Performance comparison of centrifugal

pump in variation of impeller blade number

through numerical simulation
Cite as: AIP Conference Proceedings 2403, 040013 (2021); https://doi.org/10.1063/5.0074344
Published Online: 02 December 2021

Joko Waluyo and Rahmat Waluyo

AIP Conference Proceedings 2403, 040013 (2021); https://doi.org/10.1063/5.0074344 2403, 040013

© 2021 Author(s).
Performance Comparison of Centrifugal Pump in Variation
of Impeller Blade Number through Numerical Simulation
Joko Waluyo1, a) and Rahmat Waluyo2,b)
1
Department of Mechanical and Industrial Engineering, Faculty of Engineering, Universitas Gadjah
Mada,Jl. Grafika No. 2, Yogyakarta,QGRQHVLD
2
Alumnae of Mechanical Engineering, Department of Mechanical and Industrial Engineering, Faculty of
Engineering, Universitas Gadjah Mada, Jl. Grafika No. 2, Yogyakarta,QGRQHVLD
a)
Corresponding author: [email protected]
b)
[email protected]

Abstract. Centrifugal pump is indispensable equipment for various industrial applications which mainly used to move
fluid by supplying head at specified flow rate. Pump performance is influenced by occurrence of losses which consist of
volumetric, hydraulic and mechanical losses which further quantified by corresponding efficiency. In its application, real
operating pump performance is quite difficult to predict without referring to manufacture data report which measure pump
performance directly during operation of the pump This research is aimed to investigate numerically the pump performance
variation with respect to the number of blades. A specified pump has operating condition of head 12.08 m, flowrate of
0.0015 m3/s, rating rotation of 2800 rpm with 7 blades is investigated. The performance analysis is performed through
numerical simulation using ANSYS. Performance comparison is carried in varied blade number of 5,7 and 9. The
performance of centrifugal pump is determined in terms of head (H) and flow rate (Q), as well as pump hydrodynamic
efficiency (h). It is found that the higher head occured in the case of 9 blades up to a certain limit. On the other hand,
hydrodynamic efficiency in the case of 7 blade is higher than the other two cases for simulated flow rate range. From the
head and flow rate relationship, it is noted that the pattern produced by plotting head and efficiency against flow rate
indicate that there is an optimum number of impeller blade for centrifugal pump.

Keyword: centrifugal pump, performance, blade number, numerical study.

INTRODUCTION

Centrifugal pumps have been used in various industrial applications, including industrial petrochemical, oil and
gas plant, and domestic application. Working principle of centrifugal pump consist of increasing fluid dynamic head
at rotating impeller and convert it to static head at volute passage. Head produced by centrifugal pump is influenced
by occurrence of losses which could be quantified in term of volumetric, hydraulic, and mechanical efficiency.
Furthermore, pump performance is also affected by geometric configuration of the pump such as dimension of the
blade, impeller, and volute. Currently, pump performance is conventionally measured during pump operation thus
make it impossible to predict the real performance exactly during the design and manufacture stage. However, the
advance of commercial CFD code opens possibility to predict pump performance numerically which lead to numerous
investigations employing numerical simulation as a complementary works to experimental study.
As an attempt to improve pump performance various works had been conducted to see the effects of geometric
variation of pump components of its hydraulic efficiency. An experimental study was carried out to analyze the
fluctuating pressure field in volute by measuring the pressure distribution along impeller periphery which shows that
flow rate variation was the main factor that affect pressure fluctuation in the volute, primarily in volute tongue region
[1]. The effects of blade curvature were investigated experimentally in the pump at specified capacity, with various
arrangements ranging from conventional smooth blade to staggered configuration [2]. Experimental and numerical

THERMOFLUID XI
AIP Conf. Proc. 2403, 040013-1–040013-8; https://doi.org/10.1063/5.0074344
Published by AIP Publishing. 978-0-7354-4167-5/$30.00

040013-1
study on the effect of different impeller dimensions with same blade number were carried out using two impellers
with different diameter size. Results show that head produced by bigger impeller is higher than the smaller one [3].
Blade wrap angle effects to pump performance were studied both in experiment and simulation both on- and off-
design flow rate [4]. Another study was addressed to investigate the effect of volute design variation on pump
performance and generated radial force. Volute geometry that was being studied consist of concentric volute and
multi-volute variants. Results indicated that modification at volute passage could reduce radial thrust force while at
the same time reducing efficiency [5,6]. Pump performance is also influenced by post-manufacture alteration such as
impeller trimming which has been studied experimentally. Study result show that generated head decrease as impeller
diameter become smaller due to trimming [7]. The effect of blade geometry modification on pump performance were
studied numerically by modifying blade angle and blade tip shape and blade outlet angle [8, 9,10]. A numerical study
was carried to investigate the effect of varied blade number on the pump performance [11]. However, the pump used
in this previous study had rectangular discharge flange which is uncommon in industrial application. Present study
aimed to investigate the influence of blade number on pump performance, using pump with circular discharge flange
so the result could be implemented on common application.

