You are on page 1of 18

1

Sludge System Pumping Losses: UK Industry Standard is flawed


Dawson, M.K., Heywood, N.I., Alderman, N.J., Papachristou, C. and Ioannou, P.
BHR Group Limited, The Fluid Engineering Centre, Cranfield, Bedfordshire, MK43 OAJ.
Corresponding Author: Dawson, M.K., Tel. 01234 756501, Email: mdawson@bhrgroup.co.uk
http://www.bhrgroup.com

Abstract
Pump selection and sizing for transport of sewage sludge and other non-Newtonian biosolid slurries has
important capital and operating cost implications. Under-sizing results in an inability to achieve required process
throughput whilst over-sizing leads to excessive capital cost and energy consumption. Selection and sizing are
dependent on the system losses, which comprise static head losses and frictional losses through straight pipe,
fittings and equipment on the suction and discharge sides of a pump.

This paper provides a critique of the UK water industry standard calculation technique for sewage sludge
pumping system pressure losses (based on WRc Report TR 185 ‘How to Design Sewage Sludge Pumping
Systems’). The approach described in TR 185 is a pragmatic and simplified one which enabled design
calculations at a time when there were no PCs, spreadsheets or programmable calculators. However, this
approach is fundamentally flawed and more reliable methods for predicting friction pressure losses for pipeline
flows of sludges exhibiting yield-shear thinning behaviour have been added to the literature since the 1980s.

The implications of the shortcomings in TR 185 are illustrated using a series of case studies. The outputs from
the TR 185 approach are compared with those from a fully rigorous up-to-date calculation technique embedded
in BHR Group’s System Losses Tool (SLOT). The magnitude of errors resulting from the TR 185
approximations and how the errors are influenced by system and sludge characteristics are discussed.

The BHR Group System Losses Tool has been developed as part of the WWM research programme, funded by
the water industry, in order to replace TR 185 based calculation methods.

Keywords
Sewage Sludge, Non-settling Slurries, Pumping System Losses, Pipeline Flows, Sludge Rheology.

Introduction
It is vital to know how the processed sludge flows when sizing or selecting a pump or other process equipment
for sludge applications. Sludges often exhibit non-Newtonian behaviour where the shear rate and shear stress do
not vary proportionally as for a Newtonian fluid such as water. Owing to the different composition of sludges,
their rheological behaviour varies between sludge types and between sites. The most common non-Newtonian
2
behaviour found with sludges is known as ‘shear-thinning’, where viscosity reduces with increasing shear rate
and hence position within the pipeline or pump. In addition, some sludges exhibit a yield stress.

Sludge viscosity is given by

μ a  τ/γ
 Equation 1
where:
a = viscosity [Pas]
γ
 = shear rate [s-1]

 = shear stress [Pa]

There are two flow curve models that are commonly used to characterise sludge rheology. These are the
Herschel-Bulkley model (or Generalised Bingham Plastic):

τ  τ y  kγ n Equation 2

and the Power Law model:

τ  kγ
n Equation 3

where:
y = yield stress (Pa)
k = consistency coefficient (Pasn)
n = flow behaviour index (-)

In BHR Group’s Water & Wastewater Mixing (WWM) research programme, an MS Access Sludge Rheology
Database (‘SRDB’) was established. This enabled improved property prediction for a much wider range of
sludge types (currently 500+ rheograms) than those from the obsolete WRc Report TR 185 (Frost, 1983). The
new SRDB has been established from a combination of existing data and sludge samples obtained from WWM
members and measured using the BHR rheology laboratory. The SRDB includes entries for sludge type and
origins, temperature, dry solids content, viscometer type, shear stress vs shear rate data, and fitted flow curve
model parameters. The SRDB has enabled predictive sludge rheology correlations to be built up for different
sludge types.

System pressure losses are a function of the pipe size, length, roughness, fittings, net elevation changes, etc. as
well as the sludge rheological properties and desired flowrates. Design engineers need to estimate total pressure
losses on the suction and discharge sides of a pump for alternative pipe diameters and pipe velocities as well as
determining whether flow is laminar or turbulent. This leads to pump selection and sizing, derating the pump
3
performance for sludge properties and calculating pump power (volume flowrate x total pressure loss). The
system can then be optimised using sets of pump characteristics and system curves.

