Hindawi Publishing Corporation

Advances in Mechanical Engineering

Volume 2014, Article ID 803972, 10 pages
http://dx.doi.org/10.1155/2014/803972

Research Article
3D Blade Hydraulic Design Method of the Rotodynamic
Multiphase Pump Impeller and Performance Research
Yongxue Zhang, Jinya Zhang, Hongwu Zhu, and Shujie Cai
College of Mechanical and Transportation Engineering, China University of Petroleum, Beijing 102249, China
Correspondence should be addressed to Jinya Zhang; zhjinya@cup.edu.cn
Received 23 August 2013; Revised 31 December 2013; Accepted 1 January 2014; Published 26 February 2014
Academic Editor: Junwu Wang
Copyright 2014 Yongxue Zhang et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
A hydraulic design method of three-dimensional blade was presented to design the blades of the rotodynamic multiphase pump.
Numerical simulations and bench test were conducted to investigate the performance of the example impeller designed by the
presented method. The results obtained from the bench test were in good agreement with the simulation results, which indicated
the reasonability of the simulation. The distributions of pressure and gas volume fraction were analyzed and the results showed
that the designed impeller was good for the transportation of mixture composed of gas and liquid. In addition, the advantage of
the impeller designed by the presented method was suitable for using in large volume rate conditions, which were reflected by the
comparison of the head performance between this three-dimensional design method and another one.

1. Introduction
With the exploitation of the offshore oil and remote oilfields,
gas-liquid multiphase technology has attracted attention of
the worlds major oil companies for its significant economic
benefits. In recent years, some technologies have emerged, for
example, gas-oil mixture pumping [1, 2]. So far, several kinds
of positive displacement and rotodynamic multiphase pump
have emerged for gas-liquid multiphase flow in the petroleum
industrial processes.
The rotodynamic multiphase pump which has larger flow
and higher sand toleration is predicted as the ideal choice of
the oil exploitation in the subsea [3]. Large numbers of rotodynamic multiphase pumps have been installed all over the
world since the first pump was tested successfully in the oil
field in the early 1990s [4, 5]. However, the state of fluid flow
in the impeller of a multiphase pump is complicated and the
two-phase flow regime is greatly influenced by the flow rate of
the mixture and inlet gas volume fraction (IGVF) as well as
the pump speed. Because of these reasons, new requests and
challenges have been put forward for designers.
With regard to the design method of the multiphase
pump impeller, the velocity coefficient method and the
similarity design theory were used [612]. The former
method depends heavily on the designers knowledge and

experience, and the latter needs impeller models with excellent performance. Given that most studies of this kind of
pump have so far suffered from a lack of large-scale empirical
data, the latter method is not popular. Based on this reason, a
new method is proposed to design the blades of the impeller
of a rotodynamic multiphase pump in this paper, in which,
the gas-liquid two-phase meridional velocity gradient equation and the two-phase coupled equations solved by a
quasiorthogonal method are deduced. The medium is the
gas-liquid multiphase flow, and the blade mean camber line
equation is integrated with the point-by-point integral
method, which ensures that the blade shape satisfies the gasliquid two-phase flow pattern as well as enhances the rationality of the flow passage of impeller. The blade in this paper
is calculated by 2030 stream surfaces; it ensures the smoothness and accuracy of the blade transition between impeller
shroud and hub.

2. Main Design Parameters of

the Impeller and Blade
A multiphase pump, mainly rotodynamic multiphase pump,
is a kind of multistage pump which includes the suction unit,
the compression cells and the exit unit [13], as is shown in

Downloaded from ade.sagepub.com by guest on March 31, 2015

Advances in Mechanical Engineering

Figure 1. A compression cell, which is also called one stage, is

composed of an impeller and a diffuser. Gas-liquid mixture
gains kinetic energy when it flows through the impeller.
When the mixture flows through the diffuser, its speed will
drop and part of the kinetic energy will be converted to
pressure energy. The structure of the impeller is shown in
Figure 2.
Some main parameters which will influence the geometric construction of this pump impeller are the diameters of
hub at inlet and outlet (1 and 2 ), the hub ratio at inlet
(1 / ), the cone angle of hub (2), the diameter of shroud
, the axial length of impeller (), blade inlet angle 1 , blade
outlet angle 2 , the number of blades, the wrap angle of the
blade, and so on.