METHODOLOGY
The methodology used in this paper is presented as a flowchart in FIG. 1. Mesh quality is checked by considering
both its orthogonal quality and skewness. The convergence is evaluated using residuals value. To eliminate numerical
error due to spatial discretization, grid independency test is carried out by simulating the mesh through increasing
refinement level. Simulation results were validated with experimental measurement from existing literature.

START

Grid independency test

Pump dimension
and Validation
and operating
condition

Simulation results is valid

3D modelling of pump
fluid domain
Variation

Mesh setup z1 = 5 z2 = 7 z3 = 9

Simulation on various
Orthogonal quality and
skewness is adequate mass flow rate

Plotting H vs Q curve
and h vs Q curve
Simulation

Results and discussion

Simulation results is
converged

END

FIGURE 1. Research Steps.

After being validated, the 3D model is modified to accommodate the variation of blade number at 5, 7, and 9. A
whole performance characteristics of the pump are obtained through simulation on flow rate variation. The
performance is evaluated in term of flow rate and head curve, as well as its hydraulic efficiency. A pump with simple
geometry is used in the present study to investigate the effect of changing blade number to its performance. The volute
has a circular shape with constant diameter of 22 mm, with spiral passage from tongue to discharge flange. The

040013-2
geometrical configuration of the centrifugal pump is shown in FIG 1, whereas impeller and volute dimensions are
tabulated in TABLE 1.

TABLE 1. Dimensions of impeller and volute

Dimension Value
Number of Blade (z) 5, 7, 9
Impeller inlet diameter (D1) 45 mm
Impeller outlet diameter (D2) 105 mm
Impeller inlet width (b1) 6 mm
Impeller outlet width (b2) 6 mm
Impeller inlet blade angle (b1) 220
Impeller outlet blade angle (b2) 260
Suction side diameter (DSU) 30 mm
Discharge side diameter (DSCH) 22 mm
FIGURE 2. View of fluid domain of studied pump

The performance parameters, namely head and hydraulic efficiency are defined as the following. Head is calculated
from difference of total pressure supplied by the pump.
P2total ΫP1total
H α (1)
ߩǤg
Hydraulic efficiency is then calculated by dividing hydraulic energy flow rate to energy flow by shaft work which
computed by multiplying torsion and angular velocity.
ߩǤ ݃Ǥ ܳǤ ‫ܪ‬
ߟ௛௬ௗ ൌ  ሺʹሻ
ʹߨ ൈ ݊
ܶǤ
͸Ͳ

RESULTS AND DISCUSSION

1. Numerical Simulation Model
Simulations of centrifugal pump were carried out by using commercial CFD software ANSYS Fluent 19.1. The
simulation uses k-e RNG turbulence modelling and SIMPLE algorithm are applied to solve Reynolds Averaged
Navier-Stokes Equations. Multi reference frame were employed to simulate different frame motion between rotating
impeller and static volute. Residuals plot through iterations is given in FIG 3 which show velocity residual level after
1000 iterations is below 5 x 10-3. The details of simulations set up are given in TABLE 2.

FIGURE 3. Scaled residuals in simulations

040013-3
 TABLE 2. Numerical simulation set up
Parameter Value or Description
Commercial code software ANSYS Fluent 19.1
Fluid Water
Mesh Mixture of structured and unstructured
Inlet Boundary condition Mass flow rate (kg/s)
Outlet Boundary Condition Exit pressure (Pa)
Wall Boundary Condition No slip
Turbulence model k-e RNG
Cell domain 3, suction zone (stationary), impeller (rotating), volute (stationary)
Reference frame Multi reference frame
Turbulence intensityty, Tu 5%
Solver algorithm SIMPLE

Meshing generation is performed using 3 domains, namely suction zone, impeller, and volute. The suction zone
and volute are stationary while impeller is rotating with rotational speed of 2800 rpm. The mesh orthogonal quality
and skewness were inspected to minimize error due to bad mesh quality as shown in TABLE 3. A mesh dependency
test was carried out to attain suitable level of refinement which accurate solution is obtained.
TABLE 3. Mesh Quality
Parameter z=5 z=7 z=9
Skewness 0,21244 0,2138 0,21413
Orthogonal Quality 0,78972 0,78907 0,78802
Mesh type Mixed Mixed Mixed
Number of cells 457.589 455.812 454.972

To ensure that numerical solution is independent of element size, different levels of mesh refinement are generated:
coarse mesh, intermediate, and fine mesh. The solution started with coarse mesh and then refined gradually to finer
mesh until unchanged result obtained. FIG 4 shows simulation results for various grid refinement at mass flow rate
1.0345 kg/s. The result indicates that a mesh refinement from grid 2 (455,999 cells) to grid 3 (777,347 cells) produce
small difference compared to grid 1. Accordingly, grid 2 is used in present work considering accuracy and available
computing resources. Sample of generated mesh for 7 blades pump is shown in FIG 5.