Currently, as with rheological property prediction, system loss design approaches are based on WRc’s TR 175
(Frost, 1982) and TR 185 (Frost, 1983) technical reports. The frictional pressure losses calculation approach
used in TR 185 is a simplified one and this critique will show that it is flawed. There was therefore a need to
update the TR 185 design methods with the latest published know-how on non-Newtonian slurry pumping,
integrate the calculations with the BHR Sludge Rheology Database (SRDB), and incorporate the calculation
approach into a software tool. This has been addressed in the WWM 6 research programme with the production
of the BHR System Losses Tool (SLOT).

Potential benefits of applying the new BHR System Losses Tool are capital cost saving through improved pump
selection and system design, operating cost savings through lower energy costs, lower labour costs, and higher
throughput.

Critique of TR 185 Rheological Terminology


TR 185 introduces a number of dubious terms not seen elsewhere in the rheological literature, either before or
since 1983. These include:
(1) ”laminar viscosity”, L;
(2) “turbulent viscosity”, T;
(3) “static resistance”, R.

Appendix C5, page 71 of the TR 185 report outlines the procedures for how values of ”laminar viscosity”, L,
“turbulent viscosity”, T and “static resistance”, R, are extracted from a flow curve which is modelled by the
Herschel-Bulkley model.

“Laminar Viscosity”

The use of the ”laminar viscosity”, L implies automatically that the sludge of interest is behaving as a Power
Law fluid as opposed to exhibiting Herschel-Bulkley behaviour, as shown in Fig C8 of the report (reproduced as
Figure 1). On page 74 of TR 185 (Appendix C5), it is stated that “the term L is obtained by approximating the
true flow curve to a simpler power law model over a limited shear rate or 8V/D” where V is the mean flow
velocity of the pipe and D is the internal pipe diameter. The range of 8V/D for this approximation was from 10
to 200 and it was suggested that a “maximum error of 5%” would be introduced using this approximation.
However, no evidence was provided to support this maximum error, which would seem to be significantly
underestimated. The use of a fixed range of 8V/D regardless of the pipeline application is incorrect.

“Turbulent Viscosity”

The “turbulent viscosity”, T is stated on page 74 of TR 185 (Appendix C5) as being “empirically related to the
viscous behaviour in laminar flow by
4
η T  1.77 η w  τ y  K(100) 
n 0.64
Equation 4

where w is the viscosity of water”. No explanation is given in TR 185 regarding how and why this expression
was derived. However, it is believed that it is an over-simplification so that the user can use the Newtonian
Moody chart for predicting friction in turbulent pipeflow and the criterion for laminar flow breakdown for
Newtonian flow.

The use of the term “turbulent viscosity” is reminiscent of “effective viscosity” and is to be equally discouraged.
Viscosities should only ever be defined for laminar flow conditions.

“Static Resistance”

The “static resistance” R is defined on page 10 of TR 185 as “a discrete force that has to be applied before any
movement or flow occurs”. Page 74 of TR 185 (Appendix C5) equates this “static resistance” to the yield stress
parameter of the Herschel-Bulkley model that was used to fit the upper bound flow curve. Thus it is not a
“force”, but a shear stress. To equate the “static resistance” to the yield stress parameter is contradictory as the
two quantities are defined quite differently.

The yield stress parameter is, in fact, the “dynamic” yield stress parameter as the flow curve data were obtained
using a controlled rate viscometer and were extrapolated to zero shear rate. Although the TR 185 recognises that
the sludges do exhibit a yield stress, it does not use this at all in the pipeline calculations for laminar flow.
Ironically, the yield stress parameter appearing in the Herschel-Bulkley model is used in a correlation of the
“turbulent viscosity” which is applied to turbulent pipeflow conditions only through Eqn 4.

