Vous êtes sur la page 1sur 8

SPE 95272

SPE 95272 Progressing Cavity Pump (PCP) Behavior in Multiphase Conditions C. Bratu, SPE , PCM Pompes

Progressing Cavity Pump (PCP) Behavior in Multiphase Conditions

C. Bratu, SPE, PCM Pompes

Copyright 2005, Society of Petroleum Engineers

This paper was prepared for presentation at the 2005 SPE Annual Technical Conference and Exhibition held in Dallas, Texas, U.S.A., 9 – 12 October 2005.

This paper was selected for presentation by an SPE Program Committee following review of information contained in a proposal submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to a proposal of not more than 300 words; illustrations may not be copied. The proposal must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract In operation, the Progressing Cavity Pumps (PCP) boosting gaseous multiphase mixture faces reliability issues that can be summarized as follows: the pump pressure is developed by the stages nearest the discharge, then the high pressure gradient causes heat to build up which commonly results in premature failure of the elastomeric stator. This paper first describes our recent testing program performed with an industrial PCP in multiphase conditions, and then presents a new analytical model that describes the thermo-hydraulic-mechanic processes which occur when attempting to pump multiphase mixtures. The objectives of our program are to: (1) examine the PCP system operating at high gas volume fraction, (2) analyze the pumping performance (delivered pressure, flow rate, and gas void fraction) and, (3) describe the overheating generated when the gas is compressed. This paper also describes our PCP experimental program and presents resulting experimental datas which clearly show the correlation between the pressure distribution and the heat along the pump. We formally explain this phenomenon by proposing a new analytical model that evaluates the stator degradation and failure risk. This paper should help operators and manufacturers to design pumps with better system’s Mean Time Between Failure (MTBF) performance.

Introduction The PCP is a positive displacement pump type that can be used to pump a wide range of multiphase mixtures, including high viscosity fluids with entrained gas and solid particles in suspension. However, PCP has a reduced ability to handle high gas-liquid ratio due to limitations of its elastomeric stator material required to overcome thermo and mechanical effects. This paper studies PCP’s behavior in multiphase flow conditions with high Gas Volume Fraction (GVF) (i.e., a GVF from 0 to 90%) and liquid (water and viscous oil). Results from our experimental testing program carried out with

industrial pumps, are presented in this paper. The objectives of this experimental program was to: (1) examine the PCP system operating in multiphase flow conditions, (2) analyze the pumping performance (delivered pressure, flow rate and gas void fraction), (3) describe the thermo-hydro-mechanical process and overheating generated when the gas is compressed and finally deduce the correlation between the pressure distribution and developed temperature Based on the multiphase fluid equations (momentum and mass conservation equations, fluid state function and gas thermodynamic laws), we propose a new analytical model of pressure-temperature distribution and compare it with data obtained from the experimental testing program. We find that the pump’s performance can be predicted using our deduced formulas and that they help to predict more accurately the risk of stator degradation and failure. Also, based on these results, the pump’s design parameters can be better selected to comply with production conditions and increase the pump’s performance as well as the pump’s reliability. Previous studies dealt with PCP performance in liquid and multiphase mixture. The inventor of this pump system, Moineau [1] have proposed to use Hagen-Poiseuille equation for modeling internal slip flow and deduce the delivered pressure in liquid flow. Recent studies (Vetter [2], Robello and Saveth [3], and Gamboa [4]) introduce others effects such as rotor rotation, liquid viscosity, rotor-stator gap variation and mechanical properties of stator material (elastomer and metal). Experimental work and analytical approache have shown that multiphase boosting with positive displacement pumps ( PCP and twin – screw ) is strongly affected by gas handling (Vetter [2], Wietstock [5], Toma [6], Prang [7]). Extensive experimental studies (Kenyery et al., [8] and [9]) dedicated to PCP’s with two-phase flow show the pressure and temperature distribution along the pump stator under different sets of variables (gas void fraction, rotational speed and delivered pressure). The objectives of these experiments were to estimate how flow conditions can be affecting the pump’s life and its lift performance.

Behavior of PCP in Multiphase Flow It is common knowledge that PCP comprises a rigid metallic rotor and a stator made of flexible material such as elastomer. Between the rotor and stator there is a compressive fit, which results in sealed cavities (or stages) of constant volume. As the external pressure increases, the pump delivered pressure also must increase, which is obtained by adding a number of stages.