3. Hydraulic Design of Impeller

In order to design the blade of the multiphase pump impeller,
the 3D blade design method is adopted. From the numerical
solution of the meridian flow net on the meridional plane and
blade mean camber lines on the stream-surfaces of revolution, the blade shape is obtained. The new hydraulic design
method of multiphase pump impeller is shown in Figure 3.

Compression cell

Suction unit Impeller Diffuser

Exit unit

Figure 1: Flow passage of a rotodynamic multiphase pump.

and . Figure 7 shows the relationships between the quasiorthogonal line, the meridian streamline, and the line of
wetted perimeter.
3.3. Coupling Equations of Gas-Liquid Two-Phase. The GVF
and the density of mixture are different everywhere because
the pressure in the impeller varies and the gas is compressible,
so the GVF and density of mixture mix can be expressed as
GVF =

g,
mix,

g,
g, + l

(2)

mix = GVF gas + (1 GVF) liquid ,

3.1. The Axial Profile of the Impeller Blade. The axial profile of
the impeller blade has a great influence on the flow ability of
multiphase pump, so it is the key point in the process of the
reasonable design. The axial profile designed in this paper is
shown in Figure 4. One obvious characteristic is that the cross
section gradually reduces for the compressibility of the
medium.
3.2. The Meridional Velocity Gradient Equation. Figure 5
shows an infinitesimal hexahedron which is comprised of a
series of stream surfaces. And after a series of processing and
analysis, the meridional velocity gradient equation can be
obtained as

= () ,

gas

(1)

cos ,

where is meridional velocity; is blockage coefficient

which is the area ratio of the effective flow and the geometric
flow in the pump; 1 is the angle between the meridian
streamline and the axis of rotation; 2 is the angle between the
line of wetted perimeter and the axis of rotation. The details
are shown in Figure 6.
The relationship between angles 1 and 2 is 2 = 1
(/2). The line of wetted perimeter is unknown, so the quasiorthogonal line is introduced in order to avoid the derivation of . In addition, = + (/2) 1 , where is the
angle between and and stands for the angle between

(3)

where , , () stand for the static pressure in the impeller,

the polytropic exponent of gas, and the integral constant
depending on the number of the flow line which represents
the entropy, respectively. Consider
gas1 gas1 = gas2 gas2 ,

sin 1
1
= [( 1 1 sin )

cos

ln ln mix

] sin

where the subscripts g, l, mix, and stand for the gas, the liquid, the mixture, and the working pressure, respectively; gas
and liquid represent the density of gas and liquid, respectively.
The density of gas gas can be calculated as

(4)

where gas stands for the volume flow of gas and the subscripts 1 and 2 represent the inlet and outlet, respectively.
Equations (2) to (4) are gas-liquid two-phase coupling
equations, which play a major role in the computational process. They are also the main difference between the design
method of multiphase pump and that of the common rotodynamic pump.
3.4. Calculation of Blade Mean Camber Lines and the Shape
Surface of Blade. The parameters of meridian streamline can
be obtained by iterative calculations. And then, the flow field
is divided into 60 axial planes and the blade mean camber
lines are integrated with the point-by-point integral method.
Based on the principle of B-Spline surface, all of the blade
mean camber lines are combined, so the shape of the impeller
blade surface is formed. According to the relationship between meridian streamline and the blade mean camber line as
(5), the blade thickness along the meridian streamline can be

Downloaded from ade.sagepub.com by guest on March 31, 2015

Advances in Mechanical Engineering

dh2

dh1

Dt

Figure 2: Impeller of a rotodynamic multiphase pump.

Start

Selection of main parameters

Determining the quasiorthogonal

Mesh generation and

determining the initial meridian flow net
No

Solving the meridional velocity gradient

equation of gas-liquid two-phase

Satisfied?

Flow balance?

Yes
Drawing impeller blades

Blades thickened

End

Figure 3: Design procedures of impeller blades.

obtained through the Cubic Spline Method in the curve of

= ():

,
sin

(5)

where , , and stand for the length of the meridian

streamline, the real thickness of the blade, and the blade angle
in each calculated point, respectively.