The use of the unique rheological terms L, T and R in TR 185 makes it difficult for engineers to apply the
method to sludge flow curves fitted using the more common and rigorous Herschel-Bulkley model.

Critique of TR 185 Basis for Frictional Pressure Loss Estimation for Laminar Pipeflow
In laminar flow, the assumption is made that the flow curve of all types of sewage sludges is modelled using the
power law model over the range 10 < 8V/D < 200 s -1 where V is the mean flow velocity in the pipe and D is the
internal pipe diameter.

This is in spite of the yield-shear thinning behaviour exhibited by various sludges as shown by the example flow
curve plot given in Fig C10, page 74 of TR 185 (reproduced as Figure 2). These flow curves require a yield
stress parameter for the flow curve model and should not be ignored in the subsequent analysis. This is because
a flow curve representation is required over as wide a shear rate range as possible, in order to incorporate low
shear rate data. This is needed to predict frictional pressure loss in laminar flow since the shear rate in a pipe
reduces from a maximum at the wall to zero at the pipe axis.
5
The use of a fixed range of 8V/D of 10 to 200 s-1 independent of the pipe diameter and the mean flow velocity
application is also flawed. The justification for using a fixed range of 8V/D was attempted in Fig 2, page 12 of
the TR 185, (reproduced as Figure 3) which is a plot of 8V/D versus flow rate for a range of pipe diameters.
However, this justification is not correct since the whole of the shear stress vs 8V/D curve in Figure 3 is needed
to describe the flow property for the sludge correctly. As the shear stress is zero at the pipe centre and maximum
at the pipe wall, much of the shear stress vs 8V/D curve is needed in order to minimise the level of uncertainty
between the yield stress parameter and the shear stress corresponding to the lowest 8V/D chosen. It should also
be noted that 8V/D is not the actual wall shear rate for a non-Newtonian sludge but is an approximation.

Also flawed are the fixed values of the so-called “laminar viscosity” which are correlated with sludge type and
suspended solids (SS) in Table 5, page 18 of TR 185 (reproduced in Figure 4). This “laminar viscosity” is
defined as

n
 (1  3n) 
ηL  K  Equation 5
 4n 

where K is the consistency coefficient in Pa s n and n is the flow behaviour index. With n typically varying from
0.2 to 1, it follows from Eqn 4 that L is effectively K. This is because when n = 1 (Newtonian case), L = K,
and when n = 0.2, L = 1.15K.

The corresponding laminar flow Reynolds number is defined as

1 n
ρVD  8V 
Re L  Equation 6
η L  D 

This power law model Reynolds number is used in Fig 3, page 13 of TR 185 to estimate the Fanning friction
factor, f, and therefore the frictional pressure gradient, Pf/L, according to

ΔPf 2ρ V 2 f
 Equation 7
L D

This is appropriate, provided that

(a) the flow curve can be adequately represented over the relevant shear rate range using the power law model
(often, it cannot be, and the Herschel-Bulkley model or the Bingham plastic model should be used instead),
and
(b) the power law model parameters, K and n can be taken as independent of the nominal pipe wall shear rate
range, 8V/D, of 10 to 200 s-1; which is generally not valid.
6
Critique of TR 185 Basis for Frictional Pressure Loss Estimation for Turbulent Pipeflow
In turbulent flow, the effect of the non-Newtonian property of the sludges is underestimated in TR 185, as the
calculations involve the use of a Moody chart (reproduced as Figure 5) that is for Newtonian fluids only.
Because of these assumptions, there is a loss of clarity through not using the correct flow curve model and the
appropriate frictional pressure loss calculations.

TR 185 would appear to make estimates of the frictional pressure gradient in turbulent flow through a “turbulent
viscosity” (sic). Such a concept is flawed, as viscosity should be defined only for laminar flow conditions. This
“turbulent viscosity” is not adequately defined throughout TR 185 but it would appear that experimental values
of it have been inferred as follows:

(1) frictional pressure gradient measurements made for various sludge types in turbulent flow in possibly
different pipe sizes,
(2) the corresponding Fanning friction factors calculated,
(3) the corresponding (turbulent) pipe Reynolds number read off the Moody chart (reproduced as Figure 5) for
a given relative pipe wall roughness, e/D,
(4) “turbulent viscosity”, T, calculated using this Reynolds number via

DVρ
Re T  Equation 8
ηT

These values of “turbulent viscosity” were then correlated with sludge type and sludge suspended solids (SS)
and listed in Table 5, page 18, of TR 185 (reproduced as Figure 5).

