Proposed60MGDWaterworksandAncillaryFacilitiesat

LowerSeletar

Report:Pumpsumpmodellingmembranefeedpumpstation

147216SECM170012rev0

PreparedforPUB

EXECUTIVE SUMMARY

1.

There is a requirement to carry out hydraulic sump model study for the membrane
feed pumping station at Lower Seletar Waterworks (LSWW).

2.

The physical model of the pump sump is constructed at the model-to-prototype scale
of 1:15. Froudian modelling criteria are used for the flow conditions and dimensions
of model. The physical model system is setup in the Maritime Research Centre,
Nanyang Techanological University.

3.

Based on the findings and observations, the hydraulic performance of the pump
sump is generally satisfactory at top water level but not acceptable at low water
level.

4.

The influent flow is turbulent and fluctuant at low water level. If two or more pumps
on duty in one tank, the strong inflow cause splashing and bubbles. The bubbles can
be sucked into nearby pump intakes, such as A1, A2, and A5, which can lead to
severe pump problems.

5.

No obvious bubbles are observed at top water level because the inflow is submerged.

6.

Due to the asymmetrical inflow into the narrow tank, the approaching flows to
different pumps are not parallel and straight. Such asymmetrical flow pattern can
cause circulations in the tank and pre-rotation around pipe intakes, especially at low
water level.

7.

At low water level, surface vortex can be generated and air can be sucked into
pumps, for example A3 in Tank 1 and A4 in Take 2.

8.

At top water level, there is no surface vortex observed because the pipe intake
submergence in increased.

9.

The submergence of pipe intakes at low water level is insufficient to avoid formation
of surface vortices.

10. The submergence of pipe intakes at top water level is sufficient to avoid formation of
surface vortices.
11. The pipe intake clearances from the floor and back wall are adequate to avoid
formation of vortices for all the pumps.

12. The submergence of influent flow at low water level is insufficient to avoid
formation of strong surface perturbations and air bubbles.
13. The submergence of influent flow at top water level is sufficient to avoid formation
of air bubbles.
14. Do not operate the pumps at design low water level (118.5 mRL);
15. The minimum level should be high enough to have a sufficient submergence of
inflow pipes as well as intake pipes. The recommended LWL is 120.4 mRL;
16. Higher water level is always more favourable for hydraulic performance in the sump.

CONTENTS
EXECUTIVE SUMMARY .............................................................................................. 1
1.

INTRODUCTION ..................................................................................................... 3

2.

SCOPE AND METHODOLOGY ............................................................................ 8

3.

PUMP SUMP AND TEST SCENARIOS ................................................................ 9

4.

FLOW PATTERNS ................................................................................................. 11

5.

RESULTS AND DISCUSSIONS ............................................................................ 68

6.

CONCLUSION ........................................................................................................ 70

1.

INTRODUCTION
3

There is a requirement to carry out hydraulic sump model study for the membrane
feed pumping station at Lower Seletar Waterworks (LSWW).

The membrane feed pumping station comprises of two equal tanks to allow one tank
to be taken out for cleaning without interrupting the flow to the membrane plant. The
inflow from the clarifiers is discharged by two DN 1600 mm pipes to the two tanks.
There will be seven membrane feed pumps to be installed in the two tanks, including two
small pumps of capacity 25 Ml/d and five big pumps of capacity 75 Ml/d. Three big
pumps and one small pump are to be installed in Tank 1. Two big pumps and one small
pump are to be installed in Tank 2. These pumps are of the double suction split casing
type with vertical spindle. The layout and dimensions of the tanks is shown in Figure 1.

Tank 1

Tank 2

Figure 1 Design drawings of the membrane feed pumping station at LSWW: Plan view
(left) and cross-sectional view (right)

The objective of this study is to determine the necessity of any baffle walls or any
other undesirable flow mitigation structures as a result of cross-flow from abstracting
water through the two interconnection pipes. The physical model of the existing and new
4

pump sumps are constructed at the model-to-prototype scale of 1:15. Froudian modelling
criteria are used for the flow conditions and dimensions of model. The physical model
system is setup in the Maritime Research Centre, Nanyang Technological University, as
shown in Figure 2.
Table 1. Prototype and Model Parameters
Item
Sump
Dimensions

Pipe
Diameter

Flow rate

Velocity

Prototype

Scale

Model

Sump width

4m

1 : 15

0.266 m

Sump length

36.07 m

1 : 15

2.405 m

Sump depth

4.5 m

1 : 15

0.3 m

Diameter Influent pipe

DN 1600
(ID 1529 mm)

