You are on page 1of 6

Proceedings

Proceedings ofof ASME


ASME FEDSM2003:
FEDSM’03
2003 ASME
th Fluids Engineering Division Summer Meeting
4 ASME_JSME Joint Fluids Engineering Conference
July 6-10. 2003, Honolulu, Hawaii
Honolulu, Hawaii, USA, July 6-10, 2003

FEDSM2003-45702
FEDSM2003-45702

PERFORATED PLATE PRESSURE LOSSES WITH


IMPROVED INLET AND OUTLET FLOW HOLE FEATURES

Kirkland D. Broach Michael E. Conner


Jeffery L. Norrell, Ph.D. Carter E. Lunde

Westinghouse Nuclear Fuel


5801 Bluff Road
Columbia, SC 29250

ABSTRACT restriction, such that the sensitivity of the hole geometry could
This paper describes the tests and studies performed to be investigated with different flow boundary conditions.
better understand the geometric factors affecting pressure loss To reduce the pressure drop of the perforated plate, various
in a perforated plate. In this study, the impact of a specific fundamental hole geometries, including edge chamfers and
perforated plate flow hole geometry on pressure drop was edge radii, were considered. These flow hole geometries were
investigated. The methodology established in this paper to first tested in a small-scale hydraulic loop to evaluate the
investigate this hole geometry can be extended to other fundamental pressure drop response of the features, with and
components with orifice type perforated plates. without a downstream restriction. Based on the small-scale
To reduce the pressure drop of the perforated plate, various results, several plate designs were selected for full-scale testing
fundamental hole geometries, including edge chamfers and to obtain a scaling term for this geometry.
edge radii, were considered. Results from various edge Two Westinghouse hydraulic loops were used: VISTA
treatments are provided in this study, including separate effects (Vibration Investigation of Small-scale Test Assemblies) and
for inlet and outlet hole geometries. Specific trends, such as the FACTS (Fuel Assembly Compatibility Test System). The
effect of increasing edge geometries on the hydraulic losses, are VISTA loop is a closed-loop, isothermal, room temperature,
presented. Additionally, a correlation between small-scale and hydraulic test loop designed for testing of small-array rod
full-scale pressure loss coefficients was found and is defined. bundles (typically 5x5) with actual fuel rod diameter and pitch.
The FACTS loop is a closed-loop, hydraulic test system
KEYWORDS: perforated plate, pressure drop, loss coefficient designed for testing of full-scale fuel assemblies (14x14 up to
18x18 square pitch rod arrays) at elevated temperatures.
INTRODUCTION In the small-scale testing, loss coefficients of the
In a pressurized water nuclear reactor (PWR), perforated perforated plates were measured with and without a rod bundle.
plates with orifice holes are used in various components: core Test results without a rod bundle were then compared to
entrance and exit, fuel assembly entrance and exit, steam fundamental hydraulic losses for perforated plates as provided
generator inlet, etc. As new components are designed with by standard handbooks such as Idelchik [1] and Marks’ [2].
enhanced features that tend to increase the hydraulic pressure Testing with a rod bundle was performed to provide appropriate
drop of the system, reducing the pressure losses of these small-scale test data for comparison with full-scale test results.
perforated plates is desirable. Typically, these perforated plates The impacts of radiussed and chamfered flow hole geometries
have structural and functional requirements that prevent were characterized in configurations of single effect inlet/outlet
enlarging the flow hole bore diameters or changing the flow features as well as combined inlet and outlet features.
hole patterns. Within these constraints, the present study was In the large-scale testing, select combinations of the inlet
performed by Westinghouse Nuclear Fuel to investigate the and outlet features were investigated. The test results of the two
possible reduction in pressure drop of an orificed plate design. loops were compared, and a correlation was developed between
The inlet region of one Westinghouse PWR fuel assembly the small-scale VISTA and full-scale FACTS testing. This
design was chosen for this study. This allowed for convenient correlation allows for the accurate prediction of full-scale, high
testing of perforated plates with and without a downstream Reynolds number loss coefficients from small-scale (VISTA)
test results.

