateexperimentsby
differentauthors.
Theentrainmentmechanismsuggestedbytheaboveauthorsisbasedontheinteractionbetweenthe
shear stress at the interface, i*, and the surface tension forces. The liquid film thickness, , is the
characteristiclengthwhichappearsintheexpressionofthecapillarypressure,2/.Thustherateof
entrainmentcouldbecharacterisedintermsofthedimensionlessnumberi*/.
OrdinatesinFigure1414areconcentrations,C,ofdropletsinthegascorewhicharedefinedas:
mE mE
C G
mE mG mG [1466]

L G
These concentrations appear under equilibrium conditions, i.e., when the rate of entrainment of
droplets from the liquid film is equal to the rate of deposition of droplets into the film. The
equilibrium requires long distances along the duct (L/D > 150) and not too large pressure gradients
(p/p/L>20m).HutchinsonandWhalleydisregardeddatapointswhichdonotfulfiltheselimitations.
AlsoshowninFigure1414aredatapointstakenfromTable141,clause14.3.2.8.Thoseamongthem
which fulfil the abovementioned low pressure gradient limitation fall close to the continuous curve
andallarewithinexperimentalscattering.
ThedatagiveninFigure1414couldbeusefulforpurposesofclarifyingtheentrainmentprocess,but
can hardly be used for the calculation of the pressure loss, sincethey relate three a prioriunknown
variables,namelymE/mG,dp/dx(throughi*)and.
14.3.2.7 Summary
Wewillsummariseheretheapproachusedfortheestimationofthepressuredropalongahorizontal
cylindrical duct when ENTRAINMENT EFFECTS are taken into account. This clause is parallel to
clause14.3.1.6fortheIDEALannularflow.
It is assumed that the film thickness, , does not change along the duct (hydrodynamic equilibrium
hypothesis).
335
ECSSEHB3101Part13A
5December2011
Themomentumbalanceintheliquidfilmandinthecore(whichnowconsistsofgasandentrained
droplets)givestheshearstressesatthewall,w,andattheinterface,i*.
dp D
w [1467]
dx 4
dp D 2
i* [1468]
dx 4
wherei*isusedinsteadofitoindicatethatentrainmenteffectsaretakenintoaccount.
TheshearstressesarerelatedtothefrictionfactorsfFandfGi*,respectively:
1
w LVF2 f F [1469]
2
whereVFisthemeanliquidfilmvelocityVF=(mLmE)/L(D),mListheliquidtotalmassflowrate
andmEthatoftheliquidentrainedinthecore.Ontheotherhand,
1
i* cVG2 f Gi* [1470]
2
withthemeancorevelocity
4mG
VG
G D 2 2
[1471]
The mean velocities of entrained droplets and gas in the core are assumed to be equal (noslip
assumption).Inadditioncisameancoredensity(cmG=G(mG+mE)).
IntermsoffFandfGi*thepressuregradientalongtheductbecomes
2mL mE f F
2
dp
[1472]
dx L 2 D 2 D 2
dp 32mG2 mG mE *
f Gi [1473]
dx G D 2
2 5
mG
336
ECSSEHB3101Part13A
5December2011
Now we introduce the friction factor fc* which is that friction factor corresponding to the
homogeneouscoremixture,densityc,flowingaloneintheductwiththegasviscosityG.fc*isrelated
intheusualwaywithRec=4(mG+mE)/DG.
We now recall the friction factors fSG and fSG which appear in the LockhartMartinelli formulation
concerning the noentrainment case. The equation relating the Martinelli parameter, X, to the film
thickness,/D,isnow
2
f f F mL mE mG
X SG*
2
F [1474]
fc f SL mL mG mE D
whereF(/D)isaknownfunctionof/D.
Although X2 in Eq. [1462] is known, since it depends on mL, mG, L and G, neither the factor
multiplyingX2(whichdependsonmG+mE,andonmLmE)nor/Dareknown.
Experimentaldataonentrainment,suchasFigure1413,willprovidearelationbetweenXandmEto
closethesystem.
14.3.2.8 Worked example

Inordertosubstantiatetheprocedureforthesolutionofannularflowproblemswithentrainment,let
usconsiderthefollowingexample:
AliquidmassflowratemL=0,0656kg.s1andagasmassflowratemG=0,0266kg.s1areflowingalong
a duct of 0,0254 m in diameter. We aim at calculating the pressure gradient along the duct and the
liquidentrainedinthegascore.Theinletpressureis105Paandthetemperature293K.
Thedensityandviscosityofairandwateratp=105Pa,T=293Kare,respectively:
G=1,21kg.m3,
L=998kg.m3,
G=1,85x105Pa.s,
L=1,003x103Pa.s.
TheReynoldsnumbersofthephasesflowingaloneare,
4mG
ReG 72000 [1475]
DG
4 mL
ReL 3280 [1476]
D L
therefore,theregimesareturbulentinbothphases.
UsingBlasiusformula(clause14.3.1.4)wehave:
fSG=0,079x(72000)0,25=0,0048
fSG=0,079x(3280)0,25=0,0104
337
ECSSEHB3101Part13A
5December2011
TheLockhartMartinelliparameter,X,is(clause14.3.2.5),
2
mL f SL
X G
2
0,160 , X 0,126 [1477]
L mG f SG
WithX=0,126andFigure1413,theentrainmentparameter,RE,resultstobe
RE=2,2x106m2.s.
AccordingtoEq.[1428],thepressurelossfortheliquidflowingaloneis,
dp 32mL2 f SL
13,8 Pa.m 1 [1478]
dx SL L D
2 5
andthepressurelossforthegasflowingalone,
dp 1 dp
2 863 Pa.m 1 [1479]
dx SG X dx SL
Themassflowrateentrainedinthegascoreis,Eq.[1464]:
mL dp 1 0,635
mE RE L kg.s 1 [1480]
G mG dx SG We We
TakingWe=22,wehave
mE=0,0289kg.s1.
TheReynoldsnumberfortheliquidfilmis
4mL mE
ReF 1850 [1481]
D L
Theflowcorrespondstotheintermediateregime(seeFigure148andthen
fF=0,0178.
ThecoreReynoldsnumberis,
4mG mE
ReC 150000 [1482]
DG
338
ECSSEHB3101Part13A
5December2011
andthecorrespondingfrictionfactorfc*,
fc*=0,079x(150000)0,25=0,00401.
Now,wededuceF(/D)fromEq.[1462]
2
f f m mE mG
F X 2 SG* F L
D f c f SL mL mG mE [1483]
0,00492
WiththisvalueofF(/D)andFigure1412weobtain

0,00917 [1484]
D
Once/Disknown,Eq.[1465]furnishes
L2=402.
Fromthisvalueoftheliquidpressurelossmultiplier,L2,andfrom(dp/dx)SLwededuce
dp dp
2L 5550 Pa.m 1 [1485]
dx dx SL
TheresultsofsimilarproblemsaresummarisedinTable141.Thevaluesof L, G, L, G,DandWe
arethesameasintheaboveexample.
339
ECSSEHB3101Part13A
5December2011
Table141:TypicalPressureLossesinAirWaterAnnularFlowwithEntrainment
(D=25,4x103m,p=105Pa,T=293K).
mLx103 65,6 65,6 65,6 65,6 131 131 131 131 131
[kg.s1]
mGx103 30,7 26,6 19,9 16,0 62,7 55,3 48,4 42,8 36,4
[kg.s1]
ReGx103 83,2 72,1 53,9 43,4 170 149 131 115 98,6
ReLx103 3,28 3,28 3,28 3,28 6,54 6,54 6,54 6,54 6,54
fSGx103 4,7 4,8 5,1 5,5 3,9 4,0 4,2 4,3 4,5
fSLx103 10,4 10,4 10,4 10,4 8,8 8,8 8,8 8,8 8,8
X 0,111 0,126 0,162 0,197 0,109 0,121 0,137 0,153 0,175
(p/L)SGx103 1,11 0,864 0,520 0,355 3,87 3,11 2,46 1,99 1,50
[Pa.m1]
REx106 1,79 2,34 3,99 5,94 1,72 2,17 2,78 3,48 4,68
[m2.s]
mEx103 35,0 30,7 23,6 19,3 119 107 94,0 84,9 73,1
[kg.s1]
ReFx103 1,53 1,74 2,10 2,31 0,566 1,20 1,80 2,30 2,89
Recx103 178 155 117 95,7 494 439 388 345 296
fFx103 20,0 17,8 15,6 14,0 33,0 22,1 17,0 14,8 13,0
fc*x103 3,85 3,97 4,26 4,50 2,97 3,06 3,16 3,25 3,38
/Dx103 7,88 9,07 11,7 13,9 2,40 4,16 5,60 7,00 8,70
L2 428 374 279 222 298 308 289 268 242
(p/L)x103 5,90 5,13 3,82 3,66 13,7 14,1 13,4 12,5 11,1
[Pa.m1]
340
ECSSEHB3101Part13A
5December2011
14.3.2.9 The ideal annular and the annular with entrainment models
AcomparisonofEqs.[1434]and[1462]indicatesthattheidealannularmodelisvalidprovidedthat
2
f SG f F mL mE mG
* 1 [1486]
fc f SL mL mG mE
This expression depends on the gas and liquid mass flow rates, mG and mL, on the entrained liquid
flowrate,mE,andonthefollowingReynoldsnumbers.
1. Reynoldsnumberofthewholegasflowingintheduct,
4mG
ReG [1487]
DG
2. Reynoldsnumberofthefluidinthecore,intermsoftheductdiameter,
4mG mE
ReC [1488]
DG
NotethatinanycaseRecReG.
3. Reynoldsnumberoftheliquidinthefilm,intermsoftheductdiameter,
4mL mE
ReF [1489]
D L
4. Reynoldsnumberofthewholeliquidflowingintheduct,
4mL
ReL [1490]
D L
AgainReLReF.
Evenifweleaveasidetransitionalconfigurations,9versionsofEq.[1486]willresultdependingon
thelaminarorturbulentcharacteristicsofthefourflows.Thepossiblecombinationsarearrangedin
thefollowingmatrix.vandtareasinclause14.2.1.
341
ECSSEHB3101Part13A
5December2011
Theshadowedrowandcolumndenoteimpossibleconfigurationssince2cannotbevwhen1istand3
cannotbetwhen4isv.
Thetwosimplestcasesappearwhenthefourflowsareeitherlaminarorturbulent.
WhenthefourflowsarelaminarEq.[1486]becomes
mE
1 [1491]
mL
Whenthefourflowsareturbulent
7/4
mE
1
m L
1
3/ 4 [1492]
mE
1
mG
whichcanbeapproximatelyexpressedas
7 mE 3 mE
1 [1493]
4 mL 4 mG
Similarexpressionsresultwhenthegasonlyandcoreflowsarebothlaminarorbothturbulentand,in
addition,theliquidonlyandliquidfilmarebothlaminarorbothturbulent.
InthefiveremainingcasesD,andeitherGorL(orboth)cannotbeeliminatedfromEq.[1486].
The main limitation of a discussion of this type is the lack of data for calculating mE, such as those
presented in Figure 1413, which only refers to airwater flows. In practice mechanical devices are
usedtothrowthecoredropletstowardthefilmintheperiphery.
342
ECSSEHB3101Part13A
5December2011
14.4 Condensation in ducts

Theflowinductsoutofevaporatorswouldbedominatedbycondensation.Normalgravityenhances
condensation because body forces help drive the condensed liquid. Thence, condensation in ducts
underreducedgravityismuchlessefficientthanthatinsidehorizontalductsundernormalgravity.
Theflowpatternsineithercasediffersubstantially,Figure1415.
Figure1415:Flowgeometrywhengravityisdominant(a)ornegligible(b).The
voidfraction(vaporfractionalareainthefigure)isthesameinbothcases.
Acomparisonofcondensingflowsinsideductsundernormalandunderreducedgravityconditions
has been attempted by Keshock & Sadeghipour (1983). The basic difference between the two cases
resultsfromthestratifiedcondensatelayerwhichappearswhengravityisdominant(Figure1415a)
andwhichisalmostinactiveintheheattransferprocess,whereasthethincondensatefilmformedin
theductisveryeffective.Whengravityeffectsarenegligible(Figure1415b)theannularflowpattern
prevailsandthemuchthickerlayerislesseffectivethanthethincondensatewhichappearedinFigure
1415a
14.4.1 Condensing flow model

In this clause a simple unified model of condensing flow inside a duct is introduced. Later on this
modelwillbeparticularisedtothetwoextremecaseswhichhavebeensketchedinFigure1415.
Themainsimplifyingassumptionsofthemodelare:
1. Theductishorizontal.
2. Thestaticpressureisuniformineachcrosssection.
3. Theliquidflowisassumedtobeuniformineachcrosssectionaccountingfortheeffectof
filmthickening.
4. Thevapourflowisalsouniformineachcrosssection.
Regardingtheparticularisationtothestratifiedflowcasetwoadditionalfeaturesofthe
modelshouldbementioned.
5. Liquidfilmthickeningimpliesanaxialcomponentofhydrostaticpressure.Thiseffectis
accounted for ina way which resembles thatin open channel hydraulics although here
thevapourpressurecouldchangealongtheduct(Rufer&Kezios(1966)[151]).
6. The condensate formed in the film flows down the duct wall and joins the bulk liquid
flowing along the bottom of the duct. Vapour contacts the wall where the condensate
leavesit.
343
ECSSEHB3101Part13A
5December2011
An additional simplifying assumption must be introduced in connection with the fluid
thermodynamicandtransportproperties.
7. Thewalltemperature,Tw,iskeptconstant.Ontheotherhand,thepressurelossisvery
smallcomparedtothesaturationpressurelevel.Thenthefluidsaturationpressureand
temperature are assumed to be constant along the duct for calculating fluid
thermodynamicandtransportproperties.
Figure1416summarisestheflowgeometry.Althoughparticularisedtothestratifiedflowcase,itis
sufficientlygeneralforthepresentpurposes.
Figure1416:Geometryusedinthemodelofstratifiedflowcondensation.
Theconservationequationsinthecontrolvolume,whichenclosesliquid,L,andvapour,G,are:
Masspreservationequation
d
mL mG 0 [1494]
dx
Momentumbalanceequation
d
dx
dp
L ALVL2 G AGVG2 AFL
dx
[1495]
L g H z
d dAL
dz L wL G wG
dx AL dz
wherethelefthandsidegivesthechangeinmomentum.Thefirsttermintherighthandsideisthe
resultant of the static pressure on the fluid (liquid and vapour) per unit axial length of the control
volume. AFL is the internal crosssectional area of the duct. The second term is the resultant of
hydrostatic pressures acting on planes x and x+dx which enclose different liquid areas. This term
disappearsintheannularflowcase.Thelasttwotermsarethewallfrictionoftheliquidandvapour,
respectively, per unit axial length, here w are average shear stresses at the wall. L and G are the
correspondingwettedperimeters.
Thermalenergyequation
hm Tsat Tw A hfg
dmG
[1496]
dx
344
ECSSEHB3101Part13A
5December2011
Thisequationmerelyindicatesthattheheattransfertothewall,theareaofwhichisA= Dperunit
length,equalstheenthalpychangefromsaturatedvapourtocondensate.hmisanaverageheattransfer
coefficient. Tsat is the saturated vapour temperature, Tw the wall temperature, and hfg the corrected
valueoftheheatofcondensation.
The film temperature is nonuniform and slightly below saturation temperature because of partial
subcoolingoftheliquid,then
hfg h fg 0,68c pL Tsat Tw [1497]
wherecpListheliquidspecificheat.
14.4.1.2 Static pressure loss

We will relate the static pressure loss to a friction factor via a mixture approximation correlation,
whichisstrictlyvalidwhenaccelerationtermscanbeneglected.
Dukler,Wicks&Cleveland(1964)[39]suggesteddifferentapproximationsofthetwophaseflowby
thesinglephaseflowofaliquidofaveragepropertiesrunningfullintheductofdiameterD.Inany
casethepressurelossisrelatedtothefrictionfactorbyEq.[1428],clause14.3.1.4.
dp 32m 2 f TP
[1498]
dx TP 2 D 5
where fTP is a function of the mixture Reynolds number of the Poiseuille or Blasius type, see clause
14.3.1.4,andthesubscriptTPindicatestwophase.Thetwophaseflowisconsideredaslaminarwhen
ReTPislessthan2400andturbulentotherwise(asitoccursinthesinglephaseflow,clause7.2.2).The
mixtureReynoldsnumber,ReTP,isdefinedas:
4m
ReTP G Re [1499]
DTP TP
whereasTP,TParedefinedas
TP 1 L G [14100]
TP 1 L G [14101]
beingthecrosssectionalvoidfraction.
The equations for the calculation of TP and TP are those of SPECIAL CASE I of Dukler, Wicks &
Cleveland (1964) [39]. They correspond to the noslip restriction, VL/VG = 1. This restriction is very
difficulttoappraisebutitgivesthecorrectbehaviourinthetwoextremecases,liquidonlyandvapour
only.
Notethat,ingeneral,ReTPaswellasfTPwilldependon.
345
ECSSEHB3101Part13A
5December2011
14.4.1.3 Friction terms
InordertoevaluatethefrictiontermsintherighthandsideofEq.[1495]wewillintroducetheusual
frictionfactorsatthewallfortheliquidandvapour.
1 1
wL LVL2 f L , wG GVG2 f G [14102]
2 2
wherefLandfGwilldependonReLandReGrespectively.
Theaverageliquidandgasvelocitiesare:
mL 1 w m
vL
L AL 1 L AFL
[14103]
mg m w
vG
G AG G AFL
wbeingthevapourquality.
TheReynoldsnumbersbasedontherespectivehydraulicdiameters,DEL=4AL/LandDEG=4AG/G,
canbewrittenas
G
ReL 1 w Re , ReG w Re [14104]
L L G
Finally,theexpressionsofthewallfrictionperunitductlengthare,respectively:
1 1 w m2 f L
2
L wL L
2 1 L AFL
2
[14105]
2
1 w m2 fG
G wG G
2 G AFL
2
Now Re, and then f, will depend both on the geometry of the cross sectional liquidvapour
configuration,throughL/orG/,andonthevapourquality,w.ItisconvenienttoexpressfLandfG
as:

f L L , w f L L ,0 1 w L
L

[14106]
G G LG
fG , w fG ,1 w

whereL=1whenthecorrespondingflowislaminarorL=0,25whenitisturbulent.
346
ECSSEHB3101Part13A
5December2011
14.4.1.4 Momentum equation
BringingEqs.[1498][14106]toEq.[1495]wereach:

1 w2 L w2

d 1 G
D
dx dAL
H z dz dz
1

FrAFL D AL
[14107]
L 1 w
2 L
L
L

2 fTP 2 f L L ,0
TP 1 2

w2 L G

2 L G f G G ,1
G 2

whereFrisaFroudenumberdefinedas:

m2
Fr 2 2 [14108]
L AFL Dg

Severalterms,whichdependontheparticularmodelunderconsideration,aregiveninTable142.

Table142:GeometryandFlowDependentTermsinEqs.[14107]and[14109]
TYPEOFFLOW STRATIFIED ANNULAR
Configuration
sin 2 2
2

1
2 D
L/ / 1
G/ 1/ 0
L 1 L 1
f L L ,0 16 16
G Re G Re
347
ECSSEHB3101Part13A
5December2011
TYPEOFFLOW STRATIFIED ANNULAR
0 , 25 0 , 25
1 1
0,079 L 0,079 L
G Re G Re
1 0
fG G ,1 16 1
Re
0 , 25 0
1
0,079 1
Re
LL 1 1
0,25 0,25
LG 1
0,25
1
H z dAL dz 1 2 3 sin 2 0

FrAFL D AL dz
2Fr 3
sin cos

2
4 Nu L 1/ 4
4 2 G L 1 w 4 Nu
2
1
PrReSf G 1 PrReL Sf
3Fr L G Pr 3 Re 2 Sf 3
G
(since ReL 1 w Re )
L
3/ 4
4
0
1/ 3
sin d
3
CommentsonNu hmhasbeenexpressedbymeansofNusselts Nu=NuSL/(1)

resultfortheoutsideofhorizontaltubes 1 fL
Nu SL PrReL
(Ruffer&Keizos(1966)[151],Butterworth
(1981)[17],p.295).
T 2
,
Eq.[14129]
14.4.1.5 Dimensionless energy equation

Equation[1496]indimensionlessformbecomes:

dw 4 Nu L
D [14109]
dx PrReSf G

whereNuistheNusseltnumber:
348
ECSSEHB3101Part13A
5December2011
hm D
Nu [14110]
kL
PrtheliquidPrandtlnumber:
c pL L
Pr [14111]
kL
ReistheReynoldsnumberasinclause14.2.1(basedonthegasviscosity)withAFL=D2/4:
4m
Re [14112]
DG
andSftheStefannumber:
hfg
Sf
c pL Tsat Tw
[14113]
AmongthesefourdimensionlessgroupsonlytheNusseltnumber,Nu,dependsonthegeometryand
onthemodelused.SeeTable142.
14.4.2 Variation of the vapor quality along the duct in the stratified
model
EliminationofthegeometricvariablexfromEqs.[14107]and[14109]yieldsafirstorderdifferential
equationrelatingtow.Thisdifferentialequation,withtheboundarycondition
w=1,=1
hasbeennumericallyintegratedintypicalcases.Numericalvaluesforthecomputationsaregivenin
Table143andTable144andtheresultsaresummarisedinFigure1417.
Table143:CondensationinDucts.TypicalFluidProperties.
Assumedvalues:Tsat=300K,TsatTw=10K
Fluid (G/L)x103 L/G Pr Sf
Freon11 5,30 43,6 4,60 21,7
Freon12 30,1 20,6 3,54 15,0
Freon22 39,4 14,4 3,47 15,1
Water 0,03 93,7 5,81 59,1
349
ECSSEHB3101Part13A
5December2011
Table144:CondensationinDucts.ParametersDependingonmandD.
Assumedvalues:mhfg=1,5x103W,D=16,1x103m.Stratifiedcase
Fluid mx103 Frx103 Rex103 4 Nu / a L NTDb

103
[kg.s1] PrReSf G
Freon11 8,29 4,93 60,7 3,68 0,309
Freon12 10,2 10,1 67,6 3,19 0,352
Freon22 8,29 7,51 40,7 3,74 0,713
Water 0,615 0,058 5,34 21,9 3,62

3/ 4
4
a a sin1/ 3 d , a (0) 2,53 .
3 0
b NTDistheTaitel&Duklerdimensionlessgroupwhichwillappearinclause14.4.3.
Figure1417:Liquidfraction,1,vs.vaporquality,w.Forstratifiedcondensing
flowofseveralliquidalonghorizontalducts.Calculatedbythecompiler.
The model is no longer valid in the neighbourhood of w = 0 where the void fraction, , does not
vanish. Arriving at this situation implies large pressure variations which conflict with simplifying
assumption7inclause14.4.1.Theeffectispurelylocalandhasnotbeeninvestigatedfurther.
Oncearelationbetweenandwhasbeenobtained,integrationofEq.[14109]with
x=0,w=1
givestheevolutionofthevapourquality,w,alongtheduct.ResultsaregiveninFigure1418.
350
ECSSEHB3101Part13A
5December2011
Figure1418:Vaporquality,w,vs.dimensionlessdistancealongtheduct,x/D,for
stratifiedcondensingflowforseveralliquidsalonghorizontalducts.Calculated
bythecompiler.
14.4.3 Limits of validity of the stratified model

Theabovestratifiedflowmodelisbasedontheassumptionthattheliquidvapourinterfaceisstable.
ThismeansthatinanycrosssectionmG/AFLvs.mL/AFLmustbeenclosedinthatdomainofafiguresuch
asFigure142,clause14.1,whichcorrespondstostratifiedflow.
The limits of the stratified flow region have been expressed in analytical form by Taitel & Dukler
(1976) [174]. The expressions given by these authors are more convenient for the present purposes
thanaflowpatternmap.
Theinterfacestabilityrequires,inourvariables,
w2 1 w
NTD [14114]
2 1
where
G G L 1
NTD 400 1 [14115]
L L G ReFr
isadimensionlessgroupalreadytabulatedinTable144,clause14.4.2,forseveraltypicalcases.
Curves of constant values of the parameter w2(1 w)/2(1 ) in the w vs. (1 ) plane have been
representedinFigure1418.
351
ECSSEHB3101Part13A
5December2011
Figure1419:TheTaitelandDuklerlimitforstratifiedflowinthewvs.(1)
plane.Numbersonthecurvesrepresentconstantvaluesoftheparameter
w2(1w)/2(1)whichappearsinthelefthandsideofEq.[14114].
14.4.4 Annular flow model

Now we will introduce a similar approach for condensation with annular flow. The mathematical
formulationisquiteparalleltothatinthepreviousclauses.Nowthegeometricalparameteristhefilm
thickness,.
Theconservationequationsarenow:
Masspreservationequation
d
mL mG 0 [14116]
dx
Momentumbalanceequation
d
dx

dp
L ALVL2 G AGVG2 AFL D wL
dx
[14117]
ThephysicalmeaningofthedifferenttermsinEq.[14117]isthefollowing:
The left hand side is the change in momentum. The first term in the right hand side is the static
pressureterm,andthelastthefrictionterm.
The thermal energy equation is again Eq. [1496] although the heat transfer mechanism is different
here.
Theaverageheattransfercoefficienthmcanbeexpressedas(Butterworth(1981)[17])
352
ECSSEHB3101Part13A
5December2011
hm hSL L [14118]
wherehSListheheattransfercoefficientwhichwouldexistiftheflowrateofliquidwereflowingalone
throughthetube.ListheLockhartMartinellipressurelossmultiplier(clauses14.3.1and14.3.1.5).
WrittenindimensionlessformEq.[14117]becomes:
d 1 w L w2
2
D
dx 1 G
1 w
2 L
L L

2 f TP 2 L f L L ,0 [14119]
TP 1 2

w 2 L
G

2 L G f G G ,1
G 2

Equation [14109] is still valid as thermal energy equation in dimensionless form, although in the
annularflowcase(recallEqs.[1435]and[14118],
Nu SL
Nu [14120]
1
whereNuSGistheNusseltnumberiftheliquidmassflowratewereflowingalonethroughthetube.
Thissimpleexpressionbehavescorrectlyatbothendsoftheprocess.When /D<<1thedefinitionof
theliquidNusseltnumberintermsofthehydraulicdiameteroftheliquidannuluswouldbe:
hm 4
Nu SL [14121]
kL
then
Nu SL
Nu [14122]
4 / D
whichagreeswiththepreviousexpressionofNuwhen/D<<1.
Theagreementattheotherendisobvious,when0,NuNuSG.
14.4.4.1 Heat transfer coefficient in annular flow

Differentcorrelationsareavailableforcalculatingtheheattransfercoefficienthminsidehorizontalor
slightly inclined tubes in the annular flow regime. It would be interesting to compare them before
selectingthemostappropriateforourpurpose.
1)Ductaverageheattransfercoefficient,hm,
353
ECSSEHB3101Part13A
5December2011
Boyko&Kruzhilin(1967)[16]onthebasisoftheirextensivedatawith1,2x106Pato9x106Pasteam
over a liquid Reynolds number range of 6 x 103 to 3 x 105 suggest the following correlation for the
averageNusseltnumber, Nu
0,024
0 ,8 1 l / G
Nu Pr 0 , 43
Re [14123]
1 w0 ,8 L

2

Thiscorrelationisalsorecommendedforotherthansteamwatersystemsinthesame(liquid)Prandtl
numberrange(0,9to3).
2)CorrelationsforhSL.
ThesecorrelationsarejustifiedonthebasisofEqs.[14118]and[14120]andhavebeenusedwidelyin
liquidflowheattransfercalculations.
2,1)Kosky&Staub(1971)[108]givethelocalNusseltnumberintheliquidalonecaseintermsofthe
frictionvelocity
u* w / [14124]
and in terms of the dimensionless temperature function T+(+). Taking into account Eq. [143] the
frictionvelocity,u*,becomes:
D dp
u* [14125]
4 L dx
whereasthedimensionlesstemperature,T+(+),isdefinedas:
L u *c pL
T

hm
[14126]
+=u*/isadimensionlessfilmthickness.
This correlation is based on the Martinellianalogybetween turbulent momentum and heat transfer,
whichhasbeenquotedinclause8.2.4.
T+(+)isgivenbythefollowingsequenceofequations:

Pr , for 5

T
5 Pr ln 1 Pr 1 , for 30 5 [14127]
5

5 Pr ln1 5Pr 0,495 ln , for 30

30
354
ECSSEHB3101Part13A
5December2011
andthedimensionlessfilmthickness,+,isgivenby
ReL
, for ReL 1000
2

[14128]
0,0504 Re 7 / 8 for ReL 1000
L ,
TheliquidalonelocalNusseltnumber,NuSG(x),canbewrittenas
Nu SL x PrReL
1 fL
T

2
[14129]
wheretheliquidalonefrictionfactoris(clause14.3.1.4)
16
, for ReL 1000
f L ReL [14130]
0,079 Re 0, 25 for 1000 ReL 10 5
L ,
2,2)Correlationintroducedinclause6.
Whentheliquidfilmmotionislaminar(ReL<1000)
NuSL=3,66,seeFigure61,clause6.3.1.1.
Thethermalentrylength,i.e.thedistancerequiredtoreachthefullydevelopedvalue,NuSL=3,66,isof
theorderof50diametersand,thence,ithasbeenneglected(seeFigure66,clause6.3.1.1).
The Nusselt numbers calculated on the basis of values given in Table 144, clause 14.4.2, which
correspondtothestratifiedcase,arenearlytwoordersofmagnitudelargerthanthosewhichappear
intheannularlaminarflowcase.
Whentheliquidfilmmotionisturbulent(ReL>1000)itisadvisabletowritedowntherighthandside
ofEq.[14109]intermsofReL=(1w)(G/L)Re,sinceNuSL/PrReLisgiveninFigure614,clause6.3.1.1.
Thisfigurehasbeendrawnbyuseofthefollowingmathematicalformulae(Petukhov&Roizen(1975)
[143]):
Nu SL x L / 8

[14131]
PrReL k 12,7 L / 8 Pr 2 / 3 1
with
1
L
1,82 log10 ReL 1,642 [14132]
and
355
ECSSEHB3101Part13A
5December2011
900 0,63
k 1,07 [14133]
ReL 1 10 Pr
fortherangeofnumbers4x103ReL<6x105and0,7Pr5x105.
DifferentcurvesoftheliquidaloneNusseltnumber,NuSG,givenbytheabovecorrelationsareshown
asfunctionsofReLfortypicalvaluesofPrand L/Gand,whenrequired,of(1 )/(1w)0,8inFigure
1420.
Figure1420:LiquidaloneNusseltnumber,NuSL,vs.liquidReynoldsnumber,ReL,
asgivenbydifferentcorrelationsintypicalcases.Calculatedbythecompiler.
It can be seen in the figure that Eqs. [14129] and [14131] give almost identical results except in a
region of intermediate liquid Reynolds numbers, where Kosky & Staubs correlation (Eq. [14129]
furnishes a smoother transition than correlation [14131]. Thus, Eq. [14129] will be used in the
computationswhichfollow.
14.4.5 Variation of the vapor quality along the duct in the annular
model
Proceeding as in clause 14.4.2 we reach a first order differential equation which relates to w. This
equationshouldbeintegratedwiththeboundaryconditionw=1, =1toyieldthecurvesofFigure
1421.
356
ECSSEHB3101Part13A
5December2011
Figure1421:Liquidfraction,1,asafunctionofvaporquality,w,forannular
flowofseveralliquidalongducts.Calculatedbythecompiler.
A difficulty appears, however, near the origin (w = = 1), where the liquid is still absent and the
vapourwetsthewallsoftheduct.Itisclearthatbothliquidandvapourfrictionterms,whichappear
intherighthandsideofEq.[14119],shouldbetakenintoaccount.Sincetheonsetofcondensationis
notincludedinthemodel,theproblemsassociatedwiththetransitionof G/from1to0areeluded
startingthenumericalintegrationwith G/=1, L/=0andthenswitchingto G/=0, L/=1
withcontinuousd/dwderivative.
Oncearelationbetween andwhasbeenobtained,integrationofEq.[14119]withx=0,w=1,and
thesamesharptransitionbetweenonlyvapourandonlyliquidwettingthewall,givestheevolution
ofthevapourquality,w,alongtheduct.
TypicalresultsaregiveninFigure1422.AcomparisonoftheresultsinFigure1418andFigure1422
clearly indicates the large enhancing effect of gravity on condensation. This effect results to be
considerablylargerthanthatpreviouslypredicted(Keshock&Sadeghipour(1983)[104]).
357
ECSSEHB3101Part13A
5December2011
Figure1422:Vaporquality,w,asafunctionofdimensionlessdistancealongthe
duct,x/D,forannularflowofseveralliquidsalongducts.Calculatedbythe
compiler.
358
ECSSEHB3101Part13A
5December2011
15
Two-phase thermal transport systems
15.1 General
This clause is devoted to the description of twophase thermal transport systems which could be
typical of a Thermal Management System (TMS) for the Space Station. Singlephase loops are also
considered,butonlyascurrentstatesystemsforcomparisonpurposes.
15.1.1 Evolution of thermal transport systems