4. Design Example
By using the method above, an example of the rotodynamic
multiphase pump impeller is designed according to the

parameters showed in Table 1. The 3D impeller model is

shown in Figure 8.

5. Numerical Simulation of
the Design Example
The characteristics of the designed multiphase pump are
investigated by numerical simulations with the commercial
software ANSYS-CFX 14.0 in this paper.
5.1. Calculation Model and Mesh. As shown in Figure 9, three
parts compose the calculation model. The first part is the front
extension which is once as long as the impellers diameter,

Downloaded from ade.sagepub.com by guest on March 31, 2015

Advances in Mechanical Engineering

Hub

z (axis of rotation)

Inlet

Shroud

d2
dl +

B
2

d2

l
d

d +
dl
l

d1

Outlet

Axis of rotation

dl

Figure 4: Axial profile of the impeller blade.

l
d1

Figure 6: Geometric meaning of 1 and 2 .

(rd)
d F

z (axis of rotation)

Cm
d
d O

d +

dl
l

dl +

l
d

u
C
rd

Cm +

(Cm )
dl
l

z (axis of rotation)

rd +
d

A
dl
1

dS

dl

l
A

Figure 7: Relationships between the quasiorthogonal line, the

meridian streamline, and the line of wetted perimeter.

Figure 5: Flow path differential unit.

Table 1: Design parameters of the design example.
Parameters
Rotation speed ()
Flow rate ()
IGVF
Medium
Differential pressure per
impeller ()

Value

Unit

2900
100
30
Water and ideal
compressible air

rpm
m3 /h
%

0.16

MPa

the second part is the impeller, and the third part is the back
extension which is three times the length of the impellers
diameter.
5.2. Mesh Independence. For the reduction of computation
time and the improvement of accuracy, the optimum number
of grid cells in the simulation has been investigated. In
Figure 10, it is observed how the differential pressure and
hydraulic efficiency reach an asymptotic value as the number
of cells increases. According to this figure, when the number
of cells is more than 867180, the change of differential pressure is less than 0.2 104 MPa and the hydraulic efficiency is

less than 0.15%. So, the Grid E (867180 cells) is considered to

be sufficiently reliable to ensure mesh independence.
The grid numbers of each part are 147670 on the front
extension, 867180 on the impeller, and 193,600 on the back
extension. Figures 11 and 12 show the mesh generated on
calculation field and impeller, respectively.
5.3. Simulation Solution and Boundary Conditions. In the
process of 3D flow simulation of the pump, the Reynolds
Averaged Navier-Stokes equations are adopted and the finite
volume method is used to solve the control equations. The
- SST turbulence model is adopted to deal with the water,
and the dispersed phase zero equation model is applied to the
ideal compressible gas. In addition, the chasing method of
finite differences of implicit scheme is applied. Within the
computational domain, the multiphase flow calculation utilizes the two-fluid model which supposed that the bubbly flow
occurred in the multiphase pump [14].
For all simulations, the boundary conditions are as
follows.
(1) Inlet Conditions. Before the mixed fluid flowed into the
multiphase pump, liquid and gas had been mixed in a buffer

Downloaded from ade.sagepub.com by guest on March 31, 2015

Advances in Mechanical Engineering

5
Outlet

Inlet
Front Impeller
extension

Back
extension

Figure 11: Calculation mesh.

Figure 8: 3D impeller model of the design example.

Inlet

Outlet

Front
extension Impeller

Back
extension

Figure 9: Calculation model.

48.99

0.181

48.30

0.180
B

0.178

47.61
46.92

46.23

0.177
B

0.176

0.174

0.179

0.175

Figure 12: Mesh of impeller flow field.

