Analysis and Prediction of

Fluid Flow Behavior in

Eissa Al-Safran1
Mem. ASME
Petroleum Engineering Department,
Progressing Cavity Pumps
Kuwait University,
Kuwait City 13060, Kuwait A progressing cavity pump (PCP) is a positive displacement pump with an eccentric
e-mails: [email protected]; screw movement, which is used as an artificial lift method in oil wells. Downhole PCP
[email protected] systems provide an efficient lifting method for heavy oil wells producing under cold pro-
duction, with or without sand. Newer PCP designs are also being used to produce wells
Ahmed Aql operating under thermal recovery. The objective of this study is to develop a set of theo-
Petroleum Engineering Department,
retical operational, fluid property, and pump geometry dimensionless groups that govern
Kuwait University,
fluid flow behavior in a PCP. A further objective is to correlate these dimensionless
Kuwait City 13060, Kuwait
groups to develop a simple model to predict flow rate (or pressure drop) along a PCP.
e-mail: [email protected]
Four PCP dimensionless groups, namely, Euler number, inverse Reynolds number, spe-
cific capacity number, and Knudsen number were derived from continuity, Navier–Stokes
equations, and appropriate boundary conditions. For simplification, the specific capacity
Tan Nguyen and Knudsen dimensionless groups were combined in a new dimensionless group named
Petroleum Engineering Department,
the PCP number. Using the developed dimensionless groups, nonlinear regression mod-
New Mexico Institute of Mining and Technology,
eling was carried out using large PCP experimental database to develop dimensionless
Socorro, NM 87801
empirical models of both single- and two-phase flow in a PCP. The developed single-
e-mail: [email protected]
phase model was validated against an independent single-phase experimental database.
The validation study results show that the developed model is capable of predicting
pressure drop across a PCP for different pump speeds with 85% accuracy.
[DOI: 10.1115/1.4037057]

Introduction tolerance and resistance properties [9]. However, metallic stator

PCPs are less efficient due to fluid slippage caused by necessary
A progressing cavity pump (PCP) is a positive displacement
internal clearance between the rotor and the stator [10]. Therefore,
pump [1] used as an effective artificial lift method, especially in
PCPs with elastomeric stators are more efficient, especially in
heavy oil production wells [2,3]. A typical oil well PCP is a sur-
applications where thermal, mechanical, or chemical degradation
face driven, single lobe pump that consists of a single external
is not an issue [11].
helical rotor turning eccentrically inside a double internal helical
In conventional elastomeric stator pumps, high differential
stator. The rotor creates a pair of sealing lines 180 deg apart inside
pressure typically develops across the pump, which increases with
the stator to form a parallel series of cavities extending from the
fluid viscosity and leads to contraction of the rotor/stator interfer-
intake to the discharge end of the pump. During rotor eccentric
ence fit and increase in the elastomer deformation. As differential
rotation, the trapped fluid in cavities is pushed from pump inlet to
pressure between adjacent cavities rises to exceed interference
pump outlet in a nonpulsation smooth flow [4]. Among all artifi-
seal capacity, a gap is formed between the rotor and the stator,
cial lift methods, the PCP is considered high efficiency pump that
allowing internal fluid slippage to take place, which decreases
can handle heavy oil, free gas, sand [5,6], and scale [4]. While the
pump efficiency. Conversely, in metallic stator pumps, a constant
rotors are typically metallic with some type of hardened coating
rotor–stator clearance exists (i.e., within manufacturing toleran-
material (e.g., chrome), PCPs are differentiated based on the stator
ces), which is independent of differential pressure along the
material, namely, elastomeric or metallic. The elastomeric PCP
pump [12].
stator is typically sized with a small negative internal clearance
between the rotor and stator for normal operation conditions,
which results in an interference or compression fitting between
the rotor and stator, which serves to minimize slippage and thus Literature Review
maximize pump efficiency [7]. However, the elastomer material is Moineau [10] modeled slip flow in a PCP using the
exposed to thermal, mechanical, and chemical degradation due to Hagen–Poiseuille equation. Gamboa et al. [2] identified and
pumping erosive and/or corrosive multiphase flow mixtures. It is evaluated the physical parameters affecting slippage across metal-
observed that flow of viscous fluid with high gas fraction, and sus- to-metal PCP seal lines. Furthermore, their single-phase experi-
pended sand particles causes a chaotic pressure distribution along mental study revealed that the relationship between flow capacity
the pump cavities, leading to both thermal and mechanical degra- and differential pressure is linear in the laminar flow regime for
dation, in addition to elastomer swelling [8]. On the other hand, metallic stator PCP (i.e., constant clearance). However, a nonlin-
metallic stators are mechanically, chemically, and thermally more ear behavior was observed when water was pumped under turbu-
durable than elastomeric stators due to their high degradation lent flow conditions. It was concluded that PCP performance is
affected by fluid viscosity, rotor–stator interference (pump fit),
and pressure-induced deformation of pump stator. Pessoa et al.
1
Corresponding author. [13] used Colebrook’s friction factor equation to predict turbulent
Contributed by the Fluids Engineering Division of ASME for publication in the
JOURNAL OF FLUIDS ENGINEERING. Manuscript received February 2, 2017; final
fluid flow behavior in a PCP. Karthikeshwaran and Samad [14]
manuscript received May 28, 2017; published online August 28, 2017. Assoc. modeled metal-to-metal pump behavior with viscous oil in an
Editor: Wayne Strasser. approach similar to Gamboa et al. [2]. Gamboa et al. [15] found