1 Copyright © 2003 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 04/15/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Table 1: Description of Test Articles
NOMENCLATURE
A Total flow area through perforated plate
Tests Performed
AC Area correction term; required because area at taps is Plate Inlet Outlet
Without Bundle With Bundle
different than area A
A - - X X
B Power fit exponent
Bnd Denotes testing with bundle B Ri/D = 0.49 - X X
C Power fit coefficient C Ri/D = 0.82 - X X
D Diameter of perforated plate flow hole bore D Ri/D = 2.63 - X X
Dh_p Hydraulic diameter of test plate E Ri/D = 10.53 - X X
Dh_FH Hydraulic diameter of flow housing, upstream of plate F Li/D = 0.39 - X X
KFACTS Full-scale loss coefficient (FACTS conditions) B Rev - Ro/D = 0.49 X X
K fr Loss coefficient of rod friction C Rev - Ro/D = 0.82 X X
Kin Loss coefficient of test flow plate/inlet region D Rev - Ro/D = 2.63 X X
KVISTA Small-scale loss coefficient (VISTA conditions) E Rev - Ro/D = 10.53 X X
K500,000 Loss Coefficient extrapolated to Re = 500,000 G Ri/D = 0.08 Ro/D = 0.16 X X
Lh Length of flow hole (plate thickness) H Li/D = 0.39 Lo/D = 0.45 X X
Li Flow hole inlet chamfer depth I Ri/D = 0.82 Ro/D = 0.49 - X
Lo Flow hole outlet chamfer depth I Rev Ri/D = 0.49 Ro/D = 0.82 - X
Re Reynolds number based on area A J Ri/D = 10.53 Ro/D = 2.63 - X
Ri Flow hole inlet radius J Rev Ri/D = 2.63 Ro/D = 10.53 X X
Ro Flow hole outlet radius
V Velocity of the fluid in area A
x Independent variable
y Dependent variable
∆P Pressure drop measured in test
ρ Fluid density

FLOW HOLE GEOMETRIES


The perforated plate of interest in this study is highly
constrained by functional requirements and structural
interfaces. The plate thickness and flow hole bore size are fixed
by structural design requirements and the hole pattern in the
plate is set by interface requirements relative to the downstream
rod bundle. Therefore, the scope of this study was limited to the
inlet and outlet features of the flow hole geometry. Figure 1: Flow Hole Geometries
With an Lh /D ratio of 2.95, the test articles in this study are
classified as thick plates. Fundamental treatments to reduce
hydraulic losses of flow holes in a thick-plate are edge VISTA LOOP TESTING
chamfers and edge radii. A number of such geometries were The VISTA loop was designed to allow separate effects
selected for this study. These edge conditions were applied to vibration testing on small-scale 5x5 test bundles. The loop
the inlet separately, the outlet separately, and combined setup also allows for accurate pressure drop measurements. In
together on small-scale test articles for testing in the VISTA this hydraulic study, the thick perforated plates were tested with
loop. Table 1 provides a list of the test articles discussed in this and without a downstream restriction (vertically oriented rod
paper. Plates described as “Rev” were designed to be bundle with axial flow conditions). The VISTA test setup in
reversible, allowing the edge radii to be oriented as inlet or Fig. 2 presents the test configuration with the downstream
outlet features. Figure 1 illustrates the fundamental edge restriction. In this configuration, the perforated plates were
chamfer and radius geometries. attached to four rods of the test bundle. When tested without
the rod bundle, the perforated plates were suspended from the
top plate by four slender support rods and all 5x5 support grids
were removed.
A flow straightener plate was used at the entrance of the
flow housing in the VISTA loop to minimize inlet region flow
effects on the tests. Additionally, taking advantage of the
vertical height of the flow housing, the test flow plates were
positioned approximately 15 hydraulic diameters (based on the
empty flow housing) above this flow straightener. While not
prototypical of the flow field experienced by perforated plates
in a PWR, this testing configuration removed any loop inlet
effect from the pressure drop data.