2

SPE 95272

Typical pressure distribution along the pump is the result of slippage across the seal lines between the cavities, from the pump discharge (high pressure) to the inlet section (low pressure). Therefore, when there is pressure in a cavity to permit the slippage into the second cavity, the pressure is partially transmitted, due to pressure loss. Hence, the process is limited to a reduced number of cavities and the entire differential pressure is developed by the stages closer to the discharge end of the pump. Multiphase fluids cause special problems for the PCP due to the compressibility of the gas phase. The volume of the multiphase mixture, which enters the cavity, is determined by the inlet pressure. Therefore, due to the increasing internal pressure towards the discharge end, the gas volume does compress. Close to the discharge end of the pump, the slippage flow rate compensates the gas compressed volume and the pressure is transmitted between the cavities. However, typical PCP has low slippage which can only compensate the compressed gas volume of few discharge cavities (stages). Consequently, tests have shown that a disproportionate amount of the pressure is developed by the discharge stages, and the entire delivered pressure gets concentrated on only the last stages of the pump. This excessive pressure distribution in the discharge stages causes significant thermo – mechanical problems which commonly results in premature pump failure. This thermo-hydraulic process is a result of well known gas laws, which state that as the pressure increases the gas volume decreases and the temperature increases. Since the cavity volume is constant, the gas volume cannot decrease and the high pressure gradient causes excessive heat buildup. Due to the temperature behavior of stator elastomeric materials, the thermo- hydraulic process limits the PCP’s ability to pump gaseous fluids. Another thermo-mechanical process occurs as a result of disproportionate pressure distribution. As the differential pressure between two contiguous cavities increases, the stator flexing material is deformed and inside the low pressure cavity the stator is strained. This process is particularly apparent in liquid or multiphase flow, when the cavities differential pressure is relatively large. Then, the compression stress of the rotor on the strained stator increases, and the rotor friction (or viscous torque) increases the failure risk. Thus, the increase of friction temperature and viscous torque substantially decrease the pump’s reliability. Moreover, the deformed stator will create high local velocity fluid jets that could potentially cut the stator material. Accordingly, in multiphase flow conditions the excessive pressure distribution in the discharge stages causes both thermo-hydraulic and thermo-mechanical physical processes, which reduce the pump reliability and performance. Furthermore, the PCP reliability is related to the pressure distribution which depends on the production conditions (i.e., gas flow rate, delivered pressure, pump rotational velocity). In order to comply with the field production requirements, the PCP design is a compromise between the reliability (MTBF parameters) and the pump’s hydraulic performance. This study introduces a new approach of the relationship between the pressure distribution and the reliability parameters in multiphase and liquid flow.

The thermo-hydraulic process. The slip flow rate (S)

compensates the compressed gas volume and therefore transmits the pressure (P) to the last stages (Fig. 1). Moving towards to discharge (d, from section 1 to 2), increasing the cavity’s pressure P, and letting heat to build-up. Consequently, the thermodynamic laws for polytropic process (Eq. 1) associates temperature gradient (T) to the pressure gradient

(P) and gas compressed density (

ρ

G

) :

T

T

2

1

= ⎜

2

1

P

P

k

1

k P

2

⎛ρ

= ⎜

G2

P 1

ρ G1

k

(Eq.

1)

where, k is the polytropic coeficient. In multiphase flow, the polytropic coefficient k and the mixture density ( ρ ) are both functions of gas void fraction ( α ):

ρ =αρ×

G

+−α ×ρ

G

(1

)

ρ are respectively the gas and the liquid

densities. The pressure-temperature relation (Eq. 1) along the pump stages will be compared with the measured experimental results.

where,

ρ

G

and

L

P2 P1 T1 T2 1 2 d S
P2
P1
T1
T2
1
2
d
S

Fig.1 PCP pressure-temperature relationship due to gas compression

process. Another effect of

disproportionate pressure gradient is related to the rotor and the stator frictional-viscous forces. We present a simple approach to analyze the stator reaction to differential pressure between the cavities 1 and 2 ( Fig. 1). It is assumed that a plane element of the stator is compressed by the differential pressure (P) between the cavities 1 and 2 ( Fig.2 A). Based on Boussines equation the stator deformation (Y) is:

The

thermo-mechanical

Yx() = P ×φ()x

with

φ

( x

) Pa

2

b

ax

e

×

(cos

ax

+ sin

ax

)

a

4 b

=

4EI

and

p = b Yx() = P bφ ()x

(Eq. 2)

where, a and b are material elastic characteristics and p is the induced stress. Therefore, the stator compression is the maximum in the cavity 2, while the surface stator is strained inside the cavity 1 (Fig. 2 A). Also, the stress developed by the material is a compression (p+) inside the cavity 2 and the surface of the cavity 1 is under tension (p-) .

SPE 95272

3

of slippage multiphase flow through rotor-stator contact. For compressible flow, the slippage flow compensates the gas compressed volume and transmits the pressure inside the cavities. Therefore, an iterative process is described: pressure distribution is related to slippage flow, but the slippage flow

rate compensates the cavities gas compressed volume which is function of pressure distribution. To find the solution, an analytical procedure is described and practical simplifications are used.