Journal of Fluids Engineering Copyright V

C 2017 by ASME DECEMBER 2017, Vol. 139 / 121102-1

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 11/06/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 that the effects of fluid viscosity, pump fit, and internal elastomer volume method with a moving mesh, where the gas–liquid mix-
deformation on pump performance are related to the stator ture is simulated using the Eulerian–Eulerian approach. The Aze-
material properties. In metallic stator PCPs, although stator vedo et al. model predictions showed significant discrepancy
deformation does not occur, the effects of fluid viscosity and when gas void fraction exceeded 20%. It is speculated that the dif-
rotor–stator clearance are significant. ferences observed in the cavity pressure distribution between the
Robello and Saveth [16] developed a design methodology for Azevedo et al. model and the Olivet et al. data can be attributed to
PCP performance based on geometric variables such as stator gas compression effect and the selected meshing approach.
pitch length and helix angle. However, they assumed a leak-free To simulate the interaction between the fluid and structure
pump and neglected the effect of stator material variation. Zhou parameters of a PCP, Lima et al. [12] introduced a simplified 3D
and Yuan [4] used NODALTM analysis to predict the performance computational model to simulate fluid–structure interaction inside
of PCP pumped wells. They developed a method to determine the single-lobe elastomer stator PCP. Element-based finite volume
optimum pump rotational speed, production flow rate, and pump method was used to solve the unsteady-state flow equation of both
intake pressure in terms of well inflow and outflow performance. laminar and turbulent flow regimes. Relative motion between the
In addition, they introduced an effective method to correct pump rotor and stator, as well as stator deformation, was taken into
catalog performance curves to account for fluid viscosity. How- account. Since the latter depends on the differential pressure
ever, Zhou and Yuan [4] reported that the theoretical slip flow is across the pump, an iterative solution procedure was required at
not a function of pump rotational speed, which contradicts the each time step. This method accounts only for radial deformation
experimental study of Olivet et al. [7]. Nguyen et al. [17] modeled of the stator, neglecting axial and azimuthal deformations. The
PCP volumetric capacity using the three-dimensional (3D) vector Lima et al. [12] model can be extended to nonisothermal and/or
theory and the theory of hypocycloid. A theoretical factor character- multiphase flow in elastomeric stator pumps. Using CFD model-
izing PCP behavior was derived as a function of pump eccentricity, ing, Strasser [28] investigated the effect of fluid properties, geo-
stator lobe semicircle diameter, number of lobes, and pitch length. metrical and operational condition of gear pumps. Using the
R
Martin et al. [18], Olivet et al. [7], Bratu [8], and Zhang et al. Eulerian mixture model in FLUENTV 6.2.16 solver, the coefficient
[19] have experimentally investigated multiphase flow inside of spatial variation was used as an indicator of mixing efficiency.
metal-to-metal PCPs. The effects that different gas void fractions Nguyen et al. [29] developed a simulation tool to predict the
and pump rotational speed have on the pressure and temperature single-phase flow in PCP, which combines two existing analytical
profiles along a PCP and motor were investigated. For example, models, namely, theoretical pump performance model and slip-
Bratu [8] investigated the effect of a multiphase flow mixture with page model. Their validation study revealed reasonably accurate
high gas void fraction on the thermal and mechanical pump predictions.
behavior, demonstrating that the pump temperature increased in It is important to recognize, as identified from the above stud-
response to increase in discharge pressure and/or rotor speed, thus ies, that fluid flow and seal conditions inside PCP affect both pres-
promoting degradation of pump run life. Zhang et al. [19] devel- sure and temperature profiles. This process is somewhat complex
oped a multiphase flow model for PCP, and presented a procedure since with high interference and/or high speeds, the internal
to integrate PCP modeling with a grid-based network model. energy developed by the cyclic elastomer deformation generates
Belcher [20] analytically approximated internal pump slippage heat (i.e., energy is expended to deform elastomer and only a
as a flow between two infinite parallel plates. Behzadi et al. [21] portion of that energy is recovered when the elastomer rebounds),
proposed a mathematical model accounting for both homogenous giving rise to hysteresis, defined as the difference between
and nonhomogeneous multiphase flows in PCP. Andrade et al. expended and recovered energy, i.e., internal friction. Further-
[22] applied the lubrication theory to develop an asymptotic more, a rise in temperature of the elastomeric material over a por-
model, simulating single-phase, incompressible, Newtonian fluid tion of the stator leads to thermal expansion of the material
flow inside metallic PCP with constant internal clearance. A vali- (causing cavity distortion in extreme cases) and a corresponding
dation example based on pump performance showed agreement increase in the rotor–stator fit, which causes more internal heat
between their experimental data and a 3D transient commercial generation and expansion. Note that a portion of generated heat
model. goes into the fluid passing through the pump, which affects the
Paladino et al. [3,23,24], Lima et al. [25], Berton et al. [26], viscosity of heavy oil fluids. This behavior of stator elastomer has
and Azevedo et al. [27] developed computational fluid dynamics been observed and linked to elastomer fatigue life in downhole
(CFD) models to simulate fluid behavior in metallic PCPs. How- drilling motor, which is a PCP with a reversed operation in which
ever, their models are extremely sophisticated, expensive, and fluid flow is used to generate power [30].
time-consuming because of the complexity of accurately meshing The comprehensive literature review presented above shows
the pump cavity geometry. Paladino et al. [3] proposed a 3D CFD that the previous work aimed to establish a more comprehensive
model for predicting rigid stator PCP performance based on geo- understanding of the fluid flow behavior in PCP, and to develop
metrical parameters and operating conditions. The finite volume predictive tools to simulate a variety of flow conditions. However,
method was implemented to discretize the governing flow equa- none of the previous studies provided a simple and physical-based
tions through the CFX11 (ANSYS, 2008) software package. Their model that accounts for operational condition, fluid physical prop-
results agreed with the experimental results of Olivet et al. [7] and erties, and pump geometry. Therefore, this study aimed to provide
Gamboa et al. [2,15]. However, some deviations were observed a simplified visual, physical, and mathematical modeling to
due to fluid data uncertainty and turbulence effects. The Paladino improve both understanding and prediction of fluid flow behavior
et al. model is the first to simulate fluid flow in entire pump geom- in PCP.
etry, using mesh generation process. Berton et al. [26] have used
the “rigorous V&V” approach in the LEMMATM CFD software Progressing Cavity Pump Geometrical and
package to develop a numerical simulation tool for non-
Newtonian fluid flow in a PCP. Validation against experimental Kinematics Description
data showed accurate model predictions. Azevedo et al. [27] In this study, a 3D geometrical model of a pump stator and
proposed a CFD model for unsteady, 3D fluid flow in metallic rotor was developed by integrating pump cross-sectional images
PCP, where the gas phase was modeled using ideal gas equation in the axial direction as shown in Fig. 1, using the AUTOCADTM
of state, assuming a bubbly flow pattern. The model was then vali- software package. The integration of these cross-sectional images
dated with the experimental data obtained by Olivet et al. [7] and generated a double helix stator that rotates 360 deg per stage
Gamboa et al. [2,15] for different gas volume fractions. In their around its stationary center, while being extended in the axial
study, the commercial software ANSYS-CFXTM was used to solve direction. Furthermore, the resulting 3D geometrical model was
the multiphase governing equations, using element-based finite animated to visualize the motion of both the rotor and the fluid

121102-2 / Vol. 139, DECEMBER 2017 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 11/06/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 1 PCP 3D image: (a) successive two-dimensional cross
sections and (b) 3D pump image