2 Copyright © 2003 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 04/15/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Ri/D = 0.49. This trend is independent of both (1) the
presence/absence of the rod bundle, and (2) Re. Therefore, for
inlet radiussing, a relatively small radius is just as effective as a
very large radius in pressure drop reduction of the flow plate.

1.0

0.9

0.8

Loss Coefficient
0.7

0.6

0.5

0.4

0.3

0.2
0 2 4 6 8 10 12
Ri/D

21,738 Re - Bnd 36,230 Re - Bnd 50,722 Re - Bnd


21,738 Re 36,230 Re 50,722 Re

Figure 3: Inlet Radii Effects

Figure 4 presents pressure loss data for perforated plates


with edge radii on the flow hole outlets only. Similar
conclusions can be made as for inlet radiussing: the trend of
Figure 2: VISTA Test Setup
loss coefficients is independent of both the presence/absence of
the rod bundle and Re. However, the sensitivity of pressure
SMALL-SCALE TEST RESULTS losses to the outlet features is different. While for the inlet, any
Pressure drop measurements of the test bundle inlet region radius caused a substantial reduction in loss coefficient, a larger
were taken with the small-scale test plates listed in Table 1. outlet radius (Ro/D = 2.63) is required to significantly affect
This pressure data was obtained for Reynolds numbers ranging pressure losses.
from approximately 20,000 to 50,000. For testing without the A comparison of Figures 3 and 4 also shows that the
rod bundle, the pressure drop was simply that across the test reduction in loss coefficient is equivalent for inlet only or outlet
flow plate. For testing with the rod bundle, the measurement only radii. The inlet loss coefficients are observed to decrease
also included the pressure drop associated with support grids with increasing Reynolds numbers up to 36,230. The inlet loss
and rods. The pressure drop values were converted to loss coefficients remain relatively constant with further increases of
coefficients using the following equation: Re. However, the outlet loss coefficients are observed to
decrease with increasing Re throughout the range tested.
∆P
K in = − K fr + AC (1)
1 ρV 2
2 1.0

0.9

Separate Inlet/Outlet Edge Testing 0.8

A series of perforated plates with radiussed inlet/outlet


Loss Coefficient

0.7
features were tested to quantify the effect of edge radius on
0.6
plate pressure drop. Note that a plate with no flow hole edge
0.5
treatment (straight bore perpendicular to top and bottom
surfaces) was also tested as a benchmark. 0.4

Figure 3 presents pressure loss coefficients for radiussing 0.3

of the flow hole inlet edges only. Data obtained by testing with 0.2
and without the rod bundle are provided at the minimum, 0 2 4 6 8 10 12

median, and maximum test Reynolds numbers. Figure 3 shows Ro/D

that the inlet region loss coefficient with the rod bundle is
21,738 Re - Bnd 36,230 Re - Bnd 50,722 Re - Bnd
higher than without the rod bundle. This is expected since 21,738 Re 36,230 Re 50,722 Re
losses associated with the support grids and rods are included in
the inlet region measured pressure drop and loss coefficient for Figure 4: Outlet Radii Effects
the rod bundle case. A consistent trend seen in Fig. 3 is a large,
initial sensitivity to the radius size which rapidly dissipates with The data for inlet only edge conditions is plotted as a
increasing inlet radii. Although a significant effect was function of Re in Fig. 5 (without rod bundle) and Fig. 6 (with
observed with the smallest inlet radius tested, the loss rod bundle). Data from an inlet chamfer edge condition (Plate
coefficients are relatively constant for inlet radii larger than F) is also included. These figures show that the trends plotted in
Fig. 3 are consistent over the entire Re range tested.
3 Copyright © 2003 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 04/15/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


Additionally, these figures show that the inlet chamfer (Plate F)
1.0
provides a lower bound to all of the radiussed designs.
0.9