Thus, to estimate the frictional pressure gradient in turbulent pipeflow, the reverse process is applied. A
“turbulent viscosity” is estimated from Figure 4, and a turbulent Reynolds number calculated from Eqn (8). For
a given pipe wall relative roughness, e/D, the Fanning friction factor is estimated from the Moody chart, and the
frictional pressure gradient calculated from the friction factor.

This is an inappropriate approach as it ignores any non-Newtonian property the sludge may exhibit, its detail,
and its influence in determining frictional pressure loss in turbulent pipeflow.

Critique of TR 185 Basis for Laminar Flow Breakdown and the Onset of Fully-Developed
Turbulent Pipeflow
The criterion used for laminar flow breakdown in TR 185 uses the definition of power law model Reynolds
number given by Eqn (6). However, this definition is for a power law fluid only and the critical value has been
shown to depend upon the flow behaviour index, n (Ryan and Johnson, 1959). This is ignored on page 12 of TR
7
185, where a constant critical value of Reynolds number of 2300 is taken. Depending on the value of n, critical
Reynolds numbers can range from 1800 to 2400.

There are several methods that are much more reliable than TR 185 in determining the laminar flow breakdown
for a non-Newtonian sludge:

(1) “Intersection method”: plots of frictional pressure loss calculated for laminar and turbulent flow conditions
are represented on the same figure, and the pipe velocity at which the two curves intersect is the critical
pipe velocity for laminar flow breakdown;
(2) Various methods (most of which agree well with each other) based on a critical Bingham Reynolds number
for Bingham plastic model sludges (Hanks, 1963; Hanks and Pratt, 1967; Wasp et al, 1977; Slatter and
Wasp, 2000)
(3) Methods for the Herschel-Bulkley model (Slatter, 1999).

The onset of fully turbulent flow is defined in TR 185 when the Reynolds number for laminar flow given by
Eqn (6), the “laminar” definition is greater than 4000. If found to be the case, the value of the friction factor is
read from the Moody chart, Figure 4, using Eqn (8), the “turbulent definition”, to predict the friction factor.

Critique of TR 185 Basis for Prediction of Valve and Fitting Loss Coefficients
The prediction of the loss coefficients for laminar and turbulent flow of three different types of sewage sludges
in various fittings and valves is outlined in Section 3.2.2 of the TR 185 report. This approach is not referenced
within the report or elsewhere in the literature. Table 4, page 15 of the report lists the K factors for turbulent
flow only given in the Chemical Engineers’ Handbook (Perry et al, 1973) for the different fittings and valves
used in the water industry. Whilst recognising that these factors are applicable to turbulent flow of water only,
TR 185 attempts to extend its applicability to laminar and turbulent flow of non-Newtonian fluids.

A plot of the fitting losses factor, F versus the pipe Reynolds number, Re as defined by Eqn (6) is given in Fig 5,
page 5 of the TR 185 (reproduced as Figure 6) for primary, activated and digested sludge. No reference was
made within the report regarding how this plot was obtained. A curve fit through the dataset was made and this
was found to be

2000
F  1.0  Equation 9
Re

with Re defined by Eqn (6). This Re is the Power Law Reynolds number definition for power law fluids.

The factor “F” is defined as the ratio of the pressure loss for the sludge divided the pressure loss for water under
the same flow conditions, i.e., same velocity through the fitting.
8
Using Eqn (9) together with the k-factor values from the Chemical Engineers’ Handbook (Perry et al 1973) for
turbulent flow of water, the additional head loss of sludge across the fitting is given in TR 185 as

V2
h fitt  F k Equation 10
2g

The product of F.k versus Reynolds number plot will show typical behaviour relating to the laminar and
turbulent regimes for a fitting or a valve.