Fig. 2 Video footage of 3D configuration of rotor motion in two

pump stages

flow in the PCP. A video clip of the developed model for two
pump stages (i.e., a pump stage is equal to one stator pitch length)
is shown in Fig. 2.
Using the PCP animated AUTOCADTM model, Fig. 3 was devel-
oped, which illustrates fluid motion as well as rotor position with
Fig. 3 3D video footage of rotor motion showing variations in
respect to stator along one pump stage. In this study, these images pump cross section
were used in understanding the kinematics of PCP, the spatial var-
iation of the seal lines, and the flow boundary conditions during
rotor motion. This understanding enabled our modeling of fluid
flow behavior in PCP. parallel cavities, respectively. Figure 6 illustrates the direction of
Four geometrical parameters describe the geometry of PCP, both cross-sectional and longitudinal slip flows across PCP
namely, rotor diameter (dr), eccentricity (e), pump clearance (w), cavities.
and stator pitch (Ps). At any pump cross section, the rotor has two It is important to note that the cavity geometry can change con-
simultaneous motions during operation, namely, an angular siderably between different pump models. In general, flow and
motion and a transitional motion. Figure 4 shows the angular and slippage vary considerably based on cavities geometry, i.e., wide-
translational motions of a fixed point (P) on rotor surface through short pitch length versus narrow-long pitch length. This is due to
one complete cycle (or turn). The rotor center of mass (Cm) rotates rubber distortion in axial direction as well as overall length of the
around the stator center (Cs), forming a circle with a radius repre- seal lines. Longer pitch pump tends to have limited end-to-end
sentative of the pump eccentricity (e). Rotor motion starts with cavity seepage. Such analyses assume uniform dimensional and
rotor center (Cr) periodically moving (translating) back and forth material properties for stator elastomer, which is not the case in
along an axis that passes through centers of the stator cusps. real pumps due to manufacturing/curing process.
Therefore, all points on the rotor circumference move in com- The understanding of fluid flow behavior in PCP requires an
bined motions, namely, 360 deg angular motion, and translate understanding of the pump motions and cavities formation.
along the distance between stator cusps. Figure 7 shows one full cavity on one side of a centered rotor.
As pump rotor turns, cavities are formed between the rotor and Figure 7(a) shows seal lines along which longitudinal slip flow
stator, which move continuously from pump inlet to outlet as occurs. These longitudinal lines extend over a longer distance
shown in Fig. 5. These cavities are circumscribed by the seal lines, compared to the seal lines over which the cross-sectional flow
which are formed at the close proximity locations between the occurs. The changing boundaries of the fluid cavity make it a
rotor and stator where the clearance is minimum or nil. In metal- challenge to simulate fluid flow in PCP by solving the governing
to-metal PCP, the alternately changing seal lines circumscribing equation of motion. Figure 7(b) shows the locations at which a
pump cavities do not generate complete separated cavities, result- single fluid cavity exhibits the minimum and maximum flow
ing in fluid back flow between cavities, where a rotor–stator clear- areas, which increase gradually and then decrease from one end of
ance is present. This back flow (internal slippage) is eliminated in the cavity to the other as shown by the arrows in Fig. 7.
negative clearance pumps, where the compression fit of the rotor
within the stator serves to create isolated pump cavities. The inter-
nal slip flow is divided into two types, namely, cross-sectional, Theoretical Modeling
and longitudinal, where cross-sectional slip flow and longitudinal The fluid flow governing equations in a PCP are presented in
slip flow occur between end-to-end cavities, and adjacent or this section, along with a dimensional analysis of these equations

Journal of Fluids Engineering DECEMBER 2017, Vol. 139 / 121102-3

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 11/06/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Governing Conservation Equations. The isothermal, steady-
state, incompressible, Newtonian fluid flow is governed by the
conservation laws of mass and momentum through the continuity
equation and Navier–Stokes equation, respectively. In this study,
the differential form of the continuity and Navier–Stokes equa-
tions in vector notations were dimensionally analyzed using a
cylindrical coordinate system with relevant boundary conditions.
The developed dimensionless groups were then scaled to pump
geometrical and operational parameters to characterize the fluid
flow behavior within the PCP. The differential form of the
continuity equation for incompressible steady-state flow in vector
notation is

rv¼0 (1)

The differential form of the momentum equation in vector nota-

tion is given as
 
Dv
q ¼ rp þ lr2 v þ qg (2)
Dt

Due to the absence of free surfaces within the PCP cavities, which
are isolated by seal lines, the gravity term in Eq. (2) is negligible,
reducing Eq. (2) to
 
Dv
q ¼ rp þ lr2 v (3)
Dt

Fig. 4 PCP rotor translational and rotational motion in one PCP theoretical flowing area is given by Nguyen et al. [17] as
cycle
Af ¼ pwðds  wÞ þ 4eds (4)

where Af is the flowing cross-sectional area, ds is the diameter of

the semicircle at the stator cusp, w is the pump clearance, and e is
the pump eccentricity. The axial velocity component in a PCP (vz )
is defined in terms of the actual flow rate (qa) and the pump cross-
sectional flow area (Af) as
qa qa
vz ¼ ¼ (5)
Af pwðd  wÞ þ 4ed
Fig. 5 3D images of pump stator, rotor, and fluid cavities
Progressing Cavity Pump Scaling Parameters. Andrade
et al. [22] have mathematically analyzed fluid flow in a PCP and
completed using appropriate boundary conditions and scaling analytically solved the Navier–Stokes equations, making several
parameters. Four dimensionless groups are initially developed and simplifying assumptions based on the lubrication theory. In
later reduced to three for numerical and mathematical efficiency. addition, they carried out a geometrical analysis to develop appro-
An empirical PCP model is proposed based on the dimensionless priate boundary conditions in terms of the pump geometrical
groups, and it is subsequently statistically evaluated and parameters. They described the developed boundary conditions in
experimentally validated. terms of independent geometrical parameters, which characterize

Fig. 6 Cross-sectional and longitudinal slip flows in PCP

121102-4 / Vol. 139, DECEMBER 2017 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 11/06/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 7 3D configuration of fluid cavity in PCP

the fluid flow behavior in a PCP. These independent parameters Dividing Eq. (10) by qLr =t2r gives
are stator cusp radius (rs), rotor radius (rr), clearance between
  " 2#  
rotor and stator (w), pump eccentricity (e), rotor angular velocity Dv pr tr ltr
(x), and stator pitch length (Ps). In this study, these parameters 
¼ 2
ðr p Þ þ ðr2 v Þ (11)
Dt qLr qLr
were used to scale the dimensionless groups established through
a dimensional analysis of the continuity and Navier–Stokes
equations. Dimensionalizing and simplifying Eq. (4) gives
 