Comparedwithpresentstateofthearttechnology,representedbytheSpaceTransportationSystem
(STS), advanced TMSs should reject 10 times more heat to space (up to 100 kW) , transport heat 10
times farther (up to 50 m) maintaining isothermal heat load interfaces, should require low electrical
power, have a larger system flexibility and multiyear mission duration (up to 10 years with
maintenance).
Twophase heat transport systems are well suited to meet these requirements. One is tempted,
however,tostartthisclausebyexploringtowhichextentandatwhichexpenseinmassandpumping
power the TMS design requirements can beachieved usingavailable STS technology before dealing
withmoreadvancedconfigurationsbasedonpumpedtwophaseflowsystems,heatpipenetworksor
mechanicallyaugmentedheatpipesystems.
Changing trend during the last 10 years can be seen in Table 151 where the SPACELAB Active
ThermalControlSystem(ATCS)requirementsarecomparedtothoseofatypicallaboratoryofaspace
station.
Table151:EvolutionofRequirements
SpacelabATCSa SpaceStationLaboratoryb
HeatRejectioninOrbit[kW] 8,5(Average) 30
12.(Peak)
GrowthCapability[percent] 25
Isothermality[K] 2885 2392,5
Crew 3 8(formorethan30d)
MissionTime 7d 10yr(withmaintenance)
a FromOwen,Sessions&Walker(1976)[139],Valerani,DegliEsposti&Sessions(1978)[181].
b FromSadunas,Lehtinen&Parish(1985)[153].
359
ECSSEHB3101Part13A
5December2011
TheATCSwasdevisedtotransfertheheatloadsgeneratedwithintheModuleandonthePalletby
SPACELABsubsystems,experimentsandmetabolictotheOrbiterheatrejectionsystem.Asketchof
theATCSloopisgiveninFigure151.
Figure151:SchematicofATCSfluidloopintheModulePalletmode.From
Owen,Sessions&Walker(1976)[139].
TheTMSrepresentativeloopusedinthiscomparisonisdescribedwithsomedetailinclause15.2.
15.1.2 Two-phase loop general layout

The general layout of a singlephase loop has been presented in clause 13.2. Here we intend to
introducethemaincomponentsofatwophaseloopthermalcontrolsystem.
A twophase loop basically consists of fluidcarrying ducts connecting heat sources to sinks, and a
pump to overcome the pressure losses. The heat sources interface the loop through cold plates
(evaporators)andtheheatsinksthroughcondensers(heatexchangersorradiators).
Thefluidlinescouldcarryliquidonly,vapouronlyofliquidvapourtwophasemixtures.
The loop also comprises accumulators for accommodating changes of fluid volume due to
temperaturechanges.
Fluid temperature can be controlled via a reservoir where fluid is kept under saturated pressure
temperatureconditions.Eitherthepressurecanbeincreasedbyconnectiontoapressuretankorthe
temperaturechangedbythermostaticallycontrolledheatersorbyradiationtospace.Additionallythe
pressurelevelcanbechangedbyapressureregulatingvalveplaceddownstreamoftheevaporators.
Controls (sensors, control logic, actuators, valves, diverters, phase separators, ...) are required to
providetheappropriatemassflowrateateachmomentwithvapourqualitieswithinasuitablerange.
Three general types of loops are possible, namely: parallel circuit, series circuit and hybrid or
series/parallelcircuit,seeFigure152,eachonewithitsownadvantagesanddrawbacksasindicated
inTable152.
360
ECSSEHB3101Part13A
5December2011
Figure152:Schematicofthreedifferenttypesoftwophaseflowloops.a)Parallel
circuit.b)Seriescircuit.c)Series/parallelcircuit.
Table152:TwoPhaseLoopLinePatterns
Circuit Description Advantages Drawbacks
Parallel Separated liquid Complete separation Aliquidflowcontrol

(Figure only and vaporonly between the liquid device must be
152a) filledlines. andvaporflows. provided at each
evaporator.
Liquid flow rate to No gdependent
the evaporator is flow through the Modules cannot
controlled so that it duct provided that work both as heat
completely the right flow rate is sources or heat sinks
evaporatesthere simulated. unlessareturnliquid
lineisadded.
Moderate impact of
one evaporator on Liquid flow control
the operation of can be achieved by
another. capillary pumping.
Seeclause15.6
Extendable and
modular.
Series Saturated twophase No flow control The pressure drop

(Figure fluid flows through devicesarerequired. depends on the flow
152b) inseries pattern and thus on
Extremesimplicity.
evaporators. gravity.
361
ECSSEHB3101Part13A
5December2011
Circuit Description Advantages Drawbacks
Heat exchange All modules can be Optimal flow

within the modules used for either patterns should be
involves evaporation heating or cooling, sought.
or condensation and withtheexceptionof
Questionable Earth
thus changes in the first module
simulation.
vaporquality. whenfedwithliquid
only. Strong mutual
influence between
consecutive
modules.
Reduced growth
capability.
Hybrid Module bypass Sameasaboveatthe Temperature

(Figure provided to decrease cost of greater operated flow valves
152c) the mass flow rate complexity. are required to
through the adjust the fluid flow
Reduced pressure
evaporators of the rate to the heat load
drop in the
seriescircuit. on each evaporator.
evaporators.
Thisaddscomplexity
A phase separator
tothesystem.
increases the liquid
concentration of the
fluid entering the
evaporator
NOTE PreparedbythecompileraftervanOost&Mathieu(1983).
Figure153showsabreakawayviewofa25kWtechnologydemonstratorthermalbus.
Figure153:Physicallayoutof25kWtechnologydemonstrator(thermalbus).From
Carlisle/Nolan(1987)[18].
362
ECSSEHB3101Part13A
5December2011
15.1.3 About the nomenclature of this clause

Thefollowingnomenclatureisused:
1. ThermalManagementSystem.ItisthewholeactivethermalcontrolsystemoftheSpace
Station. It consists of 1) Interfacing devices; 2) Heat transport subsystem; and 3) Heat
rejectionsubsystem.Veryoften(notinthisclause)itisknownasThermalUtility.
2. Singlephasesystem.Itreferstoapumpedliquidloopthetemperaturegradientofwhich
(betweencoldplateandheatrejector)iszeroornegative.
3. Twophase system. This heat transport system is often known as thermal bus. It can
accept heat loads of various magnitudes and locations without affecting the whole
system.
15.2 Tms trade-off study

The study baseline Space Platform Configuration is presented in Figure 154. It consists of the
unpressurised experiment truss and resource (power) module and five pressurised (and manned)
modules: Habitat, Logistics, Multiple Berthing Adaptor (MBA) and Laboratories No. 1 and 2. Solar
arraysgeneratetheelectricalpower,withregenerativefuelcellsprovidingpoweronthedarksideof
theorbit.Thecurrentfullyequippedplatformhas75kWelectricalpoweratthebus.Themoduleheat
rejectionloadsaregiveninTable153,andtheTMSdesignrequirementsinTable1114.
Figure154:Spaceplatformconfiguration.FromSadunas,Lehtinen&Parish(1985)
[153].
363
ECSSEHB3101Part13A
5December2011
Table153:PowerDissipationbyModule
Module HeatSource Sunside(kW) Darkside(kW) OrbitalAverage
(kW)
Resource Equipment 13,41 11,46 12,68

FuelCell 0,0 46,96 15,65
MBA Equipment 6,0 6,0 6,0

Metabolica 0,295 0,295 0,295
Logistics Equipment 1,0 1,0 1,0

Metabolica 0,295 0,295 0,295
Habitat Equipment 5,0 5,0 5,0

Metabolica 0,59 0,59 0,59
Laboratory1 Equipment 10,0 10,0 10,0

Laboratory2 Equipment 30,0 30,0 30,0

Experiment Equipment 21,0 21,0 21,0

Truss
Subtotal Equipment 86,41 84,46 85,68

FuelCell 0,0 46,96 15,65
TOTAL 88,77 133,78 103,69

a Representtypicaldistributiontomeetsafehavenrequirement.Eachpressurizedmoduleissizedfor2,36kW.
NOTE FromSadunas,Lehtinen&Parish(1985)[153].
364
ECSSEHB3101Part13A
5December2011
Table154:TMSDesignRequirements
Parameter Requirement
HeatRejection 100150kW
GrowthCapability 25%
Metabolic Equipment FuelCell
HeatSourceTemperature[K] 277 293 343
ColdPlateTemperatureRange[K] 277288 290,5295,5 343
MeteoroidprotectionNASATMX82478 1%probabilityofpenetration
Safety,pressurizedmodules Notoxicorflammablefluids
Safehavenrequirement Crewofeightforminimumof30days
LoadSharingCapability 20%
TMSpowerrequirement(Target) 1%ofheatrejectionload
Systemlife 10yearswithmaintenance
Singlepointfailure Designtoprecludesinglepointfailure;nominal
levelheatrejectionwithanycomponentfailedor
subsysteminactivatedformaintenance
Thefollowingassumptionsareconsideredinthisstudy:
OrbitalEnvironmentLoads:Orbitalaverageenvironmentloadsareusedtosizetheradiators.
Centralised Radiator Orientation: The fixed radiators are assumed to be oriented in the YZ plane
(Figure154),whichcorrespondstotheminimumorbitalaverageenvironmentload.
ThermalStorage:PCcapacitors(seeECSSEHB3101Part10)areincludedinthethermaltransport
systems. Orbital average loads are used for radiator sizing. The benefits and characteristics of PC
capacitorsthermalstorageasappliedtoTMSdesignsareexaminedinclause15.3.1.
Thermal Coating Degradation: A 10 year endoflife (EOL) sof 0,3 is assumed, based on s = 0,1 at
beginningoflife (BOL) and a degradation rate of 0,02 per year. Also, a constant emittance, = 0,78
overthe10yeardesignlifeperiodisassumed.
SpaceTemperature:Theradiatingsurfacesareassumedtoradiateto0Kspace.
MetabolicHeatLoad:Ametabolicheatloadof295Wpercrewmanisused.Thiscorrespondstoan
activestate.
PumpPowerEfficiency:Anoverallefficiencyforelectricaltomechanicalpumpingpowerof20%is
used.
Theobjectivenowistoperformananalysisofdifferentthermalmanagementoptions.Theapproach
takenistodefineastudybaselineTMSwhichrepresentsthecurrentstateofthearttechnology,and
365
ECSSEHB3101Part13A
5December2011
toanalyzeandcomparethisbaselinewithalternateconceptswhichreflectincrementalimprovements
tothebaselinesystem.
15.2.1 TMS study baseline

ThestudybaselineschematicispresentedinFigure155.Thesystemconsistsofacentralizedpumped
liquid (Freon 21) loop transport with a centralized twostage radiator. The radiator provides 277 K
coolant flow to the metabolic heat exchangers and 290,5 K coolant flow to the equipment heat
exchangers.Partofthefluid,beforereturningtothemainradiator,passesthroughthefuelcellheat
exchangerraisingthefluidtemperatureofthatflowbranchto343K.
Figure155:Studybaselinecentralizedliquidloopexternalbus.FromSadunas,
Lehtinen&Parish(1985)[153].
Due to the pressurised module safety requirement, all of the Freon lines and heat exchangers are
located outside the pressure vessel. The intermodule fluid transfer is affected through fluid line
connections in the docking ports. The external bus thermal transport system interfaces with the
internalwaterloopsthroughheatexchangers.
The external thermal bus is driven by a single redundant. pump package located in the resource
module.Asmallauxiliarypumpisincludedinthefuelcellbranchtominimizetotalpumpingpower.
15.2.2 TMS design concepts

ThedesignoptionsconsideredarelistedinTable155wheretheglobalcharacteristicsofthedifferent
designconceptsareincluded(ConceptIisthebaselinesystem).
366
ECSSEHB3101Part13A
5December2011
Table155:TMSDesignConcepts
DesignOption DesignConcept
I II III IV V
Heat InterfaceHeatExchanger x x x x x
Acquisition
Heat Centralized x x
Transport
Decentralized(3loops) x x x
PumpedLiquid Freon21 x
Thermal
Transport Ammonia x x
TwophaseFluid(Ammonia) x x
ThermalTransport
Heat Radiator HeatPipe x x x x

Rejection
PumpedLiquid x
Loop
Fixed x x x x
Steerable x
ThermalStorage x x x x
NOTE PreparedbythecompilerafterSadunas,Lehtinen&parish(1985)[153].
Thechangesfromoneconcepttothenextareaimedatimprovingtheoverallperformance.Thedesign
goalsofeachconceptarepresentedinTable156.
Table156:TMSDesignGoals
DESIGN DESIGNGOALS
CONCEPT
I Currentstateofthearttechnology
II Toreducepumpingpowerandoverallsystemmass
III Toenhancemodulethermalautonomy,safehavenprovisioncapabilityand
compatibilitywiththephasedspaceplatformsbuildupinspace
IV Toreducepumpingpowerandoverallsystemweight,andtoenhance
isothermalitycapability
V Toreduceoverallsystemmassandtoeliminatetheneedforthermalstorage
NOTE PreparedbythecompilerafterSadunas,Lehtinen&Parish(1985)[153].
367
ECSSEHB3101Part13A
5December2011
ConceptIIisidenticaltothebaselinesystemexceptthattheliquidloopradiatorsarereplacedbyheat
piperadiators,andtheworkingfluidisAmmoniainsteadofFreon21.
Concept II consists of a decentralised pumped liquid loop system. The schematic of this system is
presented in Figure 156. This concept has the following features: pressurised modules (except the
Multiple Berthing Adaptor (MBA)) utilising body mounted radiators for metabolic heat rejection;
MBA module metabolic heat rejection affected through a resource module deployed radiator;
centralisedexternalliquidloopwithdeployedradiator;andautonomousfuelcellpumpedliquidloop
andheatpiperadiator.
Figure156:ConceptIII.Decentralizedliquidloop.FromSadunas,Lehtinen&
Parish(1985)[153].
ConceptIVisidenticaltoConceptIIIexceptthatthepumpedliquidAmmoniatransportisreplaced
withatwophaseAmmoniatransport.TheschematicofthissystemispresentedinFigure157,where
it can be seen that the thermal storage is now integral with the internal and external bus heat
exchangersandcentralbuscoldplates.
Figure157:ConceptIV.Decentralizedtwophasetransport.FromSadunas,
Lehtinen&Parish(1985)[153].
368
ECSSEHB3101Part13A
5December2011
ConceptVisidenticaltoConceptIVexceptthatthefixedequipmentradiatorhasbeenreplacedbya
steerable radiator. Because of the added complexity of steerable radiators (see clause 15.5.3), the
benefits are considered insufficient to replace the fixed fuel celland MBA modules metabolic
radiators.
TheinternalwaterloopiscommontoalltheTMSconceptsinvestigated.
15.2.3 Evaluation of tms concepts

The relative merits and performance of each TMS design concept are assessed on the following
criteria: 1) system mass; 2) power requirement; 3) reliability; 4) modularity; 5) growth capability; 6)
isothermalitypenalty;7)safehavenprovision.
Table 157 presents the evaluation of the five TMS design concepts defined above. A pictorial mass
summaryoftheseconceptsispresentedinFigure158.
369
ECSSEHB3101Part13A
5December2011
Table157:EvaluationofConcepts
CRITERION CONCEPTI CONCEPTII CONCEPTIII CONCEPTIV CONCEPTV
Systemmass,M M=10900kg M=6900kg M=8440kg M=7830kg M=5690kg

RadiatorMass:6000kg Radiatormass:2500kg Decentralizedsystems Ammoniamass:18 Centralbusradiator
ThermalStorageMass:990 RadiatorHeatExchanger weighmore. kg massreductionof1110
kg mass:770kg Radiatormass:3310kg Reductionof91,6% kgfromConceptIV
Freonmass:940kg ThermalStorageMass:1060 RadiatorHeatExchanger ThermalStorage
kg mass:920kg reductionof1040kg
Ammoniamass:210kg ThermalStorageMass: fromConceptIV
1720kg
Ammoniamass:210kg
Power P=4,2kW0,04Q P=1,7kW0,016Q P=1,7kW0,016Q P=0,01kW104Q

Requirement,P 3%abovedesigngoal 60%reductionfrom Samepowerrequirement Greatpumpingpowerreduction
ConceptI asincentralizedsystems
Reliability Moderate. High.Heatpiperadiators High. LowerthaninIV.

Themostprobablefailure aremuchlesssensitivethan EnhancedascomparedtoConceptII. Thesteerableradiator
causecouldbethe pumpedliquidloop Intermodulefluidconnectionsarevalvesare addscomplexitytothe
intermoduleandradiator radiatorstometeoroid eliminated. system.
fluidconnectors. penetration. Probabilityofleakageisdiminished(lowerbus Steeringmechanismand
Leakdetectionandflow pressure). sensors,leakdetectors,
bypassesarerequiredat valvesandcontrolsto
eachconnector. bypassarerequired.
Modularity Feasible. Feasible.

Modularityrequirementsdictatedtheprovisionof EnhancedascomparedtoConceptII.
additionalfluidflowpathsandconnectorsatthe Decentralizedsystemsenhancemodulethermalautonomy.
modules,andredundantpumpingcapacity.
370
ECSSEHB3101Part13A
5December2011
CRITERION CONCEPTI CONCEPTII CONCEPTIII CONCEPTIV CONCEPTV
Growth Feasible. Metabolicsubsystem:Limited:

Capability Provisionsforgrowthcapabilityrequireoversizedfluid Limitedto1,75kWbytheBodyMountedradiatorarea.
linesandredundantpumpingcapacity. Equipmentsubsystem:Feasible:
Systemgrowthcanbeaccommodatedinthedesignbyincreasingflowrates
andpumpingpower,orbyoversizingthefluidlines.
Isothermality 540kg0,05M. 227kg0,033M. 227kg0,027M. Nil.

Penalty The5KIsothermality 58%reductionfrom Isothemalsystem.
requirementaffects ConceptI
transportsystemmass.
Massofradiatorheat
exchangerandcoldplate
remainsunchanged.
Safehaven CrewofEight. CrewofEight.

Provision Thesafehavenprovisionismetthroughredundant EnhancedascomparedtoConceptII,duetotheautonomousnatureofthe
distributedmetabolicheatrejectioncapabilitybetween metabolicloopofeachmodule.However,the1,75kWcapacityisprobably
modulesandredundantintermodulefluidconnections. adequateonlyforshorttermsafehavenuse.
NOTE PreparedbythecompilerafterSadunas,Lehtinen&Parish(1985)[153].
371
ECSSEHB3101Part13A
5December2011
Figure158:Tradestudymasssummary.FromSadunas,Lehtinen&Parish(1985)
[153].
15.3 Design for orbital average load

The TMS design is complicated by a 10year minimum life requirement, minimum onorbit
maintenance, mass constraints, and the fact that the system will operate in low earth orbit which
reducestheeffectivenessoftheradiators.Inordertooptimisetheoverallsystemdesign,withrespect
tomassandpowerrequirements,theTMSissizedfororbitalaverageheatrejectionloadsandorbital
average environments. This generates the requirement for incorporation of thermal storage (TS)
withintheTMS.
The benefits of TS in the context of TMS optimisation and as a means of compensating for radiator
optical coating degradation (an idea already discussed in ECSSEHB3101 Part 10 clause 8.5) were
analysedbyLehtinen&Sadunas(1985)[114];theirresultswillbesummarizedinclause11.3.1.)
TS isimplemented in the TMS by means of phase change capacitors (seeECSSEHB3101 Part 10),
usuallyinseriesintheloop.TheparticularconfigurationofinseriesPCcellshasnotbeendisclosed,
butforthetimebeingtheycouldbeenvisagedasashellandtubeheatexchanger,Figure114,where
theshellfluidisthePCmaterialandthetubefluidthatflowingthroughtheloop.
In the following clause the characteristics of PC capacitors for both single and twophase transport
systemswillbediscussed.
15.3.1 Phase change capacitor performance

Twodifferentsystemswereconsideredfortheanalysisofthecharacteristicsandspecificationsofthe
requiredPCcapacitors.
System 1. Decentralised pumped liquid loop system. This system corresponds to Concept III
(clause15.2.2).Figure159ashowsaschematicofanidealisedpumpedliquidloopsystemwith
TS.
372
ECSSEHB3101Part13A
5December2011
System 2. Decentralised twophase transport system. This system corresponds to Concept IV
(clause 15.2.2). The schematic of an idealised twophase system with TS is shown in Figure
159bThephasechangecapacitorisintegralwiththeheatsource,thus,forthisarrangementthe
heat source continuouslydumps heat into the capacitor. The system will have a varyingflow
ratewhichiscontrolledbytheradiatorheatrejectioncapacity;this,inturn,isafunctionofthe
radiatorenvironmentalheatfluxhistoryshowninFigure1510.
Figure159:DecentralizedsystemswithTS.a)Pumpedliquidloop.b)Twophase
transport.FromLehtinen&Sadunas(1985)[114].
Figure1510:Radiatorabsorbedenvironmentalheatflux,qe,vs.dimensionless
time,t/tp.Orbitalperiod,tp=90min.Radiatorlimit:Blackbodyemissivepower.
FromLehtinen&Sadunas(1985)[114].
ThecharacteristicsofsingleandtwophaseTMSswithTSaregiveninTable158.
373
ECSSEHB3101Part13A
5December2011
Table158:CharacteristicsofSinglePhaseandTwoPhaseTMSswithPC
Capacitor.
Characteristic SinglePhaseSystem TwoPhaseSystem
Modeofoperation ControlsHXflowinlettemperatureto Capacitoractsasanintermediate

prescribedvaluewithvaryingradiator storageforsourceheatrejectionwith
outlettemperatures.Capacitoraddsor heatbeingcontinuouslyremoved
removesthermalenergyfromthefluid fromthecapacitorbyevaporationof
streamwithvaryingloads. thefluidatthecapacitorHX.
Capacitor Downstreamofradiator. Integralwithheatsource.

Location
MassFlow Constant,determinedbythespecified Varyingwithandcontrolledbythenet

ThroughHX heatexchangerinletandoutlet instantaneousheatrejectioncapacity
temperaturesandbythemaximum oftheradiator.
heatrejectionload.
System ConstantatcapacitorHXoutlet, Essentiallyconstanttemperature

Temperature varyingatcapacitorHXinlet(radiator throughoutthesystemandradiator
outlet). withvaryingloadandenvironment.a
PCTemperature PCmaterialwithphasechange PCmaterialwithphasechange

temperatureequaltodesiredHXinlet temperatureequaltothenominal
temperature. selectedsystemtemperature.
a Whenoperatingatthedesignheatrejectionloadandenvironmentcondition.
NOTE FromLehtinen&Sadunas(1985)[114].
InordertodeterminetheTScharacteristics,transientanalyseswereperformed.Theprincipalareasof
interestare:
1. Thermalcapacitorenergystoragerate,and
2. Thermalcapacitorstoredenergyhistory.
The equations used by Lehtinen & Sadunas (1985) [114] are summarised in Table 159. Figure 1511
shows the performance of the PC capacitor in three different subsystems, namely: Metabolic
Equipment and Fuel Cell. These results show no substantial differences between single and two
phasesystems.
374
ECSSEHB3101Part13A
5December2011
Table159:PCCapacitorPerformance
SinglePhaseSystem TwoPhaseSystem
RadiatorArea,AR[m2] Q Q
AR AR
E R qe E R qe
Q:HeatRejectionRate[W] Q :AverageHeatRejectionRate
[W]
qe :AverageEnvironmentalHeatFlux[W.m2]
ER:RadiatorEmissivePower[W.m2]
ER=2TR4
:StefanBoltzmannConstant[W.m2.K4]
:RadiatorEmittance
:RadiatingEffectivenessoftheradiator
TR:RadiatorTemperature[K]
To5 Ti 5 TR=TM
TR4
5To Ti
TM:CapacitorPhaseChange
Temperature[K]
To:AverageRadiatorOutlet
Temperature[K]
Ti:RadiatorInletTemperature[K]
MassFlowRate,m[kg.s1] Qmax AR ER qe t
m mt
c p Ti TM

h fg
Qmax:MaximumHeatRejection qe(t):EnvironmentalHeatFlux
Load[W] [W.m2]
cp:SpecificHeatoftheWorking hfg:HeatofVaporization[J.kg1]
Fluid[J.kg1.K1]
ThermalCapacitorEnergy Qc(t):mcp{To(t)TM} Qc(t)=Q(t)m(t)hfg

StorageRateQc(t)[W] To(t):RadiatorOutletTemperature Q(t):HeatRejectionLoad[W]
[K]
ThermalCapacitorStored t
EnergyJc(t)[J] J c t Qc t dt
o
NOTE PreparedbytheCompilerafterLehtinen&Sadunas(1985)[114]
375
ECSSEHB3101Part13A
5December2011
Figure1511:ThermalperformanceofPCcapacitorsindifferentsubsystems.a)
Metabolic.b)Equipment.c)Fuelcell.Thermalperformanceisgivenintermsof:1)
energystoragerate,Qc;and2)netstoredenergy,Jc.FromLehtinen&Sadunas
(1985)[114].
ThespecificationsofTMSwithPCcapacitorsaresummarisedinTable1510.
376
ECSSEHB3101Part13A
5December2011
Table1510:SinglePhaseandTwoPhaseTMSCapacitorSpecifications
SinglePhase TwoPhase
Parameter Metabolic Equipment FuelCell Total Metabolic Equipment FuelCell Total
AverageHeatLoad 2,36 88,24 21,92 112,52 2,36 88,24 21,92 112,52

Q x10 [W]
3
AverageEnvironment 306,7 306,7 306,7 306,7 306,7 306,7

qe [W.m2]
InletTemperature 277 290,5 340,5 279,5 293 343

Ti[K]
OutletTemperature 282 295,5 345,5 279,5 293 343

To[K]
PhaseChange 277 290,5 340,5 279,5 293 343

TemperatureTM[K]
RadiatorArea 18,8 410,0 33,3 462,1 15,3 409,4 32,6 457,3

AR[m2]
Flowrate 103,6 3869,7 2083,6 6056,9 1,9 74,7 20,8 97,4

mx103[kg.s1]
EnergyStorageRate 3,71 84,08 17,88 105,67 2,36 88,24 18,89 109,49

Qcx103[W]
StoredEnergya 7,16 161,93 36,86 205,95 5,69 174,85 38,95 219,49

Jcx106[J]
a TabulatedresultsarelargerthanthosegiveninFigure1511.
NOTE FromLehtinen&Sadunas(1985)[114].
377
ECSSEHB3101Part13A
5December2011
Acomparisonofspecificstoragecapacity,Jc/Qc,forsingleandtwophasetransportsystemsisshown
inFigure1512.Itcanbeseenthatthesinglephasesystemhasslightlyhigherstoragerequirementat
low temperatures. The reason for this behaviour, which is modeldependent, can be understood as
follows:whenthesystemoperatingtemperatureislow,theheatrejectiontospaceceasesforawhile
(see shadowed zones in Figure 1510). For the twophase system, once the liquid is fully vaporised,
and provided that the sensible heat of thevapour is neglected, the heat transfer to the thermal
capacitor will reach a top plateau, Figure 1511a. For the singlephase system, on the contrary, the
amountofheattobetransferredtothecapacitoriscontrolledbythesensibleheatoftheliquidinthe
loop(Table159)andexceedsthatcorrespondingtothetwophasesystem.
Figure1512:Requiredspecificstoragecapacity.Jc/Qc,oftypicalTMSsvs.
temperature,T.FromLehtinen&Sadunas(1985)[114].
15.4 Off-design operation

InthestudyofThermalManagementSystems,sizingisusuallymadeformaximumheatrejectionand
worstcaseradiatorenvironment,whichoccursneartheendofdesignlifebecauseofradiatoroptical
coatingdegradation.TheTMS,then,willberequiredtooperateforaconsiderablelengthoftimewith
reduced heat rejection loads, that is, at offdesign conditions. In this clause, the temperature control
duringoffDesignoperationisanalysed.
Oneofthethreesubsystemsstudiedinclause15.3.1,namelythefuelcellloop,isselectedhere.Asin
thatclause,twodifferentthermaltransportsystemsareconsidered:
System1.Pumpedliquidloop(ConceptIIIofclause15.2.2).
System2.Twophasetransportsystems(ConceptIVofclause15.2.2).
ThesketchesofthesetwosystemsareshowninFigure1513.
378
ECSSEHB3101Part13A
5December2011
Figure1513:SchematicofthefuelcellloopTMS.a)Pumpedliquidloop.b)Two
phasetransport.FromSadunas,Lehtinen&Parish(1986)[153].
The selected fuel cell thermal control loop has a peak waste heat rejection load of 47 kW. Since the
allowablefuelcellinlettemperaturerangeisquitebroad,anoptimumfuelcellheatexchangerinlet
temperature is determined by minimising radiator, pumping and wet plumbing weight. This
optimum HX inlet temperature (322 K) also represents the thermal capacitor phase change
temperatureinthesinglephasesystem.Inthetwophasesystemthephasechangetemperatureisthe
maximumHXoutlettemperature(343K).
Asmentionedinthepreviousclause,aPhaseChangecapacitorhasbeenincorporatedinordertosize
theradiatorfororbitalaverageheatrejectionandexternalenvironment(insteadofusingpeakvalues).
In the twophase transport system the PC capacitor (integral with the heat source) is in thermal
contact with the fluid through an evaporator. Subcooled liquid is vaporised in the evaporator and
recondensedintheradiatorcondenser.Theliquidissubcooledtopreventboilingintheaccumulator
andcavitationinthepump(seeFigure1513b)Theliquidflowratetotheevaporatoriscontrolledbya
flow metering valve. Vapour pressure(and boiling temperature) in the evaporator can be controlled
eitherbymeansofamechanicallydrivenaccumulator,orbyapressureregulatingvalvedownstream
oftheevaporator(thissolutionreduces,oreliminates,thecyclingoftheaccumulator).
Asummaryoffunctionalrequirementsofthefuelcellloopandradiatorenvironmentsispresentedin
Table1511.
379
ECSSEHB3101Part13A
5December2011
Table1511:FuelCellLoopDesignRequirements
Parameter Value
Orbitalcharacteristics range52a
Period94,2min
FuelCellCycle(Darkside) 35,6minat=0
27,5minat=52
FuelCellHeatRejection 47kWpeak
17,7kWaverageat=0b
12,5kWaverageat=52
RadiatorSolarAbsorptance s=0,1BOL
s=0,3EOL
Radiator DesignCase 301W.m2averageat=0withs=0,3b

Environment
ColdestOffDesignCase 144W.m2averageat=52withs=0,1
a Hereis90minustheanglebetweenthelinetothesunandthenormaltoorbitplane(seethelistof
symbolsofECSSEHB3101Part15)
b Thesevaluesslightlydifferfrom21,9kWand306,7W.m2giveninTable1510,clause15.3.1.
NOTE FromSadunas,Lehtinen,Nguyen&Parish(1986)[154].
15.4.1 Temperature control

Under offdesign conditions an excess radiator area will exist and, in the absence of control, the PC
material will solidify and subcool causing the heat exchange inlet temperature to fall below the
minimumdesign(322K).Schemesforsystemtemperaturecontrolwillbeanalysedinthefollowing.
15.4.1.1 Pumped liquid loop system

Thecontrolschemeshouldmaintainthefuelcellheatexchangerinlettemperatureascloseaspossible
to322K.Twooptionalschemesareavailable.
1. Respondtodemand scheme. This scheme is shown in Figure 1514. The fuel cell heat
exchangerinlettemperature,T1,iscontrolledbymixingpartofthefluidexitingthefuel
cell heat exchanger at temperature T2 with the fluid exiting the TS heat exchanger at
temperatureT3.Becauseoftheexcesscoolingcapacity,thePCmaterialremainssolidand
thetemperatureT3belowthephasechangetemperature.
380
ECSSEHB3101Part13A
5December2011
Figure1514:RespondtodemandtemperaturecontrolschemeFromSadunas,
Lehtinen,Nguyen&Parish(1986)[154].
IfxmisthemassflowratethroughtheradiatorandPCcapacitor,T1willbegivenbythe
followingequation
T1 1 x T2 xT3 [151]

whichgivesxonceT1,T2andT3areknown.
Themaindrawbacksofthisschemeare:
o Bypassvalveiscontinuouslyoperating.
o Radiatortemperatureexperimentssharptimechanges.
o PCcapacitorisinoperative.
2. Orbitalaveragescheme.Themassflowratethroughtheradiatorheatexchangerissetto
theorbitalaverageconditions.SeeFigure1515foraschematicofthissystem.Theheat
exchanger inlet temperature is now constant, equal to the constant TS temperature (the
PCmaterialisnowinatwophasestate).