45.54
44.85

44.16

A
6

Hydraulic efficiency (%)

Differential pressure (MPa)

0.182

10

Number of cells 10

11

12

13

Differential pressure
Hydraulic efficiency

Figure 10: Influence of grid cells on impeller performance ( =

2900 rpm, = 100 m3 /h, and IGVF = 15%).

tank. So the distribution of IGVF is considered as homogeneous, and the inlet total mass flow and IGVF are set as the
inlet boundary condition.
(2) Outlet Conditions. According to the estimated differential
pressure, the outlet static pressure is set as the outlet condition.
(3) Interfaces Boundary Conditions. The data on interfaces
between inlet extension, outlet extension, and impeller deliver each other.
(4) Wall Boundary Conditions. The flow pattern is set to noslip condition and the wall function is applied for dealing with
two-phase fluid in the near-wall region.
5.4. Calculation Results and Discussions
5.4.1. Pressure Field Characteristics. When the IGVF is 30%,
the static pressure distribution on the blade to blade sur-

faces is shown in Figure 13 and the average static pressure

distributed along the impeller meridional surface is shown in
Figure 14.
From Figure 13, it can be seen that the static pressure
gradient is uniform on the blade to blade surfaces in different
spans of the impeller, which is good for the transportation of
mixture composed of gas and liquid except for the appearance
of pressure fluctuation on the surface of 0.9 span near the
outlet.
The results obtained from Figure 14 are as follows.
(1) Average static pressure increases uniformly from the
inlet to the outlet. The lowest pressure is about
0.02 MPa and the highest is about 0.24 MPa.
(2) There is a high pressure region near the shroud of outlet, which is probably due to the separation of mixture
composed of gas and liquid.
5.4.2. The Distribution Characteristics of GVF. Figure 15
shows the contour plots of the GVF distribution on blade to
blade surfaces. Figure 16 shows the average GVF distribution
along the impeller meridional surface.
The details obtained from Figure 15 are as follows.
(1) The GVF decreases uniformly from the inlet to the
outlet for the compressibility of gas.
(2) It is inevitable that the GVF is lower on the surface of
0.9 spans. This is due to the centrifugal force caused by
the rotating impeller. However, this phenomenon is
very faint in the design example which affirms the
design method presented in this paper.
As is shown in Figure 16, the higher GVF that appeared
near the hub is about 20% and the lower that appeared near
the shroud is about 5%. On the whole, the GVF decreases
gradually from the inlet to the outlet as the gas is compressed,

Downloaded from ade.sagepub.com by guest on March 31, 2015

Advances in Mechanical Engineering

2.400e + 005

2.200e + 005
2.000e + 005
1.800e + 005
1.400e + 005
1.200e + 005
1.000e + 005
8.000e + 004

Pressure (Pa)

1.600e + 005
Inlet

Outlet

6.000e + 004
4.000e + 004
2.000e + 004
0.000e + 000
2.000e + 004
(a) 0.1 span (b) 0.3 span (c) 0.5 span (d) 0.7 span (e) 0.9 span

Figure 13: Static pressure distributions on the blade to blade surfaces.

Shroud
Pressure (Pa)

2.400e + 005
2.200e + 005
2.000e + 005
1.800e + 005
1.600e + 005
1.400e + 005
1.200e + 005
1.000e + 005
8.000e + 004
6.000e + 004
4.000e + 004
2.000e + 004

Inlet

Outlet

Hub

Figure 14: Average static pressure distribution along the impeller meridional surface.

which reflects that the gas-liquid separation is not significant

which is good for the transportation of gas-liquid mixture.

6. Experiment of the Designed Example

6.1. Experiment Facility. According to the requirement of
experiment operated on the designed multiphase pump, a test
bench has been set up and its schematic diagram is shown as
in Figure 17, which consists of open-type gas circuit and
closed-cycle water-flow loop. Limited to the test conditions,
water and air are selected as the test medium. Power is provided by the variable-frequency high-speed motor and speed
is controlled by the frequency converter. First, air and water
flow into the buffer mixture and form uniform multiphase
flow that bubbles are uniformly distributed in it. Then, the
mixture flows into the multiphase pump and is pressurized.
Finally, it enters the open tank through the outer tubes. Water
will be recycled and air will be discharged directly into the
atmosphere.
6.2. Experiment Results
6.2.1. Performance Test. In order to test the differential pressure and hydraulic efficiency of the pump produced based

on the 3D design method, three rotational speeds are investigated. Figure 18 shows the differential pressure and efficiency
change with the total volume flow when the IGVF keeps 15%
and the speeds are 2600 rpm, 2900 rpm, and 3200 rpm. Some
conclusions can be obtained as follows.
(1) It is obvious that the differential pressure changes with
the increase of rotational speed.
(2) The maximum efficiency point of the pump and high
efficient area changes with the increase of rotational
speed. The best volume flow is improved with the
increase of rotational speed.