Moreover this approach makes the tacit assumption that the adjusted loss coefficient, F.k vs Reynolds number
relationship for non-Newtonian fluids matches exactly that found for Newtonian fluids. Research by various
authors (e.g. Edwards et al, 1985; Ma, 1987; Pienaar, 1998) on a variety of Newtonian and non-Newtonian
fluids since the publication of TR 185 has shown that this is indeed the case provided the appropriate Reynolds
number is used for the fluid under test.

However, the use of turbulent loss coefficients in combination with Figure 6 to generate a loss coefficient versus
Reynolds number plot covering both laminar and turbulent regimes for that fitting or valve is unproven in terms
of whether its usage is correct.

Research since 1983 when TR 185 was published on various fittings with Newtonian and non-Newtonian fluids
has shown the loss coefficient vs Reynolds number relationship is very dependent on the type and size of fitting
used. This, in turn, has led to different critical Reynolds numbers for distinguishing laminar flow from turbulent
flow for different fluids in various fittings. Figure 6 suggests there is only one critical Re value for all fittings
which is incorrect and highly misleading.

Given the extensive research conducted since TR 185 on the estimation of frictional pressures for flow of
Newtonian and non-Newtonian fluids through fittings and valves, the approach used in TR 185 should not be
used anymore, and the large research dataset used instead.

BHR Group’s System Losses Tool (SLOT)


The System Losses Tool developed by BHR Group Limited was created to provide a calculation tool for
WWM6 members. The tool calculates:

 Friction losses in straight pipes,


 Friction losses through a wide range of pipe fittings,
 Losses or gains in pressure due to elevation changes.

Sludge rheological properties are entered using the Herschel-Bulkley model parameters. Sub-models of the
generalized Herschel-Bulkley model such as power law and Bingham Plastic model are provided as options.
9
Both the laminar and turbulent regimes for non-Newtonian flows in straight pipes are covered. The option of
choosing from various different turbulent flow pressure loss models is given. For sludge flow curves fitted using
either the Herschel-Bulkley or power law rheological models, BHR recommends that the user adopts the
Wilson-Thomas model (Wilson and Thomas, 1985) in turbulent flow. Another recent turbulent flow model that
is supported by SLOT for the Herschel-Bulkley sludges is the Chilton-Stainsby model (Chilton and Stainsby,
1998).

The laminar flow breakdown in straight pipes is based on the critical Reynolds number according to Ryan-
Johnson model (Ryan and Johnson, 1959) for power law fluids and the Slatter model (Slatter, 1996) for
Herschel-Bulkley fluids.

There is a three-stage plan for implementing fitting loss coefficients in the BHR SLOT. The current version of
SLOT uses an approach for turbulent flow of Newtonian fluids. A wide range of fitting types are covered
including various geometries of elbows, tees, reducers, expanders, valves, tank entries and exits, orifices and
user defined fittings. BHR SLOT is being upgraded to provide approaches for the prediction of loss coefficients
of laminar flow of Newtonian fluids as well as laminar and turbulent flow of non-Newtonian fluids. The BHR
SLOT uses the Hooper 2K method (Hooper, 1981) to determine the critical Reynolds number for laminar flow
breakdown in different pipe fitting types.

When using SLOT a series of sludge flowrates can be entered in order to generate a system curve, pump
pressures can also be entered to generate a pump curve.

Ongoing work at BHR Group to develop SLOT further is focusing on:


 Fitting losses for non-Newtonian fluids
 Validation of calculations using measured non-Newtonian pressure loss data
 Manifolds & branches
 Pump start-up calculations
 Derating pumps
 Fouling factors
 Gas-liquid flows

Comparison between TR 185 & BHR SLOT


A series of comparisons of pumping system loss calculation results have been carried out to determine the effect
of the erroneous assumptions and approximations used in TR 185. The comparisons have been based on the
following:

 Identical sludge rheology values input to TR 185 & SLOT based on Herschel-Bulkley model parameters.
 Identical pipeline system and fitting characteristics input to TR 185 & SLOT.
 Identical sludge flowrates input to TR 185 & SLOT
10

Three sludges were selected for the comparison with varying yield stress and shear-thinning characteristics.