Figure1515:Orbitalaveragetemperaturecontrolscheme.FromSadunas,
Lehtinen,Nguyen&Parish(1986)[154].
Therequiredradiatorflowfraction,x,wouldbegivenby
1 x T2 xTR TM [152]
381
ECSSEHB3101Part13A
5December2011
where TM is the phasechange temperature. The fuel cell heat exchanger average outlet
temperature, T2 , and the radiator average temperature, TR , can be obtained by
expressingthatbothsubsystemsareabletomanagetheaverageheatrejectionload.
InthisschemethefluidflowratethroughbothradiatorandPCcapacitorheatexchanger
arekeptconstant,andthetemperatureexcursionsoftheradiatoraresmallprovidedthat
thefuelcelloperatingconditionschangeslowlywithtime.
Table1512presentscomputedvaluesofxforarangeofoffdesignoperatingconditions.
Alsogiveninthistablearetheorbitalaverageradiatortemperatures, TR .
Table1512:OrbitalAverageSchema.ValuesofxforOffDesignOperation
s qe Q x103 x TR
[deg] [W.m2] [W] [K]
0 0,1 162 17,7 0,36 308

0 0,3 301 17,7 1,00 322
52 0,1 144 12,5 0,14 287
52 0,3 247 12,5 0,21 301
NOTE FromSadunas,Lehtinen,Nguyen&Parish(1986)[154].
Itshouldbenotedthatforthecoldestoperatingcase(=52, s=0,1)only14%oftheflowpasses
throughtheradiatorheatexchanger.
15.4.1.2 Two-phase transport system

Themodeofoperationduringpredictableoffdesignconditionsofatwophasetransportsystem(see
Figure1513b)isthesameasduringoperationatthedesignpoint,exceptthat,inthepresentcase,TRis
settoalowervaluebasedonorbitalaverageheatrejectionload, Q ,andenvironmentalheatflux, qe .
The radiator operating temperatures for a range of offdesign operating conditions are presented in
Table1513.Forthecoldestoperatingcasetheradiatortemperatureresultstobe299Kcomparedto
335Kwhenoperatingatthedesignpointconditions.
Table1513:TwoPhaseSystem.RadiatorTemperaturesforOffDesignOperation
s qe Q x103 TR
[deg] [W.m2] [W] [K]
0 0,1 162 17,7 322

0 0,3 301 17,7 335
52 0,1 144 12,5 299
52 0,3 247 12,5 311
NOTE AssumedRadiatorArea,AR=27,3m2.
FromSadunas,Lehtinen,Nguyen&Parish(1986)[154].
382
ECSSEHB3101Part13A
5December2011
15.4.2 Instrumentation requirements

Three schemes for system temperature control under offdesign conditions have been considered in
clause 15.4.1. The respondtodemand and the orbitalaverage schemes are used for pumped liquid
loop systems( clause 15.4.1.1), and a variant of the orbitalaverage scheme for twophase transport
systems(clause15.4.1.2).
TheinstrumentationrequirementsforthesethreecontrolschemesaresummarisedinFigure1516.
Figure1516:Instrumentationofthedifferentcontrolschemes.a)Respondto
demandpumpedliquidloopscheme.b)Orbitalaveragepumpedliquidloop
scheme.c)Twophasetransportloopscheme.FromSadunas,Lehtinen,Nguyen&
Parish(1986)[154].
The orbitalaverage and the twophase system control schemes are rather complicated, and since
they operate on an orbital average basis, they will be sensitive to cumulative measurement errors,
requiringprovisionforperiodiccorrections.
For the respondtodemand control scheme, on the other hand, instrumentation requirements are
much simpler. In addition, since this system operates on an instantaneousresponse approach it is
insensitivetocumulativemeasurementerrors.
383
ECSSEHB3101Part13A
5December2011
Theorbitalaveragecontrolschemeisthepreferredapproachfromthethermalpointofview(lower
temperatureexcursionsandfulluseofthePCcapacitor),despitethenecessityforalargernumberof
measurementsandmorecomplicatedcontrolalgorithms.Therespondtodemandschemecouldbe
incorporatedasabackupoftheorbitalaveragecontrolschemebecauseofitssimplicity.
15.5 Radiator-loop interaction

TheradiatoristhemostmassivecomponentofaTMS.Itisalsothemostexposedportionofthespace
station.
Table1514givesthecomponentmassofatypicalpumpedliquidloopsystemfordifferentvaluesof
theheatrejectionrate,Q.TheratioMR/M,ofradiatortosystemmassisshowninthelastcolumn.This
ratioincreasesasQincreases.
Table1514:ComponentMassofaTypicalPumpedLoopTMS
Qx103 Mass,M[kg] MR/M
PowerSystem Radiator Pump Othera Total
1 5,9 42 0,91 7,2 56 0,75

2 9,1 84 1,80 13,1 108 0,78
5 15 210 3,60 31 260 0,81
10 22 420 6,30 62 510 0,82
20 32 840 10,4 128 1010 0,83
50 54 2100 21 320 2500 0,84
100 80 4210 35 630 4960 0,85
a Includescoldplateandwetlines.
NOTE FromDexter&Haskin(1984)[34].
Minimising the size of the radiators would save mass and cost, and would simplify design, vehicle
attitude control and stability, maintenance, payload space sensor viewing, resupply vehicle docking
andmicrometeoroidprotection.
AlthoughradiatorshavebeendealtwithinECSSEHB3101Part9,wewilldiscussinthefollowingp
clauses several waysfor decreasing radiator area and mass, since these ways will affect the overall
configurationoftheTMS.)
384
ECSSEHB3101Part13A
5December2011
15.5.1 Boosting radiator temperature with a heat pump

TherangeofapplicabilityofpumpedliquidloopsandactiveheatpumpsforlongtermTMSshasbeen
studiedbyDexter&Haskin(1984)[34].TheirmainideaissummarisedinFigure1517.
Figure1517:Evolutionoftheheatrejectionrateandconfigurationofspacecraft
thermalcontrolsystems.FromDexter&Haskin(1984)[34].
Pumpedliquidactivethermalcontrolsystemsareusefulinthe3x103Wto30x103Wheatrejection
raterange,abovewhichheatpumpaugmentedsystemscouldbeemployed.
The heat pump raises the temperature of a circulating fluid through compression of the evaporated
workingfluid.Theworkingprincipleisthesameasforarefrigerator,Figure1518.
Figure1518:Schematicoftheheatpumpworkingprinciple.Adaptedbythe
compilerafterRye&Steen(1986)[152].
Thefluidcirculatesthroughaclosesystem.Whenheatfromthesourceistransferredtothefluid,in
liquid phase and at low temperature and pressure, the liquid is evaporated. Then, vapour pressure
and temperature are raised at the compressor. The vapour passes to a condenser where heat is
rejected.Theresultingliquidpassesthroughathrottlevalvewhereitregainstheinitialtemperature
andpressurebeforeinitiatingthecycle.
Theideaisnotnew.Itisbeingincreasinglyusedforindustrialanddomesticheating.SeeHeap(1979)
[81] for a recent account of heat pump performance and applications. A heat pump for Spacelab
385
ECSSEHB3101Part13A
5December2011
applicationinthe15W20Wheatrejectionraterangehasbeenconsidered(Berner&Savage(1981)
[11]).
Inthepresentinstancetheheatpumpaimsatboostingradiatortemperature,thencereducingradiator
areaandmass.
TherelevantstepsinthetradeoffperformedbyDexter&Haskin(1984)[34]aresummarisedinTable
1515.RepresentativeresultsaregiveninFigure1519toFigure1522.Theseresultsarelessstriking
thanthosepresentedbythementionedauthorsbecauseofonemainreason:theyassumeafairlylarge
equivalent temperature of the surrounding space, Te = 283 K, which apparently would be
representative of unfavourable low altitude orbits. This means that a very modest increase in the
radiator equilibrium temperature, TR, will sharply increase the factor TR4Te4 of the radiative term in
theradiatorthermalbalanceequation.InthepresentcaseitisassumedthatTe=0.
Table1515:HeatPumpAugmentedvs.ClassicalFluidLoopTradeOff
HeatPumpAugmentedFluid ClassicalFluidLoop
Loop
BasicAssumptions LoopTemperature,Tc=277K LoopTemperature,Tc=277K

RadiatorEquilibrium Workingfluid:Freon21,See
Temperature,TR clauses15.5and15.5.1.
T=11KacrosstheinterfaceHX
p=1,4x105Pa.Pressureloss
throughtheloop
RadiatorMass,MR[kg] P
MR 1 Q
sTR4 Q
:RadiatorAreaDensity.=5kg.m2
:5,6697x108W.m2K4.StefanBoltzmannConstant
:0,82.RadiatorEmittance
:0,80.RadiatingEffectivenessoftheRadiator
P:Powersuppliedtotheheatpumportothepump[W]
Q:Heatload[W]
P TR Tc TR TC

Q cTc P 1 1
p
Tc Q p c p T

TR Tc Seeclause15.3
p=0,25.OverallEfficiencyofthe
CarnotCoefficientof
pump.cpandarerespectivelythe
Performance
specificheatanddensityofthe
c=9,948[1,8(TR273)+32]0,612
workingfluid.
FractionofCarnotEfficiency
SolarPanelMass,MSP[kg] P 0,562
M SP 4,67Q 1 1
0 , 562
(chargeabletothefluid
driver) Q
Thisexpressioncorrespondstoaconservativeestimateoftheoverall
efficiencyofaphotovoltaicsystem.
386
ECSSEHB3101Part13A
5December2011
HeatPumpAugmentedFluid ClassicalFluidLoop
Loop
FluidDriverMass,MP[kg] MP=20,68{2+9,4x103P5/6 0 , 75
Q
(nx104)1,25 M P 2 0,80 10
3
c T
p
Factor2hasbeenintroducedtoaccountforredundancy.
ColdPlateMass,MCP[kg] MCP=1,87x103Q
(Wet)PlumbingMass,MWL MWL=(1/10)(MR+MP+MCP)
[kg]
TotalSystemMass,M[kg] M=MR+MSP+MP+MCP+MWL
NOTE PreparedbytheCompilerafterDexter&Haskin(1984)[34].
Figure1519:Radiatormass,MR,offourtypicalsystems(seetextabove)for
differentvaluesoftheheatrejectionrate,Q.Calculatedbythecompiler.
387
ECSSEHB3101Part13A
5December2011
Figure1520:Massbreakdownofaheatpumpaugmentedsystemfordifferent
valuesoftheheatrejectionrate,Q.TR=300K.Calculatedbythecompiler.
388
ECSSEHB3101Part13A
5December2011
Figure1521:Massbreakdownofaheatpumpaugmentedsystemfordifferent
valuesoftheheatrejectionrate,Q.a)TR=340K.b)TR=380K.Calculatedbythe
compiler.
389
ECSSEHB3101Part13A
5December2011
Figure1522:Ratiooftotalmass,M,oftheheatpumpaugmentedsystemtothe
totalmass,Mr,ofthereferenceclassicalpumpedliquidloopvs.theinverseCarnot
coefficientofperformance[TRTc]/Tcfordifferentvaluesoftheheatrejectionrate,
Q.Calculatedbythecompiler.
SystemsconsideredinFigure1519toFigure1522are:
1. 1.Baselinesystem.AclassicalpumpedliquidloopwithameantemperatureclosetoTc=
277K.
2. 2.Aheatpumpaugmentedloopwithalowtemperaturesectionat277Kandaradiator
at temperature TR. Three different values of TR have been assumed in Figure 1519 to
Figure1521,namely:TR=300K,340Kand380K.
Resultingmasssavingsaresmall(ornil).OnlyforafairlyreducedTRrange,nearTR Tc,wherethe
pump efficiency is large, a mass reduction can be achieved (Figure 1522). In the high TR side the
energyrequiredtooperatetheratherinsufficientheatpumpis,atpresent,toocostlyintermsofadded
solar panel mass (see the increasing trend of the solar panel mass, MSP, in Figure 1520 and Figure
1521).
15.5.2 Thermal-storage assisted radiator

Ithasbeenalreadyshowninclause15.3thattwobasicTMSdesignoptionsexist.1)Designwithout
thermalstorage(TS)andsizetheradiatorforpeakheatrejectionloadsandpeakenvironmentsor2)
include TS and size the radiator for orbital average heat rejection and external environment. In any
caseradiatorsizinghastoaccountforsdegradationoverthedesignmissionfile.
Theradiatorspecificarea,AR/Q,isrelatedtotheincidentsolar,qs,Earthinfrared,qir,andEarthalbedo,
qa,fluxes,totheradiatoremissivepowerER,andtotheopticalpropertiessandby
390
ECSSEHB3101Part13A
5December2011
AR 1

Q ER s qs qa sqir
[153]
FortheTMSdesignwithoutTS,the(qs+qa)termrepresentsthepeakorbitalvalue.Forthedesignwith
TS,itrepresentstheorbitalaveragevalue.Inbothcases,theradiatorissizedforEOLs.Notethatthe
qirtermcanbeassumedtobeconstantbecausetheradiatorplaneremainsnormaltotheorbitalplane.
The effect of the s/ ratio on radiator specific area, for three different radiator temperatures is
presented in Figure 1523. These temperatures correspond to the metabolic radiator (279,5 K),
equipmentradiator(293K)andfuelcellradiator(343K)whichwereconsideredinclause15.3.1.
Figure1523:Effectofs/onradiatorspecificarea,AR/Q.FromLehtinen&
Sadunas(1985).
ThedatashowninFigure1523indicatethat:
SubstantialradiatorareasavingscanbeachievedbyincorporatingTSintheTMSdesign.
Theradiatoraregrowthrequirementwith sdegradationissubstantiallydecreasedfordesign
withTS.
Alimitings/valueexistforeachradiatortemperature.
In general, radiator area will be sized to meet EOL requirements, that is allowance is made for s
degradation.InthisstudyBOLandEOL s/valuesare0,12and0,38respectively.ThedatainFigure
1523 show that for the 279,5 K metabolic radiator, the EOL requirement cannot be met since the
limitingvalueofs/is0,3.Thisisalsotrueforthe293Kequipmentradiator,whichhasalimitings/
value of 0,38 (this corresponds to an infinitely large radiator). Therefore, for these two cases, the
designer has two options, design for shorter life (lower EOL s) with provision for coating
refurbishment, or design with TS. The 343 K fuel cell radiator data show that, at this higher
temperature, the radiator area reduction resulting from TS is very small for s/ ratios up to 0,5.
Becausethefuelcellloophasacyclicheatrejectionload,withthemaximumoccurringonthedarkside
of the orbit, a third option exists for this subsystem, namely, size the radiator on the peak darkside
heatrejectionloadwithoutTS.TheresultsabovearesummarisedinTable1516.
391
ECSSEHB3101Part13A
5December2011
Table1516:EOLRadiatorAreafor10yearLife
Subsystem Radiator RadiatorArea,AR[m2]
Temperature[K] WithoutTS WithTS
Metabolic 279,5 * 18,8

Equipment 293 * 410,0
FuelCell 343 56,7** 33,3
NOTE *RadiatortemperaturewillexceeddesignallowableatEOL.
**Sizedforpeakheatrejectionloadontheorbitdarkside.
FromLehtinen&Sadunas(1985).[114]
15.5.2.1 Coating degradation and radiator life

The preceding discussion showed that for low temperature radiator subsystems, EOL s values
dictated the incorporation of TS into the TMS system. Conversely, designs without TS and hence
lower EOL s values, require provision for periodic refurbishment of the radiator coatings. Figure
1524presentstheestimatedradiatorlife/refurbishmentfrequencyforTMSdesignswithandwithout
TS.Usingthesedata,radiatorareafor5yearand10yeardesignlife(orrefurbishmentfrequency)are
computedforthebaselineconfigurationandaresummarisedinTable1517.Thesedatashowthatthe
metabolicsubsystemTMSdesignwithoutTScanbeachievedifthedesign/refurbishmentfrequencyis
reducedto5years.Evenso,theradiatorarepenaltyissubstantial.ThenTSisincludedinthedesign,
10yearlife(orrefurbishmentfrequency)isachievedatamodestradiatorareapenaltyof6,7m2.For
allrefurbishmentfrequencies,considerableradiatorareasavingsareachievedwithTS.
Figure1524:Radiatorspecificarea,AR/Q,vs.refurbishmentfrequency,t.From
Lehtinen&Sadunas(1985)[114].
392
ECSSEHB3101Part13A
5December2011
Table1517:RadiatorAreavs.DesignLife
Subsystem RadiatorArea,AR[m2]
5yearlife 10yearlife
WithoutTS WithTS WithoutTS WithTS
Metabolic 41,7 12,1 * 18,8

Equipment 615,6 297,7 * 410,0
FuelCell 53,0 28,8 53,0 33,3
Total 710,3 338,6 462,1

NOTE *Designtemperatureexceeded.
FromLehtinen&Sadunas(1985)[114].
Similar conclusions are drawn for the equipment radiator with the exception that radiator area
refurbishmentfrequencypenaltiesarelowerthanthoseforthemetabolicradiator.
The fuel cell radiator data show that the radiator area increases very gradually with refurbishment
frequencyforthedesignwithTS,whichissizedfortheorbitalaverageloadandenvironment.Thatis,
therefurbishmentfrequencycanbeextendedatasmallcostofadditionalradiatorarea.Ontheother
hand,thedesignwithoutTSisindependentofrefurbishmentfrequency,sinceitisbasedontheorbital
darksideheatrejectionloadandenvironment(i.e.,independentofsdegradation).
Insummary,theseresultsdemonstratethatTSisessentialforlonglife,lowtemperaturedesignwhen
s has to be considered. For all cases, radiator area is reduced and coating refurbishment frequency
extendedbyincludingTS.
15.5.3 Steerable radiators

Considerable savings in radiator area can be achieved by steering the radiator to minimise incident
solar flux, earth albedo and earth IR flux. Also steerable radiators partially prevent coating
degradationbyavoidingsolarradiationimpingement.Thisreducedsurfacedegradationratereduces
therequiredradiatorareaprojectedforEOLperformanceandmaysignificantlyextendtheusefullife
oftheradiatorelements.Typicalvariationofsteerabletofixedradiatorspecificheatrejectionratiosis
presentedinFigure1525.
393
ECSSEHB3101Part13A
5December2011
Figure1525:Steerabletofixedradiatorspecificheatrejectionration,
(Q/AR)steer/(Q/AR),vs.radiatortemperature,TR.FromSadunas,Lehtinen&Parish
(1985)[154].
AcomparisonoffixedandsteerableradiatorareasfortheTMSunderstudy(seeConceptsIVandVof
clause 15.2.2) is shown in Table 1518. Also shown in the table are radiator areas corresponding to
peak and orbital average loads along with the amount of TS needed in the latter case. The 277 K
metabolicradiatorshowsthelargestrelativereductionwhensteered,butbecauseofitssmallarea,the
added complexity of steering would probably not be justified. In practical terms, only the 293 K
equipment radiator should be steered resulting in a net savings of about 180 m2 of radiator area
comparedtothefixedradiator.Sincethedifferencebetweenthepeakandorbitalaveragedesignareas
issmall(27,1m2forthe293Kradiator)andtheorbitalaveragedesignrequiresTS(61,2kgforthe293
K radiator), the trade is 27,1 m2 of radiator area versus 61,2 kg of thermal storage. Thus, steering
considerablereducestheradiatorareaandpracticallyeliminatestheneedforthermalstorage.
Table1518:ComparisonofFixedandSteerableRadiatorAreas
Radiator Heat RadiatorArea,AR[m2] TS
Temperature Rejection Fixeda Steerableb [kg]
TR[K] Q[kW] WithTS WithoutTS
277 2,36 21,4 8,0 10,5 2,3

293 88,24 409,9 222,1 249,2 61,2
348c 15,65 30,3 24,2 25,4 6,1
a YZplaneradiator.
b Zaxissteering.
c Thisvaluediffersfrom343KgiveninTable1514,clause15.2.
Finally,duringoperationatlowheatrejectionrates,externalheatingoftheradiationsystemcouldbe
beneficialtoavoidfreezingorcontractionoftheworkingfluidwithintheradiatorelements.Insuch
situations it would be convenient to steer the radiators toward a warmer thermal environment to
maintainanacceptableoperatingtemperature.
394
ECSSEHB3101Part13A
5December2011
The main subsystem of a steerable radiator to be discussed here is the rotary thermal coupling to
achievetheheattransferacrosstheradiatorinterface.
15.5.3.1 Rotary thermal couplings

Several concepts can be imagined. The selection of the best one is based on many different
considerations regarding temperature gradients, operational sensitivity or radiator configuration
(fluidlooporheatpipe).Thetwomainrequirementstobefulfilledare:highthermalconductanceand
smallsteeringtorque.Rotarythermalcouplingconceptscouldbeclassifiedintothreemaincategories:
1. Rotarycontactcoupling.Heattransferisachievedthroughthejointthermalconductance
ofmovablesolidparts.
2. Rewindmechanismswithflexiblehosesconnectingtheinboardandoutboardfluidlines.
3. Continuouslyrotatablefluidtransfermechanismsusingdynamicalseals.
TherequirementsofaCategory3thermalcouplingbeingdevelopedfortheSpaceStationarelistedin
Table1519.
Table1519:RequirementsforaRotaryCouplingOnboard.SpaceStation.
OperationalRequirements
360degreecontinuousrotation
Oneaxisofrotationpercoupling
Reversible,variablespeedrotation
Lowmaintenance
PhysicalRequirements
Ammoniacompatible
Thermal/Vacuumcompatible
Zeroleakagetospace
Accommodatetwofluidloopsatdifferenttemperatures
Redundancycapability
Onorbitreplaceable
MinimumupdatingrequiredwithSpaceStationgrowth
LifeandReliabilityRequirements
10yearlifewithoutreplacement
Indefinitelifewithreplacement
99%Reliability
NOTE FromHeizer,Goo,Rhodes,Thoreson&Parish(1986)[83].
Table1520presentsapreliminaryevaluationofjointconceptswithfocusonnearterm(Columbus)
applications.Heatrejectionloadis1,1kWpeak,operatingtemperaturerangefrom278Kto298Kand
lifetimeofatleast4years.Mostoftheconceptsaresuitedtodeploymentpurposesonly.Thestateof
theartlooksfarfromsatisfactory.
395
ECSSEHB3101Part13A
5December2011
Table1520:JointsforSteerableRadiators
Concept Fig. Description Advantages Drawbacks Evaluation
(1987)
Rotatable Figure1526 Radialflowheatpipe.Theevaporatorconsistsofa Goodthermal Appropriateleak Inconceptualstage.
Radial Figure1527 finegauzewichattachedtotheinnertube.The performance: tightendcaps Nonearterm
HeatPipe condenserisplacedradiallyoutwardsandseparated hc500W.K1.m1. required. application.
fromtheoutertubewallbyasmallclearance. Lowrotationtorque.
OscillatingHydrody Figure1528 Tworeservoirsatdifferenttemperaturesconnected Conductancecanbe Powerrequired. Nonearterm
namicThermalJoint viaa(flexible)capillarybundle.Verylargeaxialheat controlledbythefluid Fatiguelimits application.For
fluxesachievedwhenthefluidoscillatesaxiallyat oscillation. lifetime. deploymentonly.
highfrequencyandverylargeamplitudes. hcmax104W.K1.m1. Poorstateoftheart.
ArticulatedHeatPipe Figure1529 Asplittedheatpipewithasphericalorcylindrical Minimumtemperature Sealingproblems Unmaturestateofthe
Figure1530 jointandaflexiblearterialwick.Sealingis drop. limittheoperating art.
Figure1531 particularlydifficultinthehemispherical T=5K. pressureand
configuration.Inthecylindricalconfigurationthis temperature,and
problemisminimized,butonlyaxialrotationis requirecareful
possible. manufacture.
Hightorques.
FlexibleHeatPipe Figure1532 Rigidevaporatorandcondenserconnectedwitha Minimumtemperature Temperaturedrop Limitedsteerability.
Figure1533 flexiblehosewhichincorporatesaflexiblewick. drop.Available increaseswith Mainlyfor
Severalmodelsdevelopedwithworkingfluidssuch technology. increasedbendangle. deployment.
asammonia,acetoneorliquidmetalsandlengthsup Q.leff2x103W.m. Largeradiirequired.
to1m. Heatpipeextendstothe Veryhightorque.
radiator. Fatiguelimits
lifetime.
HPtoHPwithVacuum Figure1534 Afinnedheatexchangerthermallycouplestwo Minimumtorque. Inadequatethermal Unfeasible.
Interstice coaxial,independentlysealedheatpipes.The performance.
intersticeisopentospace.
HPtoHPwithLow Figure1534 Asabovebuttheintersticeisfilledwithlowpressure hc100W.K1.m1. Highmanufacturing Promising.
396
ECSSEHB3101Part13A
5December2011
Concept Fig. Description Advantages Drawbacks Evaluation

(1987)
PressureHeFilled (<7x103Pa)heliumgassealedwithinajoint. Acceptableperformance. cost.
Interstice Lowtorque. Sealingproblems.
HPtoHPwithLow Figure1534 Asabovebuttheinterseticeisfilledwithalow Moderatelyhighthermal Material Possiblefallback
MeltingPointAlloy meltingpointmetallicmaterial(typicallyNaK). transport. compatibility. concept.
Interstice Extraheatingisrequiredtomelttheinterstitial hc190W.K .m .
1 1 Freezing.Torque
material. uncertain.Longterm
materialmigration.
HPtoHPwithGrease Figure1534 Asabovebuttheintersticeisfilledwithagrease hc120W.K1.m1. Outgassing. Promising.
FilledInterstice Figure1535 (typicallyEcotherm).Extraheatingrequired. Acceptabletorquefrom Freezing.
0,8to2N.minthe250K
294Ktemperaturerange.
BraidedSolidConductor Figure1536 Flexiblearrangementoffoilsorinterwovenstrandsof hc50W.K1.m1. Limitedrotation. Fordeploymentonly.
thermalconductingmaterials.Itthermallycouples Simple,inexpensive, Hightorque.Fatigue
movingandfixedinterfacesallowingacertaindegree predictableperformance. limitslifetime.
ofrotation. Massive.
ClampedJointContact Figure1537 Thermalcontactbetweenaheatpipeandaconcentric Goodthermalcontact. Complexity Nonearterm
Conductor heatexchangerisenhancedbygasorhydraulic hc4x10 W.K .m at2x introducedbythegas application.Non
3 1 1
pressureagainstadiaphragm.Contactconductanceis106Papressure. orhydraulic steerabe.Originally

pressuredependent. pressurization developedforonorbit
system.Needsextra assembly.
retracting
mechanism.
RollableMembrane Figure1538 Flatarrayofjoinedflexibleheatpipesthatcanbe Smallmass. Poorstateoftheart. Nonearterm
Radiator deployedeitherpassivelybytheinternalpressureor Lowsensibilityto Fatiguelimits application.Mainlyfor
activelybyspringsorcords.Limitedsteerabilitycan meteoroiddamage. lifetime. deployment.
beachievedinthelattercase.
NOTE PreparedbythecompilerafterDelil(1987a)[31].
397
ECSSEHB3101Part13A
5December2011
Figure1526:Rotatableradialflowheatpipe.FromDelil(1986a)[33].
Figure1527:Thermaljointbasedonaradialflowheatpipe.FromHinderer&
Savage(1978)[88].
398
ECSSEHB3101Part13A
5December2011
Figure1528:Heattransferenhancementbyflowoscillationdemonstrator.From
Kurzweg&Zhao(1984)[113].
Figure1529:Hemisphericalheatpipejunction.FromDelil(1986a)[33].
Figure1530:Cylindricalheatpipejunction.FromDelil(1986a)[33].
399
ECSSEHB3101Part13A
5December2011
Figure1531:Heatpipeinheatpipejoint.Thisconceptisanoffspringofthe
cylindricalheatpipejoint.FromShaubach(1985)[165].
Figure1532:Felxibleheatpipeconsistingofa0,15mlongrigidevaporatoranda
0,20mlongrigidcondenser.FromDelil(1986)a[33].
Figure1533:Flexiblecooper/acetoneheatpipe.Itfeaturesa0,381mlongadiabatic
sectionbetweenanevaporatorandacondenserboth0,178mlong.Insidediameter
is0,016m.FromDelil(1986a)[33].
Figure1534:Finnedheatexchangers.Intersticeeithervacuumfilledorfilledwith
agas,alowmeltingpointoragrease.FromFrench(1985)[68].
400
ECSSEHB3101Part13A
5December2011
Figure1535:Greasefilledheatpipeheatpipejoint.Hingejointcoaxialwiththe
internalheatpipeandparalleltotheplaneoftheradiatorheatpipe.FromDelil
(1986a)[33].
Figure1536:Braidedconductorflexiblethermaljoint.FromDelil(1987a)[31].
401
ECSSEHB3101Part13A
5December2011
Figure1537:Clampedjointcontactconductor.Heatpipepenetratesheat
exchanger.Fluidpressureenhancedthermalcontact.FromEllis&Rankin(1983)
[42].
Figure1538:Selfdeployedmembraneheatpiperadiator.Bothdeployedand
undeployedconfigurationsshowninthefigure.FromDelil(1986a)[33].
15.5.3.2 Rotatable fluid transfer coupling

A continuously rotatable fluid transfer mechanism to be used on board Space Station has been
developed(Heizeretal.(1986)[83]).Operational,physicalandliferequirementsforthistypeofrotary
couplinghavebeensummarisedinTable1519.
The rotatable fluid transfer coupling consists, Figure 1539, of a housing 0,56 m long and 0,20 m in
diameter, made of 304 stainless steel (AISI designation: Felow C0,19 Cr0,10 Ni) and modular in
construction.
402
ECSSEHB3101Part13A
5December2011
Figure1539:Internaldetailsoftherotatablefluidtransfercoupling.Dimensions
areinmm.FromHeizer,Goo,Rhodes,Thoreson&Parish(1986)[83].
Four modules, each with a fluid inlet and outlet port, are bolted together. Two modules will
accommodatetworefrigerantloops(at275Kand294Krespectively)fromthemainTMS,leavingthe
remainingtwomodulesasbackup.Whenonemodulefailsthefluidisdivertedtoanunusedmodule
andthefailedsystemshutdown.
Theaftendoftherotarycouplingissealedwithanendcaptoallowonlyonepossiblepathtospace.
To prevent leakage at this point a ferrofluid low pressure (105 Pa) magnetic seal is used. For a
descriptionoftypicalferrofluidssealsseePerry(1978)[141].
The shaft is constructed in one piece of 174 PH stainless steel (commercial designation: Fe0,17 Cr
0,04Ni0,04Cu)witha12x104mmrmsfinish.
Theoverallsizeofthecouplingis0,86minlengthand0,36mindiameter.Overallmassis50kg.
Theestimatedrequiredtorqueis270N.m.
The housingis mounted on the stationary portion of the spacecraftstructure.The rotatingshaft has
mountingprovisionsthatallowattachmenttotherotatingportionoftheradiatorboom.
The vapour from the heat source will enter the larger of the two radial ports (Figure 1540a) of one
active module and will flow around an annulus (Figure 1540b) up to the inlet port of the rotating
shaft. The vapour will then pass through an axial tube (Figure 1539) out to the radiator heat
exchanger.
Figure1540:Rotatablefluidtransfercoupling.a)Generalviewshowingtheliquid
andvapourradialports.b)Channelgeometry.FromHeizer,Goo,Rhodes,
Thoreson&Parish(1986)[83].
403
ECSSEHB3101Part13A
5December2011
The condensate from the radiator will flow through the smaller diameter axial tube of the second
activemoduletowardthesmallerportoftherotatingshaft,andwillflowaroundtheannulusuntilit
encountersthesmallerdiameterouterport.Itthenreturnstothemainloop.Fluidductcrosssectional
diametersare0.016mfortheliquidlineand0.032mforthevapourlines.TheseductsizesareSpace
Stationcompatible.
Pressures losses through the ducts have been predicted at one position of the coupling shaft for a
rangeofheatrejectionrates,Q(Figure1541).
Figure1541:Pressuredrop,p,alongtheductatonepositionoftherotary
couplingfordifferentvaluesoftheheatrejectionrate,Q.FromHeizer,Goo,
Rhodes,Thoreson&Parish(1986)[83].
Thispressuredropwillaffecttheperformanceofboththevapourandliquidlegsofthefluidloops.In
addition, assuming that the vapour is supplied to the coupling in saturated condition, the pressure
losswillresultinalowersaturationtemperature,whichwillaffecttherequiredradiatorarea.Thisis
not the case in the present instance where the values quoted in the reference T 0,2 K) are really
small.
Asealleakagemanagementsubsystem(SLMS)isdesignedtomaintaintheinterstitialsealspaceat105
Pa,thelevelofpressuretheferrofluidsealcontains.
TheSLMSisdesignedtomonitorandmeasuresealleakage,removeleakagefluidfromtheinterstitial
seal space and store that leakage in a reservoir. It consists of (1) pressure taps to monitor the seal
status;(2)avalvesystemtocheckeachinterstitialspaceand(3)areservoirtocontaintheleakage.
15.5.4 Radiators coupling