6.2.2. The Influence of Volume Flow on the Performance.

Figure 19 shows the performance change with the increase
of IGVF under different operational conditions, that is, =
2900 rpm, = 90 m3 /h; = 2900 rpm, = 100 m3 /h; =
2900 rpm, = 110 m3 /h. Some obtained details are as
follows.
(1) The differential pressure of the impeller declines gradually with the increase of IGVF, but there is no sudden
failure. At the same IGVF, the maximum differential

Downloaded from ade.sagepub.com by guest on March 31, 2015

Advances in Mechanical Engineering

9.500e 001
9.000e 001

7.000e 001
6.000e 001
5.000e 001
4.500e 001
4.000e 001
3.500e 001
3.000e 001
2.500e 001
2.000e 001
1.500e 001
1.000e 001
5.000e 002

Air volume fraction

8.000e 001

Inlet

Outlet

(a) 0.1 span (b) 0.3 span (c) 0.5 span (d) 0.7 span (e) 0.9 span

Figure 15: GVF distributions on blade to blade surfaces.

9.000e 001

8.000e 001
6.000e 001
5.000e 001
4.500e 001
4.000e 001
3.500e 001
3.000e 001
2.500e 001
2.000e 001
1.500e 001
1.000e 001
5.000e 002

Shroud

Air volume fraction

7.000e 001

Outlet

Inlet

Hub

Figure 16: Average GVF distribution along the impeller meridional surface.

Water
injection

Gas
access

Inline pump

Buffer tank

Transducer

Water
tank

Multiphase
pump

ADC
DAC
Date acquisition control system
Pressure transmitter
Globe valve
Check valve

Figure 17: Schematic diagram of the multiphase pump bench test.

Downloaded from ade.sagepub.com by guest on March 31, 2015

Advances in Mechanical Engineering

0.18

41.6

0.16

40.0
38.4

0.14

36.8

0.12

35.2

0.10

33.6
32.0

0.08
80

90

100

110

46

0.20

44
42

0.18

40

0.16

38

0.14

36

0.12

34
32

0.10

30.4
70

48

0.22

70

120

80

90

Differential pressure
CFD
Exp.

Efficiency
2600 rpm
2900 rpm
3200 rpm

Differential pressure (MPa)

0.20
0.15
0.10
0.05
0.00

20

Differential pressure
Q = 90 m3 /h
Q = 100 m3 /h
Q = 110 m3 /h

40
60
IGVF (%)

80

49.0
47.6
46.2
44.8
43.4
42.0
40.6
39.2
37.8
36.4
35.0
100

Hydraulic efficiency (%)

Figure 18: Differential pressure and hydraulic efficiency with

different (IGVF = 15%).

0.05

100

110

120

30

Q (m3 /h)

Q (m3 /h)

Differential pressure
2600 rpm
2900 rpm
3200 rpm

Hydraulic efficiency (%)

43.2

Differential pressure (MPa)

0.20

50

0.24

44.8
Hydraulic efficiency (%)

Differential pressure (MPa)

0.22

Efficiency
Q = 90 m3 /h
Q = 100 m3 /h
Q = 110 m3 /h

Figure 20: Comparisons between CFD results and experiment

results.

is 2900 rpm. From Figure 20, it can be seen that the simulation results are slightly overestimated as compared with
the experiment ones. The experiment results are in a good
agreement with the numerical simulation results, which
indicates that the performance of the multiphase pump can be
precisely evaluated by the experimental techniques and the
computational method implemented in this study.
6.2.4. Comparison of the Results Obtained by the Design
Method Presented in This Paper and Another Design Method.
In order to evaluate the 3D design method, it is necessary to
compare the test results with ones obtained by another design
method [15], which has finished the simulations and experiments to research its performance. Comparisons can be
made by investigating the change of head coefficient () with
the flow coefficient ().
Consider,
=

Figure 19: Differential pressure and hydraulic efficiency with

different IGVF.