Sludge A: Primary poly-thickened (7% DS), laminar pipeflow, yHB = 15.5 Pa; K = 10.4 Pa sn: n = 0.45

Sludge B: Digested (3% DS), laminar & turbulent pipeflow, yHB = 0.17 Pa; K = 0.37 Pa sn: n = 0.36

Sludge C: Digested (6% DS), laminar & turbulent pipeflow, yHB = 1.1 Pa; K = 1.7 Pa sn: n = 0.32

Comparisons were made for a range of flowrates and pipe or fitting diameters. Representative samples of the
results are illustrated in Figures 7 to 10.

The plots show the ratio of pressure drop calculated using SLOT divided by pressure drop calculated using TR
185. In almost all cases, TR 185 pressure drop values were higher than the values calculated using SLOT. In
other words, TR 185 overestimated frictional pressure losses by from 5% to 60% compared with the up-to-date
and rigorous SLOT.

Figure 7 shows straight pipe, laminar flow pressure drop comparisons. TR 185 pressure drop estimates were
5%-15% higher depending on pipe velocity. The variation in magnitude of the yield stress between sludges did
not seem to affect the SLOT/TR 185 pressure drop ratio.

Figure 8 shows straight pipe, turbulent flow pressure drop comparisons based on the SLOT Wilson Thomas
model option (Wilson and Thomas, 1985). TR 185 pressure drop estimates were 10%-55% higher depending on
pipe velocity. Comparisons based on the SLOT Chilton-Stainsby model option (Chilton and Stainsby, 1998)
ranged from 10% lower to 15% higher. The Wilson Thomas model features some experimentally validated non-
Newtonian friction factors which are significantly lower than the Newtonian friction factors used in TR 185.

Figures 9 and 10 show an example fitting type (90 0 long radius elbow) for which a direct comparisons could be
made. Frictional losses calculated using the more advanced SLOT Hooper 2K method (Hooper, 1981) were only
50% to 60% of those calculated using TR 185 in both laminar and turbulent flow.

Other fitting types for which direct comparison was possible were

45-degree elbow long radius: TR 185 11%-20% greater than SLOT


Gate valve fully open: TR 185 35%-40% greater than SLOT
Diaphragm valve fully open: TR 185 10%-85% greater than SLOT
11
Conclusions
The TR 185 approach to estimating pipe frictional pressure losses can be summarised as follows. The summary
illustrates the inconsistent methodology that it involves.

(1) Assume that the sludge will exhibit laminar flow property which can be modelled by the Herschel-Bulkley
model;
(2) However, ignore the yield stress parameter in this model and approximate the flow curve using the two-
parameter (K, n) power law model in order to be able to predict a “laminar viscosity” as a function of
sludge suspended solids (SS), as given in Table 5 of TR 185 (reproduced in Figure 4).
(3) Use this “laminar viscosity” in the power law Reynolds number definition to estimate the Fanning friction
factor, and therefore frictional pressure loss, in laminar flow.
(4) Check that the Re value is less than 2300.
(5) If Re > 4000, estimate the frictional pressure loss using the Newtonian Moody chart (even though the
sludge will usually be non-Newtonian) which involves a “turbulent viscosity” in the Re definition.
(6) Obtain an estimate for the “turbulent viscosity” as a function of sludge suspended solids (SS) again using
Table 5 of TR 185 (reproduced in Figure 1), noting that this “turbulent viscosity” has been correlated with
the three parameters (including the yield stress parameter) appearing in the Herschel-Bulkley model!

If TR 185 had been explicit in setting out the steps to be taken in this way, the flawed approach taken would
have been more apparent.