Eastandwestradiatorpanelsofathreeaxisstabilisedgeostationarycommunicationssatellitecanbe
coupled,viapassivelycontrolledvariableconductanceheatpipes(VCHPs,seeECSSEHB3101Part
8, clause 8.1), with the aim of sharing the heat load being radiated to space. This can be illustrated
withreferencetoFigure1542aThespacecraftexperiencesarelative360diurnalrotationofthesolar
vector in the east/earth/west/antiearth plane. When the solar vector is at 0 the earth facing side is
directlyilluminated,whiletheeastandwestsidesareinshadow.Asthesolarvectorstartsrotatingto
the east the equivalent temperature of the space surrounding that radiator increases and that in the
westradiatorsimultaneouslydecreases.
404
ECSSEHB3101Part13A
5December2011
Figure1542:a)Schematicofathreeaxisstabilizedspacecraftingeosynchronous
orbit.FromChalmers&Pustay(1986).b)Typicalthermalloadsharingofeastwest
faces.FromWise(1986)[192].
Bothradiatorsarethermallycoupledtothepayloadplatformbutthetotalheatradiatedisdifferently
shared among them (Figure 1542b) because the VCHPs adjust to the changing environment. When
theheatrejectioncapabilityoftheeastradiatorisreduced,thesourcetemperatureincreasesandthe
heattransfercapabilityofwestVCHPsalsoincreases.
Thelimitingdesigncase(withinanorbitalperiod)occurswhentheeastsideisdirectlyilluminated(at
90)andmostoftheheatisrejectedthroughthewestradiator.
The remainder of the diurnal cycle produces a gradual reduction in the east radiator surrounding
temperature(between90and180)asthesolarvectormovestowardstheantiearthspacecraftside,
atwhichtime(180)theheatrejectionrateofbotheastandwestradiatorsisequalised.
Thehalfcycleisthenrepeated(between180and360),thistimewiththewestsideilluminatedand
theeastsideinshadow.
ThehotcasecorrespondstoEOLwintersolstice.Sufficientdesignmargin(typically25%inradiator
area)shouldbeallowedtoaccommodatethisconditions.
DesignoptimisationoftheVCHPinvolves:
1. Calculation of the volume ratio, i.e. reservoir volume under maximum operating
conditionsoverinactivecondenservolumeintheminimumoperatingconditions.
2. Determinationofreservoirandsystemmassforaspecifiedheatpipe.
3. Assessment of the diode response of the VCHP when the sink temperature rises above
thesourcetemperature.
Detailed thermal and systemlevel trade analyses based on a RCAs series 4000 communication
satellite with a generic payload indicate (Chalmers & Pustay (1986) [22]) that, compared with a
conventionalSeries4000satellite,thevariableeast/westconceptallowsa43%increasedheatrejection
ratewitha70%increaseinradiatorareawithinacostincreaseof100%.
Aschematicofaspacecraftincorporatingeat/westradiatorscoupledwithVCHPsisshowninFigure
1543 a. A second option based on Capillary Pumped Loop (CPL) systems (see clause 15.6) is
introducedinFigure1543b
405
ECSSEHB3101Part13A
5December2011
Figure1543:Eastwestradiatorcoupling.a)BasedonHPtechnology.From
Chalmers&Pustay(1986).b)BasedonCPLtechnology.FromChalmers,Pustay,
Moy&Kroliczek(1986)[23].
TwoindependentloopsareusedinthesecondconceptofFigure1543.Onelooptransportspayload
heattotheeastradiatorwhilethesecondlooptransportscomponentwasteheattothewestradiator.
The temperature of each loop is controlled by reservoirs where fluid is kept in saturated pressure
temperature conditions (see clause 15.6.3.4). Because both loops operate independently, a Sun
illuminated radiator can be effectively shut down while the opposite radiator, viewing deep space,
opensup.
Compared with HPs, CPL technology is particularly well suited to this application because of the
followingtworeasons:
4. CPL evaporator pumps can handle much higher radial heat fluxed than conventional
HPs (see clause 15.6.3.1). When component thermal dissipation increases, massive heat
spreadersarerequiredifHPsareused,andaheatfluxisachievedabovewhichtheCPL
ismoremasseffectivethantheHP.
5. Heat exchangers between payload panels and radiator panels are required with HP
systems.Theresultingtemperaturedropgreatlyreducestheeffectivenessoftheradiator.
Several important design issued need to be evaluated in detail before a CPL system can be
incorporated into a threeaxis stabilised satellite, namely, reservoir location, startup characteristics
andonegthermaltesting.
Forneartermspacecraftapplications,smallerCPLpumpsthanthosebeingdevelopedfortheSpace
Station(seeFigure1546)willbeverymasseffective.
15.6 Capillary pumped loop (cpl) technology

CPLsystemscanbeusedtotransfer,passively,heatloadsbelow10kWoverdistancesoftheorderof
10m.
TheCPLoperatingprinciplecanbeunderstoodonthebasisofFigure1544wherethesameschematic
loopisrepresentedintwodifferentsituations:atlowoperatingheatload,Figure1544a,andathigh
operatingheatload,Figure1544b.
406
ECSSEHB3101Part13A
5December2011
Figure1544:SchematicofCPLengineeringmodel.Ina)theliquidfillsmostofthe
loop,whereasinb)partofthisliquidhasbeenevaporated.FromChalmers,
Pustay,Moy&Kroliczek(1986)[23].
When the evaporatorpumps are heated, the liquid from the saturated wick evaporates and flows
toward the condenser where heat is removed. The condensed fluid flows again toward the
evaporatorpumpthroughasubcooledliquidreturnline.Thewickstructureintheevaporatorpump
providestherequiredpumping,bycapillaryactionforconcurrentflowofvapourandliquid.
Wickedisolators,alsosketchedinFigure1514,isolatelocalfailuresinindividualpumpspreventing
thedeprimeofasinglepumpfromspreadingtotheotherpumps.Theyalsorestrictthemovementof
anynoncondensablegasintothepumps.
NASA in house, NASA sponsored and privately sponsored efforts in the field are summarised in
Table1521.
407
ECSSEHB3101Part13A
5December2011
Table1521:NASAdevelopmenteffortsinCPLtechnology
EXPERIMENTOR MATERIALS FEATURES HEATLOAD TRANSPORT COMPLETION
PROGRAM QX103[W] DISTANCE DATE
L[m]
NASA/Lewis Stainlesssteel/water Feasibilitydemonstration. 1966
NASA/Goddard
GroundDemonstration Two0,3mpumps.Two1,8m
Eng.Model Aluminium/Freon conds.
0,35 10 Sept.1981
Eightpumps.Sixcondensers.

CPLIPrototype Isolators.
CPLIIPrototype Aluminium/Ammoniaa Asabove.Welded.Thermal 6,3 10 March1984
Aluminium/Ammoniaa vacuumtest.
7 10 July1985
NASA/Goddard
FlightExperiments
CPL/GASb Aluminium/Ammonia Two0,15mpumps.
0,2 1 June1985
CPL/Hitchhikerc Aluminium/Ammonia RefurbishedCPL/GAS
0,5 1 Jan.1986
NASA/GoddardDynatherm Two0,3mpumps. 3 10 1985
CPLEvaporator
NASA/GoddardRCAOAO Aluminium/Ammonia Fourcoldplateseach3x102m2. 3 July1986
PCCPd Six0,3mpumps.Figure
CAPLeFlight Aluminium/Ammonia 1547.
Experiment
Twocoldplateseachwith0,3m 1,5 Late1986
pumps.Fourcondensers.

10
OAO 0,6mpump. 4 Sept.1985
EvaporatorPump Indesigndefinitionphase
Late1986
HPSTMfCPL (1986).
408
ECSSEHB3101Part13A
5December2011
Initiated(1986).
CPLScaling Late1986
OAOIR&Dg Two0,3mpumps. 2x2,8 ~7m Jan.1985
CPLDirectCondensation 1x3,6m
Radiator radiator
a 105mPorexWickStructure.Porousplasticwick(highdensitypolyethylene).
b GAS:GetAwaySpecial.ExperimentflownonShuttleFlightSTS51GinJune1985.
c HitchhikerG.ExperimentflownonShuttleFlightSTS61CinJanuary1986.
d PCCP:PrototypeCapillaryColdPlates.
e CAPL:CapillaryPumpedHeatTransportLoop.Shuttlemountedflightexperiment.
f HPSTM:HighPowerSpacecraftThermalManagement.
g IR&D:InternalResearchandDevelopment.
NOTE FromChalmers,Pustay,Moy&Kroliczek(1986)[23].
409
ECSSEHB3101Part13A
5December2011
15.6.1 Advantages of cpl systems

CPL technology represents a potentially powerful tool for thermal control of high power spacecraft
whenhighcomponentreliabilityisrequired.Themorequotedadvantagesofthesesystemsare:
1. Passiveoperation.
2. Minimumpowerrequirements.Electricalheatingmayberequiredforstartupbyheating
one of the evaporatorpump and for thermal control of the reservoir (saturated
temperaturepressurecontrolled).
3. Heatistransferreddirectlytothevaporisingsurface.
4. Pumpingactionisselfadjustedtotheheatload.
5. Inherenttwophaseliquidvapourseparation.
6. Isolated evaporator wicks are insensitive to small noncondensable gas or vapour
bubbles.
7. Heatloadscanbesharedamongindividualcapillarypumps.Typicallyheatloadsharing
ofupto90%ofappliedload.
8. Easily expandable. Pumps can be added in parallel to reach large heat transport
capabilities.
9. Readilyconfigurabletootheruses(seeFigure1543bforCPLtechnologyradiators).
10. Processingandtestingcanbebasedonavailableheatpipetechnologystandards.
15.6.2 CPL performance constraints

PerformanceconstraintsofCPLsystemsresultfromtheircapillarypumpingaction.Mainlimitations
arelistedbelow.
1. Heattransportcapabilitieslimitedtoabout7kW(at300K)for10mdistances.Individual
evaporatorpumpcapabilitynear2kW.
2. Power densities limited to about 26 x 104 W.m2. This is, however, a fairly large value,
compareforexamplewithclause11.7.
3. Capillarypumpinglimitoftheorderof3,2x103Pa(for105mPorexwicksandAmmonia
at293K).
4. PrecautionsofpressurecontainmentwhenAmmoniaisused(Ammoniaisappropriateat
temperaturesaround300K).
5. CPLsarenotasisothermalasmechanicallypumpedtwophaseflowsystems.
6. Reservoir size and performance under reduced gravity operation could be critical if
upsizingisconsidered.
15.6.3 CPL basic system concept

OnthebasisoftheexperiencederivedfromexistingCPLbreadboardandengineeringmodelslistedin
Table 1521, a typical CPL system is shown in Figure 1545. It consists of four main subsystems
namely:heatacquisition,heattransport,heatrejectionandcontrols.
410
ECSSEHB3101Part13A
5December2011
Figure1545:SchematicofabasicCPLsystem.FromChalmers,Pustary,Moy&
Kroliczek(1986)[23].
15.6.3.1 Heat acquisition

Figure1546showsastandardevaporatorpump3x102mindiameterand0,3mlong.Figure1546b
indicateshowtheheatistransferreddirectlytotheevaporatingsurfaceincontrastwithothersystems.
ThissignificantfeatureoftheCPLsystems,alreadymentionedinclause15.6.2,isthereasonwhyCPL
evaporatorpumpscanhandlemuchhigherradialheatfluxesthan,f.i.,conventionalheatpipes.
Figure1546:a)StandardCPLevaporatorpump.FromChalmersetal.(1986)[23].
b)HeatflowinacrosssectionofatypicalCPLevaporator.FromWise(1986)[192].
Heattransferfromtheheatsourcetotheevaporatortakesplacethroughastandardmountingplateto
which a bunch of pumps in parallel are metallurgical bonded. Figure 1547 shows the payload
mountingplatedevelopedinthePCCPprogram(Table1521)fortheNASAThermalTestBed.More
massoptimised and systemcompatible capillary cold plates are also under development. Two of
them,whichusecapillaryforcestodistributetheliquidovertheevaporatingcondensingsurface,are
presentedinclause15.7.1.
411
ECSSEHB3101Part13A
5December2011
Figure1547:Prototypecapillarycoldplate(PCCP)design.Dimensioninmm.
FromChalmers,Pustay,Moy&Kroliczek(1986)[23].
15.6.3.2 Heat transport

The ducts connecting evaporator to condenser are smooth walled tubes selected on the basis of
minimumpressureloss.Asageneralrule,linesaredevisedsothattotalpressurelossdoesnotexceed
10% of the capillary pumping limit. This indicates that valves, fittings, flowmeters and pressure
transducersinthelinesshouldalsobedesignedforminimumpressureloss.
Line diameters are chosen by means of a trade between minimum system mass and minimum
pressureloss.Strategiestominimisethetotalmassofatwophasefluidloopsubjecttoageneralloop
pressurelossconstrainthavebeenformulated(Schember(1986)[156]).
15.6.3.3 Heat rejection

CondenserRadiator.Severalcondenserdesignsshouldbeuseddependingonthesystemapplication.
Figure 1548 shows a Direct Condensation Radiator (DCR) where the condenser tubes are directly
attached to a radiator. This in turn can be separated into different elements which can operate
independentlyenablingmodularadditionorremoval.
Figure1548:CPLtechnologyradiators.a)Directcondensationradiator.b)Heat
exchangerheatpiperadiator.FromChalmers,Pustay,Moy&Kroliczek(1985)[23].
412
ECSSEHB3101Part13A
5December2011
In the second type of radiator, heat from the system fluid is transferred through an interface heat
exchanger to a number of independent heat pipes so that failure of one of them only influences a
single section of the radiator. The interface heat exchanger results in a reduced radiator operating
temperatureand thence,in an increased radiator mass.Whether or not thisadded massis balanced
withthereductionintherequiredmeteoroidprotectiondependsonthemission.
15.6.3.4 Controls
Temperature control of the CPL systems is accomplished via a reservoir where fluid is kept in
saturated pressuretemperature conditions. Temperature control of this reservoir is achieved by
thermostaticallycontrolledelectricalheatersandbyradiationtoouterspace.
Liquidacquisitionbafflesandacentralwickstructureinthereservoirserveasliquidflowconductsto
the reservoir line, which is connected to the rest of the CPL system. Volume requirements are at
presentmodestbutcouldresultindevelopmentneedsforlargevolumesystems.
Electricalheatingcouldbealsorequiredforstartingupthepumps.Aminimumamountofpowerhas
tobesuppliedtoanevaporatortoensurefluidflowinthesystem.
15.7 Components
This clause is intended tointroduce the main components of a twophase thermal transport system.
Thecontentofthisclauseisquiteopenatthiswriting.
15.7.1 Pumping systems

Thepumpingsystemfurnishesthepowerrequiredforfluidmotion.
PowerrequirementshavebeendiscussedinClause7,regardingclassicalsinglephasefluidloopsand
inClause10,fortwophasesystems.
Different pumping systems have been or currently are being considered under various NASA
contracts.Abriefdescriptionofthesesystemsfollows,whichisclosedbyasummaryinTable1522,
clause15.7.1.6.
15.7.1.1 Monogroove heat pipe

Thisheatpipe(Alario,Haslett&Kosson(1981)[1])consistsoftwoparallelducts,oneforliquidand
one for vapour, connected by a small monogroove slot, Figure 1549. Since the two fluids flow
independently, pressure losses are minimised. The small slot separating the channels creates a high
capillary pressure difference which, coupled with the minimal flow resistance of the two separate
channels, gives the high axial heat transport. The probability of liquid becoming entrained in the
vapourisminimal.
413
ECSSEHB3101Part13A
5December2011
Figure1549:Monogrooveheatpipe.FromAlario,Haslett&Kosson(1981)[1].
Preliminarytestsindicatethataheattransportfactor(Qleff)of25000W.mcanbeachieved.
Initially the monogroove heat pipe was devised to provide both heat transport and pumping. This
idea was dropped due to the wall wicking limitation of the heat pipe in the temperature range of
interest,sothatatpresentthemonogrooveheatpipeisconsideredasapurelypumpingsystem.
15.7.1.2 Capillary pump

Seeclause15.6.
15.7.1.3 Vapour compressor

Seeclause15.5.1.Thisistheconceptrequiringthegreatestamountofexternalpoweramongthesix
pumpingdevicesunderconsideration.
15.7.1.4 Mechanical pump

Usuallythepumpisplacedintheliquidportionofthesystem(seeFigure152)andfromthispointof
viewitdoesnotdifferfromthepumpsintroducedinClause8.Themajordifferenceisthatnowthe
powerrequirementsaredrasticallyreducedbecausethemassflowrateforagivenheatloadismuch
smaller when the heat is transferred by phase change than by sensible heat of the liquid (compare
ConceptsIIIandIV,inclause15.2.2).Althoughthisrepresentsasignificantadvantage,theadditionof
amechanicalpumpresultsinanoverallreductioninsystemreliabilityandanincreaseintotalsystem
mass.Thelargestportionofthisincreaseoccursasaresultofthelargesurfacearearequiredforthe
radiatorandisnotassociatetothepumpitself.Inaddition,provisionsforactivecontroloftheflow
rateshouldbetaken.
15.7.1.5 Osmotic pump

When two fluids, a solventsolute mixture and a pure solvent are separated by a semi permeable
membrane (impermeable to the solute), the solvent will flow through the membrane to attain
equilibriumandthismotionwillpersistaslongassoluteconcentrationgradientsarehigh.
This osmotic effect can create pressure differences orders of magnitude greater than the capillary
actioninconventionalheatpipewicks.
The solvent is the working (heat transfer) fluid which will be evaporated at the cold plate, leaving
behind the solute. From the evaporator the vapour solvent flows toward the condenser, reaching
againthemembraneasaliquid,Figure1550.
414
ECSSEHB3101Part13A
5December2011
Figure1550:Osmoticallypumpedheattransfersystem.FromTanzer,Fleischman
&Stalmach(1982)[175].
A wick structure is employed in the cold plate to prevent solute from being carried over with the
solventvapour,whichwouldcontaminatethesystem.
Undernormalgravityconditionstheconcentratedsolutionwillflowtowardthemembranebygravity
actioninordertoincreasethesoluteconcentrationinitsvicinity,withtheaimofpromotingenhanced
osmoticpressure.Somesolutioncirculationmethodshouldbedevisedtoperformthesameoperation
underreducedgravity(Peterson(1987)[142]).
Osmoticpumpingpresentsseveralmajordrawbacks.Ononeside,thepumpingratedoesnotdepend
ontheheatloadand,thence,anadditionalcontrolisrequiredasinmechanicalpumps.Ontheother
side,water,whichisthesolventusedfortesting,doesnotfitthetemperaturerangeofinterest.Finally,
availablemembranesareincapableofsustainedoperationattemperaturesabove350K.
15.7.1.6 Biomorph pump

It is a deformablewall (peristaltic) pump. The wall consists of two layers of ferroelectric crystals or
ceramicmaterialswhichhavebeenbondedtogether,Figure1551.
Figure1551:Biomorph(Biomorph)pump.FromPeterson(1987)[142].
The two deformable elements are attached to separate electrical leads. When an AC potential is
applied across the leads, the internal stresses (piezoelectric effect) results in repeated flexing of the
walls,whichcombinedwithappropriatevalvemotionproduceapumpingaction.
Control of the mass flow rate is also required in this case but it can be easily achieved through a
variationineithervoltageorfrequency.
415
ECSSEHB3101Part13A
5December2011
Thistypeofpumpisstillinitsdevelopmentstage.Itsreliabilityhasnotyetbeenassessed.
Table1522belowpresentsasummaryofthePumpingSystemswhichhavebeendescribedabove.
Table1522:ComparativeSummaryofPumpingSystems
Criteria Monogroove Capillary Vapor Mechanical Osmotic Biomorph
HeatPipe Pump Compressor Pump Pump Pump
External None None Large Small Small Small

Power
Required
Priming Severe Severe None None Minimal Minimal

Limitations
Control Unknown Minimal Consider Considerablea Unknown Unknown

Necessary able Moderateb
Acceptand No No Yes Yes Yes Yes

RejectHeat
Life Unknown Excellent Fair Goodc Poorc Poorc

Expectancy
Technology Unknown Adequate Good Excellent Poorc Poorc

Level
System Notfeasible Minimal Complex Fair Complexc Complexc

Complexity
a Pumpassisted.
b Pumpedtwophase.
c Atcurrentstateoftheart.
NOTE FromPeterson(1987)[142].
15.7.2 Mounting plates

TheinterfacebetweenadedicatedinstrumentorpayloadcomponentandtheTMScouldbe:a)Acold
platemountingsurfaceinthermalcontactwiththeTMSfluid.b)AfluiddiversionoftheTMSwhich
internallyorexternallycontactstheinstrumenttobecooled.c)Ashroudorcontainerenclosingoneor
more instruments and in thermal contact with the TMS. And d) an independent thermal control
systeminthermalcontactwithalocalorremotecoldplate.SeesketchesinFigure1552.
416
ECSSEHB3101Part13A
5December2011
Figure1552:InstrumentTMSinterfaces.FromAlmgrenetal.(1981)[2].
AsignificantdesignparameteroftheTMSisthethermalresistancebetweentheinstrumentforwhich
operating temperature limits have been set and the TMS fluid. If this resistance is large, the
temperaturedifferencewillbelargeandthiscouldforceloweringthefluidlooptemperaturelevelof
theTMSforafixedlevelintheinstrument.
A standard mounting plate based on CPL technology has been introduced in Figure 1547, clause
15.6.3.1.Twodifferentcapillaryplates,whicharebeinginvestigated,aredescribedinthefollowing.
ThemodelinFigure1553,whichispurelyconceptual,exhibitsagroovedcoldplate/hotplatewhich
can operate as an evaporator or as a condenser. This model has been used in an analytical study
aimingattheoptimisationoftheevaporatorgroovegeometry.Intheevaporator(coldplate)modeof
operationthecondensateflowsbycapillaryactionfromaliquidreservoirthroughacommunication
(bridging)wick.Theresultingvapourflowsoutofthecoldplatetoaseparatecondenserwherethe
heatisrejected.Thereservoirisreplenishedatintervalsfromavalvedexternalliquidsupplyline.
Figure1553:Schematicofagroovedcoldplate/hotplateinthecoldplatemode.
FromHwangbo&McEver(1985)[94].
Inthecondenser(hotplate)modeofoperation,heatisdeliveredtotheheatsinkbycondensationof
vapoursfromthevapourline.Asliquidbuildsupinthecapillarygroovesitiscarriedtothereservoir
whereitisperiodicallyremovedbyacondensatereturnline.
ResultsofananalyticalstudyperformedbyHwangbo&McEver(1985)[94]anticipateheatfluxesas
largeas9x104W.m2.
ThemountingplatesketchedinFigure1554isbeingdevelopedbyMcDonnellDouglasunderNASA
contract. Again it can operate either in the evaporator mode or in the condenser mode. In the
evaporatormodeliquidispumpedthroughasystemofsixfeedtubes.Eachtubehasaninterfacewick
insertedintoaslotandkeptincontactwiththegroovedfaceoftheheatedsurface.Theliquidinthe
417
ECSSEHB3101Part13A
5December2011
feedtubeisdrawnbythecapillaryactionofthegroovedsurface.Onlytheamountofflowneededto
transfertheheatloadispulledthroughthewick.Theexcessliquidflowsouttheendoftheexittube
whereitmixeswiththevapourandexitsthemountingplate.Becauseofthisflowthroughdesign,the
platecanoperatewithinawiderangeofflowratesandflowqualities.
Figure1554:TwoPhaseMountingPlate(TPMP)developmentunit.TheLexan
windowhasbeenincorporatedtoobservetheflowduringtesting.Alldimensions
areinmm.FromGrote&Swanson(1985)[77].
In the condenser mode of operation the feed tube inlet manifold is closed and the twophase fluid
enters through the evaporator mode exit. The vapour condensed and the condensate flows into the
annularregionoftheevaporatormodeinletfittings,andexitsthroughoneofthetwotubesthatareat
rightanglestothefittings.
TheunitwasdesignedtooperateeitherwithFreon11orAmmonia,buttestingwasconductedwith
Freon11.
Themountingplatewastestedintheevaporatormodefortotalheatloadsover3x103Wandlocal
heatfluxesover4x104W.m2.Inthecondensermodetheheatloadswereinthe60W400Wrange.
15.7.3 Vapour quality sensors

Sensors, control logic and actuators are incorporated at different locations on the TMS to adjust
pumping speed, working fluid reservoir content, throughput of valves, ... These sensors could be:
flowmeters,pressureandtemperaturegagesand,withinthepresentcontext,vapourqualitysensors.
Vapourqualitysensorscanbeimplemented,f.i.,atthecoldplateexitformonitoringtheliquidflow
rateand,thuspreventingevaporatordryoutandcontrollingtheflooding.
The vapour quality, w, is deduced from the void fraction, , once the pressure, temperature, and
liquid and vapour flow regimes (v or t) are known, see clause 14.2.1, Eqs. [148] and [1411]. This
requirestemperaturemonitoringplusflowpatternsidentification(possiblythroughflowcalibration).
Pressure,ontheotherhand,variesslowly.
Awidevarietyoftechniquesareavailablefortwophaseflowdiagnosis.However,severalamongthe
proven techniques and certain ontheshelf instrumentation cannot be applied on board orbital
platforms;eithertheydonotmeetthestrictsafetyrequirementsformannedplatforms,theyarevery
massiveordependongravityfortheirperformance.ForacompletereviewofthesubjectseeHewitt
(1982b)[87],whereachartfortheselectionofthemostappropriatetechniqueisgiven.
418
ECSSEHB3101Part13A
5December2011
SimilarnegativeconclusionswerereachedbyDelil(1986b)[30]afterconfrontingfourteencandidate
methodsformeasuringthevoidfractionagainstthefollowingrequirements:
1. Simple,flexibleandlowsystemimpact.
2. Smallmass,sizeandpowerconsumptions.
3. Safeandnonhazardoustoman.
4. Fastresponse.
5. Accuratewithinawidequalityrangeuptovapourqualityvaluesexceedingw=0,8.
6. Highlysensitivetoqualityvariations.
7. Suitedtotheworkingfluid(Freon,Ammonia,...)
8. Noncontaminatingtheloopandlongtermreliable.
9. Slightlydisturbingtheflow.
10. Testableunderbothnormalandreducedgravity.
11. Lowcost.
12. Lowdevelopmentrisk.
Electrical impedance methods are apparently worth of further consideration although they are not
verysensitivetothefluidflowpattern.
The electrical impedance of a twophase flow depends on the concentration and distribution of the
phases,andtheuseofimpedancemeasurementisattractivebecauseitgivesavirtuallyinstantaneous
response. Depending on the frequency of the system the impedance will be governed by the
conductance (low frequency), capacitance (high frequency) or both. Although conductance methods
are easier to use (see f.i. Gao, Zhou & Chen (1985) [70]), capacitance methods are often preferred
because the liquid dielectric permittivity varies less with temperature than does the liquid electrical
conductivity.
15.7.3.2 Capacitance methods

Capacitancemethodsarebasedonthedifferenceinrelativepermittivity, /o,betweenaliquidanda
gas.oisthevacuumpermittivityandthatofthe(dielectric)materialunderconsideration.
Theelectricalcapacitance,C,ofaparallelplatecapacitor,platearea,A,distance,d,betweenplatesand
permittivity,,oftheinterstitialmaterialis:
sA
c [154]
d
Slightly different expressions should be used for different capacitor geometries, but the above
equationsufficesforillustrativepurposes.
Estimating the electrical capacitance of a twophase fluid dielectric between suitable plates is not
simple.Foragivensteady,twophaseconfigurationandassumingthatdynamiccurrentsareabsent,
Electrostaticsallowsthecalculationoftheelectricfieldintensityfromwhichtheinternalenergy,u,of
thecapacitorcanbeexpressed(Sommerfeld(1952))
419
ECSSEHB3101Part13A
5December2011
u
1

2 V
1
2
sE 2 dv C ds
2 V
[155]
wherethefirstintegralisextendedoverthedielectricvolumeandthesecondoveritssurface.isthe
electricpotentialandE= .Theassumptionisusuallymadethatfreeandpolarchargesareabsent
from the fluid in which case .E = 0, and since E = 0, the resulting differential equation is of
Laplacetypefortheelectricpotential,=0,butthisequationmustbesolvedwiththeappropriate
boundary conditions at the duct walls and/or capacitor plates, plus jump conditions at the
(supposedly known) liquidvapour interfaces. This extraordinarily difficult the calculation of the
capacitance, C, of a twophase medium, which will only furnish an unequivocal result when the
geometryoftheflowpatterncanbedefinedbyonesinglecharacteristiclength,asintheannularflow
case.
In many instances the problem is circumvented by use of an inparalleltype expression for the
effectivepermittivity,TP,ofthefluid(Delil(1986b)[30])
1
sTP
1 [156]

sG sL
Otherexpressionsforhomogeneousdispersionsofgasbubblesintheliquid,andforliquiddroplets
dispersedinagas,againcharacterisedbyasinglelength,canbefoundinHewitt(1982b)[87].
Four types of capacitors have been sketched in Figure 1555, three radial electric field capacitors
(single coaxial capacitor, concave plate capacitor and double helix capacitor) plus an axial field
capacitor(filmthicknessgage).
Figure1555:Voidfractionsensorsofthecapacitancetype.a)Singlecoaxial
capacitor.b)Concaveplatecapacitor.c)Doublehelixcapacitor.d)Filmthickness
gage.FromDelil(1986b)[30].
420
ECSSEHB3101Part13A
5December2011
The first three types are intrusive and then will result in a pressure loss in the duct. Although the
concave plate and double helix type capacitors could be incorporated in the duct wall, the duct
diameter will then determine the interplate distance, d, presumably reducing the sensitivity. Only
the film thickness gage can be easily mounted flush with the wall duct without compromising the
optimumdistance,d.
The coaxial and concave plate capacitors can be used only in axisymmetrical flow (either
homogeneous or annular). The double helix capacitor is sensitive to any steady flow because of the
averagingeffectofthehelixelectrodes.Thefilmthicknessgagecanbeusedonlywithannularflows.
InformationonthesensitivitiesofthedifferentsystemshasbeencollectedinFigure1556andFigure
1557, which show some dimensionless functions of the measured capacitance, C, versus the void
fraction,,orthedimensionlessfilmthickness,/d.
Figure1556:Dimensionlesscapacitanceasafunctionofvoidfraction,.a)
Annularflow,singlecoaxialcapacitor.b)Annularflow,concaveplatecapacitor.
FromDelil(1986b)[30].
421
ECSSEHB3101Part13A
5December2011
Figure1557:Dimensionlesscapacitanceasafunctionofdimensionlessfilm
thickness,/d.Flatwall,filmthicknessgage.FromDelil(1986b)[30].
Thisinformationcanbesummarisedasfollows:
1. No reliable fast response method exists for measuring the void fraction in slug flow
(Figure 141, Figure 142, in clause 14.1) because of its unsteadiness. This does not
representamajordifficultyheresincethistypeofflowwillnotappearinTMSs,which
arecharacterisedbyhighvapourqualityflows(1)andreducedgravityoperation.
2. Thesensitivityofthesinglecoaxialcapacitortochangesinthehighvoidfractionsideis
poor,Figure1556a
3. Theconcaveplate(andalsothedoublehelix)capacitorexhibitsgoodperformancefigures
for annular flow, Figure 1556b Unfortunately, the absolute capacitance level and the
sensitivityaretoopoorforthehighvoidfractionsforeseenintheTMS.
4. the film thickness gage is well suited to annular, high flow because of its sensitivity,
Figure1557.Thesignalcouldbeenhancedbyoptimisingthegeometry.
The feasibility of a film thickness gage has been demonstrated (Delil (1987b) [32]). The breadboard
modelisbasedonatwostripplanarconfiguration.Thedistance,d,betweenstripsisapproximately
equaltotheanticipatedliquidfilmthickness, .Plasticsheetswereusedtorepresenttheliquid.The
implementationofmorerealistictestsdoesnotfaceinsurmountabledifficulties.
15.7.4 Fluid disconnects

AfluiddisconnectisanimportantcomponentofaTMSandhasbeenincorporatedinmanysystems
wheremaintenanceorreconfigurationmustbequicklyeffectedwithouttheaccompanyingnecessity
fordepressurization.
Therequirementsofthesedevicesshouldbeminimalpressuredropwhenfullyengagedandminimal
spillage during disengagement, low mass, reliability and ease of employment (onehand or remote
armmanipulationunderreducedgravity).
Oftheseveralavailabledesigns,theRotaryShutoff(RSO)Disconnect,manufacturedbyMOOGInc.
EastAurora,NewYork,seemstomeettheaboverequirements.
The RSO disconnect utilises, instead of traditional poppets, spherical valving elements to open and
closethefloworifice.Thedesigncanincorporateawidevarietyofmaterials.Whilestainlesssteeland
aluminium have been used to date, titanium, plastics and composites could be employed without
alteringthedesignanddegradingitscharacteristics.
422
ECSSEHB3101Part13A
5December2011
Oncethedisconnectisengaged,theflowpathisunobstructed,whichallowstheplacementinseriesof
severalofthesedeviceswithoutaffectingthepressureloss.Duringdisengagementthespillageisless
than106m3andanoptionalabilitytopurgethedevicecanreducespillagetovirtuallyzero.
TheengagementsequenceoftheRSOdisconnectissummarisedinFigure1558.
Figure1558:EngagementsequenceoftheRSOdisconnect.Disengagementis
achievedthroughthereversesequence.FromMOOG[127].
423
ECSSEHB3101Part13A
5December2011
16
Control technology
16.1 Basic definitions

