UNIVERSITY OF SOUTH AFRICA

SCHOOL OF ENGINEERING
CONTINUOUS ASSESSMENT

SOLVED EXAMPLES
 QUESTION
A closed loop water distribution network is shown in figure 1. The network has two loops
with pipes of length, L and diameters, D. Assume a friction coefficient λ of 0.072 for all
pipes. All pipes have a length of 200m and a diameter of 600mm.

3 3
0.22 m /s 0.020 m /s

A B

D C E
3
0.025 m /s

3 3
0.06 m /s X m /s

Figure 1: Pipe network

You are required to

3
1.1 Determine the flow X m /s.
1.2 Assuming initial flows through AB and AD to be equal, draw a balanced pipe
network.
1.3 Using the Hardy Cross method, carry out 1 iteration to correct flows.
1.4 Calculate the new friction losses in the pipes.
1.5 Assuming the elevation at point A to be 75 m and all points are at equal
elevation, determine the pressure at all points.
1.6 What is the closing error at point E?

a) Total inflow=Total outflow

0,22=0,02+0,06+0,025+X
X=0,115

b)
 Darcy Weisbach
c) ITERATION 1
Loop I Correcte
d flows

Pipe No. Flow(m3/ Diameter( Length(m) Area Velocit λ hf hf/Q Q1

s) m) y
AB 0,110 0,600 200,000 0,28 0,389 0,072 0,185 1,685 0,111
3
BC 0,045 0,600 200,000 0,28 0,159 0,072 0,031 0,689 0,044
3
CD -0,050 0,600 200,000 0,28 -0,177 0,072 -0,038 0,766 -0,049
3
DA -0,110 0,600 200,000 0,28 -0,389 0,072 -0,185 1,685 -0,109
3
-0,00728 4,82479
1

 hf  
 
DQ =  exp
 
h f / Q 


0,00075
4

Loop II
Pipe No. Flow(m3/ Diameter( Length(m) area velocit hf hf/Q Q1
s) m) y
BE 0,045 0,600 200,000 0,28 0,159 0,072 0,031 0,689 0,047
3
EC -0,020 0,600 200,000 0,28 -0,071 0,072 -0,006 0,306 -0,018
3

Loop I
d) Pipe No. Q1 V HF(New)
AB 0,1108 0,3919 6,5092
BC 0,0439 0,1555 1,0243
CD -0,0492 -0,1743 1,2869
DA -0,1092 -0,3866 6,3332

Loop II
Pipe No. Q1 V HF
BE 0,047 0,166 1,163
EC -0,018 -0,064 0,175
CB -0,044 -0,155 1,024

e) A 75 75
B A-hfAB 68,491
C B-HfBC 67,466
C A-hfAD- 67,380
hfDC
D A-hfAD 68,667
 E C-hfCE 67,291
E B-hfBE 67,328

f) Closing error at E 0,037

2
SOLVED QUESTIONS

QUESTION
2
(a) The head loss in a pipe can be expressed in the form hf = KQ . Two pipes having constants K1
and K2 are to be considered as a single equivalent pipe. Determine the value K3 of this single
pipe when the two are laid:

i. in series
ii. in parallel.

SOLUTION PART A

i. In series the flow is the same and total head loss is the sum of the two.
2 2 2 2 2
hf1 = k1Q hf2 = k2Q hf1 + hf2 = k3Q = k1Q + k2Q
Hence k3 = k1 + k2
ii. In parallel the friction heads are the same and the flows different.
2 1/2
hf = k1Q1 Q1 = (hf /k1)
2 1/2
hf = k2Q2 Q2 = (hf /k2)
2
hf = k3(Q1 + Q2)
h h 2 h 2h
hf  k3 f  f  k3 f  f 
h
f
k k2 k 1 k kk
1
2 1 2
1 1 2
1  k3 k  k 
kk
 1 2
1 2

k 1
3  1 1 2
k k kk
1 2 1 2

(b) When the flow rates are expressed in litres per second and the head losses in metres, K values
for the pipe systems shown are as given in the table. Under a particular set of inputs and
demands the network experienced the flow rates indicated.

The head loss in the system was considered to be

excessive and a second pipe was alongside pipe 3
so that they carried flow in parallel. The equivalent
single pipe for these two pipes has k = 0.000818
2 2
ms /litre . When the pipe had been installed the
 pipe flows shown changed but the inputs and
demands on the system remained the same.

Use the flows shown as initially assumed flows and

apply an iterative method of network analysis to

determine the changed flows in the pipes. Make only two rounds of corrections to the initial
flows.