6.2.3. Test Results Compared with CFD Results. Figure 20

shows a qualitative comparison between differential pressure
and hydraulic efficiency obtained by experiments and simulations, under differential operating conditions where changes from 70 m3 /h to 120 m3 /h when the IGVF is 15% and

,
2 2

= (1 ) l + g ,

pressure happens when the is 90 m3 /h, and the minimum happens when the is 110 m3 /h.
(2) The hydraulic efficiency of the impeller keeps high
with different IGVF. The maximum efficiency point of
the pump happens when the IGVF is about 40% in
each volume flow. When the is 100 m3 /h, the efficiency is higher than others.

Efficiency
CFD
Exp.

l =

2
2
(2 1 )l
+
,
l
2

(6)

2
2
(2 1 )g
g =
+
,
g
2

where , , l , g , 1 , and 2 are the mass flow rate of gas, the

total head of the pump, the head of water and gas, and the
absolute entrance and exit velocity, respectively. Consider
V
= 2 ,
(7)
2
where V2 and 2 are the average axial velocity in the outlet of
impeller and the tangential velocity in the shroud of the
impeller, respectively.

Downloaded from ade.sagepub.com by guest on March 31, 2015

Advances in Mechanical Engineering

Nomenclature

1.6
1.4

1.2
1.0
0.8
0.6

0.072

0.108

0.144

0.180
t

0.216

0.252

0.288

The results in the reference

The results in this paper

Figure 21: Comparison of head performance.

Figure 21 shows the comparison of head performance

obtained by the other design method and the method in this
paper. The details are as follows.
(1) The head performance obtained by the 3D design
method has the same trend of change with the reference results, which verified the reasonability of the 3D
design method.
(2) The results in the reference have large in low flow
rate. The 3D method results change smoothly in a
large flow range, so the 3D design method has an
advantage in the transportation of large discharge.

7. Conclusions
3D design method and an example impeller investigated by
CFD numerical simulations and bench test for a rotodynamic
multiphase pump are presented in this paper. A brief summary of the major findings are as follows.
(1) In the process of design, a 3D method based on the
numerical solution of meridian flow net for the impeller of rotodynamic multiphase pump is put forward.
(2) In the steady numerical prediction, the static pressure
gradient and GVF are uniformly distributed on the
blade to blade surfaces in different spans of the impeller, which is good for the transportation of mixture
composed of gas and liquid.
(3) In experimental study, the hydraulic efficiency of the
example pump is investigated in this paper. The experiment results verify the reasonability of the design
method. According to the comparison of the 3D
design method with another design method, the
advantage of the former is shown that it is suitable for
large volume flow rate. Furthermore, the results provide a good guidance for the initial design of pump
and can also be used as the basis for further research.

:
mix :
:
:
:
g , l :
:
:
:
:
1 , 2 :
:

Volumetric flow rate

Density of the gas-liquid mixture
Rotational speed
Head
Differential pressure
Volumetric flow rate of gas and liquid
Hydraulic efficiency
Peripheral velocity
The diameter of shroud
Half of the cone angle of hub
Diameters of hub at inlet and outlet
The distance between stream surfaces
OABC and GDEF
:
The distance between cross sections
OCFG and ABED
:
Angle between axial planes OADG and
CBEF
:
Blockage coefficient which is the area ratio
of the effective flow and the geometric
flow in the pump
:
Mass flow rate of gas
The head of water
l :
The head of gas
g :
Meridional velocity
:
1, 2: Circumferential component of the
entrance and exit absolute velocity of blade
The absolute entrance velocity
1 :
The absolute exit velocity
2 :
:
Radius from the axis of rotation
The angle between the meridian
1 :
streamline and the axis of rotation
The angle between the line of wetted
2 :
perimeter and the axis of rotation
Blade inlet angle
1 :
Blade outlet angle
2 :
:
The angle between and
:
Line of wetted perimeter
():
The integral constant depending on the
number of the flow line which represents
the entropy
:
The polytropic exponent of gas
:
The length of the meridian streamline
The real thickness of the blade
:
:
The blade angle in each calculated point
GVF:
Gas volume fraction
IGVF:
Inlet gas volume fraction.

Conflict of Interests
The authors declare that there is no conflict of interests
regarding the publication of this paper.