Inthisclause,terminologycommonlyusedinthisclauseaboutcontrolsystemsisdefined.
Actuatingsignal:Asignalusedtoinfluenceintheplantinfeedbackcontrolsystems.
Analog signal: A signal defined over a continuous range of time whose amplitude can assume a
continuousrangeofvalues.
AnalogtoDigitalConverter(A/D):Adevice(alsocalledanencoder)whichconvertsananalogsignal
intoadigitalsignal.
Analog transducer: A transducer in which the input and output signals are continuous functions of
time.
Automatic regulation system: A feedbackcontrolled system in which the reference input and the
desired output are constant or change slowly in time. The main task is to maintain the output at a
desiredlevel,independentlyofexistingdisturbances.
Continuoustimesignal:Asignaldefinedoveracontinuousrangeoftime.
Control system of process: An automatic regulation system where the output is a variable as
temperature,pressureflow,liquidlevelorpH.
Decoder:SeeDigitaltoAnalogConverter.
Digital control system: A control system in which a digital computer is encharged of the signal
processing.
Digitalsignal:Adiscretetimesignalwithquantizedamplitude.
DigitaltoAnalog Converter (D/A): A device (also called a decoder) which converts a digital signal
intoananalogsignal.
Digitaltransducer:Atransducerinwhichtheinputandoutputsignalsoccuronlyatdiscreteinstants
oftimeandthesignalmagnitudesarequantized.
Discretetimesignal:Asignaldefinedonlyatdiscreteinstantsoftime.
Encoder:SeeAnalogtoDiGgitalConverter
Error:Signalresultingfromthedifferencebetweentheactualoutputandthedesiredstate.
Feedbackcontrol:Anoperationwhich,inthepresenceofdisturbances,tendstoreducethedifference
existingbetweentheoutputandthereferenceinputofasystem(oradesiredstate,randomlyvaried)
anditdoesitbasedonthatdifference.
Gain:Thegainofasystemistheoutputinputratioofit.
424
ECSSEHB3101Part13A
5December2011
Perturbation:Anirregularitywhichaffectsasignal,anddeterioratesthesystemoutput.Perturbations
generatedinsidethesystemareknownasinternalperturbations.Thoseproceedingfromoutsidecan
beconsideredasinputsforthesystem.
Plant:Incontrolsystems,plantmeanstheapparatusorequipmentforacertainphysicaloperationthat
hastobecontrolled.
Process: A continuing development involving many changes, which is going to be controlled. A
processcanbechemical,physical,mechanical,economical,biological,...
SampleandHold(S/H):Sampleandholdisageneraltermusedforasampleandholdamplifier.
Itdescribesacircuitthatreceivesananaloginputsignalandholdsthissignalataconstantvaluefora
specifiedperiodoftime.
Sampleddatatransducer:Atransducerinwhichtheinputandoutputsignalsoccuronlyatdiscrete
instants of time (usually periodic), and the magnitudes of the signals, as in the case of the analog
transducer,areunquantized.
Servomechanism: A feedbackcontrolled system in which the output is a mechanical variable, as
position,velocityoracceleration.
Stability:Thenotionthatdescribeswhetherthesystemwillbeabletofollowtheinputcommand.Ina
nonrigurousmanner,asystemissaidtobeunstableifitsoutputisoutofcontrolorincreaseswithout
bound.
System: A set or arrangement of components related or connected as to form a unit performing a
specifictask.Asystemisnotconstrainedtoachievephysicalobjectives.Thisconceptcanbeappliedto
abstract and dynamic phenomena, as in economy. Therefore, the term system has to be understood
regardingphysical,biological,economical,...
System with feedback control: A system which tends to maintain a preset ratio between the output
and the reference input, by comparing both signals and using the difference between them as the
controlparameter.
Transducer:Adevicewhichconvertsaninputsignalintoanoutputsignalofanotherform,suchasa
devicethatconvertsapressuresignalintoavoltageoutput.
16.2 General description of control systems
16.2.1 Introduction
Thepurposeofthisclauseistointroducethedifferenttypesofcontrolsystems.Atypicalclassification
is:
Closedloopcontrolsystems.
Openloopcontrolsystems.
Adaptivecontrolsystems.
Learningcontrolsystems.
First,thesetypesofcontrolsystemsaredescribedandafterthatthemostusedtypes,closedloopand
openloop,arecompared.
425
ECSSEHB3101Part13A
5December2011
16.2.2 Closed-loop control systems

Aclosedloopcontrolsystemisthatinwhichtheoutputsignalaffectsdirectlythecontrolaction.
Thus, a closedloop control system is a feedbackcontrolled system. The actuating signal, u, is a
functionofthefeedbacksignal(whichisafunctionoftheoutputsignalorvariabletobecontrolled)
andofthereferenceinput.
TheblockdiagramofaclosedloopcontrolsystemisshowninFigure161.
Figure161:Closedloopcontrolsystemblockdiagram.
The actual output is measured, and a signal corresponding to this measurement is fed back to the
inputstation,whereitiscomparedwiththeinput(desiredoutput).
This comparison is performed in some device which produces a signal related to the difference
betweenthetwosignalsfedintoit.Thissignalmaybecalledtheerrorifthetwosignalscompared
arethetrueinputandoutputsignals.Frequently,thesignalfedbackfromtheoutputismodifiedin
somewaybeforecomparisonwiththeinput,andinsuchcasestheresultisnottheerror.Thesignal
thusdeterminedisusedtodrivethesystemandmayalwaysbecalledactuatingsignal.
16.2.3 Open-loop control system

Inopenloopcontrolsystemstheoutputsignaldoesnotaffectinanywaythecontrolaction.Insuch
systems,theoutputsignalisnotcomparedwiththereferenceinput,asitissketchedinFigure162.
Thus, each point of work corresponds to a different reference input. The expected output, and its
accuracy,dependsuponthecalibration,and,thus,theactualoutputmayvaryfromtheexpectedone.
Figure162:Openloopcontrolsystemblockdiagram.
Theelementsofanopenloopcontrolsystemcanusuallybedividedintotwoparts:thecontrollerand
thecontrolledprocessasshownbytheblockdiagram.Aninputsignal,orcommand,isappliedtothe
controller,whoseoutputactsastheactuatingsignalu;thissignalthencontrolstheprocess,sothatthe
controlledvariablecwillperformaccordingtosomeprescribedstandards.
Insimplecases,thecontrollercanbeanamplifier,mechanicallinkage,filter,...,dependinguponthe
nature of the system. In more sophisticated cases, the controller can be a computer such as a
microprocessor,andmorerecentlyamicrocontroller.
426
ECSSEHB3101Part13A
5December2011
Becauseofthesimplicityandeconomyofthesesystems,openloopcontrolsystemsarefoundinmany
noncriticalapplications.
16.2.4 Adaptative control systems

This type of control systems are presently under research and they are not commonly used in
spacecrafttechnology.
Thedynamiccharacteristicsofmostsystemsarenotconstant(componentsdeteriorationandwasting
during their operative life, modification in parameters or in the environment as atmospheric
behaviour or solar radiation, ...). Although feedback control systems attenuate the effects of small
changes in dynamic characteristics, when modifications in the parameters of the system or in the
environment are large enough,a system has to be able to adapt to have a satisfactory response and
behaviour.Adaptationmeanstohavecapacityofautomodificationandautoadjustingaccordingtothe
nonforeseenmodificationsoftheenvironmentandtheplant.
Controlsystemswithanycapacityofadaptingareknownasadaptivecontrolsystems.
Themaininterestofthistypeofcontrolsystemisitscapacityofadaptingnotonlytochangesinthe
system(plantandenvironment)butalsotomoderatedefectsinthedesignofthesystem,uncertainties
andminorcomponentfailures.Therefore,thesystemreliabilityishighlyincreased.
Anadaptivecontrolsystemmayconsistinthreefunctionsasfollows:
1. Toidentifythedynamiccharacteristicsoftheplant.
2. Toadoptdecisionsbasedontheplantidentification.
3. Modificationoractionbasedontheadopteddecision.
The plant identification is a process that must be performed according to the plant modification
velocity.Systemshighaccuratelyknownandveryconstantintheircharacteristicscanbecharacterised
withaninitialidentificationatthebeginningoftheiroperativelifeandaninitialmodificationofthe
decisionprocesstominimisethebehaviourindex.
When the characteristics of the system change continuously, procedures must be revised
continuously.
Figure163isatypicalblockdiagramofadaptivecontrolsystems.
Figure163:Typicalblockdiagramofadaptativecontrolsystems.
427
ECSSEHB3101Part13A
5December2011
16.2.5 Learning control system

Controlsystemsapparentlybeingopenloopare,infact,closedloopcontrolsystemsduetotheaction
ofahumancontrol,whocomparesinputandoutputsignalsandexecutesthecorrectiveactions,based
onhisappreciationoftheerrorordifference.
Itisdifficulttomathematicallyrepresentthehumanbehaviour.Themostimportantcharacteristicis
the human capacity of learning. At the beginning, the human control may seem inaccurate or even
mistaken,but,asheisbecominganexpert,hisdecisionbecomesoptimal,andthismustbetakeninto
accountintheanalysis.
Recently, computers are becoming with learning capacity and are being implemented in control
systems, and, therefore, control systems which are able to learn are known as learning control
systems.Thisisaverynewconceptandhasnotbeenexploreddeeplyyet.
16.2.6 Trade-off of open- and closed-loop control systems

Itisimportanttopointoutthat,wheninputsarepreviouslyknownanddisturbancesdonotexist,an
openloop control system is preferable. Closeloop control systems present advantages when
nonforeseen disturbances and/or variation of components performances exists. It is noted that the
servomechanismoutputpowerinfluences,partially,itscost,weightandsize.Toreducetherequired
output power of a system, it is preferable to implement an openloop control, when applicable; but
generallyforasatisfactoryworkingofthesystemanadequatecombinationofopenandclosedloop
controlsystemistheoptimalsolution.
Commonly, the use of feedback is applicated with the purpose of reducing the error between the
referenceinputandthesystemoutput.However,thesignificanceoftheeffectsoffeedbackincontrol
systemsismorecomplexandhasawidereffect.Infact,thereductionofsystemerrorismerelyoneof
the many important effects that closing the loop may have on (upon) a system. The following
paragraphsanalysetheeffectsoffeedbackonsuchsystemperformancecharacteristicsasoverallgain,
stability,sensitivityandnoise.
LetusconsiderthesimplefeedbacksystemconfigurationshowninFigure164,whereristheinput
signal, c the output signal, e the error, and b the feedback signal. The parameters G and H may be
considered as constant gains. By simple algebraic manipulations, it is easy to show that the output
inputratioofthesystemis
Figure164:Feedbacksystem
428
ECSSEHB3101Part13A
5December2011
c G
M [161]
r 1 GH
Using this basic relationship of the feedback system structure, some of the significant effects of the
feedbackcanbeuncovered.
16.2.6.1 Effect of feedback on overall gain

As it can be seen from Eq. [161], feedback affects the gain G of a nonfeedback system by a factor
1+GH.ThesystemofFigure164issaidtohavenegativefeedback,sinceaminussignisassignedto
thefeedbacksignal.ThequantityGHmayitselfincludeaminussign,sothegeneraleffectoffeedback
isthatitmayincreaseordecreasethegainG.Inapracticalcontrolsystem,GandHarefunctionsof
frequency, so the magnitude 1+GH may be greater than unity in one frequency range but less than
unityinanother.Therefore,feedbackcouldincreasethegainofthesysteminonefrequencyrangebut
decreaseitinanother.ThisissketchedinFigure165.
Figure165:Sketchofthegainofasystemasafunctionoffrequency.
16.2.6.2 Effect of feedback on stability

To investigate the effect of feedback on stability, we can again refer to Eq. [161]). If GH = 1, the
outputofthesystemisinfiniteforanyfiniteinput,andthesystemissaidtobeunstable.Therefore,
wemaystatethatfeedbackcancauseasystemthatisoriginallystabletobecomeunstable.Certainly,
feedbackisatwoedgedsword;whenitisimproperlyused,itcanbeharmful.Itshouldbepointed
out,however,thatonlythestaticcaseisbeingconsideredhere,and,ingeneral,GH=1isnottheonly
conditionforinstability.
Itcanbedemonstratedthatoneoftheadvantagesofincorporatingfeedbackisthatitcanstabilisean
unstablesystem.LetusassumethatthefeedbacksysteminFigure164isunstablebecauseGH=1.If
429
ECSSEHB3101Part13A
5December2011
weintroduceanotherfeedbackloopthroughanegativefeedbackgainF,asshowninFigure166,the
outputinputratiooftheoverallsystemis
Figure166:Feedbacksystemwithtwofeedbackloops.
c G
[162]
r 1 GH GF
ItisapparentthatalthoughthepropertiesofGandHaresuchthattheinnerloopfeedbacksystemis
unstable, because GH = 1, the overall system can be stable by properly selecting the outerloop
feedbackgainF.
In practice, GH is a function of frequency, and the stability condition of the closedloop system
dependsonthemagnitudeandphaseofGH.Thebottomlineisthatfeedbackcanimprovestabilityor
beharmfultostabilityifitisnotproperlyapplied.
16.2.6.3 Effect of feedback on sensitivity

Sensitivity considerations often are important in the design of control systems. Since all physical
elementshavepropertiesthatchangewithenvironmentandage,theparametersofacontrolsystem
cannotalwaysbeconsideredascompletelystationaryovertheentireoperatinglifeofthesystem.For
instance, the winding resistance of an electric motor changes as the temperature of the motor rises
duringoperation.Ingeneral,agoodcontrolsystemshouldbeveryinsensitivetoparametervariations
butsensitivetotheinputcommands.Nowtheeffectthatfeedbackhasonthesensitivitytoparameter
variationswillbeinvestigated.
Referring to the system in Figure 168, G is considered as a gain parameter that may vary. The
sensitivityofthegainoftheoverallsystem,M,tothevariationinGisdefinedas
M / M percentage change in M
S GM [163]
G / G percentage change in G
430
ECSSEHB3101Part13A
5December2011
where MdenotestheincrementalchangeinMduetotheincrementalchangeinG, G.ByusingEq.
[161],thesensitivityfunctioniswritten
M / M 1
S GM [164]
G / G 1 GH
Thisrelationshowsthat,ifGHisapositiveconstant,themagnitudeofthesensitivityfunctioncanbe
madearbitrarilysmallbyincreasingGH,providedthatthesystemremainsstable.Itisapparentthat
inanopenloopsystem,thegainofthesystemwillrespondinaonetoonefashiontothevariationin
G;i.e.,SGM.Itshouldagainberemindedthat,inpractice,GHisafunctionoffrequency;themagnitude
of1+GHmaybelessthanunityoversomefrequencyranges,sothatfeedbackcouldbeharmfultothe
sensitivitytoparametervariationincertaincases.
Ingeneral,thesensitivityofthesystemgainofafeedbacksystemtoparametervariationsdependson
wheretheparameterislocated.
16.2.6.4 Effect of feedback on external disturbance or noise

All physical systems are subject to some types of extraneous signals or noise during operation.
Examples of these signals are thermalnoise voltage in electronic circuits and brush or commutator
noiseinelectricmotors.Therefore,inthedesignofacontrolsystem,considerationshouldbegivento
ensurethatthesystemisinsensitivetonoiseanddisturbancesandsensitivetoinputcommands.
The effect of feedback on noise and disturbance depends greatly on where these extraneous signals
occur in the system. No general conclusions can be made; but in many situations, feedback reduces
theeffectofnoiseanddisturbanceonsystemperformance.
Referring to the system shown in Figure 167, in which r denotes the command signal and n is the
noisesignal,intheabsenceoffeedback,H=0,theoutputcis
Figure167:Feedbacksystemwithanoisesignal.
c G1G2 e G2 n [165]
wheree=r.Thesignaltonoiseratiooftheoutputisdefinedas
431
ECSSEHB3101Part13A
5December2011
Output due to signal G1G2 e e

G1 [166]
Output due to noise G2 n n
ToincreasethesignaltonoiseratiooneshouldincreasethemagnitudeofeitherG1orerelativeton.
VaryingthemagnitudeofG2wouldhavenoeffectwhatsoeverontheratio.
Withthepresenceoffeedback,thesystemoutputduetorandnactingsimultaneouslyis
G1G2 G2
c r n [167]
1 G1G2 H 1 G1G2 H
ComparisonofEq.[167]withEq.[165]showsthatthenoisecomponentintheoutputofEq.[167]is
reduced by the factor 1+G1G2H if the latter is greater than unity, but the signal component is also
changedbythesameamount.Thesignaltonoiseratiois
Output due to signal G1G2 r / 1 G1G2 H r

G1
G2 n / 1 G1G2 H
[168]
Output due to noise n
anditisthesameasthatwithoutfeedback.Inthiscase,feedbackisshowntohavenodirecteffecton
the output signaltonoise ratio of the system. However, the application of feedback suggests the
possibilityofimprovingthesignaltonoiseratioundercertainconditions.Ifitisassumedthat,inthe
system of Figure 167, the magnitude of G1 is increased to G1 and that of the input r to r, with all
other parameters unchanged, the output due to the input signal acting alone is at the same level as
thatwhenfeedbackisabsent.Inotherwords,
c G
[169]
r 1 GH GF
withG1increasedtoG1,theoutputduetonoiseactingalonebecomes
G2
c r 0 n [1610]
1 G1G2 H
whichissmallerthantheoutputduetonwhenG1isnotincreased.Thesignaltonoiseratioisnow
1 1 G1G2 H
G1G2 r Gr
G2 n / 1 G1G2 H
[1611]
n
whichisgreaterthanthatofthesystemwithoutfeedbackinafactorof(1+G1G2H).
432
ECSSEHB3101Part13A
5December2011
16.3 Basic control actions
16.3.1 Introduction
An automatic controller compares the actual value of the plant output with the reference input
(desiredvalue),determinesthedeviation,andproducesacontrolsignalthatwillreducethedeviation
tozeroortoasmallvalue.Themannerinwhichtheautomaticcontrollerproducesthecontrolsignal
is called the control action. In this clause the basic control actions used in control systems are
discussed.
Thecentralelementoftypicalfeedbackcontrolsystemsisthecontroller.Thisdevicehasasitsinput
a lowpowerlevel signal, which is a measure of a process variable, and as its output a highpower
level signal, which is related to the input signal by some specified, but usually adjustable, input
output relation. Note that ideally, the controller is a strictly unilateral device (the conditions at the
outputendofthecontrollershouldinnowayhaveanyeffectbackwardthroughthecontrolleron
the process). The output signal itself is used to manipulate the process in some way through an
actuator. Because a highpowergain device by itself can be very difficult to regulate, a common
technique in building controllers is to use a passive feedback element around the highgain device.
This reduces the overall gain considerably, but the end result is an instrument that has sufficient
powergainandiscapableofpreciseregulation.ThismethodisshownschematicallyinFigure168.
Thepowerunithasanexternalpowersupply(notshown)toprovidethepowergain,andtheinput
outputrelationiscontrolledcompletelybythepassivefeedback.
Figure168:Feedbackstructureofinstrumentsandregulators.FromOgata(1990)
[137].
Controllerscanbeclassifiedinsixgroups:
1. Twopositionoronoffcontrollers.
2. Proportionalcontrollers(P).
3. Integralcontrollers(I).
4. Proportionalintegralcontrollers(PI).
5. Proportionalderivativecontrollers(PD).
6. Proportionalintegralderivativecontrollers(PID).
433
ECSSEHB3101Part13A
5December2011
Most controllers use electricity or pressurised fluid as power sources. Controllers may also be
classifiedaccordingtothekindofpoweremployedintheoperation,suchaspneumatic,hydraulic,or
electroniccontrollers,aswillbeshowninclause16.4.
Figure 169 presents a block diagram of a typical control system, which consists of an automatic
controller,anactuator,aplant,andasensor(measuringelement).Thecontrollerdetectstheactuating
errorsignalandamplifiesittoasufficientlyhighlevel.(Thustheautomaticcontrollercomprisesan
errordetectorandamplifier.Quiteoftenasuitablefeedbackcircuit,togetherwithanamplifier,isused
toaltertheactuatingerrorsignalbyamplifying,andsometimesbydifferentiatingand/orintegratingit
toproduceabettercontrolsignal).Theactuatorisapowerdevicethatproducestheinputtotheplant
according to the control signal so that the feedback signal will correspond to the reference input
signal.
Figure169:Blockdiagramofanindustrialcontrolsystem,whichconsistsofan
automaticcontroller,anactuator,aplant,andasensor(measuringelement).From
Ogata(1990)[137].
Thesensor,ormeasuringelement,isadevicethatconvertstheoutputvariableintoanothersuitable
variable,suchasadisplacement,pressure,orvoltage,thatcanbeusedtocomparetheoutputtothe
referenceinputsignal.Thiselementisinthefeedbackpathoftheclosedloopsystem.
16.3.2 Two-position or on-off control action

In a twoposition control system, the actuating element has only two fixed positions, which are, in
many cases, simply on and off. Twoposition or onoff control is relatively simple, reliable and
inexpensive.
Lettheoutputsignalfromthecontrollerbeu(t)andtheactuatingerrorsignalbee(t).Intwoposition
control, the signal u(t) remains at either a maximum or minimum value, depending on whether the
actuatingerrorsignalispositiveornegative,sothat
u t U1 , if et 0
On off control law
u t U 2 , if et 0
[1612]
434
ECSSEHB3101Part13A
5December2011
where U1 and U2 are constants. The minimum value U2 is usually either zero or U1. Twoposition
controllersaregenerallyelectricaldevices,andanelectricsolenoidoperatedvalveiswidelyusedin
suchcontrollers(seeclause16.4.3.1.1).Pneumaticproportionalcontrollerswithveryhighgainsactas
twopositioncontrollersandaresometimescalledpneumatictwopositioncontrollers.
Figure 1610 shows two block diagrams of twoposition controllers. The range through which the
actuatingerrorsignalmovesbeforetheswitchingoccursiscalledthedifferentialgap(asitisindicated
in Figure 1610(b)). Such a differential gap causes the controller output u(t) to maintain its present
valueuntiltheactuatingerrorsignalhasmovedslightlybeyondthezerovalue(seeFigure1610(c)).
Insomecases,thedifferentialgapisaresultofunintentionalfrictionandlostmotion;however,quite
oftenitisintentionallyprovidedinordertopreventtoofrequentoperationoftheonoffmechanism.
Figure1610:a)Blockdiagramofanonoffcontroller;b)blockdiagramofanon
offcontrollerwithdifferentialgap;c)outputversustimecurve.FromOgata(1990)
[137].
16.3.3 Proportional control action (p controller)

Foracontrollerwithproportionalcontrolaction,therelationshipbetweentheoutputofthecontroller
u(t)andtheactuatingerrorsignale(t)is
u t K p et [1613]
435
ECSSEHB3101Part13A
5December2011
or,inLaplacetransformedquantities,
U s
Kp
E s
[1614]
whereKpiscalledtheproportionalgain.
In fact, a proportional controller is an amplifier with adjustable gain. A block diagram of such a
controllerisshowninFigure1611.
Figure1611:Blockdiagramofaproportionalcontroller.FromOgata(1990)[137].
16.3.4 Integral control action (i controller).

Inacontrollerwithintegralcontrolaction,thevalueofthecontrolleroutputu(t)ischangedatarate
proportionaltotheactuatingsignale(t).Thatis,
du t
K i et [1615]
dt
or
t
u t K i et dt [1616]
0
whereKiisanadjustableconstant.InLaplacesdomain,thetransferfunctionoftheintegralcontroller
is
U s K i

E s s
[1617]
436
ECSSEHB3101Part13A
5December2011
Itisimportanttopointoutthatwhenthevalueofe(t)isdoubled,then,thevalueofu(t)variestwiceas
fast. For zero actuating error, the value of u(t) remains stationary. The integral control action is
sometimescalledresetcontrol.Figure1612showsablockdiagramofsuchacontroller.
Figure1612:Blockdiagramofanintegralcontroller.FromOgata(1990)[137].
16.3.5 Proportional-integral control action (pi controller)

Thecontrolactionofaproportionalintegralcontrollerisdefinedbythefollowingequation:
t
Kp
u t K p et et dt [1618]
Ti 0
thus,thetransferfunctionofthecontrolleris
U s 1
K p 1
E s
[1619]
Ti s
whereKpistheproportionalgain,andTiiscalledtheintegraltime.TheinverseoftheintegraltimeTi
iscalledtheresetrate.Theresetrateisthenumberoftimesperminutethattheproportionalpartof
the control action is duplicated. Figure 1613(a) shows a block diagram of a proportionalintegral
controller.Iftheactuatingsignale(t)isaunitstepfunction,asshowninFigure1613(b),thecontroller
outputu(t)becomesasshowninFigure1613(c).
437
ECSSEHB3101Part13A
5December2011
Figure1613:a)Blockdiagramofaproportionalintegralcontroller;b)andc)
diagramsdepictaunitstepinputandthecontrolleroutput.FromOgata(1990)
[137].
16.3.6 Proportional-derivative control action (pd controller)

Thecontrolactionofaproportionalderivativecontrollerisdefinedbythefollowingequation:
det
u t K p et K pTd [1620]
dt
anditstransferfunctionis
U s
K p 1 Td s
E s
[1621]
whereKpistheproportionalgainandTdisaconstantcalledthederivativetime.Thederivativecontrol
action,sometimescalledratecontrol,iswherethemagnitudeofthecontrolleroutputisproportional
totherateofchangeoftheactuatingerrorsignal.ThederivativetimeTdisthetimeintervalbywhich
the rate action advances the effect of the proportional control action. Figure 1614(a) shows a block
diagram of a proportionalderivative controller. If the actuating error signal e(t) is a unitramp
function as shown in Figure 1614(b), then the controller output u(t) becomes as shown in Figure
1614(c). As may be seen from Figure 1614(c), the derivative control action has an anticipatory
character.
438
ECSSEHB3101Part13A
5December2011
Figure1614:a)Blockdiagramofaproportionalderivativecontroller;b)andc)
diagramsdepictaunitrampinputandthecontrolleroutput.FromOgata(1990)
[137].
While derivative control action has the advantage of being anticipatory, it has the disadvantages of
amplifying noise signals (thereby causing a saturation effect in the actuator) and of being effective
onlyduringtransientperiods.
16.3.7 Proportional-integral-derivative control action (pid

controller)
The combination of proportional control action, integral control action, and derivative control is
termed proportionalintegralderivative control action. This combined action has the advantages of
eachofthethreeindividualcontrolactions.Theequationofacontrollerwiththisactionis:
det
t
Kp
u t K p et et dt K T
p d [1622]
Ti 0
dt
thus,thecorrespondingtransferfunctionis
U s 1
K p 1 Td s
E s
[1623]
Ti s
where Kp is the proportional gain, Ti is the integral time, and Td is the derivative time. The block
diagramofaproportionalintegralderivativecontrollerisshowninFigure78(a).Ife(t)isaunitramp
function,asshowninFigure1615(b),thecontrolleroutputu(t)becomesasshowninFigure1615(c).
439
ECSSEHB3101Part13A
5December2011
Figure1615:a)Blockdiagramofaproportionalintegralderivativecontroller;b)
andc)diagramsdepictaunitrampandthecontrolleroutput.FromOgata(1990)
[137].
16.3.8 Summary
Inthisclausethesixbasiccontrolactionshavebeenpresented,andmodeledasblockdiagrams.Also,
theiradvantagesanddisadvantageshavebeenmentioned.
Table161summarisesthepreviousdata,inparticulartheequationofitsoutput,u(t),asfunctionof
theactuatingsignal,e(t),inthetimedomain,andthetransferfunctionintheLaplacesplane.
Table161:Summaryofbasiccontrolactions
Control Equation Transferfunction Block Parameters
Law diagram
ON/OFF u t U 1 , if et 0 NA Figure U1,U2

u t U 2 , if et 0 1610
P u t K p et U s Figure Kp
Kp
E s 1611
I t U s K i Figure Ki
u t K i et dt
E s s 1612
0
PI U s 1 Figure Kp,Ti
t
Kp
u t K p et et dt K p 1
E s
1613a
Ti 0 Ti s
440
ECSSEHB3101Part13A
5December2011
Control Equation Transferfunction Block Parameters

Law diagram
det U s
u t K p et K pTd K p 1 Td s
PD Figure Kp,Td

dt E s 1614a
PID t
U s 1 Figure Kp,Ti,Td
u t K p et Td et et dt
d 1
K p 1 Td s 1615a
dt Ti 0 E s Ti s
16.4 Implementation techniques of control laws
16.4.1 Introduction
Thisclausewillbefocusedonthecontrollerblockofacontrolsystem(seeFigure72(clause16.3.1)for
a typical block diagram). Different techniques of implementing the control actions (Paction,PD
action,PIactionandPIDaction)willbepresented.
Therecentevolutionofmicroprocessorsandmicrocomputerswhichcanbeusedforvariouscontrol
functions has established a new trend toward the inclusion of digital computers to implement the
control actions by signal processing and calculations. Before this situation, the control action was
commonly implemented by a combination of devices as nozzles, relays, valves, ..., which act
continuouslyintime,andareknownasanalogcontrollers.
Thecurrenttrendtowarddigitalratherthananalogcontrolofdynamicsystemsismainlyduetothe
availability of lowcost digital computers and the advantages found in working with digital signals
ratherthancontinuoustimesignals.Also,thedecisionmakingcapabilityandflexibilityinthecontrol
programareanadditionalmajoradvantageofdigitalcontrolsystems.
This clause will be focused on analog devices for implementing control actions, although a brief
descriptionofdigitalcontrolsystemswillbeincluded.(Inclause16.5computerscommonlyusedin
spacesystemtoimplementdigitalcontrollawswillbelistedanddiscussed).
16.4.1.1 Digital control systems

A control scheme in which a digital computer is included in a control loop to perform signal
processinginadesiredfashioniscalleddigitalcontrol.
Direct digital control of a process or plant has the following advantages over the corresponding
analogcontrol:
1. Dataprocessinginthedigitalcontrollerisstraightforward;complexcontrolcalculations
canbeperformedeasily.
2. Control programs (controller characteristics) can be changed easily if such changes are
needed.
3. Digital controllers are far superior to the corresponding analog controllers from the
viewpointofinternalnoiseanddrifteffects.
Themaindisadvantagesofdigitalcontrolare:
1. The sampling and quantizing processes tend to result in more errors, which degrade
systemperformance.
441
ECSSEHB3101Part13A
5December2011
2. Designing digital controllers to compensate for such degradation is more complex than
designinganalogcontrollersatanequivalentlevelofperformance.
Figure1616depictsablockdiagramofadigitalcontrolsystemshowingaconfigurationofthebasic
controlscheme.
Figure1616:Blockdiagramofadigitalcontrolsystem.FromOgata(1987)[135].
Figure1617showsablockdiagramofasimplifieddigitalcontrolsystem.Thecontrolleroperationis
controlled by the clock. In such a digital control system, some points of the system pass signals of
varying amplitude in either continuous time or discrete time, while other points pass signals in
numericalcode.
Figure1617:Blockdiagramofadigitalcontrolsystemshowingsignalsinbinary
orgraphicfrom.FromOgata(1987)[135].
Theoutputoftheplantisacontinuoustimesignal.Theerrorsignalisconvertedintodigitalformby
the sampleandhold circuit and the analogtodigital converter. The conversion is done at the
sampling time. The digital computer processes the sequences ofnumbers by means of analgorithm
and produces new sequences of numbers. At every sampling instant a coded number (usually a
442
ECSSEHB3101Part13A
5December2011
binary number consisting of eight or more binary digits) must be converted to a physical control
signal,whichisusuallyacontinuoustime,oranalog,signal.Thedigitaltoanalogconverterandthe
hold circuit convert the sequence of numbers in numerical code into a piecewise continuoustime
signal.Therealtimeclockinthecomputersynchronisestheevents.Theoutputoftheholdcircuit,a
continuoustime signal, is fed to the plant, either directly or through the actuator, to control its
dynamics.
16.4.1.2 Analog controllers