11.65
11.6
11.55

11.5
Head (m)

11.45

11.4
11.35

11.3

11.25

11.2

Grid 1: 288245 Grid 2: 455999 Grid 3: 777347

FIGURE 5. Mesh sample

FIGURE 4. Grid independency

040013-4
 The given boundary conditions are mass flow rate and outlet pressure in inlet and outlet respectively. The walls
are stationary for suction and volute zone while rotating in impeller zone. Surface of blade tip and interface between
stationary and rotating cell zone are set apart by distance of 2.5 mm to enable transformation from relative system to
absolute one. Value of boundary conditions are listed in TABLE 4. The computation is carried out by setting
turbulence intensity (Tu) at 5% and hydraulic turbulence length (L) at 0.005 m on velocity-inlet face to specify
turbulence value on that boundary condition.

2. Model Validation
Simulation results need to be validated with experimental data to lower uncertainties between predicted flow and
real flow in centrifugal pump. In this paper, validation was conducted by comparing numerical result with
experimental data from suitable literature. FIG 6. shows bar chart diagram of validation results taken from pump with
7 blades. It can be seen from FIG. 6 that head coefficient calculated from simulation closely approaches experimental
value. The error between numerical results and experiment is below 10% and therefore considered valid
TABLE 4. Operating conditions
Parameter Value
Mass flow (kg/s) 0.00001 0.49 0.99 1.49 1.99 2.49 2.99
Exit pressure (Pa) 176,325 161,325 141,325 11,325 111,325 111,325 111,325

1.4
1.2
Y=gH/((u2)2/2)

1
0.8
0.6 Simulation
0.4
0.2 Experiment (Stickland et. al.)
0
0.00001 0.3846 0.667 1.0345
mass flow rate (kg/s)

FIGURE 6. Validation results

Numerical analysis is carried out to characterize the pump performance using different number of blades 5, 7, and
9. Changing the number of blades alter the aspect ratio of blade passage which causes different sources of losses.
Lower blade number cause losses due to turbulence and lack of blade guidance along impeller channel. On other hand
using more blade reduce losses caused by turbulence but increase the losses due to friction since flow cross-section
along blade passage is reduced. These two contradictions are tested numerically to determine the optimum number of
blades, which give adequate guidance to fluid flow while minimizing the friction losses along the passage. These are
investigated in term of pressure distribution as well velocity contour. Further information could be inferred by
inspecting generated total pressure contour from numerical solution as shown in FIG. 7. The contour is taken midplane
of the pump with mass flow rate of 1.49 kg/s.

(a) (b) (c)

FIGURE 7. Pressure contour for (a) 5 blades, (b) 7 blades, (c) 9 blades

040013-5
 Fewer number of blades produce nonuniform pressure variation in volute whereas the increase of blade number
reduces the effect. The pressure concentration is caused by circulation flow due to lack of guidance found in impeller
with 5 blades. On the other hand, it is noted pressure gradient in impeller of 9 blades case is higher than 7 blades
which contribute to mixing losses at impeller outlet.
Circulation flow is more apparent visualized in fluid velocity vector, presented in FIG. 8. Pump with a fewer blade
number produce more vortices in blade passage as shown in 5 blades case. The whole discussion can explain the order
of loss occurrence in the three cases as follows: case of 7 blades perform best followed by case of 9 blades and finally
the 5 blades.

(a) (b)

(c)
FIGURE 8. Simulated velocity field results for pump with (a) 5 blades, (b) 7 blades, (c) 9 blades

Analysis of pump performance is carried out by considering the relationship between flow rate, head, and hydraulic
efficiency. The flow rate versus head curves which is presented in FIG. 9 show the influence of blade number on
produced head. By comparing of the head from FIG. 9, it is shown that impeller with 9 blades produce higher head
than the other two up until certain flow rate. Increasing flow rate beyond that value will further decrease pump head
even lower than pump with 7 blades. In case of 5 blades, lack of guidance by impeller blades results in development
of turbulence and circulation flow inside blade passage which cause decrease in both head and efficiency. FIG. 9 also
show comparison on hydraulic efficiency between pump with different blade numbers. It is shown that impeller with
7 blades perform with higher efficiency than the alternatives along the simulated flow rate. The highest efficiency also

040013-6
 attained by centrifugal pump with 7 blades operating at design flow rate. It is also indicated that at higher flow rate,
pump with 9 blades will experience drop of efficiency due to increasing friction losses at higher fluid velocity along
its blade passage.