Af
Dimensional Analysis. In this study, Eqs. (1), (3), (4), and (5) ¼ ðpw ðd   w Þ þ 4e d Þ (12)
are dimensionally analyzed. The reference variables are denoted Lr 2
as length (Lr), linear velocity (Vr), time (tr), and pressure (pr). The
dimensionless PCP geometrical and operational parameters are Dimensionalizing Eq. (5) and substituting Eq. (12) into it gives
defined in terms of dimensionless values as r  ¼ r=Lr , z ¼ z=Lr ,   
vr  ¼ vr tr =Lr , vh  ¼ vh tr =Lr , vz  ¼ vz tr =Lr , t ¼ t=tr , p ¼ p=pr , qa tr 1
vz  ¼ (13)
w ¼ w=Lr , d ¼ d=Lr , and e ¼ e=Lr , where r  , z , w , d , and Lr 3 pw ðd  w Þ þ 4e d
e are dimensionless radial coordinate, axial coordinate, pump
clearance, stator cusp diameter, and pump eccentricity, respec- Further details of the dimensional analysis are given in
tively. In addition, vr  , vh  , vz  , t , and p represent dimensionless Ref. [31]. The terms in brackets in Eqs. (11)–(13) are the dimen-
radial, azimuthal, and axial velocities, dimensionless time, and sionless groups, namely, Euler number NEu ¼ ½pr tr 2 =qLr 2 , inverse
1
dimensionless pressure variables, respectively. Dimensionalizing Reynolds number NRe ¼ ½ltr =qLr 2 , the square of Knudsen num-
Eq. (1) gives ber NKn ¼ ½Af =Lr 2 , and specific capacity number U ¼ ½qa tr =Lr 3 ,
which govern the fluid flow behavior in a PCP. To ensure that ther-
1   mal energy generation due to viscous dissipation is negligible, the
rv¼ ðr  v Þ ¼ 0 (6)
tr Brinkman dimensionless number ðNBr ¼ v2 lL =kL DTÞ is calculated
using the range of the experimental parameters as input data,
assuming 1  C temperature difference between fluid and rotor and
The dimensionless pressure divergent, velocity vector Laplacian, stator walls. It was found that Brinkman number range is from
and velocity substantial derivative term in Eq. (3) are given, 2.22  105 to 3.69, i.e., small. Furthermore, the ratio of Brinkman
respectively, as number to Peclet number ðNPe ¼ dqvCp =kL Þ is calculated, which
represents the relative importance of generated heat by viscous dis-
pr sipation to convective heat. Using the range of the experimental
rp ¼ ðr p Þ (7) data, the calculated ratio was found to range from 5.83  1010 to
tr
8.37  107, indicating a very small contribution of generated heat
due to viscous dissipation as compared to convective heat.
1
r2 v ¼ ðr2 v Þ (8)
tr Lr
Dimensionless Parameters Scaling. Substituting the reference
variables in the derived dimensionless groups with the elementary
 
Dv Lr Dv parameters of Andrade et al. [22] used for boundary conditions
¼ (9) (i.e., stator pitch length for length scale ðLr ¼ Ps Þ, reciprocal of
Dt tr 2 Dt
angular velocity for time scale ðtr ¼ x1 Þ, and differential pres-
sure across one pump stage for pressure scale ðpr ¼ DpÞ), and
Substituting Eqs. (7)–(9) into Eq. (3) gives the dimensionless defining x ¼ 2pN, gives the following scaled dimensionless
1
Navier–Stokes equation as groups: NEu ¼ ½Dp=qPs 2 N 2 , NRe ¼ ½l=qPs 2 N, U ¼ ½qa =P3s N,
and NKn ¼ ½Af =Ps 2 . To simplify the set of dimensionless parame-
  ters, the specific capacity and the square of Knudsen number
qLr Dv pr l
¼ ðr p Þ þ ðr2 v Þ (10) groups were combined into a new dimensionless number, named
tr 2 Dt Lr tr Lr “PCP number” given as

Journal of Fluids Engineering DECEMBER 2017, Vol. 139 / 121102-5

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 11/06/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 qa
NPCP ¼ (14)
Ps Af N

To verify the derivation accuracy of the above dimensionless

groups, the Buckingham Pi theorem was applied, which resulted
in the same set of dimensionless groups [31].

Empirical Modeling
Using the developed dimensionless groups, a relationship relat-
ing the Euler number to the inverse Reynolds number and the
PCP number is theorized as

1
NEu ¼ f ðNRe ; NPCP Þ (15)

Note that the unit of N is revolution per second. Using this theor-
ized dimensionless relationship, a nonlinear regression model for Fig. 9 Two-phase flow proposed model curve fit to experimen-
steady-state, single-phase, Newtonian fluid flow was developed tal data
using the experimental database of Gamboa et al. [2,15]. The
ranges of the dimensionless numbers of this database are
0.053–1.04 for NPCP, 44–1980 for NRe, and 0.62–25,000 for NEu. Two-Phase Homogenous Regression Model. The experimen-
Furthermore, a second nonlinear regression model was developed tal data sets of Olivet et al. [7] and Gamboa et al. [2,15] were used
for homogeneous two-phase flow in a PCP based on the two- to generate a homogenous two-phase flow regression model. The
phase experimental databases of Olivet et al. [7] and Gamboa two-phase homogenous empirical modeling required the predic-
et al. [2,15]. The ranges of the dimensionless numbers in these tion of mixture fluid properties, such as mixture density and vis-
databases are 0.27–0.96 for NPCP, 985–1980 for NRe, and cosity along the pump. Using the measured pressure distribution
1.39–4450 for NEu. along the PCP in both databases, the density and viscosity of air
and lube oil were determined, assuming an isothermal system.
The no-slip liquid holdup was also calculated at each pressure
Single-Phase Dimensionless Regression Model. Figure 8 point, from which the mixture fluid properties were determined.
shows a 3D plot of the developed dimensionless numbers using Appendix B presents the calculation of mixture fluid properties
the Gamboa et al. [2,15] data, indicating a nonlinear relationship such as density and viscosity, which are used to develop the pres-
between Euler dimensionless number, and both Reynolds and ent model.
PCP dimensionless numbers. A best-fit nonlinear empirical corre- Figure 9 shows a 3D plot of Euler number versus Reynolds and
lation representing the relationship in Fig. 9 was generated using PCP numbers developed based on the two-phase flow data sets
the TABLECURVE 3DTM software package as follows: given in Refs. [2], [7], and [15]. The trend relationship between
the dimensionless groups differs considerably from the relation-
23:5 ship of single-phase flow. Using the TableCurve 3DTM software
LnðNEu Þ ¼ 5:23  3:44ðNPCP Þ3 þ (16) package, a nonlinear regression model was generated as follows:
LnðNRe Þ
0:277 41:4
LnðNEu Þ ¼ 7:32 þ þ pffiffiffiffiffiffiffiffi (18)
To solve for actual flow rate, Eq. (16) is rearranged as LnðNPCP Þ NRe
2  31=3 To predict the actual flow rate, Eq. (18) can be rearranged as
6:84 Dp follows:
qa ¼ 6 ! þ 1:52  0:291  Ln 7 ð Ps Af N Þ
6 2 qPs 2 N 2 7
4Ln NPs q 5   
l 0:277
qa ¼ Ps Af N exp (19)
X
(17)
where X is
 