Most analog controllers use electricity or pressurised fluid as the power source. Analog controllers
maybeclassifiedaccordingtothekindofpoweremployedintheoperationasfollows:
Pneumaticcontrollers
Hydrauliccontrollers
Electroniccontrollers.
The type of controller to use should be decided based on the nature of the plant and the operating
conditions,includingsuchconsiderationsassafety,cost,availability,reliability,accuracy,weightand
size.
Beforedescribinghowcontrolactionsareimplemented,abriefdescriptionofdevicesispresentedand
theirwayofoperationanalysed.
16.4.2 Devices characterization
16.4.2.1 Pressure systems

ConsiderthepressuresystemshowninFigure1618.Thegasflowthroughtherestrictionisafunction
of the gas pressure difference pipo. Such a pressure system may be characterised in terms of a
resistanceandacapacitance.
Figure1618:Schematicdiagramofapressuresystem.FromOgata(1990)[137].
Thegasflowresistance,R,isdefinedas:
change in pressure difference d p

R [1624]
change in gas flow rate dq
andthecapacitanceofthepressurevessel,C,isdefinedas:
443
ECSSEHB3101Part13A
5December2011
change in gas stored dm

C [1625]
change in gas pressure dp
where
m=massofgasinsidethevessel,
p=gaspressure,
q=gasflowrate,beingq=dm/dt
With respect to the system shown in Figure 1618, only small deviations in the variables from their
respectivesteadystatevaluesarepermitted.Thus,letdefinethefollowingvariables:
pi=smallchangeinflowgaspressure.
po=smallchangeingaspressureinthevessel.
Forsmallvaluesofpiandpo,theresistanceRbecomesconstantandmaybewrittenas
pi p o
R [1626]
q
andforcapacitanceConehas
dpo pi po
C [1627]
dt R
whichcanbewrittenas
dpo
RC p o pi [1628]
dt
Ifpiandpoareconsideredtheinputandoutput,respectively,thenthetransferfunctionofthesystem
is
Po s 1

Pi s RCs 1
[1629]
whereRChasthedimensionoftimeandisthetimeconstantofthesystem(Ogata(1990)[137].
16.4.2.2 Valves
Consider the schematic diagram of a pneumatic actuating valve shown in Figure 1619. In the
following analysis, only small variations in the variables are considered, so that the model of the
pneumaticactuatingvalvecanbelinearized.
444
ECSSEHB3101Part13A
5December2011
Figure1619:Schematicdiagramofapneumaticactuatingvalve.FromOgata(1990)
[137].
Letpcandxbethesmallvariationinthecontrolpressureandthecorrespondingvalvedisplacement,
respectively.IftheforceduetothemassandviscousfrictioninFigure1619arenegligiblysmall,the
restoringforceduetochangesinpressureis
Apc kx [1630]
whereAistheareaofthediaphragmandkplaystheroleofaspringconstant.
InLaplacesdomain,thetransferfunctionbetweenxandpcthusbecomes
X s A
Kc
Pc s k
[1631]
whereX(s)=L[x]andPc(s)=L[pc],LbeingtheLaplacetransform.Ifqi,thechangeinflowthroughthe
pneumaticactuatingvalve,isproportionaltox,thechangeinthevalvestemdisplacement,then
Qi s
Kq
X s
[1632]
whereQi(s)=L[qi]andKqisaconstant.Thetransferfunctionbetweenqiandpcisthen
Qi s
Kc Kq Kv
Pc s
[1633]
whereKvisaconstant.
16.4.2.3 Dashpots
ThedashpotshowninFigure1620(a)actsasadifferentiatingelement.Supposeastepdisplacement
to the piston position x. Then the displacement y becomes equal to x momentarily because of the
445
ECSSEHB3101Part13A
5December2011
springforce.However,thefluidwillflowthroughtheresistanceRandthecylinderwillcomebackto
theoriginalposition.ThecurvesxversustandyversustareshowninFigure1620(b).
Figure1620:a)Dashpot;b)stepchangeinxandthecorrespondingchangeiny
plottedversust,c)blockdiagramofthedashpot.FromOgata(1990)[137].
Definethepressuresexistingontherightandleftsidesofthepistonasp1andp2respectively.Suppose
thattheinertiaforceinvolvedisnegligible.Thentheforceactingonthepistonmustbalancethespring
force.Thus
A p1 p 2 ky [1634]
where
A=pistonarea,
k=springconstant.
Theflowrateqisgivenby
p1 p 2
q [1635]
R
where
q=flowratethroughtherestriction,
R=resistancetoflowattherestriction.
Sincetheflowthroughtherestrictionduringdtsecondsmustequalthechangeinthemassoffluidto
theleftofthepistonduringthesamedtseconds,oneobtains
qdt A dx dy [1636]
446
ECSSEHB3101Part13A
5December2011
where isthedensity(ithasbeenassumedthatthefluidisincompressible).Thislastequationcanbe
rewrittenas
dx dy q p p2 ky
1 [1637]
dt dt A RA RA 2
or
dx dy ky
[1638]
dt dt RA 2
Taking the Laplace transforms of both sides of this last equation, assuming zero initial conditions,
then
sX s sY s Y s
ky
[1639]
RA 2
Thetransferfunctionofthissystemthusbecomes
Y s s

X s k [1640]
s
RA 2
LetusdefineT=RA2/kthen
Y s Ts 1

X s Ts 1 1 [1641]
1
Ts
Figure1620(c)showsablockdiagramrepresentationforthissystem.
Writingtheinverseoftherelation(16.418a)
X s Ts 1 1
1
Y s
[1642]
Ts Ts
which,infact,isthetransferfunctioncorrespondingtoaproportionalintegralcontrolaction
16.4.3 Analog-controller implementation techniques

Thisclausedescribesdifferenttechniquesofimplementingthecontrollawsthroughphysicaldevices
(analogcontrol).
447
ECSSEHB3101Part13A
5December2011
16.4.3.1 Proportional control actions
16.4.3.1.1 Pneumatic force-distance p controller

Figure 1621(a) shows a schematic diagram of such a proportional controller. The nozzle flapper
amplifier constitutes the firststage amplifier, and the nozzle back pressure is controlled by the
nozzleflapper distance. The relaytype amplifier constitutes the secondstage amplifier. The nozzle
backpressuredeterminesthepositionofthediaphragmvalveforthesecondstageamplifier,whichis
capableofhandlingalargequantityoffluidflow.
Figure1621:a)Schematicdiagramofaforcedistancetypepneumaticproportional
controller;b)blockdiagram;c)simplifiedblockdiagram.FromOgata(1990)[137].
In most pneumatic controllers, some type of pneumatic feedback is employed. Feedback of the
pneumatic output reduces the amount of actual movement of the flapper. Instead of mounting the
flapperonafixedpoint,itisoftenpivotedonthefeedbackbellows,asshowninFigure1621(a).The
amountoffeedbackcanberegulatedbyintroducingavariablelinkagebetweenthefeedbackbellows
andtheflapperconnectingpoint.Theflapperthenbecomesafloatinglink.Itcanbemovedbyboth
theerrorsignalandthefeedbacksignal.
The operation of the controller shown in Figure 1621(a) is as follows. The input signal to the two
stagepneumaticamplifieristheactuatingerrorsignal.Increasingtheactuatingerrorsignalmovesthe
flappertotheleft.Thismovewill,inturn,increasethenozzlebackpressure,andthediaphragmvalve
moves downward. This results in an increase of the control pressure. This increase will cause the
bellows F to expand and move the flapper to the right, thus opening the nozzle. Because of this
feedback,thenozzleflapperdisplacementisverysmall,butthechangeinthecontrolpressurecanbe
large.
Assumingthatallvariationsinthevariablesarewithinalinearrange,ablockdiagramforthesystem
isshowninFigure1621(b).Thetransferfunctionbetweenpc(t)ande(t)isgivenby
448
ECSSEHB3101Part13A
5December2011
a
Pc s
K
a b Kp
E s
[1643]
a A
1 K
a b Ks
being Pc(s) and E(s) the Laplace transforms of pc(t) and e(t). Equation [1643] can be simplified as
showninFigure1621(c).
16.4.3.1.2 Pneumatic force-balance p controller

Figure 1622 represents a forcebalance pneumatic proportional controller. The basic principle of
operationdoesnotdifferfromthatoftheforcedistancecontroller.Themainadvantageoftheforce
balancecontrolleristhatiteliminatesmanymechanicallinkagesandpivotjoints,therebyreducingthe
effectsoffriction.
Figure1622:Schematicdiagramofaforcebalancepneumaticproportional
controller.FromOgata(1990)[137].
InthecontrollershowninFigure1622,thereferenceinputpressurePrandtheoutputpressurePoare
fed to large diaphragm chambers. Note that a forcebalance pneumatic controller operates only on
pressure signals. Therefore, it is necessary to convert the reference input and system output to
correspondingpressuresignals.
ThetransferfunctionbetweenPcandPe=PrPois
Pc s A2 A1
Kp
Pe s
[1644]
A1
16.4.3.1.3 Hydraulic p controller

ConsiderthehydrauliccontrollerinFigure1623.
449
ECSSEHB3101Part13A
5December2011
Figure1623:a)Servomoterthatactsasaproportionalcontroller;b)blockdiagram
oftheservomotor.FromOgata(1990)[137].
TheleftsideofthepilotvalveisjoinedtotheleftsideofthepowerpistonbyalinkABC.Thislinkisa
floatinglinkratherthanonemovingaboutafixedpivot.
Thecontrolleroperatesinthefollowingway.Ifinputxmovesthepilotvalvetotheright,portIIwill
beuncoveredandhighpressurefluidwillflowthroughitintotherightsideofthepowerpistonand
forcethispistontotheleft.Thepowerpiston,inmovingtotheleft,willcarrythefeedbacklinkABC
with it, thereby moving the pilot valve to the left. This action continues until the pilot piston again
coversportsIandII.AblockdiagramofthesystemisshowninFigure1623(b).Thetransferfunction
betweenY(s)andE(s)isgivenby
dmG dmL
0 [1645]
dx dx
Notice that under normal operating conditions the relation |Ka/[s(a+b)]| > 1 holds, so that the last
equationcanbesimplifiedto
Y s b
Kp
E s a
[1646]
The transfer function between y and x becomes a constant. Thus, the hydraulic controller shown in
Figure1623(a)actsasaproportionalcontroller,thegainofwhichisKp.Thisgaincanbeadjustedby
effectivelychangingtheleverratiob/a.
16.4.3.2 Proportional-derivative control actions
16.4.3.2.1 Pneumatic pd controller

ConsiderthepneumaticcontrollershowninFigure1624(a).Assumingsmallchangesintheactuating
error,nozzleflapperdistance,andcontrolpressure,theoperationofthiscontrollercanbedescribed
asfollows:Letfirstassumeasmallstepchangeine.Thenthechangeinthecontrolpressurepcwillbe
instantaneous. The restriction R will momentarily prevent the feedback bellows from sensing the
pressure change pc. Thus the feedback bellows will not respond momentarily, and the pneumatic
actuatingvalvewillfeelthefulleffectofthemovementoftheflapper.Astimegoeson,thefeedback
bellows will expand or contract. The change in the nozzleflapper distance x and the change in the
control pressure pc, can be plotted against time t, as shown in Figure 1624(b). At steady state, the
450
ECSSEHB3101Part13A
5December2011
feedbackbellowsactslikeanordinaryfeedbackmechanism.Thecurvepcversustclearlyshowsthat
thiscontrollerisoftheproportionalderivativetype.
Figure1624:a)Pneumaticproportionalderivativecontroller;b)stepchangeine
andthecorrespondingchangesinxandpcplottedversust;c)blockdiagram.From
Ogata(1990)[137].
AblockdiagramcorrespondingtothispneumaticcontrollerisshowninFigure1624(c).Intheblock
diagram,Kisaconstant,Aistheareaofthebellows,andksistheequivalentspringconstantofthe
bellows.Thetransferfunctionbetweenpcandecanbeobtainedfromtheblockdiagramasfollows:
b
Pc s
K
a b
E s
[1647]
Ka A 1
1
a b K s RCs 1
InsuchacontrollertheloopgainKaA/[(a+b)ks(RCs+1)]isnormallyverymuchgreaterthanunity,thus
thetransferfunctionPc(s)/E(s)canbesimplifiedtogive
Pc s
K p 1 Td s
E s
[1648]
where
bk s
Kp , Td RC [1649]
aA
Thus,delayednegativefeedback,orthetransferfunction1/(RCs+1)inthefeedbackpath,modifiesthe
proportionalcontrollertoaproportionalderivativecontroller.
451
ECSSEHB3101Part13A
5December2011
16.4.3.2.2 Hydraulic pd controller
Figure 1625(a) depicts a schematic diagram of a hydraulic proportionalderivative controller. The
cylinders are fixed and the pistons are able to move. For this system, a block diagram is shown in
Figure1625(b).
Figure1625:a)Sketchofahydraulicproportionalderivativecontroller;b)block
diagram.FromOgata(1990)[137].
Byanalysingthemechanismcanbestatedthat
k y z A p 2 p1 [1650]
and,thepressurelossthroughtheresistanceRis
p 2 p1
q [1651]
R
beingtheflowmassrate,q,
dz
q A [1652]
dt
Hence,
RA 2 dz
yz [1653]
k dt
and,definingthetimeconstantTas
RA 2
T [1654]
k
so
452
ECSSEHB3101Part13A
5December2011
dz
y z T [1655]
dt
whichimpliesthatinLaplacedomainthetransferfunctionis
Z s 1

Y s Ts 1
[1656]
Taken into account the feedback branch, and the block diagram in Figure 1625(b), the transfer
functionY(s)/E(s)is:
b K
Y s ab s

E s
[1657]
a K 1
1
a b s Ts 1
and,becauseusually|aK/[(a+b)s(Ts+1)]|>>1,then
Y s
K p 1 Ts
E s
[1658]
beingKp=b/a.
Thus,thecontrollerinFigure1625(a)actsasaproportionalderivativecontroller.
16.4.3.3 Integral control actions
16.4.3.3.1 Hydraulic i controller

ThehydraulicservomotorshowninFigure1626isessentiallyapilotvalvehydraulicpoweramplifier
andactuator.Thepilotvalveisabalancedvalveinthesensethatthepressureforcesactingonitare
allbalanced.
453
ECSSEHB3101Part13A
5December2011
Figure1626:Hydraulicintegralcontroller.FromOgata(1990)[137].
The operation procedure is as follows. If input x moves the pilot valve to the right, port II is
uncovered, and so highpressure fluid enters the right side of the power piston. Since port I is
connectedtothedrainport,thefluidintheleftsideofthepowerpistonisreturnedtothedrain.The
fluid flowing into the power cylinder is at high pressure; the fluid flowing out from the power
cylinder into the drain is at low pressure. The resulting difference in pressure on both sides of the
powerpistonwillcauseittomovetotheleft.
Byanalysingthesystemperformance,thetransferfunctionis
Y s K K
1
X s As s
[1659]
where K = K1/(A), being A the piston area, , the fluid (incompressible) density, and K1 the
proportionalityconstantbetweenthefluidflowrate,q,andthepilotvalvedisplacement,x
q K1 x [1660]
Thus,thesimplehydraulicservomotorshowninFigure1625(a)actsasanintegralcontroller.
16.4.3.4 Proportional-integral control actions
16.4.3.4.1 Pneumatic pi controller

Consider the pneumatic controller shown in Figure 1627(a). The operation of this controller is as
follows:ThebellowsdenotedbyIisconnectedtothecontrolpressuresourcewithoutanyrestriction.
ThebellowsdenotedbyIIisconnectedtothecontrolpressuresourcethrougharestrictionR.Assume
454
ECSSEHB3101Part13A
5December2011
asmallstepchangeintheactuatingerror.Thiswillcausethebackpressureinthenozzletochange
instantaneously. Thus a change in the control pressure pc also occurs instantaneously. Due to the
restrictionofthevalveinthepathtobellowsII,therewillbeapressuredropacrossthevalve.Astime
goeson,airwillflowacrossthevalveinsuchawaythatthechangeinpressureinbellowsIIattains
the value pc. Thus bellows II will expand or contract as time elapses in such a way as to move the
flapperanadditionalamountinthedirectionoftheoriginaldisplacemente.Thiswillcausetheback
pressurepcinthenozzletochangecontinuously,asshowninFigure1627(b).
Figure1627:a)Pneumaticproportionalintegralcontroller;b)stepchangeineand
thecorrespondingchangesinxandpcplottedversust;c)blockdiagramonthe
controller;simplifiedblockdiagram.FromOgata(1990)[137].
Ablockdiagramofthiscontrollerundertheassumptionofsmallvariationsinthevariablesisshown
inFigure1627(c).AsimplificationofthisblockdiagramyieldsFigure1627(d).Thetransferfunction
ofthiscontrolleris
a
Pc s
K
ab
E s
[1661]
a A 1
1 K 1
a b k s RCs 1
455
ECSSEHB3101Part13A
5December2011
where K is a constant, A is the area of the bellows, and ks is the equivalent spring constant of the
combinedbellows.If|KaARCs/[(a+b)ks(RCs+1)]|>1whichisusuallythecase,thetransferfunctioncan
besimplifiedto
Pc s 1
K p 1
E s
[1662]
Ti s
where
dpo pi po
C [1663]
dt R
Transferfunction[1662]representsaproportionalintegralcontrolaction.
16.4.3.4.2 Hydraulic pi controller

Figure 1628(a) shows a schematic diagram or a hydraulic proportionalintegral controller. A block
diagramofthiscontrollerisshowninFigure1628(b).
Figure1628:a)Schematicdiagramofhydraulicproportionalintegralcontroller;b)
blockdiagram.FromOgata(1990)[137].
ThetransferfunctionY(3s)/E(s)isgivenby
b K
Y s ab s

E s 1 ka T
[1664]
a b Ts 1
Insuchacontroller,undernormaloperation|KaT/[(a+b)(Ts+1)]|>1withtheresultthat
Y s 1
K p 1
E s
[1665]
Ti s
where
456
ECSSEHB3101Part13A
5December2011
b RA 2
Kp , Ti T [1666]
a k
ThusthecontrollershowninFigure1628(a)isaPIcontroller.
16.4.3.5 Proportional-integral-derivative control actions
16.4.3.5.1 Pneumatic pid controller

A combination of controllers shown in Figure 1624(a) and Figure 1625(a) yields the PIDcontroller
shown schematically in Figure 1629(a), whose corresponding block diagram is depicted in Figure
1629(b).
Figure1629:a)Pneumaticproportionalintegralderivativecontroller;b)block
diagramofthecontroller.FromOgata(1990)[137].
Undertheassumptionofsmallperturbations,thetransferfunctionis
bK
Pc s ab

E s
1 K
a A Ri C Rd C s
[1667]
a b k s Rd C s 1Ri C s 1
Bydefining
Ti Ri C , Td Rd C [1668]
457
ECSSEHB3101Part13A
5December2011
andnotingthatundernormaloperation
KaATi Td s /a b k s Td s 1Ti s 1 1 and Ti Td [1669]
andTi>>Td,equation[1662]reducesto
Pc s 1
K p 1 Td s
E s
[1670]
Ti s
with
bk s
Kp [1671]
aA
Equation[1670]indicatesthatthesysteminFigure1629(a)isaPIDcontroller.
16.4.4 Summary
Table 162 summarises the theory presented in previous clauses. It also includes the main physical
parameters and how the parameters in the transfer function can be calculated based on its physical
characteristics.
Table162:SummaryofImplementationTechniques
Control Analog Clause Parameters Functional Block
Action Device Scheme Diagram
P Force 12.4.3.1.1 b / a ab A Figure Figure

Distance,
Kp
1621a 1621b
a ks
1 K
Pneumatic ab
a,b:levergeometry
A:bellowsarea
K:ratiobetweencontrolpressureand
nozzleflapperdistance
ks:bellowsequivalentspringconstant
Force 12.4.3.1.2 A2 A1 Figure

Balance, Kp 1622
A1
Pneumatic A2:areaofthemiddlediaphragm
A1:areaoftheupperandlower
diaphragms
Hydraulic 12.4.3.1.3 b Figure Figure

Kp
1623a 1623b
a
b,a:levergeometry
458
ECSSEHB3101Part13A
5December2011

PD Pneumatic 12.4.3.2.1 bk s Figure Figure

Kp , Td RC 1624a 1624c
aA
b,a:levergeometry
R:valveresistance
C:bellowcapacitance
ks:bellowequivalentspringconstant
A:bellowarea
Hydraulic 12.4.3.2.4 b RA 2 Figure Figure

Kp , Td 1625a 1625b
a k
:fluiddensity
A:pistonarea
R:valveresistance
k:springconstant
b,a:levergeometry
I Hydraulic 12.4.3.2.3 b RA 2 Figure

Kp , Ti 1626
a k
b,a:levergeometry
k:springconstant
A:pistonarea
:fluiddensity
R:valveresistance
PI Pneumatic 12.4.3.2.2 bk s Figure Figure

Kp , Ti RC 1627a 1627c
aA
b,a:levergeometry
R:valveresistance
C:bellowcapacitance
A:bellowarea
Hydraulic 12.4.3.2.4 b RA 2 Figure Figure

Kp , Ti 1628a 1628b
a k
b,a:levergeometry
k:springconstant
A:pistonarea
:fluiddensity
R:valveresistance
PID Pneumatic 12.4.3.3.1 bk s Figure Figure

Kp , Td Rd C , Ti Ri C 1629a 1629b
aA
b,a:levergeometry
Ri>>Rd
Ri:circuit2valveresistance
Rd:circuit2valveresistance
459
ECSSEHB3101Part13A
5December2011

C:bellowcapacitance
A:bellowarea
16.5 Hardware description
16.5.1 Introduction
This clause covers a description of the hardware generally used in fluid loops existing in spacecraft
control systems. Hardware to be described is classified from the point of view of its function in the
controlloop.Figure1630depictsatypicalblockdiagramofafluidloopmodelledasacontrolsystem.
Figure1630:Fluidloopmodelledasacontrolsystem.
TheaimofthisclauseistoenumeratetheelementsinsidethedashedsquareinFigure1630,which
aretheelementscapableofbeingadaptedoncetheplantdesignhasbeenclosed.
Asensorisadevicewhichconvertsaninputsignalintoanoutputsignalinaformatadequatetobe
understoodbythecontroller;forexample,adevicewhichmeasurestemperature,convertingitintoa
voltageoutput.
Theactuatorrespondstothecontrollersdemandandgivesanoutputwhichisamechanicalvariable
asposition,velocity,acceleration,...
Asithasalreadybeenmentioned,thelowcostofdigitalcomputersandtheadvantagesofhandling
digitalsignalsarethemainreasonsfortheuseofcomputersasthecontrollerkernel.
Fluidloopsareusuallydesignedandsizedforoperatingunderspecificenvironmentalconditionsand
component properties (coatings emittance, absorptance, ...). During operative life environmental
conditionschangeanddegradationofcomponentsoccurs;thence,offdesignsubsystemoperationis
animportanttasktobetakenintoaccount(Sadunasetal.(1986).Inadigitalcomputerthecontrollaw
can be programmed with parameters that can adapt to the actual working conditions in order to
optimise thesubsystem performance.Figure93depicts the control unit blockdiagram for the fluid
loop used for cooling Spacelab Experiments (Microtecnia (1977) [126]) in which the controller is a
computer,withinterfacestoadaptthesignalstothoserequiredbytheactuators.
460
ECSSEHB3101Part13A
5December2011
Figure1631:ControlblockdiagramofthefluidloopforcoolingSpacelab
experiments.FromMicrotecnia(1977)[126].
Due to the new trend toward the inclusion of digital computers as controllers, and since analog
controllershavebeenstudiedinclause16.4,clause16.5.2focusesondigitalcomputerswhichcanbe
usedascontrollers.
Regardingsensorsandactuators,awidevarietyofcombinationscanbechosen.Theywillbeanalysed
in clause 0 and clause 16.5.4 respectively. Figure 1632 shows two typical fluid loop configurations
(alreadypresentedinclause14);inafirstapproach,onlythedeviceshereusedwillbeanalysed.
461
ECSSEHB3101Part13A
5December2011
Figure1632:Instrumentationandcontrolsystemschematics.FromSadunasetal.
(1986)[154].
16.5.2 Controllers
16.5.2.1 Digital/analog controllers trade/off

Thegeneralrequirementsforthefluidloopcontrolunit(seeFigure1632)canbedescribedasfollows:
(a) Toreceiveelectricalsignals,correspondingtosensedphysicalquantitiesfromfluid
loopsensors.
(b) To receive commands from a remote processor or from the Ground operation
center.
(c) To process both commands and signals and generate actuating commands, to be
used by the fluid loop regulating elements in keeping the controlled variables
withinthespecifiedband.
(d) Toprovideadequateflexibilityformissiontomissionanddynamicvariations.
Theanalogcontrollermaybepneumatic/hydraulic(seeclause16.4)orelectronic.Anelectronicanalog
controllerismadeupofanalogcircuitryandmodulessuchasamplifiers,integrators,compensating
networks, analog functional modules and similar, which perform, alone or as a whole, the required
control functions and laws. A minimum of digital circuitry is also included in order to fulfil the
requirementofconfigurationselection.
462
ECSSEHB3101Part13A
5December2011
A digital controller is built around a central processing unit (CPU) or microprocessor (MPU) which
can process both numerical and logical functions, according to a preselected program. Auxiliary
hardwareconsistsofmemories,input/outputdevices,timingcircuits,...
The comparison of the main features offered by either solution is shown in Table 163. Eleven
evaluationcriteriaareusedinordertoassessthemeritfigureofeachsolution.
Table163:Controlunitphilosophytradeoff.FromMicrotecnia(1977)[126].
EVALUATIONITEM ANALOGCONTROLLER DIGITALCONTROLLER
1.Flexibility Extensivehardwaremodification Dedicatedmemorycardscan

1.1.Missionconfiguration requiredprovided beeasily
changes
1.2.Payloadconfiguration Extensivehardwareduplication Duplicationislimitedto

changeswithinthesamesession required dedicatedmemorycards
1.3.Equipmentflexibilityduring Verypoor.Hardwarerework Verygood(1).Onlysoftware

development required needstobechanged
1.4.Possibilityofmultimodal Verylimited,unlesscomplex Verygood.Instructionby

operation hardwarearrangementprovided softwarerequired
2.Growthcapability Poor Poor
3.Capabilityofadditionalsignals Poor.Additionalcircuitry Verygood(2)

processing required
4.Selfmonitoringcapability Limited Verygood
5.EaseofinterfacewithCMDs Fair.RequirescommandsD/A Good.Noconversionrequired.

conversion.Signaltransmission Digitalsignaltransmissionis
isanalog possible
6.EaseofinterfacewithECL Goorforanalogsignals.Fairfor RequiresA/Dconversionfor

sensors discretessignals analogsignals.Discretesignals
canbereadilyaccepted
7.Easeofinterfacewithactuating Fair(3) Fair(3)

elements
8.Resistancetotheenvironment Good Verygood
9.Usageofcriticalcomponents Severalcriticalcomponentsare Veryfewcriticalcomponents

foreseen areforeseen
10.Possibilityofintroducing Fair Good

redundancies
11.Needforauxiliarycircuitry Auxiliarycircuitryrequired Noauxiliarycircuitryrequired

NOTE Duringthedevelopmentphase,erasablememorymaybeused.
(1)
Providedthatthecomputingtimeandaddressablememorysizelimitationsarenotexceeded.
(2)
Dependsuponactuatorrequirements.
(3)
463
ECSSEHB3101Part13A
5December2011
16.5.2.2 Digital controllers
Astheresultofthistradeoffanalysis,itseemsthatthedigitalcontrolleroffers,generally,advantages
over the analog solution in the areas which are more important for the fluid loop applications
(Microtecnia(1977)[126]).Forthisreason,thisclauseisconcentratedontheexistingdigitaloptionfor
thecontroller.
Theavailabilityofsuitable,qualifiedhardwarehasbeenassessedbymeansofaliteraturesurvey,and
lookingfortheselectionadoptedinexistingsatellites.ThesecomputersaredescribedinTable164in
terms of CPU, memory I/O system, technology, dimensions, weight, resistance to vibrations and
temperature.AndTable93showsotherspacequalifiedCPUs.
464
ECSSEHB3101Part13A
5December2011
Table164:Spacecomputers
Name CPU Memory Power Temp. Envir. Weight Sizex103 Manufacturer
[bits] [W] [K] [kg] [m] Supplier
DACS 8086 RAM:64K 35 220 vib:14grms 10,4 217,4x168,3x266,7 SCITechnology

(8087) UVROM: to to INC
5MHz 256Kto1M 80 340
SC1E 80c86 RAM:1,5M (15) 240 vib:15grms (5.3) (454,7x250,0x76,2) SwRISouthwesternResearch

8087 EEPROM:1M 45 to 10,7 454,7x250,0x193,0 Institute
5MHz UVROM:1M 330
SC2 80c86 RAM:1Mor 30 250 vib:NA 10,8 355,6x266,7x139,7 SwRISouthwesternResearch

8087 EEPROM:1M to Institute
8MHz orEEPROM+ 350
UVROM:0,5M
Bubbleme.:4M
GSVC Honey Upto24M 15 NA vib:NA 5,4 203,2x177,8x114,3 Honeywell

well
1750A
465
ECSSEHB3101Part13A
5December2011
Table165:Availablemicroprocessoroptions
CPU Supplier Technology Bit Remarks

No.
8080A INTEL NMOS 8 Extendedtemperaturerange.ManufacturedaccordingtoMILM38510.

ScreeningperMILSTD883,levelB
PPS48 ROCKWELL PMOS 416 Extendedtemperaturerange.ScreeningperMILSTD883,levelB

p16 ROCKWELL PMOS
Z80 ZYLOG NMOS 8 Extendedtemperaturerange.ScreeningperMILSTD883,levelB
SPB9900 TEXAS I2L 16 Extendedtemperaturerange.Needsonlyonesupplyvoltage;wide

INST. tolerancetovoltagevariation.Lowpowerconsumption
IM6100 INTERSIL/ CMOS 12 Extendedtemperaturerange

HARRIS
F100 FERRANTI BIPOLAR 16 Extendedtemperaturerange.

QualificationperBS9000
80c31 MATRA/ CMOS 8 Lowpowerconsumption.Velocity12MHz

HARRIS
80386 INTEL CHMOS 32 Highvelocity.Highperformance.Lowcost
466
ECSSEHB3101Part13A
5December2011
16.5.3 Sensors
16.5.3.1 Effects of the sensor on system performance

Sincethedynamicandstaticcharacteristicsofthesensor,ormeasuringelement,affecttheindication
of the actual value of the output variable, the sensor plays an important role in determining the
overallperformanceofthecontrolsystem.Thesensorusuallydeterminesthetransferfunctioninthe
feedback path. If the time constants of a sensor are negligibly small compared with other time
constantsofthecontrolsystem,thetransferfunctionofthesensorsimplybecomesaconstant.Figure
1633 shows block diagrams of automatic controllers having a firstorder sensor, an overdamped
secondordersensor,andanunderdampedsecondordersensor.Theresponseofathermalsensoris
oftenoftheoverdampedsecondordertype.
Figure1633:Blockdiagramsofautomaticcontrollerswitha)firstordersensor;b)
overdampedsecondordersensor;c)underdampedsecondordersensor.From
Ogata(1990)[137].
Ablockdiagramofasimpleautomaticcontrolsystemmaybeobtainedbyconnectingtheplanttothe
automaticcontroller,asshowninFigure1634.Feedbackoftheoutputsignalisaccomplishedbythe
sensor. Then, equation relating the output variable C(s) to the reference input R(s) and disturbance
variableN(s)isgivenby
Figure1634:Blockdiagramofacontrolsystem.
467
ECSSEHB3101Part13A
5December2011
G1 s G2 s
C s R s
1 G1 s G2 s H s

G3 s
[1672]
N s
1 G1 s G2 s H s

16.5.3.2 Temperature sensors

Thetemperaturesensormustbemorethanjustsomethingwhichrespondstoatemperaturechange.
Its output must be capable of interpretation and measurement as simply as possible. Table 166
presentscharacteristicsofseveraltemperaturessensors.
Table166:Characteristicsofseveraltemperaturesensors.
Type Temperature Accuracy Linearity Repeatability Response PartialListof
range[K] [K] Time Suppliers
Glass 145to590 0,05to1 Linear Excellent Good Brannon
Steam Thermometers
Thermometer Ltd.
Palmer
InstrumentInc.
Taylor
Instrument
Companies
Wesksler
InstrumentCorp.
Bimetallic 210to700 0,5to10 Linearover Fair Good W.M.ChoseCo.
Thermometer mostof CooperThomas
range Co.
Engelhard
IndustriesInc.
Resistance 15to1250 0,05 NiNon Excellent Good Honeywell
Thermometer linear IndustrialDiv.
Ni,Pt Ptlinear Fischer&Porter
Co.
Taylor
InstrumentCo.
Thermistor 170to590 0,05 Nonlinear Good Good AtkinsTechnical
Inc.
Carborundum
Co.
Thermistor
InstrumentsCo.
Thermocouple 10to3030 0,1 Kmost Good Excellent Foster
T,J,K,R&S linear InstrumentCo.
Jnearly Honeywell
linear IndustrialDiv.
468
ECSSEHB3101Part13A
5December2011
Type Temperature Accuracy Linearity Repeatability Response PartialListof

range[K] [K] Time Suppliers
Omega
EngineeringInc.
Integrated 215to420 0,3 Linear Excellent Excellent AnalogDevicer
circuit Inc.
temperature
transducer
16.5.3.3 Pressure sensor

Therearethreebasicmethodsforpressuremeasurement.Thesimplestmethodconsistsinbalancing
the unknown pressure against the pressure produced by a column of liquid of known density
(Manometers).Inthesecondmethodtheunknownpressureisallowedtoactonaknownareaandthe
resultantforceismeasuredeitherdirectlyorindirectly(Deadweighttesters).Andthethirdmethod,
the great majority of pressure sensors, allows the known pressure to act on a flexible member as a
Bourbontube,astackeddiaphragm,orabellowstosensethepressure;theappliedpressurecausesa
changeintheshapeofthesensorthatisusedtomoveapointerwithrespecttoascale.
Theapplicationofelectronictechniquestomeasurethedeflectionofadiaphragmandhencetoinfer
the pressure has resulted in major improvements in both sensitivity and resolution as well as
providingmeansforcompensatingfornonlineareffects.Oneofthedevicesinwhichthesetechniques
have been applied is the Capacitance manometer. The Piezoresistive pressure sensors utilise the
piezoresistive effect in order to measure the pressure. The characteristics of most pressure
transducersusedinfluidloopsarepresentedinTable167.
Table167:Characteristicsofpressuresensors.
Name Type Pressure Linearity Operation Response Supplier
range Temperature time
x103[Pa] Range[K] x103[s]
(10%to
90%)
OEM Piezo 02000 0,25% 230to400 1,0 ICSensors
1431 resistive
Kavlico Capacitive 103500 0,3% 260to330 DresserIndustries
Setra Capacitive 0,0170000 <0,1% 270to350 DresserIndustries
PS Quartz 05500 330to390 ParoscientificInc.
resonant
16.5.3.4 Flow sensors

Flowmeasurementisatechniqueusedinanyprocessrequiringthetransportofafluidfromonepoint
toanother.Dependingonthenatureoftheprocessstreamsinvolved,severaltypesofflowsensorscan
beconsidered:displacementflowmeters,magneticflowmeters,turbinetypeflowmetersandmetering
pumpsaregenerallyutilised.SomecharacteristicofthesesensorsareshowninTable168.