1.0 Ro/D
0.8
0

0.9 0.49
0.7

Loss Coefficient
0.82
Ri/D
0.8 0.6 2.63
0
10.53
0.49
0.7 0.5
0.82
Loss Coefficient

2.63
0.4
0.6 10.53
Series20
Li/D 0.3
0.5
0.39
0.2
0.4
20000 28000 36000 44000 52000
Reynolds Number
0.3

0.2
20000 28000 36000 44000 52000 Figure 8: Outlet Feature Loss Coefficients With Bundle
Reynolds Number

Figure 5: Inlet Feature Loss Coefficients Without Bundle Combined Inlet/Outlet Edge Testing
As shown in Table 1, a series of perforated plates with
combined inlet and outlet edge conditions was also tested in the
VISTA loop. Figure 9 presents loss coefficients for the plates
1.0
with combined inlet and outlet features as tested without the
0.9 Ri/D rod bundle. Figure 10 presents loss coefficients for the same
0
plates, tested with the rod bundle. In both instances, Plate A
0.8 0.49
0.82
(without edge treatments) was included to benchmark the data.
0.7 2.63 Comparison of Figures 9 and 10 shows that, like the single
Loss Coefficient

10.53
Series20
feature effects, the loss coefficients of plates with combined
0.6
Li/D inlet and outlet features are higher when tested with the rod
0.39
0.5 bundle. Consistent loss coefficient trends are observed for these
plates whether tested with or without the rod bundle.
0.4
Significant reductions of pressure loss were measured for
0.3 the perforated plates with combined inlet and outlet features. Of
those tested, the greatest reduction was measured for Plate
0.2
20000 28000 36000 44000 52000 J_Rev with large inlet and outlet radii. At Re ~ 36,000, Plate
Reynolds Number J_Rev operates with a 68.5% lower pressure loss than the
baseline plate when tested without the bundle. With the bundle,
Figure 6: Inlet Feature Loss Coefficients With Bundle the pressure loss is 50% lower than the baseline at this Re.
The data for the outlet only edge conditions is plotted as a
function of Re in Fig. 7 (without rod bundle) and Fig. 8 (with 1.0

rod bundle). These figures support the trends and conclusions 0.9
previously discussed for the outlet only edge conditions. Ri/D_Ro/D
0.8
0_0
0.08_0.16
1.0 0.7
2.63_10.53
Loss Coefficient

Series5
0.9 0.6 Li/D_Lo/D
0.39_0.45
0.8 0.5
Ro/D
0
0.7 0.49 0.4
Loss Coefficient

0.82
0.6 2.63 0.3
10.53

0.5 0.2
20000 28000 36000 44000 52000

0.4 Reynolds Number

0.3

Figure 9: Combined Inlet/Outlet Features Without Bundle


0.2
20000 28000 36000 44000 52000
Reynolds Number

Figure 7: Outlet Feature Loss Coefficients Without Bundle

4 Copyright © 2003 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 04/15/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


1.0

COMPARISON VERSUS HANDBOOK VALUES


0.9
Fundamental relations from Refences [1], [2], and [3] were
0.8 Ri/D_Ro/D used to define the initial radius sizes tested in VISTA. These
0_0
0.08_0.16
relations suggested that smaller radii would provide significant
0.7
pressure drop reductions. While the perforated plate pressure
Loss Coefficient

2.63_10.53

0.6
Series5
drop was reduced with the smaller radii when tested in VISTA,
Li/D_Lo/D
0.39_0.45
the magnitude of the reduction was much less than that
0.5
suggested by the handbook data. Figure 12 presents a
0.4 comparison of relative loss coefficients for perforated plates
with rounded inlet orifice edges as reported by Idelchik [1] and
0.3
measured in the present study. While the trends are similar, the
0.2 losses measured in VISTA are higher than those found in the
20000 28000 36000 44000 52000
literature. This difference could be a result of the specific flow
Reynolds Number
hole pattern geometry tested in the VISTA loop.