The main shortcomings of TR 185 can be summarised as follows:

Laminar Flow
 Totally ignores the Herschel-Bulkley curve fit to the measured flow curve for predicting pipe pressure
losses in laminar flow.
 Uses a fixed range 10 < 8V/D < 200s –1 for all sludge types and flow situations which is incorrect.
 Unnecessary use of the “laminar viscosity” term – this is effectively the consistency coefficient.
 TR 185 approach for estimating frictional pressure gradient in laminar flow of sludges exhibiting power law
(not yield-shear thinning) behaviour is a tortuous one.

Turbulent flow
 Use of “turbulent viscosity” is conceptually flawed as viscosity is defined under laminar flow conditions
only.
 TR 185 approach for estimating frictional pressure gradient in turbulent flow is wrong as it uses the
Newtonian Moody chart for the pipe friction factor.
 There are many more recent, improved ways to predict frictional pressure loss in turbulent flow for non-
Newtonian sludges.
12
Laminar flow breakdown and onset of fully-turbulent pipeflow
 Use of a fixed value of critical Reynolds number (2300) for laminar flow breakdown is incorrect.

Fitting Losses
 Approach outlined is unproven and outdated.
 Very limited range of fitting types covered (8 in total).

TR 185/SLOT comparison
 Frictional pressure losses calculated using TR 185 for straight pipes and fittings in both laminar and
turbulent flow were almost universally higher than those calculated using the more rigorous and up-to-date
SLOT.
 The largest consistent differences were found when comparing fitting losses, where TR 185 values were
from 10% to 85% greater than SLOT.
 The smallest consistent differences were found for laminar pipeflow (5% to 15%) and turbulent pipeflow
using the Chilton-Stainsby model (-10% to 15%).
 These results suggest that pump specification based on TR 185 system loss calculations will result in
significant oversizing, especially if additional ‘safety factors’ are added.
 Use of SLOT for system loss calculations will therefore significantly reduce both capital and operating
costs through use of smaller pumps when compared with TR 185.

Acknowledgements
The authors would like to acknowledge the funding and guidance provided by the WWM consortium members
during the development of the BHR System Losses Tool.

The authors would also like to acknowledge and thank WRc plc for giving permission to reproduce elements of
Technical Report TR 185.

References
Chilton, R.A and Stainsby, R. (1998), Pressure loss equations for laminar and turbulent non-Newtonian pipe
flow, J. Hydr. Engng., 124 (5), 522-529.
Edwards, M.F., Jadallah, M.S.M. and Smith, R. (1985), Head losses in pipe fittings at low Reynolds numbers,
Chem. Eng. Res. Des., 63 (1), 43-50.
Frost, R.C. (1982), WRC Water Research Centre Technical Report TR 175, ‘Prediction of Friction Losses for the
flow of Sewage Sludges in straight pipes’.
Frost, R.C. (1983), WRC Water Research Centre Technical Report TR 185, ‘How to design sewage sludge
pumping systems’.
Hanks, R.W. (1963), The laminar-turbulent transition for flow in pipes, concentric annuli and parallel plates.
AIChE J., 9, 45-48.
13
Hanks, R.W. & Pratt, D.R. (1967), On the flow of Bingham plastic slurries in pipes and between parallel plates.
Soc Petrol Eng J, 1, 347.