Dp 41:4
X ¼ Ln  pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  7:32 (20)
qPs 2 N 2 NPs 2 q=l

The developed single- and two-phase flow models in Eqs. (17)

and (19) are applicable for laminar flow, because they were devel-
oped using dimensionless groups derived from Navier–Stokes
equation and the Reynolds number of the fitted data are below
2000. However, when both models were tested against the

Table 1 Single-phase and homogenous two-phase flow

models coefficients of determination
2
Model Model df Error df R2 R

Single-phase flow 2 218 0.880 0.879

Fig. 8 Single-phase proposed model curve fit to Gamboa et al. Two-phase flow 2 129 0.860 0.857
[2,15] experimental data

121102-6 / Vol. 139, DECEMBER 2017 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 11/06/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 2 Single-phase and homogenous two-phase flow models’ coefficients standard error, 95% confidence interval, t-values,
and p-values

Model Coefficient ^
Coefficient estimate b Standard error sb^ j Lower CI95% Higher CI95% t-value p-value
j

Single-phase flow b0 5.226 0.1248 4.981 5.472 41.87 <0.0000

b1 3.435 0.1263 3.684 3.186 27.20 <0.0000
b2 23.51 0.6800 22.16 24.85 34.57 <0.0000
Two-phase flow b0 7.319 0.0881 7.144 7.493 83.19 <0.0000
b1 0.277 0.0182 0.241 0.3129 15.26 <0.0000
b2 41.42 3.391 34.72 48.13 12.22 <0.0000

turbulent flow data of Ref. [7], they showed accurate predictions differential pressure, pump speed, and gas void fraction, are inves-
with a coefficient of variation (R2) of 0.98. Therefore, although tigated. For example, at a given pump type and pressure head,
the models were developed from laminar regime data, it appears Eq. (17) shows that as liquid viscosity increases, pump volumetric
they are applicable for turbulent flow conditions, but should be flow rate increases. This is mainly due to the reduction of internal
applied with caution that this finding cannot be generalized for pump slippage, i.e., higher resistance to back flow. However,
any conditions. It is also important to note that the maximum although the increase in liquid viscosity increases the friction
value of gas void fraction in the two-phase flow data used to factor; the effect of internal pump slippage is predominant. [29].
develop the two-phase flow model is 50%. Therefore, the model is Furthermore, the functional relationship in Eq. (17) between
accurate for up to 50% gas void fraction, and extrapolating it pump speed and pump volumetric flow rate is also directly propor-
beyond this value should be done with caution as it may produce tional due to the added hydraulic horsepower as the pump rota-
significant prediction error. tional speed increases. However, Arellano [32] reported that field
The fluid flow physics in PCP depends on pump operational operations have shown that for viscous oil, lower pump speed is
and geometrical conditions, as well as fluid properties, which are recommended because not only it increases pump volumetric effi-
all related to fluid slippage phenomenon. To ensure that both pro- ciency due to lower slippage but also it increases pump run life.
posed single-phase and two-phase flow models given in Eqs. (17) Conversely, Eq. (17) shows that as differential pressure across the
and (19) capture the physics of the flow, the functional relation- pump increases, pump volumetric efficiency decreases due to
ships between the pump volumetric flow rate (or volumetric effi- higher total dynamic head that the pump should overcome. Addi-
ciency) and the independent variables, such as viscosity, pump tional reason for this inverse relationship is that as differential

Fig. 10 A schematic of the PCP testing facility at the New Mexico Institute of Min-
ing and Technology Pumping Facility

Journal of Fluids Engineering DECEMBER 2017, Vol. 139 / 121102-7

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 11/06/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 3 Differential pressure validation of the proposed
single-phase model at various pump rotational speeds

Pump rotational e1 e2 e3 e4 e5 e6
speed (rpm) (%) (%) (%) (MPa) (MPa) (MPa)

200 19.33 19.38 9.89 0.11 0.11 0.05

300 6.59 6.59 0.92 0.04 0.04 0.01
400 3.97 6.12 6.28 0.03 0.04 0.05
500 1.13 2.45 2.45 0.02 0.03 0.02

confidence intervals given in Table 2 enable the developed models

to be flexible through tuning its coefficients to fit a given data.
The calculation details of the statistical parameters shown in
Table 2 are given in Ref. [31].