1 : 15

100 mm

Diameter of suction pipe for

big pump

DN 900
(ID 899 mm)

1 : 15

60 mm

Diameter of suction pipe for

small pump

DN 500
(ID 495mm)

1 : 15

40 mm

Maximum flow rate of big

pump

0.868 m3/s
(75Ml/d)

1 : 871.42

0.996 l/s

Maximum flow rate of small

pump

0.289 m3/s
(25Ml/d)

1 : 871.42

0.332 l/s

Maximum inflow rate in Tank

1

2.893 m3/s
(250 Ml/d)

1 : 871.42

3.32 l/s

Maximum inflow rate in Tank

2

2.025 m3/s
(175 Ml/d)

1 : 871.42

2.323 l/s

Maximum suction velocity for

big pump
Maximum suction velocity for
small pump
Maximum influent velocity
in Tank 1
Maximum influent velocity
in Tank 2

1.368 m/s

1 : 3.873

0.353 m/s

1.503 m/s

1 : 3.873

0.388 m/s

1.576 m/s

1 : 3.873

0.407 m/s

1.103 m/s

1 : 3.873

0.285 m/s

Figure 2(a) Overview of the model setup

Figure 2(b) Side view of the exiting sump model

6

This report presents the findings of the hydraulic performance of the sump with the
objectives to determine: (a) whether surface vortex occurs for different pumping
conditions, (b) whether there is swirl formation around pump for different pumping
conditions, and (c) the necessity of any baffle walls or any other measures to mitigate
undesirable flow patterns in the sump.

2.

SCOPE AND METHODOLOGY

This study include the following scope of work
1) Construction of a physical model of the pumping station including connection
pipes in a scale of 1:15;
2) The pumping system is modelled to reflect the equivalent flow system, and
the flow field in the sump.
3) Selected scenarios of pump operations are modeled to observe the flow
patterns and whether there are swirls, surface or submerged vortices and any
undesirable flow patterns with the sump.
4) Visualisation of the flow patterns is carried out with the aids of lighting and
dye. Relevant flow patterns are recorded with cameras; and
5) Identify and recommend possible modification or solution to avert any
undesirable flow conditions, if any.

Flow visualization is achieved by means of color dye injected through a small tube
and placed at different locations in the flow field. This technique is used extensively to
locate the possible origin of submerged vortices and to identify objectionable pumpapproach flow patterns. Flow patterns are photographed and videotaped, and selected
color photos are included herein.

The model flow rate is controlled by valve and the outflow rate of each suction pipe
is measured using volume-time method. Water surface elevations are determined and
monitored through the level marks to the tank sidewalls.

3.

PUMP SUMP AND TEST SCENARIOS

The pertinent dimensions of the sumps according to the data supplied by the client

are shown in Figure 1. Four pumps will be installed in Tank 1 and three pumps will be
installed in Tank 2. The schematic layout of the sumps and pumps and their names are
shown in Figure 3, where A1 A5 are big pumps and B1 B2 are small pumps.

Figure 3 Schematic layout of the sumps and pumps

The design capacity of the pumping station is 300 Ml/d, which can be achieved by
different pump combinations. The maximum pumping rate of Tank 1 is 250 Ml/d with 3
big pumps and 1 small pump, and the maximum pumping rate of Tank 2 is 175 Ml/d with
2 big pumps and 1 small pump. Although the critical operation conditions are generally
the case of the maximum pumping rate, other combinations of duty pumps are also
selected for test to observe the flow conditions.

The design lowest and highest water levels in the tanks are 118.5 mRL and 121
mRL, respectively. Normally the lowest water level is the worst condition. Therefore
most test runs are based on the lowest water level but the highest water level is also tested
for two selected cases. Besides, two test runs are conducted based on recommended
lowest water level. The test matrix is listed in Table 2.1 and 2.2.

Table 2.1 Test run matrix for the pump sump model
9

No.