5 Blade 7 Blade 9 Blade 5 Blade 7 Blade 9 Blade

0.5
14.000 0.45

Efficiency ( )
Head, H (m)

12.000 0.4
10.000 0.35
0.3
8.000
0.25
6.000 0.2
0.00 0.50 1.01 1.51 2.02 2.52 3.02 0.00 0.50 1.01 1.51 2.02 2.52 3.02
3
Flow rate, Q ( x 10 m /s)3 Flowrate, Q ( x 103 m3/s)

(a) (b)
FIGURE 9. Pump characteristic curve (a) H-Q curve (b) h-Q curve

CONCLUSION
A numerical study using a CFD code was carried out to investigate the effect of varying blade number in centrifugal
pump on pump performance. Numerical results were validated by comparing it with experimental data from previous
work. The studied cases show the effect of changing blade number on flow mechanism primarily on formation of
vortex, and friction losses. It is concluded that fewer blade number decrease blade’s guidance on fluid and
consequently increase circulation flow inside blade passage which will produce more vortices and reduce the hydraulic
efficiency of the pump. On the other hand, higher number of blades will increase the friction losses due to decreasing
crossflow area. This in turn, will reduce pump performance especially at higher flow rate. It is noted that optimum
number of blades that could provide adequate guidance while minimizing friction losses. In this case impeller with 7
blades perform better than the other two, which followed by 9 and 5 blades.

ACKNOWLEDGEMENT
The authors would like to express gratitude thanks to Penelitian Hibah DTMI for the financial supports.

REFERENCES
1. Stickland, M. T., and Scanlon, T. J. An Experimental Study of Unsteady Pressure distribution around the
Impeller Outlet of Centrifugal Pump. ASME 2000 Fluids Engineering Division Summer Meeting. Boston:
ASME. (2000)
2. Sadek, A., Saleh, I., Nassif, R., and Mouris, E. Experimental Study of the Effect of Blade Configurations on the
Performance of Centrifugal Pumps. Cairo: Aerospace Sciences & Aviation Technology. (2013)
3. Meakhail, T. A., Salem, M., and Shafie, I. Steady and unsteady flow inside a centrifugal pump for two different
impellers. International Journal of Energy and Power Engineering, (2014), pp. 65-76.
4. Tan, L., Zhu, B., Cao, S., Bing, H., and Wang, Y. Influence of Blade Wrap Angle on Centrifugal Pump
Performance by Numerical and Experimental Study. Chinese Journal of Mechanical Engineering, 27(1), (2013)
pp. 171-177.
5. Alemi, H., Nourbakhsh, S. A., Raisee, M., and Najafi, F. Development of new "multivolute casing" geometries
for radial force reduction in centrifugal pumps. Engineering Applications of Computational Fluid Mechanics,
(2015) pp. 1-11.
6. Kelder, J., Dijkers, R., Esch, B. V., and Knyt, N. Experimental and theoretical study of the flow in the volute of
a low specific-speed pump. Journal of Fluid Dynamic Research, (2001), pp. 267-280.
7. Savar, M., Kozmar, H., and Sutlovic, I. (2009). Improving centrifugal pump efficiency by impeller trimming.
Desalination, 654-659.

040013-7
8. Yousefi, H., Noorollahi, Y., Tahani, M., and Fahimi, R. Numerical simulation for obtaining optimal impeller's
blade parameters of a centrifugal pump for high-viscosity fluid pumping. Sustainable Energy Technologies and
Assessments, (2019), pp. 16-26.
9. Ding, H., Li, Z., Gong, X., and Li, M. The influence of blade outlet angle on the performance of centrifugal
pump Vacuum, Vol I (159), (2019), pp. 239-246.
10. Bacharoudis, E., Filios, A., Mentzos, M., and Margaris, D. Parametric Study of a Centrifugal Pump Impeller
by Varying the Outlet Blade Angle. The Open echanical Engineering Journal, 2(1), (2008), pp. 75-83.
11. Elyamin, G. R., Bassily, M. A., and Khalil, K. Y. Effect of impeller blades number on the performance of
a centrifugal pump. Alexandria Engineering Journal, 58(1), (2019), pp. 39-48.

040013-8