Figure 10: Combined Inlet/Outlet Features With Bundle


1.0

Of the plates with combined inlet and outlet features, 0.8

Plates I and J were tested in reverse orientations with the rod

Loss Coefficient
bundle. Figure 11 presents inlet loss coefficients for Plates I, 0.6

I_Rev, J, and J Rev. Pressure loss data of Plate A (without edge


0.4
treatments) was again included as a benchmark.
In all cases, comparison of the losses for combined inlet
0.2
and outlet features (relative to the benchmark) and the losses
associated with comparable single inlet/outlet features (relative
0.0
to the benchmark) indicates that the effect of combined features 0 0.2 0.4 0.6 0.8 1
is greater than just the sum of the single effects. For the plate Ri/D
thickness and flow hole diameter tested, the inlet and outlet
feature effects are interdependent. Idelchik Present Study
It is also interesting to note that the results for Plate I were
virtually the same for both test orientations while very different
Figure 12: Comparison of Idelchik and Present Study
results were observed for Plate J. The penetration depths of the
Pressure losses for Perforated Plates with Rounded Inlet
0.49 and 0.82 R/D features are very similar. Thus, the straight
Orifice Edges
bore section (between the inlet and outlet features) of the flow
holes was positioned at relatively the same distance from the
flow hole entrance for both test orientations of Plate I. FACTS LOOP TESTING
However, the 10.53 R/D feature extended approximately twice The FACTS loop was designed for single assembly
as deep into Plate J as the 2.63 R/D feature. The deep features hydraulic testing of full-scale nuclear fuel assemblies. Test
of Plate J resulted in a relatively short straight bore section assemblies are tested under controlled conditions to provide
which was closer to one plate surface. Thus, the position of the high quality pressure drop information. The FACTS loop has
straight bore section (relative to the flow hole inlet) differed by the capability to test at temperatures higher than the VISTA
a factor of 2 for the two test orientations of Plate J. loop. In this test series, pressure drop data was acquired in the
FACTS loop at 66°C and 121°C. High flowrates, relative to
1.0
actual in-core conditions, were used to extend the Re range of
0.9
the pressure loss data to approximately half of the in-core
conditions.
0.8
Ri/D_Ro/D Based on the VISTA results, Plates F, H, and J Rev were
0.7
0_0 selected for unique full-scale bottom nozzles to be tested in
0.49_0.82
FACTS. Plate F provided the lowest pressure drop in VISTA
Loss Coefficient

0.82_0.49
0.6 2.63_10.53 for an inlet feature only. Plate J Rev provided the lowest
10.53_2.63
0.5
pressure drop in VISTA for features on both the inlet and outlet
edges. Plate H represented an intermediate hole design relative
0.4 to Plates F and J Rev. Furthermore, comparison of Plates F and
H allowed for a direct comparison of the improvement of an
0.3
outlet edge feature to a flow hole design that also includes an
0.2 inlet edge feature.
20000 28000 36000 44000 52000
Reynolds Number

COMPARISON OF SMALL-SCALE VS FULL-SCALE


TEST RESULTS
Figure 11: Comparison of Combined Inlet/Outlet Features
Pressure loss data was measured in the VISTA loop over a
in Reverse Orientations (With Bundle)
range of Reynolds numbers from approximately 20,000 to
50,000. With the higher fluid temperatures and flowrates of the
5 Copyright © 2003 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 04/15/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use


FACTS loop, the fuel assembly inlet region was tested at
Reynolds numbers ranging from approximately 50,000 to 1.2

200,000. Thus, correspondence of the VISTA and FACTS


hydraulic data occurs with Re ~ 50,000. The FACTS loss 1.1 Li/D_Lo/D_Loop

coefficients for Plates F, H, and J Rev are presented with the 0.39_0_VISTA
0.39_0_FACTS
corresponding VISTA loss coefficients in Figure 13 for Re ~