Pipe 1 2 4 5
2 2
K ms /l 0.000570 0.012118 0.001698 0.006946
(Pipe 3 has K = 0.000818 in question)
SOLUTION PART B
The problem must be solved as two loops with a common pipe 3. Start with loop 1 with the flows
shown. Data is shown for initial guess. Note clockwise flow is positive.

Starting data
First iteration loop 1 (pipes 1, 2 and 3)
PIPE K Q hf hf/Q
1 0.000570 204 23.7212 0.11628
2 0.012118 104 131.068 1.260
3 0.000818 -123 -12.376 0.1006
142.41.4772
h
δQ 
∑ f 
142.4
 48.2 Correct all flows in loop 1 by subtracting 48.2
2∑hf /Q 2 x 1.4772

First correction shown above

First Iteration loop 2 (pipes 3,5 and 4)
PIPE K Q hf hf/Q
3 0.000570 171.2 23.97 0.1400
5 0.006946 -123 -105.08 -0.8544
4 0.001698 -173 -50.82 -0.2938
h -131.93 1.288
∑ f -131.93
δQ    −51.208
2∑hf /Q 2 x 1.288
Correct all flows in loop 2 by subtracting -51.2

Second correction
Second iteration loop 1
PIPE K Q hf hf/Q
1 0.000570 155.8 13.8 0.0888
2 0.012118 55.8 37.7 0.676
3 0.003272 -222.4 -40.5 0.947
h 11.1 0.947
∑ f 11.1
δQ    5.9
2∑hf /Q 2 x 0.947
Correct all flows in loop 1 by subtracting 5.9

After third correction

Second iteration loop 2
PIPE K Q hf hf/Q
3 0.000570 228.27 42.6 0.187
5 0.006946 -71.8 -35.7 0.499
4 0.001698 -121.8 -25.2 0.207
h -18.4 0.892
∑ f -18.4
δQ  −  −10.3
2∑hf /Q 2 x 0.892

Results after 2 iterations

QUESTION

(a) Compare and contrast the following two iterative calculation methods for complex networks of
pipes.
(i) the head balance method (also known as the Hardy Cross or loop method).
(ii) the flow balance method (also known as the quantity balance or nodal method.

Explain briefly how and in what situation each of the methods may be used and state which of
the correction methods shown at the end of this question is used in which method.

SOLUTION PART (a)

The nodal balance method is used for solving problems involving many pipes with a common
junction where the total flow into the junction must be zero. The correction factor used for iteration
is
∆H  ∆
2 Q

∑Q/hf
The flow balance method is used for problems with multiple loops where the total head loss around
a given loop is zero. The correction factor to be used is
∆Q  − ∑ f
h

2∑hf /Q
(b) Water is supplied from a large reservoir at A to a pipe network BCDE as shown, in the diagram.
The frictional resistances of the various pipes are given by the
K value in the table which may be used with the formula
2
hf = KQ to relate the magnitude of head loss hf in the pipeline
to the volumetric flow rate Q. Water is drawn at constant flow
rates from the network at nodes C and D. The static heads
(elevation + pressure head) at nodes B, C and D are 100m, 65m
and 61m respectively above the local datum. Calculate the
discharges at C and D and the water level in reservoir A. (The
data has been added to diagram to aid the solution)
Use no more than 3 iterations and 3 significant figures

TABLE
Pipeline AB BC CD 40 DE CE BE
K s2m5 4 110 25 25 35

SOLUTION PART (b)

The problem must be solved as two loops with a common pipe EC. First calculate the flows in
known pipes.

BC hf = 35 m Q = (hf /K)1/2 = (35/40) 1/2 = 0.935 m3/s

CD hf = 4 m Q = (hf /K)1/2 = (4/110) 1/2 = 0.191 m3/s

The solution evolves around doing a flow balance at node E.

1st ITERATION Guess hE = 80

1/2
PIPE K hf Q = (hf/K) Q/hf
BE 35 20 0.756 0.0378 (into junction)
EC 25 -15 -0.775 0.0516 (out of junction)
ED 25 -19 -0.872 0.0349 (out of junction)
Totals -0.89 0.135
h 
∑ f
2 ∆Q

2 x (-0.89)
 −13.16
∑Q/hf 0.135
2nd ITERATION Guess hE = 80 -13.16 = 66.84
1/2
PIPE K hf Q = (hf/K) Q/hf
BE 35 20 0.0294 0.0378
EC 25 -15 -0.148 0.0516
ED 25 -19 -0.083 0.0349
Totals 0.219 0.26

h  2 ∆Q 2 x (0.219)
∑ f
∑Q/hf   1.686
0.26

3rd ITERATION Guess hE = 66.84 + 1.69 = = 68.53

1/2
PIPE K hf Q = (hf/K) Q/hf
BE 35 20 0.948 0.0301
EC 25 -15 -0.376 0.106
ED 25 -19 -0.549 0.0729
Totals 0.0241 0.23

Further iterations will show only minor corrections giving of

flows 0.945, -0.388 and -0.557. If these figures are used you get
the answers given by the examiner.