Model Validation
The proposed single-phase flow model for a PCP was validated
against independent single-phase laminar experimental data acquired
using a PCP with a stator made of nitrile butadiene rubber material.
Fig. 11 Model validation results at different rotational speeds The geometrical parameters of the pump are Ps ¼ 60 mm,
e ¼ 3.5 mm, d ¼ 8.3 mm, and w ¼ 0.5 mm. Figure 10 shows the
experimental setup located at the New Mexico Institute of Mining
and Technology Pumping Facility, which includes a MicroMotionTM
pressure increases, pump interference increases, resulting in flow metering section to obtain accurate measurements of fluid
higher slippage and lower volumetric efficiency. More details on density and mass flow rate. The test section was designed to mea-
the interference increase due to the increase of differential pres- sure pressure and temperature across the pump, using gauge pres-
sure are found in the validation section of this paper. sure transmitters and temperature transducers, respectively. The
The physical effect of two-phase flow on pump volumetric 50-gal oil tank is equipped with a heating element attached at the
efficiency is significant due to possible gas lock and pump over- bottom to vary the oil temperature, and thus the oil viscosity. A
heating, which deteriorates both the pump volumetric efficiency LabVIEWTM data acquisition system is used to acquire raw volt-
and run life. This effect is captured by the proposed two-phase age data and process it into actual values of desired measurement
volumetric flow rate model given by Eq. (19) as follows: The variables. In this facility, 892 kg/m3 density mineral oil was used
presence of gas phase reduces the pump volumetric efficiency for to conduct the first stage of the experimental work in which oil
two reasons, namely, less amount of liquid in each cavity, and was pumped with the PCP operating at constant rotational speeds
additional slippage between gas and liquid phases. Equation (19) of 200, 300, 400, and 500 rpm. While the pump rotational speed
shows that as the pump flowing cross-sectional area decreases; the was maintained constant, the oil temperature in the tank was raised
pump volumetric flow rate decreases. Furthermore, the increase of gradually from 25  C to 40  C, decreasing the oil viscosity from
gas void fraction reduces the mixture density and viscosity, which 0.4 Pas to 0.19 Pas.
both increase the variable “X” in Eq. (19), resulting in low pump Figure 11 is a cross plot of the calculated differential pressure
volumetric flow rate. (DpC ) plotted versus the measured differential pressure (DpM ) for
a number of data points obtained at different rotational speeds.
Figure 11 shows that as the pump rotational speed decreases, the
Model Statistical Evaluation model underpredicts the pressure drop data. This behavior of the
This section presents the overall and coefficient statistical reli- proposed model is attributed to the fact that at low pump speed,
ability evaluation of the proposed empirical models. Table 1 sum- lower differential pressure is formed across the pump, at which
marizes the overall statistics of the regression models developed the stator is deformed and the actual cross-sectional area deviates
for both single-phase and homogeneous two-phase flow condi- from that predicted by Eq. (4) (stator diameter is greater than or
2
tions. As shown, the adjusted coefficients of determination (R ) equal to rotor diameter). This deviation in cross-sectional area
are 0.879 and 0.857 for the single- and two-phase flow models, introduces the under-prediction error of the model seen in Fig. 11.
respectively, indicating the fraction of variation in the Euler num- Conversely, at higher pump speed (i.e., higher differential pres-
ber that is explained by the variation in both the PCP and inverse sure), the cross-sectional area extends and a clearance starts form-
Reynolds dimensionless numbers. This result shows a very good ing at the sealing line as shown schematically in Fig. 12, resulting
statistical reliability of both developed regression models. in a cross-sectional area that is accurately predicted by Eq. (4),
Table 2 shows the model coefficient statistics for both the which reduces the model error. Another postulated reason for the
single-phase and the homogeneous two-phase flow models. The model deviation at high pump rotational speed is viscosity

Fig. 12 Pump interference change with respect to pump rotational speed and differential
pressure ((a) low-speed, (b) increasing speed, and (c) high-speed)

121102-8 / Vol. 139, DECEMBER 2017 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 11/06/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 variation with time during the experiment. Although this valida- R¼ gas constant, ML2/t2T; radial position, L
tion results show a satisfactory match between model prediction R2 ¼ regression coefficient of variation
and experimental measurements, further validation with other s¼ standard error
pump models, pump fits, and fluid conditions is important to t¼ time, t
assess model reliability and accuracy. T¼ temperature, T
A statistical model evaluation was carried out through the v¼ velocity vector, L/t
calculation of six statistical error parameters (e1–e6), which are v¼ velocity, L/t
presented in detail in Appendix A. Table 3 shows the statistical V¼ characteristic velocity, L/t
validation results for a comparison of the predicted differential w¼ clearance between rotor and stator, L
pressures against the corresponding experimental measurements. x¼ distance, L
These results indicate that both the absolute average percent error z¼ axial coordinate of points, L
ðe2 Þ and the absolute average error ðe5 Þ decrease as the pump rota-
tional speed increases. Consistent with the observations made
based on Fig. 11, this suggests that the model performs better at
Greek Symbols
high rotational speed conditions. b ¼ correlation coefficient
D ¼ difference
en ¼ average percent error (%); absolute average percent error
Conclusion (%); percent standard deviation (%); average error; absolute
Fluid-flow dimensional analysis and empirical modeling in average error; standard deviation
PCP are carried out in this study. A set of operational, fluid prop- k ¼ no-slip holdup
erty, and pump geometry dimensionless groups that govern fluid l ¼ fluid dynamic viscosity, M/Lt
flow behavior in a PCP were developed theoretically from conser- q ¼ fluid density, M/L3
vation laws of mass and momentum. These groups are, namely, U ¼ specific capacity
Euler number, inverse Reynolds number, specific capacity num- x ¼ rotor speed of rotation, 1/t
ber, and Knudsen number. For further simplification, the specific
capacity and Knudsen dimensionless groups were combined in a Subscripts/Superscripts
new dimensionless group, named the PCP number. These dimen- a¼ actual; air
sionless groups were then used to develop simple empirical sin- C¼ calculated
gle- and two-phase flow models to predict flow rate (or pressure Eu ¼ Euler
drop) across a PCP. The developed models can be used to predict f¼ flow
PCP performance in oil wells operating under isothermal, steady- i¼ index; intake
state, incompressible, and Newtonian fluid flow. A validation j¼ index
study of the single-phase model, using independent PCP experi- Kn ¼ Knudsen
mental database, showed that the proposed model predicts pres- L¼ liquid
sure drop across a PCP for a wide range of pump speeds with 85% m¼ mass
accuracy. The proposed model in this study provides a simple the- m¼ mixture
oretical/empirical model to characterize PCP performance consid- M¼ measured
ering important fluid, pump geometry, and flow parameters. The n¼ number
proposed two-phase flow PCP model in this study is crucial in Pe ¼ Peclet
understanding the fluid compressibility and gas void fraction r¼ rotor; radial direction; reference
effects on pump performance. R¼ relative
Re ¼ Reynolds
s¼ stator
Acknowledgment x¼ distance
The authors are grateful for Cam Mathew and Paul Skoczylas z¼ axial direction
of C-FER Company (Edmonton, AB, Canada) for reviewing and h¼ tangential direction
editing this paper, and to Jose Gamboa for providing experimental
data. Superscripts
* ¼ dimensionless
Nomenclature ^¼ unit vector; estimated
a¼ acceleration vector, L/t2 ¼ adjusted
A¼ cross-sectional area, L2
C¼ geometrical center
Cp ¼ heat capacity (or specific heat) at constant pressure, L2/t2T Appendix A
d¼ diameter, L The relative percentage error is used to calculate the first three
df ¼ degrees-of-freedom parameters, e1 , e2 , and e3 , given as:
e¼ PCP eccentricity, L; actual error; natural exponential Relative error, eR
g¼ gravity acceleration vector, L/t2
g¼ gravitational acceleration, L/t2 DpC  ðDpM Þ
k¼ number of pump stages; thermal conductivity, ML/t3T eR ¼ (A1)
DpM
L¼ length, L
M¼ mass, M
n¼ number of data points Average percent error, e1
N¼ dimensionless number; rotor rotational speed, 1/t
p¼ pressure, M/Lt2 !
P¼ stator pitch length, L 1 Xn