Hooper, W. B. (1981), The two-K method predicts head losses in pipe fittings. Chemical Engineering, Aug 24,
96–100.
Ma, T.W. (1987), Stability, rheology and flow in pipes, fittings and venturi meters of concentrated non-
Newtonian suspensions, PhD thesis, University of Illinois, Chicago .
Perry, R.H. and Chilton, C.H. (1973), Chemical Engineers Handbook, 5th Ed, McGraw Hill, 5-18 to 5-38.
Pienaar, V.G. (1998), Non-Newtonian fitting losses, M.Tech. thesis, Cape Technikon, Cape Town.
Ryan, N.W. and Johnson, M.M. (1959), Transition from laminar to turbulent flow in pipes, AIChE Journal, 5,
433-435.
Slatter, P.T. (1996), The laminar/turbulent transition – an industrial problem solved, Hydrotransport 13, pp 97-
114, BHR Group, Cranfield, UK.
Slatter, P.T. (1999), A New Friction Factor for Yield Stress Fluids. In: Proc Hydrotransport 14, Maastricht,
Netherlands, pp 255-265.
Slatter, P.T. and Wasp, E.J. (2000), “The Laminar-Turbulent Transition in Large Pipes”. 10th International
Conference on Transport and Sedimentation of Solid Particles, Wroclaw, Poland, 4-7 Sept, pp 389-399.
Wasp, E. J., Kenny, J. P. & Gandhi, R. L. (1977), “Solid Liquid Flow: Slurry Pipeline Transportation” Published
by Trans Tech Publications, New York.
Wilson, K.C. (1996), Laminar-turbulent transition locus for power law non-Newtonians, Hydrotransport 13, pp
61-74, BHR Group, Cranfield, UK.
Wilson, K.C. and Thomas, A.D. (1985), A new analysis of the turbulent flow of non-Newtonian fluids, Can. J.
Chem. Eng., 63, 539-546.
14th European Biosolids and Organic Resources Conference and Exhibition

Figures

Figure 1: Reproduction of TR 185 Figure C8

Figure 2: Reproduction of TR 185 Figure C10

www.european-biosolids.com
Organised by Aqua Enviro Technology Transfer
14th European Biosolids and Organic Resources Conference and Exhibition

Figure 3: Reproduction of TR 185 Figure 2

Figure 4: Reproduction of TR 185 Table 5

www.european-biosolids.com
Organised by Aqua Enviro Technology Transfer
14th European Biosolids and Organic Resources Conference and Exhibition

Figure 5: Reproduction of TR 185 Figure 4

Figure 6: Reproduction of TR 185 Figure 5

www.european-biosolids.com
Organised by Aqua Enviro Technology Transfer
14th European Biosolids and Organic Resources Conference and Exhibition

Laminar Flow
(Pf SLOT /L)/ (Pf TR185/L) vs V [m/s]
D=0.3 m
1.2

SLOT Sludge A

SLOT Sludge B

1.1 SLOT Sludge C

TR185
Pf SLOT /L)/ (Pf TR185/L)

1.0

0.9

0.8
0 1 2 3
V (m/s)

Figure 7: Straight Pipe, Laminar Flow, (Pf SLOT /L)/ (Pf TR 185/L) vs V [m/s]

Turbulent Flow
(Pf SLOT /L)/ (Pf TR185/L) vs V [m/s]
D=0.3 m
1.2

1.1

1.0

0.9
Pf SLOT /L)/ (Pf TR185/L)

0.8

0.7

0.6

0.5 SLOT Sludge B, Wilson Thomas

0.4 SLOT Sludge C, Wilson Thomas

0.3 TR185

0.2
1 2 3
V (m/s)

Figure 8: Straight Pipe, Turbulent Flow, (Pf SLOT /L)/ (Pf TR 185/L) vs V [m/s]

www.european-biosolids.com
Organised by Aqua Enviro Technology Transfer
14th European Biosolids and Organic Resources Conference and Exhibition

Laminar Flow
o
90 Elbow fitting, long radius, D=0.3m
1.2

1
( Pf SLOT /L)/ ( Pf TR185/L)

0.8

0.6
TR185
SLOT Sludge A Laminar
0.4
SLOT Sludge B Laminar
SLOT Sludge C Laminar
0.2
0 1 2 3

V (m/s)

Figure 9: 900 Elbow, Laminar Flow, (Pf SLOT /L)/ (Pf TR 185/L) vs V [m/s]

Turbulent Flow
o
90 Elbow fitting of long radius, D=0.3m
1.2

1
( Pf SLOT /L)/ ( Pf TR185/L)

TR185
0.8
SLOT Sludge B Turbulent
SLOT Sludge C Turbulent
0.6

0.4

0.2
1 2 3
V (m/s)

Figure 10: 900 Elbow, Turbulent Flow, (Pf SLOT /L)/ (Pf TR 185/L) vs V [m/s]

Word Count: 5007

www.european-biosolids.com
Organised by Aqua Enviro Technology Transfer

You might also like