We can now calculate Qc and Qd

0.935 + 0.376 – 0.191+ Qc = 0
3
Qc = -1.12 m /s

0.549 + 0.191 + Qd = 0
3
Qd = -0.74 m /s
3
Total flow from the reservoir is 1.12 + 0.74 = 1.86 m /s
2 2
Head loss pipe AB is hf = kQ = 4 x 1.86 = 13.8 m
Head at entrance to pipe is 113.8 m
QUESTION

(a) A pipeline of diameter D = 0.5 m has a length L = 200 m, and the value of the Darcy friction λ
may be assumed to have a constant value of 0.024. The pipeline contains two fully open valves, the
2
local head loss at each of which is 0.2v /2g, and three bends at each of which the head loss is
2
0.5v /2g where V is the velocity of water in the pipe. Calculate the value of K in the expression h =
2
K Q relating the total head loss h to the flow Q through the pipeline.
SOLUTION part (a)

Note λ = 4Cf
2 2 2
4Cf Lv (0.024)(200)v 9.6v
Straight pipe hf  2gD  2g(0.5)  2g

2 2 2 2
9.6v 2(0.2)v 3(0.5)v 11.5v
Total for pipe line hf    
2g 2g 2g 2g
v  Q  4Q  5.093Q
2
A π(.5)
2
11.5(5.093)Q 2 2 5
hf   2.985Q hence K = 2.985 s /m
2g

(b) For a network of pipelines, such as that described in part (a), show that the flow correction term
in an iterative head balance calculation is given by
− ∑hf

∆Q  2∑hf /Q
SOLUTION part (b)

n 2
Starting with hf = K Q Normally n = 2 so hf = K Q
2
2KQ dQ 2

Differentiate to get dhf  2KQdQ  Q and since K Q = hf

2hf dQ Qdhf
dhf  or dQ 
Q 2hf
2hf δQ Q δhf
If this relationship holds approximately true for finite changes then δhf  or δQ 
Q 2hf
In a balance of heads, the flow is corrected until ∆θ = 0 so the correction factor to be used for each
pipe is δQ  - Q δhf  − δhf (The correction must be to reduce the flow rate).
2h h
f 2 f
Q
− ∑hf

For a network we must total all the terms to give the total correction factor of ∆Q  2∑hf /Q
 3
(c) The diagram shows two loops of a horizontal network with inflows and outflows in m /s. The K
values of the seven pipes are given in the table. The pressure head at node A is 25 m. Calculate the
flow rate through each pipe and the pressure head at each node. No more than two rounds of
iteration are required, and final values of pressure heads may be rounded to the nearest metre.

Pipe AB BC CD DE 20 BE EF AF
K (s2/m5) 2 2 20 10 10 10

SOLUTION part (c)

The problem must be solved as two loops with a common pipe BE. First make a guess at the flow
rates. Bear in mind that the net flow is zero at all nodes.

Data shown for initial guess

Start with loop ABEFA

h
PIPE K Q hf hf/Q ∑ f 3.92
∆Q− −  −0.153
AB 2 1.9 7.22 3.8 2∑hf /Q 2 x 12.8
BE 10 0.2 0.4 2
EF 10 -0.6 -3.6 6
FA 10 -0.1 -0.1 1
Totals 3.92 12.8

Correct all flows in this loop by adding -0.153

First loop correction

Now do loop BCDEB

PIPEK Q hf hf/Q
h
∑
BC 2 0.7 0.98 1.4
CD 20 0.2 0.8 4 f 0.002
DE 20 -0.3 -1.8 6 ∆Q− 2∑hf /Q − 2 x 11.869  −0.000083
BE 10 0.04688 0.021973 0.46875
Totals 0.001973 11.86875

Correct loop 2 The initial guess was so good that the correction is minor

Second iteration of loop 1 is:

h
PIPE K Q hf hf/Q ∑ f 0.231
∆Q− −  −0.00825
AB 2 1.74688 6.10314 3.49375 2∑hf /Q 2 x 14.02
BE 10 -0.0468 -0.02189 0.467919
EF 10 -0.7531 -5.67197 7.53125
FA 10 -0.2531 -0.64072 2.53125
-0.23145 14.02417