Pumps on duty

Influent Flow Rate

(Ml/d)

Tank Water level

(m)

A1

25

118.5 (LWL)

A3

75

118.5

A2, A3

150

118.5

A1, A2, A3,

225

118.5

B1, A1, A2, A3

250

118.5

B1, A1, A2, A3

250

121.0 (HWL)

B2

25

118.5 (LWL)

B2, A4

75

118.5

B2, A5

75

118.5

10

A4, A5

150

118.5

11

B2, A4, A5

175

118.5

12

B2, A4, A5

175

121.0 (HWL)

Tank

Tank 1

Tank 2

13

Tank 1

B1, A1, A2, A3

250

120.4 (R LWL)

14

Tank 1

B2, A4, A5

175

120.4 (R LWL)

10

4.

FLOW PATTERNS
This section shows the observed flow patterns for each case. Dye visualization

technique is used to enhance visualisation of the flow patterns in the sumps and around
the intakes of the suction pipes.
Figures 4 to 17 show the snapshots of the flow patterns for Cases 1 to 14 with selfexplanatory captions. More photographs and video clips are also provided with the
attached CD.

Figure 4(a) Overview of flow pattern for Case 1

11

Figure 4(b) Flow pattern of influent flow for Case 1

12

Figure 4(c) Flow pattern around A1 intake for Case 1

Figure 4(d) Dye visualization of influent flow for Case 1

13

Figure 4(e) Dye visualization around A1intake for Case 1

14

Figure 5(a) Overview of flow pattern for Case 2

Figure 5(b) Flow pattern of influent flow for Case 2

15

Figure 5(c) Flow pattern around A3 intake for Case 2

16

Figure 5(d) Dye visualization of influent flow for Case 2

17

Figure 5(e) Dye visualization around A3 intake for Case 2

18

Figure 6(a) Overview of flow pattern of Case 3

Figure 6(b) Flow pattern of influent flow for Case 3

19

Figure 6(c) Flow pattern around A2 and A3 intakes for Case 3

20

Figure 6(d) Dye visualization around A2 intake for Case 3

21

Figure 6(e) Dye visualization around A3 intake for Case 3

22

Figure 7(a) Overview of flow pattern for Case 4

Figure 7(b) Flow pattern of influent flow for Case 4

23

Figure 7(c) Flow pattern around A1, A2 and A3 intakes for Case 4

24

Figure 7(d) Dye visualization of influent flow for Case 4

Figure 7(e) Dye visualization around A1 intake for Case 4

25

Figure 7(f) Dye visualization around A2 intake for Case 4

26

Figure 7(g) Dye visualization around A3 intake for Case 4

27

Figure 8(a) Overview of flow pattern of Case 5

Figure 8(b) Flow pattern of influent flow for Case 5

28

Figure 8(c) Flow pattern around A1, A2, A3 and B1 intakes for Case 5
29

Figure 8(d) Dye visualization around A1 intake for Case 5

Figure 8(e) Dye visualization around A2 intake for Case 5

30

Figure 8(f) Dye visualization around A3 intake for Case 5

Figure 8(g) Dye visualization around B1 intake for Case 5

31

Figure 9(a) Overview of flow pattern for Case 6

Figure 9(b) Flow pattern of influent flow for Case 6

32

Figure 9(c) Flow pattern around A1, A2, A3 and B1 intakes for Case 6

33

Figure 9(d) Dye visualization around A1 intake for Case 6

Figure 9(e) Dye visualization around A2intake for Case 6

34

Figure 9(f) Dye visualization around A3intake for Case 6

Figure 9(g) Dye visualization around B1intake for Case 6

Figure 10(a) Overview of flow pattern for Case 7

35

Figure 10(b) Flow pattern of influent flow for Case 7

Figure 10(c) Flow pattern of B2 intake for Case 7

36

Figure 10(d) Dye visualization of influent flow for Case 7

Figure 10(e) Dye visualization around B1 intake for Case 7

37

Figure 11(a) Overview of flow pattern for Case 8

Figure 11(b) Flow pattern of influent flow for Case 8

38

Figure 11(c) Flow pattern around A4 and B2 intakes for Case 8

Figure 11(d) Dye visualization of influent flow for Case 8

39

Figure 11(e) Dye visualization around A4 intake for Case 8

40

Figure 11(f) Dye visualization around B2 intake for Case 8

Figure 12(a) Overview of flow pattern of Case 9

Figure 12(b) Flow pattern of influent flow for Case 9

41

Figure 12(c) Flow pattern around A5 and B2 intakes for Case 9

Figure 12(d) Dye visualization of influent flow for Case 9

42

Figure 12(e) Dye visualization around A5 intake for Case 9

43

Figure 12(f) Dye visualization around B2 intake for Case 9

Figure 13(a) Overview of flow pattern of Case 10

44

Figure 13(b) Flow pattern of influent flow for Case 10

Figure 13(c) Flow pattern around A4 and A5 intakes for Case 10

45

Figure 13(d) Dye visualization of influent flow for Case 10

Figure 13(e) Dye visualization around A4 intake for Case 10