469
ECSSEHB3101Part13A
5December2011
Table168:Characteristicsofflowsensors.
Type Design Flowrangex Accuracy PartialSuppliers
pressure 103[m2.s1]
x103[Pa]
Magnetic upto2000 0,001to6000 0,5% BooksInstrumentDiv.

flowmeters TaylorInstrumentCompanies
Fisher&PorterCo.
Turbine upto20000 0,0001to2500 0,25% BooksInstrumentDiv.

flowmeter CoxInstrumentDiv.
DanielIndustriesInc.
Liquidpositive upto2000 0,001to1250 2% NeptuneMeterCo.

displacement BooksInstrumentDiv.
meter DressesIndustries
Metering upto10000 5to150 1% LappInsulatorCo.

pumps MecomaticInc.
(Diaphragm) YarwayCorp.
16.5.4 Actuators. Control valves

Acontrolvalveisavalvedesignedtomodifythefluidflowinpipesandusedforcontrolpurposesvia
anactuatorrespondingtoanexternalsignal.
Controlvalvesareusedwithanexternalcontrollerhavingtheflexibilityofvariouscontrolmodesand
withremotemanualoperation;theyarecompatiblewithanymeasuringsystem.
The purpose of this clause is to provide a summary of the more important characteristics of valves
usedinfluidloops,seeTable169.
Table169:Characteristicsofcontrolvalves.FromLiptak(1969)[115].
Valvebody Conventional Pinch PlugValve PlugValve PlugValve Tubeor
BallValve Valve Vported Adjustable Expanding Diaphragm
andCage Type Cylinder SeatPlate Type
Features
Valve Type Type Valve
Capacity 800 3750 1000 800 1000 750

2
106 (1)[m.s
d
]
1
Max/minOperating 17000/vacuum 1000/ 5000/ 800 5000 10000

Pressurex103[Pa] vacuum vacuum
Max/MinOperating 1250/20 450 500 475 475 330/250

Temperature[K]
Available(2) Over50:1 15:1 20:1 10:1 20:1 20:1

Rangeability
470
ECSSEHB3101Part13A
5December2011
Valvebody Conventional Pinch PlugValve PlugValve PlugValve Tubeor

BallValve Valve Vported Adjustable Expanding Diaphragm
andCage Type Cylinder SeatPlate Type
Features
Valve Type Type Valve
FlowCharacteristics Good Poor Good Fair Fair
AvailableSizeRangex 6350 3600 12600 10150 50400 25300

103[m]
Cost MediumHigh Low Low Medium Low Low

Medium
LowMaintenance Average No Yes Average Yes Average

NOTE Capacity:isthenumberof[m2.s1]of288Kwaterthatwillpassthroughthevalvewithapressure
(1)
dropof6,89KPa.
d:isthevalvesize.
Rangeability:theratioofmaximumtominimumcontrollableflows.Itestablishesthepointatwhich
(2)
theflowliftcharacteristicstartstodeviatefromtheexpected.
16.6 Control software

Themainobjectiveofthisclauseistoprovideanoverviewoncommonlyavailablesoftwareproducts
for analysing control system performances. After a brief historical note on computeraided control
systemsdesign(CACSD),varioustechniquesandtoolswillbereviewed.
Longbeforecomputerswereusedtodesigncontrolsystems,theywerebeingusedtosimulatethem.
Each simulation program consists of two primary segments: one is written to represent the
mathematicalmodelofthesystem,whiletheotheriswrittentodotheexperimentitself.Oneofthe
earliestsimulationprogramsisCSMP(ContinuousSystemsModellingProgram)whichwasapopular
simulation environment in the 1960s on various IBM main frame machines. The basic experiment
performed in most simulation runs has been the determination of the behaviour of the systems
trajectory under the influence of known signals such as step, sinusoidal, ... This experiment still
constitutesthemainfeatureofmanysimulationpackagesoftoday.
Improper data structure offered with early CAD (ComputerAidedDesign) packages was the main
defect.ThistrendwasbrokenbyMoler(1980)introducingamatrixmanipulationlaboratorysoftware
program called MATLAB. It provides a convenient environment for linear algebra and matrix
analysis. MATLAB has a rich collection of functions immediately useful to the control engineer or
systemtheorist.Controlsystemscanbemodelledastransferfunctionsorinstatespaceform,allowing
bothclassicalandmoderntechniquestobeused.Bothcontinuoustimeanddiscretetimesystems
arehandled.
FromMATLAB,severalCACSD(ComputerAidedControlSystemsDesign)packageandlanguages
werederived.Table1610showsalistofsixCACSDsoftwareprogramswhichhavebeenbasedonthe
originalMATLAB.

471
ECSSEHB3101Part13A
5December2011
Table1610:SomeMATLABDrivenCACSDSoftware.FromCellierandRimvall
(1988)[21]
SoftwareName LocationDeveloped,Year PrincipleDeveloper(s)
MATRIXx IntegratedSystems,Inc.,USA,1984 S.Shah,etal.
CRTLC SystemsControl,Inc.,USA,1984 J.N.Little,etal.
IMPACT SWISSFederalInstituteofTechnology, M.Rimvall,etal.

Switzerland,1983
CONTROL.lab UniversityofNewMexico,USA,1985 M.Jamshidi,etal.
PROMATLAB MathWorks,Inc.,USA,1985 J.N.Little,etal.
MATLABSC PhilipsResearchLaboratories,Germany,1985 M.VanbeginandP.van

Dooren
Parallelto,andevenbeforethediscoveryofMATLABbycontrolsystemsengineeringcommunity,a
number of nonMATLAB software programs were created both in North America and Western
Europe.AlistofeightnonMATLABCACSDprogramsaresummarisedinTable1611.
Table1611:SomenonMATLABCACSDSoftwarePackages.FromCellierand
Rimvall(1988)[21]
SoftwareName LocationDeveloped,Year PrincipleDeveloper(s)
KEDDC UniversityofBochum,Germany,1979 H.UnbehauenandChr.

(CADACS) Schmid
LUND LundInts.Technology,Sweden,1978 K.J.AstrmandH.

Elmqvist
LAS UniversityofIllinois,USA,1980 S.Bingulac,etal.
TIMDOM UniversityofNewMexico,USA,1983 M.Jamshidi,etal.
CC CaliforniaInst.ofTechnology,USA,1984 P.M.Thompson
TRIP DelftUniversityofTechnology,TheNetherlands, P.P.J.vandenBosch

1985
WCDS UniversityofWaterloo,Canada,1986 J.P.Aplevich
CATPAC PhilipsLaboratories,Germany,1986 D.Buenz

Itisestimatedthatover50differentpackagesofdifferentdegreesofdevelopmentexistintheworld
today.Table1612showsabriefsurveyof22CACSDpackages.Thistableisanabbreviatedversionof
a more extensive one of Cellier and Rimvall (1988) [21] with new additions from Jamshidi et et al.
[100]; 14 aspects of CACSD packages have been highlighted. The first six are considered as basic
attributes which should be available in all packages, the next five attributes are considered as more
advancedoneswhichmaynotbeavailableinallprograms;thefinalthreeattributesareamatterof
choiceonthepartofthedevelopersandpurchasersoftheparticularprogram.
472
ECSSEHB3101Part13A
5December2011
Table1612:ABriefSurveyof22CACSDPackages.FromJamshidietal.(1992)[100]
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
M C M P C C I M C L L T C T W I M P S S S S
A T A C A O M A A U A I C R C C A A A S U U
T R T T N P T D N S M I D A D A N P B N
L L R M P T A L A D D P R P S C A O S
A I A A R C A C O D E A A K P
B C X T C O T B S M S C D T
GLOBALCLASSIFICATION x L L (K C
A S E
B l C D
a D
b C)
ContinuousSystems 0 2 2 2 1 2 2 1 2 2 1 2 2 2 2 2 0 1 1 0 2 2
DiscreteSystems 0 2 2 2 1 2 2 1 2 2 1 2 2 2 2 0 2 0 1 2 0 2
TimeDomain 0 2 2 2 1 2 2 1 2 2 1 2 2 2 2 2 0 0 0 2 2 0
FrequencyDomain 0 2 2 2 1 1 2 0 2 2 1 0 2 2 2 2 2 2 1 0 0 2
SISO* 0 2 2 2 1 2 2 1 2 2 1 2 2 2 2 2 2 2 1 2 2 2
Multivariable 0 1 1 1 1 2 2 0 2 1 1 1 2 0 2 2 0 0 1 0 2 2
NonlinearSystems 0 2 2 0 1 1 1 0 1 1 1 0 0 1 0 0 0 0 2 1 0 2
AdaptiveControl 0 0 1 1 0 0 0 0 2 2 0 0 0 0 0 0 2 0 0 0 0 0
Identification 0 1 1 1 1 0 1 0 2 2 1 0 0 1 0 0 2 0 0 2 1 0
RealTimeInterface 0 0 2 0 0 0 0 0 2 2 0 0 0 0 0 0 0 0 0 0 0 0
Extendability 2 2 2 2 2 1 2 1 1 1 2 0 0 1 0 1 0 1 1 0 0 0
SourceCodeAvailability 2 0 0 1 0 2 0 0 1 1 0 2 0 2 0 0 2 1 1 0 2 0
Maintenance 0 2 2 2 1 0 0 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1
Cost 2 0 0 1 1 2 2 1 1 1 2 2 2 1 1 1 1 1 1 0 0 0
TotalNo.ofPoints 6 18 21 19 12 17 18 7 23 22 13 14 14 17 14 13 13 9 11 10 12 13
NOTE *SingleInputSingleOutput
473
ECSSEHB3101Part13A
5December2011
Among the MATLABbased programs, this particular survey reveals that MATRIXx is the most
completeCACSDpackage.
Among the nonMATLAB packages the CADACS(KEDDC) and LUND score very high. It is noted
thatamongallpackagesinTable163,CADACSofSchmid(1985)ranksonthetopCACSDprograms
availabletoday.
16.7 Existing systems
16.7.1 Space radiator system
16.7.1.1 General description

This example has been extracted from a paper by Baker et. al (1967) [3]. The study focuses on the
designoftheradiatorsystemforthevehiclecontrol/lifesupportsystem.Thevehicleconsideredisa
largemannedorbitalresearchlaboratory.Thesystemillustratestheproportionalcontrol,alongwith
conceptsassaturationsandadifferentialgap
Figure1635depictsthesystem,anditsmaincharacteristicsaresummarisedinTable1613.
Figure1635:Spaceradiatorsystem.FromBakeretal.(1967)[3].
474
ECSSEHB3101Part13A
5December2011
Table1613:Systemcharacteristic.FromBackeretal.(1967)[3]
Radiatorsystem
1. Coolant:Freon21at2000kPa.
2. Fluidininterfacingloop:water.Waterinlettemperaturetoheatexchanger=316KWateroutlet
temperaturefromheatexchanger=277K.
3. Designpointcircumferentialaveragesinktemperature=500K.
4. Freon21minimumpermissibletemperature=165K.
Interfaceheatexchanger
1. Counterflowtype
2. Plateandfinconstruction
3. AISI301stainlesssteel
m=7800kg.m3
K=14,4W.m1.K1
c=460J.kg1.K1
4. Heat exchanger headers and shell weight accounted for by the ratio 1,3 between total heat
exchangerweightandcoreweight.
Radiator
1. Circumferentiallengthfor1pass=20,7m.
2. Classicaltubeandfinconstruction
3. Tubewallthickness=2,54x103m.
4. Freon21.
5. 6061T6aluminiumalloy
Pump
1. Overallefficiency,=30%
SupplyandReturnPlumbing
1. Supplyandreturnplumbingeffectivelengths=18,3meach.
2. Tenbendsineachbranch.Pressurelosscoefficientforbends=0,7
3. Pressure loss based on length and bends multiplied by 2 to account for fittings, valves,
manifolds,andsoon.
4. 6061T6aluminiumalloy.
Sensor
1. Temperaturesensor.Timeconstant30s
Actuator
1. Bypassvalve.
Controller
1. ControlsFreonmixtemperatureinreturnplumbingto275Knominal.
2. Proportionalspeedcontroller:Deadband=0,5K.
3. Gain=1,8(sK)1(controlvalvepositionrateofchangeperFreonmixtemperatureerror).
475
ECSSEHB3101Part13A
5December2011
The316,5Kwaterinlettemperatureisindicativeofacceptableelectroniccoldplatetemperaturesand
the280Koutlettemperatureisindicativeofrequiredcabincondenserfluidinlettemperatures.The15
kW transferred to the Freon stream at the heat exchanger is indicative of expected environmental
control/lifesupportsystemheatloadsanditisconsideredasthedesignpoint.Thesystemincludesa
bypassvalveupstreamoftheradiatorandacontrollerwhichadjuststhevalvebypassflowsuchthat
themixedfluidtemperaturedownstreamoftheradiatorismaintainedatanominal275K.Thesystem
wasdesignedfornobypassflowwithsteadystateconditionsattheabovedesignheatload(15kW).
Forreducedthermalenvironmentalconditions,theradiatoroutlettemperatureisreducedduetothe
increased cooling effect of this environment. Initially the mixed radiator outlet temperature is also
reduced.Thecontrollerthenincreasesthebypassflowuntilthemixedoutlettemperatureapproaches
275 K. The reducedflowin the radiator resultsin even lower radiator outlettemperatures since the
heatrejectionrateisproportionaltotheproductofradiatorflowrateandradiatortemperaturedrop.
Fromthepointofviewofcontroltheorythesystemiscomprisedofthefollowingelements:
1. Plant.Itconsistsin:
o Interfaceheatexchanger
o Coolantpump
o Radiatorsupplyplumbing
o Radiatorreturnplumbing
o Bypassplumbing
2. Controller
3. Sensor:Temperaturesensor
4. Actuator:Bypassvalve.
Figure1636depictsablockdiagramofthesystem.
Figure1636:Spaceradiatorblockdiagram.FromBakeretal.(1967)[3].
Thetemperaturesensormeasuresthemixedradiatoroutlettemperatureandthecontrollerisrequired
to maintain it a 35 F despite variations in the system heat load, Q, and variation in the thermal
environment.Thesystemisdesignedforzeropositionofthebypassvalve(noflowthroughtheby
passpipe)atthenominalheatload.
476
ECSSEHB3101Part13A
5December2011
Theerrorsignaliscalculatedinsidethecontrollerandthen,asignalisgeneratedwhichdeterminesthe
positionrateofchange.Therefore,thecontrollerinputisanerrorintemperature(e(t)=TMIXTref)and
itsoutputisthespeedofthevalveanchor,u(t),asfollows:
u(t)=G.e(t),Gbeingthegain;inthiscase,G=1,8[s.K]1
The real control law is a proportional action with slight modifications: to prevent for an excessive
powerconsumptionadeadbandof0,5Khasbeenincluded,sothatwhenthetemperatureerroris
lessthan0,5Knooutputisgenerated.
Also,theoutputmustnotbegreaterthanthemaximumspeedwhichtheactuatorisabletoproduce,
andconstraintshavebeenaddedinthissense.
ThecontrollawimplementedinthecontrollerissummarisedinTable1614.
Table1614:ControllawfortheblockdiagraminFigure1636
Input Output Remarks
|e(t)|<1 u(t)=0 Deadband.Avoidexcessivepowerconsumption.
|G.e(t)|<UMAX u(t)=G.e(t) Proportionallaw
|G.e(t)|>UMAX u(t)=UMAX Physicalcontractonetomaximumspeedatvalve