q¼ pump flow rate, L3/t e1 ¼ eR;i  100 (A2)

n i¼1
r¼ radial coordinate, L; radius, L

Journal of Fluids Engineering DECEMBER 2017, Vol. 139 / 121102-9

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 11/06/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Absolute average percent error, e2 highest differential pressure encountered in the experiments
results in a density difference from inlet to outlet of 13.83 kg/m3.
! Air density qa at the mid of the pump is given as
1 Xn
e2 ¼ jeR;i j  100 (A3)
n i¼1 qa ¼ 5:9  104 Dppump þ 3:64 (B4)

The homogeneous two-phase mixture density at test experiment

Percent standard deviation, e3
conditions is calculated as
0sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1
n  
@ 1 X e1 2 A qm ¼ qa ð1  kL Þ þ qL kL
e3 ¼ eR;i   100 (A4)
n  1 i¼1 100 ¼ 5:9  104 Dppump ð1  kL Þ þ 862:43kL þ 3:64 (B5)

Using the actual error, the other three parameters, namely, e4 , where qm is the mixture density in kg/m3, qL is the liquid density
e5 , and e6 are calculated and expressed in pressure units (MPa) as of 866.07 kg/m3, and kL is the homogeneous no-slip liquid holdup.
Actual error, e (MPa) Gas viscosity is constant at 1.98  105 Pas in experimental data
sets; thus, the homogeneous two-phase viscosity is calculated as
ea ¼ DPC  ðDPM Þ (A5)
lm ¼ la ð1  kL Þ þ lL kL ¼ 1:98  105 ð1  kL Þ þ 4:2  102 kL
Average actual error, e4 (MPa)
(B6)
!
1 Xn
where lm is the mixture viscosity in Pas and lL is the liquid
e4 ¼ ea;i (A6)
n i¼1 viscosity of 0.042 Pas. The linear averaging of the density and
viscosity with the no-slip liquid holdup as given in Eqs. (B5) and
Absolute average actual error, e5 (MPa) (B6) is selected because it is widely used in gas–liquid two-phase
flow modeling. One example where a nonlinear (exponential or
! harmonic, for example) averaging may be used is in oil–water
1 Xn
e5 ¼ jea;i j (A7) mixtures, where emulsion may form which will increase the mix-
n i¼1 ture (apparent) viscosity exponentially when the mixture is oil
continuous, and decrease it also exponentially when the mixture is
Standard deviation, e6 (MPa) water continuous. Furthermore, some specific empirical two-phase
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi correlations recommend the use of a nonlinear averaging for mix-
uX n ture fluid properties to better fit their pressure gradient data, which
u 2
u ð ei  e 4 Þ cannot be generalized. In this study, the linear averaging of mix-
t
e6 ¼ i¼1
(A8) ture fluid properties using no-slip liquid holdup is selected due to
n1 its accuracy as supported in multiphase flow literature.

References
Appendix B [1] Cholet, H., and Horvath, M., 1997, Progressing Cavity Pumps, Editions
The experimental data sets of Olivet et al. [7] and Gamboa Technip, Paris, France.
[2] Gamboa, J., Olivet, A., and Espin, S., 2003, “New Approach for Modeling Pro-
et al. [2,15] were acquired at constant gas void fractions, using air gressing Cavity Pumps Performance,” SPE Annual Technical Conference and
and lube oil of 0.042 Pas viscosity. Temperature and intake pres- Exhibition, Denver, CO, Oct. 5–8, SPE Paper No. SPE-84137-MS.
sure of 22  C and 30 psig were maintained during the experiment. [3] Paladino, E., Lima, J., Almeida, R., and Assmann, B., 2008, “Computational
Assuming a homogenous two-phase flow and ideal gas, the Modeling of the Three-Dimensional Flow in a Metallic Stator Progressing Cav-
ity Pump,” SPE Progressing Cavity Pump Conference, Houston, TX, Apr.
change in air density with pressure is given as 27–29, SPE Paper No. SPE-114110-MS.
[4] Zhou, D., and Yuan, H., 2008, “Design of Progressive Cavity Pump Wells,”
dp SPE Progressing Cavity Pumps Conference, Houston, TX, Apr. 27–29, SPE
dqa ¼ (B1) Paper No. SPE-113324-MS.
RT [5] Revard, J. M., 1995, The Progressing Cavity Pump Handbook, PennWell
Publishing Company, Tulsa, OK.
where R is the gas constant for dry air (287.05 J/kg K1), qa is the [6] Nelik, L., and Brennan, J., 2005, Progressing Cavity Pumps, Downhole Pumps,
dry air density in kg/m3, p is the pressure in Pa, and T is the tem- and Mudmotors, Gulf Publishing, Houston, TX.
perature in K. Pressure along the pump is linear and given as [7] Olivet, A., Gamboa, J., and Kenyery, F., 2002, “Experimental Study of
Two-Phase Pumping in a Progressive Cavity Pump Metal to Metal,” SPE
Annual Technical Conference and Exhibition, San Antonio, TX, Sept. 29–Oct.
Dppump 2, SPE Paper No. SPE-77730-MS.
px ¼ x þ pi (B2) [8] Bratu, C., 2005, “Progressing Cavity Pump (PCP) Behavior in Multiphase
kPs Conditions,” SPE Annual Technical Conference and Exhibition, Dallas, TX,
Oct. 9–12, SPE Paper No. SPE-95272-MS.
[9] Guise, G., Crotte, G., Lehman, M., Limoges, B., and Robert, B., 2016, “Field
where k represents the number of pump stages, and pi and px are Performance and Technology Update of All Metal Progressing Cavity Pumps
the pump intake pressure and the pressure at a distance x from the Deployed in Thermal Processes,” SPE Middle East Artificial Lift Conference
pump intake, respectively. Substituting Eq. (B2) into Eq. (B1) and and Exhibition, Manama, Bahrain, Nov. 30–Dec. 1, SPE Paper No. SPE-
integrating gives 184175-MS.
[10] Moineau, R., 1930, “A New Capsulism,” Ph.D. dissertation, The University of
Paris, Paris, France.
Dppump [11] Zhanga, J., Lia, W., Wub, Y., Zhanga, S., Nairb, M., and Lic, X., 2013, “A
qx ¼ 1:18  105 x þ qi (B3) Study on a Novel PCP’s Structure Using Finite Element Analysis,” SPE
kPs Progressing Cavity Pumps Conference, Calgary, AB, Canada, Aug. 25–27, SPE
Paper No. SPE-165645-MS.
[12] Lima, J., Paladino, E., Almeida, R., and Assmann, B., 2013, “A Computational
where qi and qx are the air density at the pump intake and at a Model for Analysis of Fluid-Structure Interaction Within Elastomeric Progress-
distance x from the pump intake. At experiment conditions, air ing Cavity Pumps,” SPE Progressing Cavity Pumps Conference, Calgary, AB,
density at the intake is 3.64 kg/m3. According to Eq. (B3), the Canada, Aug. 25–27, SPE Paper No. SPE-165650-MS.