Loss Coefficient, K
0.39_0.45_VISTA
1.0
50,000. It is apparent by observation of Figure 13 that a linear 0.39_0.45_FACTS
Series8
relationship exists between the VISTA and FACTS pressure Ri/D_Ro/D_Loop
loss data. Therefore, the small-scale VISTA data was correlated 0.9
2.63_10.53_VISTA
2.63_10.53_FACTS
to the full-scale FACTS data with the following equation. A
similar scaling relationship would be necessary if other 0.8

perforated geometries were tested in the small-scale VISTA


loop. 0.7
10000 40000 70000 100000 130000 160000 190000 220000
Reynolds Number, Re

K FACTS = 1.0951 KVISTA + 0.3641 (2)


Figure 14: Comparison of VISTA (Adjusted with
It is interesting to note that the slope of the correlation is Correlation) and FACTS Test Results With Bundle
approximately unity. Despite differing flow and boundary
conditions between the test loops, the effect on loss coefficient
is nearly a constant offset, as would be expected if the Table 2: Extrapolated VISTA and FACTS Loss Coefficients
additional losses are due to form losses caused by other
geometric features of the inlet region of the FACTS test. The Plate F Plate H Plate J Rev
VISTA tests are shown to predict the full scale tests with a VISTA FACTS VISTA FACTS VISTA FACTS
constant bias. C 2.1828 1.9995 2.1530 2.1397 1.9395 1.9864
The adjusted VISTA (scaled from Eq. (2)) and measured B -0.0618 -0.0539 -0.0754 -0.0751 -0.0772 -0.0789
FACTS loss coefficients are compared in Figure 14. In general, K 500,000 0.971 0.986 0.801 0.798 0.704 0.705
good agreement is observed between the small and full-scale
% Error 1.52% 0.38% 0.14%
data trends for each test. However, the first two data points
(lowest Re) of Plate H are not consistent with the remaining
VISTA and FACTS data for this plate. These anomalous points
were neglected in the comparison of VISTA and FACTS. CONCLUSIONS
To examine how the VISTA data behaved when The sensitivities of perforated plate hydraulic pressure
extrapolated to higher Re, the hydraulic data was fit to a power losses to changes in inlet and exit flow hole geometries was
law equation of the form: presented. The trends presented are applicable to other
perforated plate geometries. Confidence has been established
in a small-scale testing technique to estimate pressure loss
y = Cx B (3) sensitivities of full-scale components, allowing multiple
concepts to be quickly tested prior to full scale confirmatory
Table 2 presents the power law extrapolations for the VISTA testing. A methodology has been established to develop the
and FACTS data presented in Figure 14. This data also shows a appropriate scaling relationship between the small-scale and
comparison at Re=500,000. In all instances, the extrapolated full-scale hydraulic losses. The use of small-scale testing
VISTA data is within 2% of the FACTS data at this high Re. techniques to obtain applicable trends for scoping design
concepts provides significant savings (time, test article cost,
testing cost, etc.) by eliminating the need for full-scale testing
1.2
of every concept.

1.1 REFERENCES
Y = 1.0951 X + 0.3641
[1] Idelchik, I.E., 1994, Handbook of Hydraulic Resistance,
FACTS Loss Coefficient, K

1.0 Third Edition, CRC Press, Boca Raton, pp. 519 – 520.
Plate F [2] Avallone, E.A., Baumeister, T, III, Mark’s Standard
Plate H
Plate J Rev
Handbook for Mechanical Engineers, Tenth Edition, McGraw-
0.9
Linear Fit Hill, New York, pp. 3-50 – 3-51.
[3] White, F.M., 1994, Fluid Mechanics, Third Edition,
0.8
Mcgraw-Hill, New York, pp. 338-339.

0.7
0.3 0.4 0.5 0.6 0.7 0.8
VISTA Loss Coefficient, K

Figure 13: Correlation of VISTA and FACTS Pressure Loss


Data With Bundle for Re ~ 50,000

6 Copyright © 2003 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 04/15/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

You might also like