Acknowledgments
This work was supported by the National Science Foundation
of China (no. 51209217) and the Science Foundation of China
University of Petroleum, Beijing (no. YJRC-2013-10).

Downloaded from ade.sagepub.com by guest on March 31, 2015

10

Advances in Mechanical Engineering

References
[1] T. Rausch, T. Vauth, J. U. Brandt, and D. Mewes, A model for the
delivering characteristic of multiphase pumps, in Proceedings of
the 4th North American Confon Multiphase Technology, pp. 313
327, Banff, Canada, 2004.
[2] K. Rabiger, T. M. A. Maksoud, J. Ward, and G. Hausmann,
Thermo- and Fluid Dynamic Model of a multiphase Screw Pump,
operating at Very High gas Volume fractions, Series of the
University of Applied Sciences, 2006.
[3] K. Rabiger, T. M. A. Maksoud, J. Ward, and G. Hausmann,
Investigation of the fluid dynamic and thermodynamic behaviour of multiphase screw pumps handling liquid/gas mixtures
with very high gas volume fractions, in Proceedings of the 13th
International Conference on Multiphase Production Technology,
pp. 91104, Edinburgh, UK, June 2007.
[4] B. E. Davis, C. Kelly, K. Kierulf et al., Deep Multiphase Boosting
Made Possible, Offshore Technology Conference, Houston, Tex,
USA, 2009.
[5] S. L. Scott, J. Xu, and C. Lenz, Subsea multiphase pressure
boosting and a new approach for speed control using a hydrodynamic variable-speed drive, in Proceedings of the SPE Annual
Technical Conference and Exhibition (ATCE 06), pp. 47624774,
San Antonio, Tex, USA, September 2006.
[6] Q. Li, D. Xue, W. Lu, Y. Ban, and Y. Wang, Some discussions
about the design of rotodynamic multiphase oump, Journal of
Engineering Thermophysics, vol. 20, no. 1, pp. 6164, 1999.
[7] M. L. Corradini, Multiphase flow: gas/liquid, in Handbook of
fluid dynamics, R. W. Johnson, Ed., CRC Press LLC/Springer,
Boca Raton, Fla, USA, 1998.
[8] S. Cao, G. Peng, and Z. Yu, Hydrodynamic design of rotodynamic pump impeller for multiphase pumping by combined
approach of inverse design and CFD analysis, Journal of Fluids
Engineering, Transactions of the ASME, vol. 127, no. 2, pp. 330
338, 2005.
[9] Q. P. Li, W. Q. Lu, and D. S. Xue, Three- dimensional numerical
analysis of gas bubbles behavior in multiphase pump impeller
using boundary elemental method, in Proceedings of the 2nd
International Symposium on Multiphase, Non-Newtonian and
Reacting Flows, pp. 313317, Hangzhou, China, 2004.
[10] J. S. Anagnostopoulos, A fast numerical method for flow analysis and blade design in centrifugal pump impellers, Computers
and Fluids, vol. 38, no. 2, pp. 284289, 2009.
[11] W.-Q. Lu, Q. Li, and D. Xue, Boundary element numerical simulation of gas-liquid two-phase flow in multiphase pump impeller, Advances in Fluid Mechanics, vol. 29, pp. 1322, 2001.
[12] D. E. Goldberg, Genetic Algorithms in Search, Optimisation, and
Machine Learning, Addison-Wesley, Reading, Massachusetts,
1989.
[13] J. Zhang, H. Zhu, C. Yang, Y. Li, and H. Wei, Multi-objective
shape optimization of helico-axial multiphase pump impeller
based on NSGA-II and ANN, Energy Conversion and Management, vol. 52, no. 1, pp. 538546, 2011.
[14] Z. Y. Yu and Y. Liu, Characteristic analysis of unsteady gasliquid two-phase flow in a multiphase rotodynamic pump,
Journal of Agricultural Machinery, vol. 5, pp. 6664, 2013.
[15] Z.-Y. Yu, S.-L. Cao, and G.-Y. Wang, Numerical model for 3-D
gas-liquid two-phase flow in vane pumps, Transaction of
Beijing Institute of Technology, vol. 27, no. 12, pp. 10571064,
2007.

Downloaded from ade.sagepub.com by guest on March 31, 2015