Final solution is

Pipe AB BC CD DE BE EF AF
Q 1.75 0.7 0.2 -0.3 ±0.05 -0.75 -0.25

Head at A is 25 m and rounding off the hf values

Head B is 25 - 6 = 19 m
Head at C is 19 – 1 = 18 m
Head at D = 18 - 1 = 17 m
Head at E = 17 + 2 = 19 m or 19 - 0 = 19 m
Head at F = 19 + 6 = 25 m or 25 -
1 = 24 m error due to rounding off
values.
QUESTION
3
It is required to pump water at a rate of 0.0160 m /s against a total head of 30.5 m. Four
geometrically similar pumps, whose sizes are 100 mm, 125 mm, 225 mm and 300 mm, are
available.

The characteristics of the 100 mm size pump, tested at 150 rad/s, are tabulated below.
3
Discharge 0 0.0076 0.0151 0.0226 0.0302 m /s
Head 43.9 46.1 43.9 34.2 14.6 m
Efficiency 0 48 66 66 45 %

(a) Determine which pump should be used, and the speed at which it should be driven, so that
maximum possible efficiency is obtained.

(b) If, temporarily, only the 125 mm pump is available, determine the speed of operation and
the input power from the motor, necessary to satisfy the head and discharge requirements.

By plotting the data for the 100 mm pump we can determine that the optimal point (for max
3
efficiency) is when Q = 0.0188 m /s and H = 40 m. The peak efficiency is 68%

3
For the 100 mm pump H = 40 m Q = 0.0188 m /s N = 150 rad/s
1/2
N Q1/2 150 x 0.0188
Ns  1 1
  1.293 rad/s (12.34 rev/min)
3/4 3/4
H 40
For the required condition
1/2
1.293  N x 0.016 Hence N = 131 rad/s (1251 rev/min)
3/4
30.5
For the optimal size, remember that condition (1) is the optimal condition of the pump
and condition (2) is the actual operating conditions.
1/ 3
QN 0.016 x 150 1/ 3
D2 2 1
Equating Flow Coefficients we get D  QN  1
1 1 2 0.0188 x 131
Equating head coefficients we get we get D2  N1 H2  150 30.5  1
D N
1 2 H
The 100 mm seems to be the best. 1 131 40
(b) 125 mm pump at the same speed
The larger pump must slower to obtain the same flow. First calculate the corresponding flow
and head for the 100 mm pump.

3
D2 125 3
For the same Flow coefficient Q  0.016  Q Q  1.953Q  1.953 x Q
2 1 D 1100 1 1
1
2 2
D2 125
For the same Head coefficient H 40H H  1.562H
2 1 D 1 100 1
1
3
Q1 0 0.0076 0.0151 0.0226 0.0302 m /s
H1 43.9 46.1 43.9 34.2 14.6 m
Efficiency 0 48 66 66 45 %
Q2 0 0.0148 0.0295 0.0441 0.059
H2 68.6 72 68.6 53.4 22.8
Plotting H2 and Q2 gives the curve shown. It is assumed that the efficiency is unchanged.

As can be seen we cannot obtain the required operating point at 150

rad/s. For the same flow coefficient between at two different speeds
QB QA N
3  QB QA B
NB DB NA D A 3 NA
For the same Head Coefficient at two different speeds
g HA g HA NB 2 NB 2
2 2 2 2 HB HA  HA
NA D A  NA D A NA
2
NA
2

Substitute NB  QB to eliminate the speed

NA QA
2 2
QB H QA
A
H H
B A

Q
 HB

Q Where A and B corres pond to di fferent s peeds .

A B

For the case in hand let B be the values at the new speed and A the values at 150 rad/s
Q A 2 2
HA  30.5  119141QA

0.016
Calculate the flows at the new speed for the 125 mm pump.

Efficiency 0 48 66 66 45 %
QA 0 0.0148 0.0295 0.0441 0.059
HB 0 26 104
Plotting HB we get the result shown. We require the speed to produce operating point B for the
same size (125 mm).
From the Flow Coefficient between points A and B.
QB QA 3

NB DB 3  N A D A 3 QA = 0.025m /s and HA = 70 m
0.016  0.025
NB 150
NB = 96
rad/s

Check by repeating the process with the head coefficient.

g HB g HA
 NBNA HB  150 30.5  99
2 2 2 2
NB DB NA DA HA 70
The efficiency at this point is 62%

Water Power = mgH = 16 x 9.81 x 30.5 =

4787 W Power Input = WP/η = 4787/0.62

=7720 W