actuator.
|G.e(t)|<UMAX u(t)=UMAX Physicalconstraintduetominimumspeedof

valveactuator.
477
ECSSEHB3101Part13A
5December2011
Bibliography
[1] Alario,J.,Haslett,R.,Kosson,R.,TheMonogrooveHighPerformanceHeatPipe,AIAA
PaperNo.811156,AIAA16thThermophysicsConference,PaloAlto,Calif.,June2325,
1981.AlsopublishedinSpacecraftRadiativeTransferandTemperatureControl,
ProgressinAstronauticsandAeronautics,Vol.83,M.Summerfield,Ed.,AIAA,New
York,1982,pp.305324.
[2] Almgren,D.W.,Fowle,A.A.,Bartoszek,J.T.,Ollendorf,S.,McIntosh,R.,TheUseofa
ThermalUtilitywithSpacePlatformMountedInstruments,AIAAPaperNo.811114,
AIAA16thThermophysicsConference,PaloAlto,Calif.,June2325,1981.
[3] Barker,R.S.,Stephens,H.M.,Taylor,B.N.,DesignandTransientPerformanceofaLiquid
CoolantSystem,SAEPaperNo.670838,SAEAeronautics&SpaceEngineeringMeeting,
LosAngeles,California,October28,1967.
[4] Barker,R.S.,Stephens,M.M.,Taylor,B.N.,DesignandTransientPerformanceofa
LiquidCoolantSystem,SAEPaperNo.670838,SAEAeronautics&SpaceEngineering
Meeting,LosAngeles,California,October28,1967.
[5] Baum,J.R.,ThermalDesignConsiderationsforPackagingElectronicEquipment,in
HandbookofElectronicPackaging,1sted.,C.A.Harper,Ed.,McGrawHillBook
Company,NewYork,1969,Chap.11,pp.143.
[6] Bellomo,M.,PerformanceoftheInterloopHeatExchangerfortheSpacelabT.C.L.,
MicrotecnicaS.p.A.,Torino,Italy,TRNo.352,June1975.
[7] Bergles,A.E.,RecentDevelopmentsinConvectiveHeatTransferAugmentation,Appl.
Mech.Rev.,Vol.26,No.6,June1973,pp.675682.
[8] Bergles,A.E.,SurveyandEvaluationofTechniquestoAugmentConvectiveHeatand
MassTransfer,inProgressinHeatandMassTransfer,1sted.,Vol.1,U.GrigullandE.
Hahne,Eds.,PergamonPress,Oxford,1969,pp.331424.
[9] Bergles,A.E.,Lee,R.A.,Mikic,B.B.,HeatTransferinRoughTubeswithTape
GeneratedSwirlFlow,TransactionsoftheASME,JournalofHeatTransfer,Vol.91,
SeriesC,No.3,August1969,pp.443445.
[10] Berner,F.,Oesch,M.,Goetz,K.,DevelopmentofanEngineeringModelDynamics
Cooler,F+WDocumentNo.5275/82/NL/PB(SL),1987.
[11] Berner,F.,Savage,C.J.,DevelopmentofaVaporCompressionHeatPumpforSpace
Use,AIAAPaperNo.811113,AIAA16thThermophysicsConference,PaloAlto,Calif.,
June2325,1981.
[12] Berner,F.,Scheider,M.,StudyofFluidSystemsforUseinSpacecraftThermalControl,
FWDocumentNoFO1332,1976.
478
ECSSEHB3101Part13A
5December2011
[13] Berner,F.,Schleicher,M.,StudyofFluidLoopSystemsforUseinSpacecraftThermal
Control,EidgenossischesFlugzeugwerkEmmen,Switzerland,F+WDocumentNo.FO
1332,preparedunderContractNo.2447/75forESA/ESTEC,October,1976.
[14] Berner,F.,Schleicher,M.,Oesch,H.,Nyfeler,H.,Steinhauser,H.,DesignStudyofan
ExperimentCoolingSystemwithLiquid/AirHeatExchangerforSpacelabPayloads,
EidgenossischesFlugzeugwerkEmmen,Switzerland,F+WDocumentNo.FO1378,
preparedunderContractNo.2886/76forESA/ESTEC,August,1977.
[15] Bolla,G.,DeGiorgio,G.,Pedrocchi,E.,HeatTransferandPressureDropComparisonin
TubeswithTransverseRibsandwithTwistedTape,EnergiaNucleare,Vol.20,No.11,
November1973,pp.604613.
[16] Boyko,L.D.,Kruzhilin,G.N.,HeatTransferandHydraulicResistanceDuring
CondensationofSteaminaHorizontalTubeandinaBundleofTubes,Int.J.HeatMass
Transfer,Vol.10,1967,pp.361373.
[17] Butterworth,D.,Condensers:BasicHeatTransferandFluidFlow,inHeatExchangers,
1sted.,S.Kakac,A.E.Bergles,F.Mayinger,Eds.,HemispherePublishingCorporation,
Washington,1981,pp.289313.
[18] Carlisle,R.F.,Nolan,M.,ApplicationofAdvancedTechnologytoaPermanently
MannedSpaceStation,SpaceTechnology,Vol.7,No.4,1987,pp.327336.
[19] Carnavos,T.C.,SomeRecentDevelopmentsinAugmentedHeatExchangers:Design
andTheorySourcebook,1sted.,N.AfganandE.U.Schlunder,Eds.,McGrawHillBook
Company,NewYork,1974,Clause21,pp.441489.
[20] Carson,M.A.,LiquidPumpforAstronautCooling,inNASATMX58106,pp.181185,
November1972,presentedatthe7thAerospaceMechanismsSymposium.Houston,
Texas,September78,1972.
[21] Cellier,F.E.,Rimvall,C.M.,ComputerAidedControlSystemsDesign:Techniquesand
Tools,inSystemsModelingandComputerSimulation,N.A.Kheir(ed.),AppendixA,
NewYork,MarcelandDekker,1988.
[22] Chalmers,D.R.,Pustay,J.J.,AdvancedCommunicationsSatelliteThermalDesignUsing
East/WestRadiators,AIAAPaperNo.860065,AIAA24rdAerospaceSciencesMeeting,
Reno,Nevada,Jan.69,1986.
[23] Chalmers,D.R.,Pustay,J.J.,Moy,C.B.,Kroliczek,E.J.,ApplicationforCapillaryPumped
LoopHeatTransportSystemstoLargeSpacecraft,AIAAPaperNo.861295,
AIAA/ASME4thJointThermophysicsandHeatTransferConference,Boston,Mass,June
24,1986.
[24] Chapter,J.J.,Johnsen,G.W.,ThermoelectricDeviceApplicationtoSpacecraftThermal
Control,AIAAPaperNo.73722,AIAA8thThermophysicsConference,PalmSprings,
California,July1618,1973.
[25] Chisholm,D.,ATheoreticalBasisfortheLockhartMartinelliCorrelationforTwoPhase
Flow,Int.J.HeatMassTransfer,Vol.10,1967,pp.17671778.
[26] Cichelly,M.T.,Boucher,D.F.,DesignofHeatExchangerHeadsforLowHoldup,Chem.
Engng.Progr.,Vol.52,No.5,May1956,pp.213218.
[27] Colburn,A.P.,King,W.J.,HeatTransferandPressureDropinEmpty,Baffled,and
PackedTubes.IIIRelationshipbetweenHeatTransferandPressureDrop,Ind.Engng.
Chem.,Vol.23,No.8,August1931,pp.919923.
479
ECSSEHB3101Part13A
5December2011
[28] Cowans,K.W.,ACountercurrentHeatExchangerthatCompensatesAutomaticallyfor
MaldistributionofFlowinParallelChannels,inAdvancesinCryogenicEngineering,
1sted.,Vol.19,K.D.Timmerhaus,Ed.,PlenumPress,NewYork,1974,pp.437444.
[29] Csanady,G.T.,TheoryofTurbomachines,1sted.,McGrawHillBookCompany,New
York,1964,Chap.1,pp.142.
[30] Delil,A.A.M.,ConsiderationsConcerningaThermalJointforaDeployableorSteerable
BatteryRadiatorfortheColumbusPolarPlatform,NLRTR86055U,1986.
[31] Delil,A.A.M.,FeasibilityDemonstrationofaSensorforHighQualityTwoPhaseFlow,
NLRTR87009U,Amsterdam,Jan.1987.
[32] Delil,A.A.M.,MoveableThermalJointsforDeployableorSteerableSpacecraftRadiator
Systems,NLRMP87016U,Amsterdam,Feb.1987.
[33] Delil,A.A.M.,SensorsforaSystemtoControltheLiquidFlowintoanEvaporativeCold
PlateofaTwoPhaseHeatTransportSystemforLargeSpacecraft,NLRTR86001U,
1986.
[34] Dexter,P.F.,Haskin,W.L.,AnalysisofHeatPumpAugmentedSystemsforSpacecraft
ThermalControl,AIAAPaperNo.841757,AIAA19thThermophysicsConference,
Snowmass,Colorado,June2528,1984.
[35] Dipprey,D.F.,Sabersky,R.H.,HeatandMomentumTransferinSmoothandRough
TubesatVariousPrandthNumbers,Int.J.HeatMassTransfer,Vol.6,No.5,May1963,
pp.329353.
[36] Domingos,J.D.,AnalysisofComplexAssembliesofHeatExchangers,Int.J.HeatMass
Transfer,Vol.12,No.5,May1969,pp.537548.
[37] DOWCORNING,InformationaboutSiliconeFluidProductsfromDowCorning,Dow
Corning,ChemicalProductsDivision,Midland,Michigan48641,Bulletin05061,
November1963.
[38] DOWCORNING,InformationaboutSiliconeFluids,DowCorningInternationalLTD,
ChaussedelaHulpe177,B1170,Brussels,Belgium,Bulletin22039a01June1972.
[39] Dukler,A.E.,WicksIII,M.,Cleveland,R.G.,FrictionalPressureDropinTwoPhase
Flow:AnApproachthroughSimilarityAnalysis,A.I.Ch.E.Journal,Vol.10,No.1,
January1964,pp.4451.
[40] Dunn,P.,Reay,D.A.,HeatPipes,1sted.,PergamonPress,Oxford,1976,Appendix1,
pp.263269.
[41] Ebersole,R.,BiosatelliteEnvironmentalControlCoolantLoopSystemDesign,in
ProceedingsoftheSymposiuminThermodynamicsandThermophysicsofSpace
Flight,PaloAlto,California,March21,1970,pp.133149.
[42] Ellis,W.E.,Rankin,J.G.,HeatBusesWillOperateLikeaPublicUtility,Astronautics&
Aeronautics,Vol.21,No.3,March19983,pp.5661.
[43] Engel,J.,Walter,H.,DesignofaSmallCentrifugalPumpforSubcooledandBoiling
Liquids(77<T<300K),FifthInternationalCryogenicEngineeringConference,Kyoto,
Japan,May710,1974,PaperNo.11.
[44] ESA,EnvironmentalandThermalControlSystemsforSpaceVehicles,ESASP200,
Toulouse,SessionIII,1983.
[45] ESANASA,SpacelabPayloadAccomodationHandbook,SLP2104,Iss.2,Rev.
14/AppendixC,Rev.5,1988.
480
ECSSEHB3101Part13A
5December2011
[46] ESDU66027,FrictionLossesforFullyDevelopedFlowinStraightPipes,Engineering
SciencesDataUnit,4HamiltonPlace,London,WI,September1966.
[47] ESDU67040,PressureLossesforIncompressibleFlowinSingleBends,Engineering
SciencesDataUnit,4HamiltonPlace,LondonWI,November1967.
[48] ESDU68006,ForcedConvectionHeatTransferinCircularTubes.PartII:Datafor
LaminarandTransitionalFlowsIncludingFreeConvectionEffects,EngineeringSciences
DataUnit,251259RegentStreet,LondonWI,February1968.
[49] ESDU68035,InteractionFactorsforCalculatingPressureLossesforIncompressibleFlow
inSomeCombinationsofTwoBendsinSeries,EngineeringSciencesDataUnit,251259
RegentStreet,LondonW1R7AD,July1968.
[50] ESDU69004,ConvectiveHeatTransferDuringForcedCrossflowofFluidovera
CircularCylinderIncludingFreeConvectionEffects,EngineeringSciencesDataUnit,
251259RegentStreet,London,W1R7AD,May1969.
[51] ESDU69022,PressureLossesinValves,EngineeringSciencesDataUnit,251259Regent
Street,London,W1R7AD,October1969.
[52] ESDU72009,PressureDropinDuctsacrossRoundWireGauzesNormaltotheFlow,
EngineeringSciencesDataUnit,251259RegentStreet,LondonW1R7AD,June1972.
[53] ESDU72010,PressureLossesacrossPerforatedPlates,OrificePlatesandCylindrical
TubeOrificesinDucts,EngineeringSciencesDataUnit,251259RegentStreet,London
W1R7AD,June1972.
[54] ESDU72011,FlowthroughaSuddenEnlargementofAreainaDuct,Engineering
SciencesDataUnit,251259RegentStreet,LondonW1R7AD,June1972.
[55] ESDU73022,PressureLossesinThreeLegPipeJunctions:DividingFlows,Engineering
SciencesDataUnit,251259RegentStreet,LondonW1R7AD,October1973.
[56] ESDU73023,PressureLossesinThreePipeJunctions:CombiningFlows,Engineering
SciencesDataUnit,251259RegentStreet,LondonW1R7AD,October1973.
[57] ESDU73031,ConvectiveHeatTransferDuringCrossflowofFluidsoverPlainTube
Banks,EngineeringSciencesDataUnit,251259RegentStreet,LondonW1R7AD,
November1973.
[58] ESDU77008,PressureLossesinCurvedDucts:SingleBends,EngineeringSciencesData
Unit,251259RegentStreet,LondonW1R7AD,May1977.
[59] ESDU77009,PressureLossesinCurvedDucts:InteractionFactorsforTwoBendsin
Series,EngineeringSciencesDataUnit,251259RegentStreet,LondonW1R7AD,May
1977.
[60] EURAMOPOMPESSALMSON,EuramoXA15R,102/7//06/77ed.1.EuramoMX,
102/8/09/77ed.1.DivisionElectroHydrauliqueLMT,4647,QuaiA.LeGallo,Boulogne
Billancourt,France,1977.
[61] Evans,L.B.,Churchill,S.W.,TheEffectofAxialPromotersonHeatTransferand
PressureDropInsideaTube,Chem.Engng.Progr.SymposiumSer.,Vol.59,No.41,
1963,pp.3646.
[62] Evans,S.I.,Sarjant,R.J.,HeatTransferandTurbulenceinGasesFlowingInsideTubes,J.
Inst.Fuel,Vol.24,No.5,September1951,pp.216227.
481
ECSSEHB3101Part13A
5December2011
[63] Filippi,F.,Arcidiacono,A.,DesignStudyofaLiquidHeatTransferLoopforThermal
ConditioningofSpacelabExperimentPayloads,MICROTECMICADocumentNo.RT
598,1977.
[64] Filippi,F.,Guerra,L.,ScreeningandSelectionofCoolantFluidsforE.C.L.,
Microtecnica,Turin,Italy,TRNo584,preparedunderESTECContract
2995/76/NL/PP(SC),15December1977.
[65] Fleming,R.B.,TheEffectofFlowDistributioninParallelChannelsofCounterflowHeat
Exchangers,inAdvancesinCryogenicEngineering,1sted.,Vol.12,K.D.Timmerhaus,
Ed.,PlenumPress,NewYork,1967,pp.352362.
[66] Fortescue,P.,Stark,J.,SpacecraftSystemsEngineering,JohnWiley&Sons,Baffing
Lane,Chichester,1991.
[67] Fraas,A.P.,Ozisik,M.N.,HeatExchangerDesign,1sted.,JohnWiley&Sons,Inc.,New
York,1965,Chap.6,p.98.
[68] French,E.,ThermalDesignofRotatableCryogenicHeatPipeJoints,AIAAPaperNo.
850068,AIAA23rdAerospaceSciencesMeeting,Reno,Nevada,Jan.1417,1985.
[69] Gambill,W.R.,Bundy,R.D.,Wansbrough,R.W.,HeatTransfer,BurnoutandPressure
DropforWaterinSwirlFlowthroughTubeswithInternalTwistedTapes,Chem.Engng.
Progr.SymposiumSer.,Vol.57,No.32,1961,pp.127137.
[70] Gao,H.H.,Zhou,F.D.,Chen,X.J.,VoidFractionMeasurementUsingtheConductance
ProbeTechniqueinAirWaterTwoPhaseFlow,inTwoPhaseFlowandHeatTransfer,
X.J.Chen&T.NejatVeziroglu,Eds.,HemispherePublishingCorporation,Washington,
1985,pp.5761.
[71] GELBER,GelberPumps7778,GelberPumpCo.,5806NorthLincolnAvenue,Chicago,
Illinois60659,1976.
[72] Gilmour,C.H.,NoFoolingNoFouling,Chem.Engng.Prog.,Vol.61,No.7,July1965,
pp.4954.
[73] Goldstein,S.,ModernDevelopmentsinFluidDynamics,3rded.,Vol.II,Oxfordatthe
ClarendonPress,1950,Chap.XV,pp.646660.
[74] Gomelauri,V.I.,MethodsandResultsofanExperimentalInvestigationintoProcessesof
IntensificationofConvectiveHeatTransfer,ThermalEngineering(Teploenergetika),
Vol.211,No.9,Sept.1974,pp.15.
[75] Govier,G.W.,Aziz,K.,TheFlowofComplexMixturesinPipes,1sted.,VanNostrand
ReinholdCompany,NewYork,1972.
[76] Greene,N.D.,ConvairAircraft,privatecommunicationtoW.R.Gambill,quotedby
Bergles(1969),May1960.
[77] Grote,M.G.,Swanson,T.D.,DesignandTestofaPumpedTwoPhaseMountingPlate,
AIAAPaperNo.850919,AIAA20thThermophysicsConference,Williamsburg,Virginia,
June1921,1985.
[78] Hall,W.B.,HeatTransferinChannelsHavingToughandSmoothSurfaces,J.Mech.
Engng.Sci.,Vol.4,No.3,March1962,pp.287291.
[79] HAMILTONSTANDARD,GSEHeatExchangerPackage,CEINo.MC25000010025
ItemNo.7.5.FreontoWaterInterchangerPackage,CEINo.MC25000010040Item
No.7.7.AvionicsCoolingAssembly,CEINo.MC62100080002ItemNo.3.70.
482
ECSSEHB3101Part13A
5December2011
HamiltonStandard,DivisionofUACElectronicSystem,BradleyFieldRoad,Windsor
Locks,Connecticut06096.
[80] Hanjalic,K.,Launder,B.E.,FullyDevelopedAsymmetricFlowinaPlaneChannel,J.
FluidMech.,Vol.51,Part2,25January1972,pp.301335.
[81] Heap,R.D.,HeatPumps,1sted.,E&F.N.SponLtd.,London,1979.
[82] Heaton,H.S.,Reynolds,W.C.,Kays,W.M.,HeatTransferinAnnularPassages.
SimultaneousDevelopmentofVelocityandTemperatureFieldsinLaminarFlow,Int.J.
HeatMassTransfer,Vol.7,No.7,July1964,pp.763781.
[83] Heizer,B.L.,Goo,S.D.,Rhodes,G.D.,Thoreson,D.W.,Parish,R.C.,ASteerableRadiator
forSpacecraftApplication,AIAAPaperNo.861298,AIAA/ASME4thJoint
ThermophysicsandHeatTransferConference,Boston,Mass,June24,1986.
[84] Helenbrook,R.G.,Antony,F.M.,Fisher,R.M.,SelectionofSpaceShuttleThermal
ProtectionSystems,AIAAPaperNo.71443,AIAA6thThermophysicsConference,
Tullahoma,Tennessee,April2628,1971.
[85] Hetsroni,G.,HandbookofMultiphaseSystems,1sted.,HemispherePublishing
Corporation,Washington,1982.
[86] Hewitt,G.F.,LiquidGasSystemsFlowRegimes,inHandbookofMultiphase
Systems,1sted.,G.Hetsroni,Ed.,HemispherePublishingCorporation,Washington,
1982,Chap.22,pp.23to243.
[87] Hewitt,G.F.,MeasurementofVoidFraction,inHandbookofMultiphaseSystems,1st
ed.,G.Hetsroni,Ed.,HemispherePub.Co.,Washington,1982,Chap.10,pp.1021to10
33.
[88] Hinderer,B.,Savage,C.J.,DevelopmentofaMovable,ThermallyConductingJointfor
ApplicationtoDeployableRadiators,inSpacecraftThermalandEnvironmentalControl
Systems,ESASP139,Paris,Nov.1978,pp.449452.
[89] Hinze,J.O.,Turbulence,1sted.,McGrawHillBookCompany,Inc.,NewYork,1959,
Chap.7,pp.537553.
[90] Holmberg,R.B.,HeatTransferinLiquidCoupledIndirectHeatExchangerSystems,
TransactionsoftheASME,JournalofHeatTransfer,Vol.97,SeriesC,No.4,November
1975,pp.499503.
[91] Hong,S.W.,Bergles,A.E.,AugmentationofLaminarFlowHeatTransferinTubesby
MeansofTwistedTapeInserts,TransactionsoftheASME,JournalofHeatTransfer,Vol.
98,SeriesC,No.2,May1976,pp.251256.
[92] Hu,M.,Chang,Y.P.,OptimizationofFinnedTubesforHeatTransferinLaminarFlow,
TransactionsoftheASME,JournalofHeatTransfer,Vol.95,SeriesC,No.3,August1973,
pp.332338.
[93] Hutchinson,P.,Whalley,P.B.,APossibleCharacterizationofEntrainmentinAnnular
Flow,Chem.Engng.Sci.,Vol.28,1973,pp.974975.
[94] Hwangbo,H.,McEver,W.S.,HighThermalCapacityColdPlate/HotPlate,AIAAPaper
No.850921,AIAA20thThermophysicsConference,Williamsburg,Virginia,June1921,
1985.
[95] HYDRAULICINSTITUTE,HydraulicInstituteStandardsforCentrifugal,Rotary&
ReciprocatingPumps,13thed.,HydraulicInstitute,1230KeithBuilding,Cleveland,Ohio
44115,1975,pp.11,133134,173174.
483
ECSSEHB3101Part13A
5December2011
[96] Ibragimov,M.H.,Nomofelov,E.V.,Subbotin,V.I.,HeatTransferandHydraulic
ResistancewiththeSwirlTypeMotionofLiquidinPipes,(inRussian),Teploenergetika,
Vol.8,No.7,July1961,pp.5760.
[97] Idelcik,I.E.,MementodesPertesdeCharge,TranslatedfromtheRussianbyM.Meury,
1sted.,EyrollesEditeur,Pairs,1969,Chap.II,pp.5580,Chap.III,pp.9596.Chap.VII,
pp.233287,Chap.VIII,pp.312314.
[98] Ignationis,A.J.,Mitchell,K.L.,MSFCSortieLaboratoryEnvironmentalControlSystems
(ECS).PhaseBDesignStudyResults,NASATMX64832,March1974.
[99] Jakob,M.,HeatTransfer,6thed.,Vol.1,JohnWiley&Sons,Inc.,NewYork,1958,Chap.
11,pp.207243.
[100] Jamshidi,M.,Tarakh,M.,Shafai,B.,ComputerAidedAnalysisandDesignofLinear
ControlSystems,PrenticeHall,Inc.,EnglewoodCliffs,NewJersey,pp.360364,1992.
[101] Kays,W.M.,ConvectiveHeatandmassTransfer,1sted.,McGrawHillBookCompany,
NewYork,1966,Chap.8,pp.118133.Chap.9,pp.169194.
[102] Kays,W.M.,London,A.L.,CompactHeatExchangers,2nded.,McGrawHillBook
Company,NewYork,1964,Chap.1,p.3,Chap.2,pp.1063,Chap.4,pp.8691.Chap.6,
pp.103119.Chap.9,pp.137159.Chap10,pp.161230.AppendixA,p.232.
[103] Kemeny,G.A.,Cyphers,J.A.,HeatTransferandPressureDropinanAnnularGapwith
SurfaceSpoilers,TransactionsoftheASME,JournalofHeatTransfer,Vol.83,SeriesC,
No.2,May1961,pp.189198.
[104] Keshock,E.G.,Sadeghipour,M.S.,AnalyticalComparisonofCondensingFlowsInside
TubesunderEarthGravityandSpaceEnvironments,ActaAstronautica,Vol.10,No.7,
July1983,pp.505511.
[105] Kissner,G.,ComparisonofHeatTransportFluidsforSpaceLabActiveThermalControl
Systems,DORNIERSYSTEM,SLBDS0076,Friedrichshafen,Germany,1973.
[106] Klaczak,A.,HeatTransferinTubeswithSpiralandHelicalTurbulators,Transactionsof
theASME,JournalofHeatTransfer,Vol.95,SeriesC,No.4,November1973,pp.557559.
[107] Knudsen,J.G.,Katz,D.L.,FluidDynamicsandHeatTransfer,1sted.,McGrawHillBook
Company,Inc.,NewYork,1958,Chap.15,pp.407455.
[108] Kosky,P.G.,Staub,F.W.,LocalCondensingHeatTransferCoefficientsintheAnnular
FlowRegime,AIChEJournal,Vol.17,No.5,September1971,pp.10371043.
[109] Kreith,F.,Margolis,D.,HeatTransferandFrictioninSwirlingTurbulentFlow,1958
HeatTransferandFluidMechanicsInstitute,HeldatUniversityofCalifornia,Berkeley,
California,June1921,1958,pp.126142.
[110] Kroeger,P.G.,PerformanceDeteriorationinHighEffectivenessHeatExchangersDueto
AxialHeatConductionEffects,inAdvancesinCryogenicEngineering,1sted.,Vol.12,
K.D.Timmerhaus,Ed.,PlenumPress,NewYork,1967,pp.363372.
[111] Kubair,V.,Kuloor,N.R.,HeatTransfertoNewtonianFluidsinCoiledPipesinLaminar
Flow,Int.J.HeatMassTransfer,Vol.9,No.1,January1966,pp.6375.
[112] Kuo,B.C.,AutomaticControlSystems,PrenticehallInternational,Inc.,1991.
[113] Kurzweg,U.H.,Zhao,L.d.,HeatTransferbyHighFrequencyOscillations:ANew
HydrodynamicTechniqueforAchievinglargeEffectiveThermalConductivities,Phys.
Fluids,Vol.27,No.11,Nov.1984,pp.26242627.
484
ECSSEHB3101Part13A
5December2011
[114] Lehtinen,A.,Sadunas,J.,ThermalStorageAnalysisforLargeMannedSpacePlatforms,
AIAAPaperNo.850066,AIAA23rdAerospaceSciencesMeeting,Reno,Nevada,Jan.14
17,1985.
[115] Liptak,B.G.,InstrumentEngineersHandbook,ChiltonBockCompany,1969.
[116] Lockhart,R.W.,Martinelli,R.C.,ProposedCorrelationofDataforIsothermalTwo
Phase,TwoComponentsFlowinPipes,Chem.Engng.Progr.,Vol.45,No.1,January
1949,pp.3948.
[117] London,A.,ThermalControloftheNIMBUSSatelliteSystem,GeneralElectricMSD,
AIAAUnmannedSpaceMeeting,1965.
[118] London,A.L.,LaminarFlowGasTurbineRegeneratorsTheInfluenceofManufacturing
Tolerances,TransactionsoftheASME,JournalofHeatTransfer,Vol.92,SeriesA,No.1,
January1970,pp.4656.
[119] London,A.V.,FactorsintheSelectionofPumpsforProcessandChemicalDuties,
PumpsPompesPumpen,No.88,January1974,pp.2531.
[120] Lopina,R.F.,Bergles,A.E.,heatTransferandPressureDropinTapeGeneratedSwirl
FlowofSinglePhaseWater,TransactionsoftheASME,JournalofHeatTransfer,Vol.91,
SeriesC,No.3,August1969,pp.434442.
[121] Lundgren,T.S.,Sparrow,E.M.,Starr,J.B.,PressureDropDuetothepp.620626.
[122] Martinelli,R.C.,Nelson,D.B.,PredictionofPressureDropDuringForcedCirculation
BoilingofWater,TransactionsoftheASME,Vol.7,No.6,August1948,pp.695702.
[123] Masliyah,J.H.,Nandakumar,K.,HeatTransferinInternallyFinnedTubes,Transactions
oftheASME,JournalofHeatTransfer,Vol.98,SeriesC,No.2,May1976,pp.257261.
[124] Mason,J.L.,HeatTransferinCrossflow,Proc.Appl.Mechanics,2ndUSNational
Congress,1954,pp.801803.
[125] McAdams,W.H.,HeatTransmission,3rd.Ed.,McGrawHillBookCompany,Inc.,New
York,1954,Chap.7,p.180.
[126] MICROTECNICA,HeatExchangersfromMICROTECNICA,Microtecnica,Torino,
Italy,privateCommunicationbyG.Lapidari,February1977.
[127] MOOG,RotaryShutoff(RSO)Disconnect,CAT58510872M.MOOGSpaceProducts
Division,MoogInc.,EastAurora,NewYork,140520018.
[128] Mori,Y.,Nakayama,W.,StudyonForcedConvectiveHeatTransferinCurvedPipes,
Int.J.HeatMassTransfer,Vol.8,No.1,January1965,pp.6782.
[129] Nachtigl,C.L.,InstrumentationandControl.FundamentalsandApplications,John
Wiley&Sons,1990.
[130] NASA,SpaceandPlanetaryEnvironmentCriteriaGuidelinesforUseinSpaceVehicle
Development(1982Revision),NASATM82478,Vol.1,1983.
[131] Nast,T.C.,Barnes,C.B.,Wedel,R.K.,DevelopmentandOrbitalOperationofaTwo
StageSolidCryogenCoolerforInfraredDetectorCooling,AIAAPaperNo.76475,
AIAA11thThermophysicsConference,SanDiego,California,July1416,1976.
[132] Nekrasov,B.,HydraulicsforAeronauticalEngineers,TranslatedfromtheRussianbyV.
Talmy.1sted.,MIRPublishers,Moscow,1969,ChapterVI,pp.8087,ChapterXII,pp.182
221.
485
ECSSEHB3101Part13A
5December2011
[133] Noltingk,B.E.,JonesInstrumentTechnologyMechanicalMeasurements,Butterworths,
1985.
[134] Nunner,W.,WrmebergangundDruckabfallinRahuenRohren,VDI
Forschungsheft455,AusgabeB,Band22,1956,pp.139.
[135] Ogata,K.,DiscretetimeControlSystems,PrenticehallInternational,S.A.,Inc.,1987.
[136] Ogata,K.,IngenieradeControlModerna,PrenticehallHispanoamericana,S.A.,
Mexico,1988.
[137] Ogata,K.,ModernControlEngineering,PrenticehallInternational,Inc.2ndEdition,
1990.
[138] Owen,P.R.,Thomson,W.R.,HeatTransferacrossRoughSurfaces,J.FluidMech.,Vol.
15,Part3,March1963,pp.321334.
[139] Owen,R.H.,Sessions,B.W.,Walker,D.L.,SpacelabEnvironmentalControlSystem
ThermalStudies,TeledyneBrownEngineering,Huntsville,Alabama,InterimReport
ESDEPO11993,May1976.
[140] Palen,J.W.,PrinciplesofGasLiquidTwoPhaseFlowCalculations,ReportTPG1,Heat
TransferResearch,Inc.,Alhambra,California,June1976.
[141] Perry,M.P.,ASurveyofFerromagneticLiquidApplications,inThermomechanicsof
MagneticFluids,B.Berkovsky,Ed.,HemispherePublishingCorporation,Washington,
1978,pp.219230.
[142] Peterson,G.P.,ThermalControlSystemsforSpacecraftInstrumentation,J.Spacecraft,
Vol.24,No.1,JanFeb1987,pp.713.
[143] Petukhov,B.S.,Roizen,L.I.,GeneralizedDependencesforHeatTransfer3,MayJune
1974,pp.565569.
[144] Pollak,F.,Cruger,C.O.,ComparisonofApplicationsandCharacteristicsofPositive
DisplacementandCentrifugalPumps,PumpsPompesPumpen,No.96,September
1974,pp.400407.
[145] Pucci,P.F.,Howard,C.P.,PiersallJr.,C.H.,TheSingleBlowTransientTestingTechnique
forCompactHeatExchangerSurfaces,TransactionsoftheASME,Journalof
EngineeringforPower,SeriesA,Vol.89,No.1,January1967,pp.2940.
[146] Rabald,E.,CorrosionGuide,2nded.,ElsevierPublishingCompany,Amsterdam,1968,
pp.153156.
[147] Raznjevic,K.,TablesetDiagrammesTermodynamiques,TranslatedfromSerbo
CroatianbyR.Podhorsky.1sted.,EditionsEyrolles,Paris,1970.
[148] Rehme,K.,TurbulenceMeasurementsinSmoothConcentricAnnuliwithSmallRadius
Ratio,J.FluidMech.,Vol.72,part1,11November1975,pp.189206.
[149] Reynolds,A.J.,TurbulentFlowsinEngineering,1sted.,JohnWiley&Sons,Inc.,
London,1974,Chap.4,pp.181240.Chap.5,pp.241285.
[150] Rittenhouse,J.B.,Singletary,J.B.,SpaceMaterialsHandbook,3rded.,NASASP3051,
1969,pp.655658.
[151] Rufer,C.E.,Kezios,S.P.,AnalysisofTwoPhase,OneComponentStratifiedFlowwith
Condensation,TransactionsoftheASME,JournalofHeatTransfer,Vol.88,SeriesC,No.
3,August1966,pp.265275.
486
ECSSEHB3101Part13A
5December2011
[152] Rye,H.,Steen,JE.,SeawaterasanEnergyResourceforHeatPumpsinNorway,in
WaterforEnergy,J.Standbury,Ed.,BHRA,Cranfield,Bedford,England,1986,pp.203
216.
[153] Sadunas,J.A.,Lehtinen,A.,Parish,R.,ThermalManagementSystemOptionsforHigh
PowerSpacePlatforms,AIAAPaperNo.851047,AIAA20thThermophysicsConference,
Williamsburg,Virginia,June1921,1985.
[154] Sadunas,J.A.,Lehtinen,A.M.,Nguyen,H.T.,Parish,R.,SpacecraftActiveThermal
ControlSubsystemDesignandOperationConsiderations,AIAAPaperNo.861267,
AIAA/ASME4thJointThermophysicsandHeatTransferConference,Boston,Mass,June
24,1986.
[155] Sadunas,J.A.,Lehtinen,A.M.,Nguyen,M.T.,Parish,R.,SpacecraftActiveThermal
ControlSubsystemDesignandOperationConsiderations,AIAAPaperNo.861267,
AIAA/ASMEConference,Boston,Mass,June24,1986.
[156] Schember,H.R.,NetworkOptimizationofLineDiametersforTwoPhaseManagement
Systems,AIAAPaperNo.861247,AIAA/ASME4thJointThermophysicsandHeat
TransferConference,Boston,Mass,June24,1986.
[157] Schlichting,H.,BoundaryLayerTheory,TranslatedfromtheGermanbyJ.Kestin.4th
ed.,McGrawHillBookCompany,Inc.,NewYork,1960,Chap.XX,pp.502531.
[158] Schmid,C.,KEDDCAComputerAidedAnalysisandDesignPackageforControl
Systems,inComputerAidedControlSystemsEngineeringM.JamshidiandC.J.Herget,
(eds.),NorthHolland,Amsterdam,pp.159180,1985.
[159] Schmidt,F.W.,Zeldin,B.,LaminarHeatTransferintheEntranceRegionofDucts,Appl.
Sci.Res.,Vol.23,No.12,October1970,pp.7394.
[160] Scobie,G.,SelectthePumpthatMeetsYourNeeds,Chart.Mech.Eng.,Vol.21,No.5,
May1974,pp.5963.
[161] Scott,A.W.,CoolingofElectronicEquipment,1sted.,JohnWiley&Sons,Inc.,New
York,1974,Chap.5,p.126.Chap.8,pp.215226.
[162] Seban,R.A.,McLaughlin,E.F.,HeatTransferinTubeCoilswithLaminarandTurbulent
Flow,Int.J.HeatMassTransfer,Vol.6,pp.387395,1963.
[163] Settles,G.S.,Hamrick,J.T.,Barr,W.J.,Summerfield,M.,Gunn,M.,EnergyEfficient
PumpUtilization,J.Energy,Vol.1,No.1,January1977,pp.6572.
[164] Shah,R.K.,London,A.L.,TheInfluenceofBrazingonVeryCompactHeatExchanger
Surfaces,DepartmentofMechanicalEngineering,StanfordUniversity,Stanford,
California,TRNo.73,November1970.
[165] Shaubach,R.M.,HeatTransferacrossStructuralBoundaries,AIAAPaperNo.850977,
AIAA20thThermophysicsConference,Williamsburg,Virginia,June1921,1985.
[166] Sheriff,N.,Gumley,P.,HeatTransferandFrictionPropertiesofSurfaceswithDiscrete
Roughnesses,Int.J.HeatMassTransfer,Vol.9,No.12,December1966,pp.12971320.
[167] Sleicher,C.A.,Tribus,M.,HeatTransferinaPipewithTurbulentFlowandArbitrary
WallTemperatureDistribution,TransactionsoftheASME,Vol.79,No.2,May1957,pp.
789797,reprintedinRecentAdvancesinHeatandMassTransfer,1sted.,J.P.Hartnett,
Ed.,McGrawHillBookCompany,Inc.,,NewYork,1961,pp.281301.
487
ECSSEHB3101Part13A
5December2011
[168] Smithberg,E.,Landis,F.,FrictionandForcedConvectionHeatTransferCharacteristics
inTubeswithTwistedTapeSwirlGenerators,TransactionsoftheASME,Journalof
HeatTransfer,Vol.86,SeriesC,No.1,February1964,pp.3949.
[169] Sparrow,E.M.,HajiSheikh,A.,FlowandHeatTransferinDuctsofArbitraryShape
withArbitraryConditions,TransactionsoftheASME,JournalofHeatTransfer,Vol.88,
SeriesC,No.4,November1966,pp.351358.
[170] Stevens,R.A.,Fernandez,J.,Woolf,J.R.,MeanTemperatureDifferenceinOne,Two,and
ThreePassCrossflowHeatExchangers,TransactionsoftheASME,Vol.79,No.1,
February1957,pp.287297.
[171] Taborek,J.,DesignMethodsforHeatTransferEquipmentACriticalSurveyofthe
StateoftheArt,inHeatExchangers:DesignandTheorySourcebook,1sted.,N.H.
AfganandE.U.Schlunder,Eds.,McGrawHillBookCompany,NewYork,1974,Clause7,
pp.4574.
[172] Taborek,J.,Aoki,T.,Ritter,R.B.,Palen,J.W.,Knudsen,J.G.,Fouling:TheMajor
UnresolvedProbleminHeatTransfer,Chem.Engng.Progr.,Vol.68,No.2,February
1972,pp.5978.
[173] Taborek,J.,Aoki,T.,Ritter,R.B.,Palen,J.W.,Knudsen,J.G.,PredictiveMethodsfor
FoulingBehavior,Chem.Engng.Progr.,Vol.68,No.7,July1972,pp.6978.
[174] Taitel,Y.,Dukler,A.E.,AModelforPredictingFlowRegimeTransitionsinHorizontal
andNearHorizontalGasLiquidFlow,A.I.Ch.E.Journal,Vol.22,No.1,January1976,
pp.4755.
[175] Tanzer,H.J.,Fleischman,G.L.,Stalmach,D.D.,OsmoticPumpedHeatPipesforLarge
SpacePlatforms,AIAAPaperNo.820902,AIAAASME3rdJointThermophysics,Fluids,
PlasmaandHeatTransferConference,St.Louis,Missouri,June711,1982.
[176] TEMA,StandardsofTubularExchangerManufacturersAssociation,Inc.,331Madison
Avenue,NewYork,1968,Section9,pp.123127.
[177] Thaller,G.J.,Brown,R.G.,AnalysisandDesignofFeedbackControlSystems,McGraw
Hill,1960.
[178] Thorsen,R.,Landis,F.,FrictionandHeatTransferCharacteristicsinTurbulentSwirl
FlowSubjectedtoLargeTransverseTemperatureGradients,TransactionsoftheASME,
JournalofHeatTransfer,Vol.90,SeriesC,No.1,February1968,pp.8797.
[179] Trusch,R.B.,ShuttleOrbiterAtmosphericRevitalizationSystemandFreonCoolant
LoopSystemTradeStudies,SAEPaperNo.740921,IntersocietyConferenceon
EnvironmentalSystems,Seattle,Washington,July29August1,1974.
[180] Trusch,R.B.,Nason,J.,CompactHeatExchangersfortheSpaceShuttle,ASMEPaper
75ENAs54,IntersocietyConferenceonEnvironmentalSystems,SanFrancisco,
California,July21124,1975.
[181] Vallerani,E.,DegliEsposti,P.L.,Sessions,B.,PresentCapabilitiesandExtension
PossibilitiesoftheSpacelabThermalControlSystems,AIAAPaperNo.781676,AIAA
ConferenceonLargeSpacePlatforms:FutureNeedsandCapabilities,LosAngeles,Calif.,
Sept.2729,1978.
[182] VanOost,S.,Mathieu,J.P.,TwoPhaseHeatTransportSystemConceptualStudies,ESA
CR(P)1873,Sabca,ChaussedeHaecht,B.1130,Bruxelles,Belgium,Aug.1983.
[183] Vargaftik,N.B.,TablesontheThermophysicalPropertiesofLiquidandGases,2nded.,
HemispherePublishingCorporation,WashingtonLondon,1975.
488
ECSSEHB3101Part13A
5December2011
[184] Wallis,G.B.,OneDimensionalTwoPhaseFlow,1sted.,McGrawHillBookCompany,
Inc.,NewYork,1969.
[185] Weast,R.C.,HandbookofChemistryandPhysics,47thed.,TheChemicalRubberCo.,
Cleveland,Ohio,1966,pp.E24andF56.
[186] Webb,R.L.,Eckert,E.R.G.,Goldstein,R.J.,HeatTransferandFrictioninTubeswith
RepeatedRibRoughness,Int.J.HeatMassTransfer,Vol.14,No.4,April1971,pp.601
617.
[187] Weimer,R.F.,Hartzog,D.G.,EffectofMaldistributiononthePerformanceof
Multistream,MultipassageHeatExchangers,inAdvancesinCryogenicEngineering,1st
ed.,Vol.18,K.D.Timmerhaus,Ed.,PlenumPress,NewYork,1973,pp.5264.
[188] Welty,J.R.,Wicks,C.E.,Wilson,R.E.,FundamentalsofMomentum,HeatandMass
Transfer,1sted.,JohnWiley&Sons,Inc.,NewYork,1969,Chap.22,pp.380383.
[189] Werth,G.R.,Vernon,R.M.,Moudy,W.D.,Hansen,W.,Reed,R.D.,Hawkins,R.S.,Waller,
C.,Warner,B.W.,StudyofThermalIsolationTechniques,LockheedMissiles&Space
Company,Sunnyvale,California,M92661,NASACR71743,preparedunderContract
No.JPL950950,March1966.
[190] WicksIII,M.,Dukler,A.E.,EntrainmentandPressureDropinConcurrentgasLiquid
Flow.1.AirWaterinHorizontalFlow,A.I.Ch.E.Journal,Vol.6,No.3,September1960,
pp.463468.
[191] Wiebelt,J.A.,DesignConsiderationsforThermostaticFinSpacecraftTemperature
Control,NASACR500,1966.
[192] Wise,P.C.,SpacecraftThermalControlTechnology:DesignChallengesintothe1990s,
ActaAstronautica,Vol.14,1986,pp.489502.
[193] WynRoberts,D.,EnvironmentalControlSystemHardwareforSpacelab.ASurveyof
SomeEuropeanAircraftEquipmentManufacturers,ESTEC,Noordwijk,The
Netherlands,InternalWorkingPaperNo.811,February1974.
[194] WynRoberts,D.,PumpsforMannedSpacecraft.AShortSurveyofSomePossible
EuropeanManufacturers,ESTEC,Noordwijk,TheNetherlands,InternalWorkingPaper
No.788,November1973.
[195] Yang,C.C.,Palen,J.W.,TwoPhasePressureDropinTubesideCondensation,Report
CT4,HeatTransferResearch,Inc.,Alhambra,California,January1977.
489

ECSS E HB 31 01 - Part13A Fluid Loops

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ECSS E HB 31 01 - Part13A Fluid Loops

Uploaded by

Copyright:

Available Formats

ECSS-E-HB-31-01 Part 13A

3 Terms, definitions and symbols..........................................................................32

4 General introduction ............................................................................................46

5 Analysis of a fluid loop ........................................................................................52

6 Thermal analysis ..................................................................................................58

7 Frictional analysis ................................................................................................92

8 Combined thermal and frictional analysis........................................................125

9 Heat transfer enhancement ...............................................................................130

15 Two-phase thermal transport systems ...........................................................359

3.1 Terms and definitions

3.2 Abbreviated terms

4.1 Fluid loops

4.2 Comparison between fluid loops and alternative

4.2.1 Passive thermal insulations

4.2.2 Thermoelectric devices

4.2.3 Phase change materials (pcm)

4.2.4 Heat pipes

4.2.5 Short-term discharge systems

5.2 Thermal performance

Q sC1 Thi Tci [51]

Q C h Thi Tho [52]

Q C c Tco Tci [53]

Q scp ch TS Tho [57]

Tci Tco Thi Tho

5.3 Power requirements

where pistheoverallefficiencyofthepump, pthepressurelossthroughtheloop,mthefluidmass

6.2 Analytical background

6.2.1 Heat transfer coefficient

6.2.2 Dimensionless groups

6.2.3 Simplifying assumptions

6.2.4 Temperature-dependence of fluid properties

6.2.5 Laminar versus turbulent fluid flow

6.2.6 Heat transfer to internal flows

6.2.7 Heat transfer to external flows

Nu f Reb , Prb , Prw , st / D, sl / D, N , L / D, [629]

6.3 Thermal performance data

6.3.1 Heat transfer to internal flow

Type Flow Heatflux Geometry InputVariables Output Fig. References

6.3.1.2 Transitional flow

6.3.2 Heat transfer to external flows

Geometry Input Output Fig. References

CircularCylinderNormaltoFlow. Re,Pr=0,7,5,3,5 Nub Figure ESDU69004[50],

CircularCylinderNormaltoFlow. 0,061<wm/b< Nu/Nub Figure ESDU69004[50],

CircularCylinderNormaltoFlow. 0,20<Twm/Tb<4,47 Nu/Nub Figure ESDU69004[50],

CircularCylinderYawedtoFlow. 40<<90 Nu/Nu90 Figure ESDU69004[50],

InLineTubeBank. Figure ESDU73031[57],

InLineTubeBank. Reb,Prb=1 Nur Figure ESDU73031[57],

InLineTubeBank. Reb,Prb=1 Nur Figure ESDU73031[57],

StaggeredTubeBank. Reb,Prb=1 Nur Figure ESDU73031[57],

InLineorStaggeredTubeBank. Prb Nu r Figure ESDU73031[57],

InLineorStaggeredTubeBank. Prb/Prw F1 Figure ESDU73031[57],

InLineorStaggeredTubeBank. N F2 Figure ESDU73031[57],

InLineorStaggeredTubeBank 2090 F3 Figure ESDU73031[57],

Geometry Input Output Fig. References

InLineorStaggeredTubeBank. 1n20 F4 Figure ESDU73031[57],

6.3.2.1 Cylindrical bodies

7.2 Analytical background

7.2.2 Fully developed flow in straight pipes

0 0,079 0,25 <105 2%oftheexperimental Blasiusfrictionformula.

0 0,046 0,20 105106 Notgiven

0,0014 0,125 0,32 <107 3%ofthelogformula

7.2.3 Temperature-dependence of fluid properties

LaminarFully Gas (Tw/Tb)n 0,45 0,30 Intherange