46

Figure 13(f) Dye visualization around A5 intake for Case 10

47

Figure 14(a) Overview of flow pattern for Case 11

Figure 14(b) Flow pattern of influent flow for Case 11

48

Figure 14(c) Flow pattern around A4, A5 and B2 intakes for Case 11
49

Figure 14(d) Dye visualization around A4 intake for Case 11

50

Figure 14(e) Dye visualization around A5 intake for Case 11

51

Figure 14(f) Dye visualization around B2 intake for Case 11

52

Figure 15(a) Overview of flow pattern for Case 12

Figure 15(b) Flow pattern of influent flow for Case 12

53

Figure 15(c) Flow pattern around A4, A5 and B5 intakes for Case 12

54

Figure 15(d) Dye visualization around A4 intake for Case 12

55

Figure 15(e) Dye visualization around A5 intake for Case 12

Figure 15(f) Dye visualization around B2 intake for Case 12

56

Figure 16(a) Overview of flow pattern for Case 13

Figure 16(b) Flow pattern of influent flow for Case 13

57

Figure 16(c) Flow pattern around A1, A2, A3 and B1 intakes for Case 13
58

Figure 16(d) Dye visualization around A1 intake for Case13

59

Figure 16(e) Dye visualization around A2 intake for Case13

60

Figure 16(f) Dye visualization around A3 intake for Case13

61

Figure 16(g) Dye visualization around B1 intake for Case13

Figure 17(a) Overview of flow pattern for Case 14

62

Figure 17(b) Flow pattern of influent flow for Case 14

63

Figure 17(c) Flow pattern around A4, A5 and B2 intakes for Case 14
64

Figure 17(d) Dye visualization of flow around A4 intake for Case 14

65

Figure 17(e) Dye visualization of flow around A5 intake for Case 14

66

Figure 17(f) Dye visualization of flow around B2 intake for Case 14

67

5.

RESULTS AND DISCUSSIONS

The previous section shows the flow patterns in the pump sumps for various

scenarios of different flow conditions and pump duties. The major observations of flow
patterns pertinent to the study objectives include the followings:
The hydraulic performance of the pump sump is generally satisfactory at top water
level but not acceptable at low water level.
The influent flow is turbulent and fluctuant at low water level. If two or more
pumps on duty in one tank, the strong inflow cause splashing and bubbles. The
bubbles can be sucked into nearby pump intakes, such as A1, A2, and A5, which
can lead to severe pump problems.
No obvious bubbles are observed at top water level because the inflow is
submerged.
Due to the asymmetrical inflow into the narrow tank, the approaching flows to
different pumps are not parallel and straight. Such asymmetrical inflow can cause
circulations in the tank and pre-rotation around pipe intakes, especially at low
water level.
At low water level, surface vortex can be generated and air can be sucked into
pumps, for example A3 in Tank 1 and A4 in Take 2.
At top water level, there is no surface vortex observed because the intake
submergence in increased.

Based on the above observations and findings, we arrived at the following

conclusions:
The submergence of pipe intake at low water level is insufficient to avoid
formation of surface vortices;
The submergence of pipe intake at top water level is sufficient to avoid formation
of surface vortices;
The pipe intake clearances from the floor and back wall are adequate to avoid
formation of vortices for all the pumps;
68

 The submergence of influent flow at low water level is insufficient to avoid

formation of strong surface perturbations and air bubbles;
The submergence of influent flow at top water level is sufficient to avoid
formation of air bubbles.

Based on the conclusions and the considerations of sump space constraints, we

recommend:
Do not operate the pumps at design low water level (118.5 mRL);
The minimum level should be high enough to have a sufficient submergence of
inflow pipes as well as pipe intakes. The recommended LWL is 120.4 mRL;
Higher water level is always more favourable for hydraulic performance in the
sump.

69

6.

CONCLUSION
It is concluded that the hydraulic performance of the pump sump is generally

satisfactory at top water level but not acceptable at low water level. Air bubbles and
surface vortex are observed for the cases at low water level. It is not recommended to
operate at low water level and the minimum operation water level should be at least 20.4
mRL.

70