121102-10 / Vol. 139, DECEMBER 2017 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 11/06/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 [13] Pessoa, P., Paladino, E., and Lima, J., 2009, “A Simplified Model for the Flow [23] Paladino, E., Lima, J., Pessoa, P., and Almeida, R., 2009, “Computational Three
in a Progressing Cavity Pump,” 20th International Congress of Mechanical Dimensional Simulation of the Flow Within Progressing Cavity Pumps,” 20th
Engineering, Gramado, Brazil, Nov. 15–20. International Congress of Mechanical Engineering, Gramado, Brazil, Nov.
[14] Karthikeshwaran, R., and Samad, A., 2011, “Leakage Analysis of Progressing 15–20.
Cavity Pump,” 11th Asian Conference on Fluid Machinery and Third Fluid [24] Paladino, E., Lima, J., Pessoa, P., and Almeida, R., 2011, “A Computational
Power Technology Exhibition, Chennai, India, Nov. 21–23. Model for the Flow Within Rigid Stator Progressing Cavity Pumps,” J. Pet. Sci.
[15] Gamboa, J., Olivet, A., Gonzalez, P., and Iglesias, J., 2003, “Understanding the Eng., 78(1), pp. 178–192.
Performance of a Progressive Cavity Pump With a Metallic Stator,” 20th Inter- [25] Lima, J., Paladino, E., Almeida, R., and Assman, F., 2009, “Mesh Generation
national Pump Users Symposium, Houston, TX, Mar. 17–20, pp. 19–31. for Numerical Simulation of Fluid-Structure Interaction Within Progressing
[16] Robello, S., and Saveth, K., 2006, “Optimal Design of Progressing Cavity Pump Cavity Pumps,” 20th International Congress of Mechanical Engineering,
(PCP),” ASME J. Energy Resour. Technol., 128(4), pp. 275–279. Gramado, Brazil, Nov. 15–20.
[17] Nguyen, T., Al-Safran, E., Saasen, A., and Nes, O., 2014, “Modeling the Design [26] Berton, M., Allain, O., Goulay, C., and Lemetayer, P., 2011, “Complex Fluid
and Performance of Progressing Cavity Pump Using 3-D Vector Approach,” Flow and Mechanical Modeling of Metal Progressing Cavity Pumps PCP,” SPE
J. Pet. Sci. Eng., 122, pp. 180–186. Heavy Oil Conference and Exhibition, Kuwait City, Kuwait, Dec. 12–14, SPE
[18] Martin, A., Kenyery, F., and Tremante, A., 1999, “Experimental Study of Two Paper No. SPE-150419-MS.
Phase Pumping in Progressive Cavity Pumps,” SPE Latin American and [27] Azevedo, V., Lima, J., and Paladino, E., 2016, “A 3D Transient Model for the
Caribbean Petroleum Engineering Conference, Caracas, Venezuela, Apr. Multiphase Flow in a Progressing-Cavity Pump,” SPE J., 21(4), pp. 1458–1469.
21–23, SPE Paper No. SPE-53967-MS. [28] Strasser, W., 2007, “CFD Investigation of Gear Pump Mixing Using Deform-
[19] Zhang, H., Li, J., Wang, W., Watters, C., and Houeto, F., 2012, “Simulating ing/Agglomerating Mesh,” ASME J. Fluids Eng., 129(4), pp. 476–484.
Progressive Cavity Pumps for Multiphase Flow and Production System [29] Nguyen, T., Tu, H., Al-Safran, E., and Saasen, A., 2016, “Simulation of
Design,” SPE Latin American and Caribbean Petroleum Engineering Confer- Single-Phase Liquid Flow in Progressing Cavity Pump,” J. Pet. Sci. Eng., 147,
ence, Mexico City, Mexico, Apr. 16–18, SPE Paper No. SPE-152841-MS. pp. 617–623.
[20] Belcher, I., 1991, “An Investigation Into the Operating Characteristics of the [30] Delpassand, S., 1999, “Stator Life of a Positive Displacement Downhole
Progressive Cavity Pump,” Ph.D. dissertation, Cranfield University, Cranfield, Drilling Motor,” ASME J. Energy Resour. Technol., 121(2), pp. 110–116.
UK. [31] Aql, A., 2016, “Dimensional and Empirical Modeling of Fluid Flow in Pro-
[21] Behzadi, A., Issa, R., and Rusche, H., 2004, “Modeling of Dispersed Bubble and gressing Cavity Pump,” M.Sc. thesis, Kuwait University, Kuwait City, Kuwait.
Droplet Flow at High Phase Fractions,” Chem. Eng. Sci., 59(4), pp. 759–770. [32] Arellano, J., 1997, “Field Comparison of Efficiency of Progressing Cavity
[22] Andrade, S., Valerio, J., and Carvalho, M., 2011, “Asymptotic Model of the 3D Pumps, Bean Units and Electric Submersible Pumps,” M.S. thesis, The Univer-
Flow in a Progressing-Cavity Pump,” SPE J., 16(2), pp. 451–462. sity of Tulsa, Tulsa, OK.

Journal of Fluids Engineering DECEMBER 2017, Vol. 139 / 121102-11

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 11/06/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use