Finally, a simple formula gives the pressure distribution along

P2 A P1 P Cavities 1 2 P p(-) Stator p(+) Y Fn B Rotor
P2
A
P1
P
Cavities
1
2
P
p(-)
Stator
p(+)
Y
Fn
B
Rotor
p(-)

x the pump in multiphase flow

P

x

P

d

L

=

1

F

(,

α

x

)

(Eq. 4)

where Px is the section x pressure, Pd is the pressure at the discharge section, α = GVF is the inlet gas void fraction and F is a polynomial function of α = GVF and x/L coordinate (Fig. 3). The value of polynomial F is easy to calculate. The new approach for modeling PCP’s internal slip in multiphase flow shows that pump pressure distribution can be analyzed with practical formulae. Therefore, it is not necessary to use complex numerical method. It is interesting to notice that the simplicity of this result is coherent with thermo- hydraulic - mechanic description. As an example , the Fig .3 shows the non dimensional pressure distribution ( Px / Pd ) along the pump , for α = GVF= 0,9 ; the analytical pressure distribution ( Eq. 4) is compared with the experimental data.

1,2 1 0,8 0,6 0,4 0,2 0 1 0,9 0,8 0,7 0,6 0,5 0,4 0,3
1,2
1
0,8
0,6
0,4
0,2
0
1 0,9
0,8
0,7 0,6 0,5
0,4 0,3 0,2 0,1
0
P x / P d

X / L

LIQUID GVF = 0.9 Calculation GVF = 0.9 Experimental

Fig.2 .Stator deformation due to pressure gradient inside the cavities (A) and rotor-strained stator contact (B).

Such strained stator surface increases the stress due to the rotor movement and the failure risk occurs ( Fig. 2B). And yet , Bowden’s approach shows that thermal energy due to the viscous friction between the rotor and stator depends on the normal force Fn, friction coefficient f and rotor velocity N. The normal force Fn is the reaction to stator deformation stress p (Fig. 2B). Following the equation 2, the pression gradient P between the cavities is the cause of stator stress p. Therefore, the viscous friction temperature T is a function of the mechanical parameters:

T PfN×

×

(Eq. 3)

The result is that the temperature T increases with the pressure gradient P, friction coefficient f and pump velocity N. Then, the viscous-friction temperature T is a measure of the force (torque) developed by the rotor in order to run over the strained stator, and in this sense T is a measure of failure risk and reliability. As for the friction coefficient f, it is equivalent of Newton’s dynamic viscosity. The thermo-mechanical process (See Eq. 3) is illustrated by experimental data.

Finally, the pressure distribution in multiphase flow which shows disproportionate pressure gradient near the discharge stage is the cause of both processes leading to temperature build up, i.e. gas fast compression (thermo-hydraulic) and rotor-stator forced contact (thermo-mechanic).

The pressure distribution in multiphase flow. We have

seen that the pressure distribution is the key for the evaluation of pump behavior in multiphase flow (see Appendix). Thus, classical momentum and mass conservation are applied to multiphase compressible flow. The pressure distribution along the pump is determined by the pressure drop

i d x L x o
i
d
x
L
x
o

Fig.3. Multiphase pressure distribution; analytical (Eq.4) and experimental for α = GVF=0.9, N= 300 rpm, P= 40 bar.

According to experimental trend, the pressure distribution in liquid flow is linear, which confirms the model aproach. In multiphase flow, with a gas content of α=GVF= 0.9 the pressure distribution curve shows a steeply gradient, roughly 4 time as big as the liquid gradient. Again, using the fact that

4

SPE 95272

pressure distribution gradient is the cause of temperature build-up in both processes, i.e. gas compression (thermo- hydraulic) and rotor-stator viscous friction (thermo-mechanic), it is clear that multiphase PCP behavior is due to pressure distribution.

The Experimental Program.

An extensive test program was performed at the PCM’s facility test rig and CREMHyG- Turbomachinery and Cavitation Research and Test Laboratory in Grenoble, France. Main objective of the experimental program was to determine the PCP’s ability to handle multiphase fluids with gas fraction up to 90%. Additionally, rotors and stators combinations were evaluated in liquid and multiphase flow. Tests in liquid and multiphase flow allow validating the physical processes and parameters related to stator reliability and pump performance. The configuration of the CREMHyG test skid is shown in the Fig. 4. Since the pressure and temperature distributions are significant factors of PCP behavior in multiphase flow, measurements were carried out at multiple locations and different intervals of time. The presented temperature values were measured during 20 minutes with multiphase flow stable conditions. In actual operations, long term pump response to thermal increase largely depends on initial temperature gradient (system invariant) and surrounding factors. Flow meters in gas and liquid lines were used to monitor the gas fraction. Additionally, pump speed and delivered pressure were also recorded. The range of measured parameters is on the table 1.

The range of measured parameters is on the table 1. Fig. 4 .The PCP test rig

Fig. 4 .The PCP test rig and the measurement installation

The pressure distribution along the pump was measured in liquid flow (GVF= 0) and multiphase conditions for various gas content (GVF), speed and discharge pressure.

Liquid flow. For example, Fig.5 shows the pressure distribution in liquid flow when the discharge pressure is P= 40 bar and 0-6 bar; Liquid pressure curve and related slippage are the result of rotor-stator interference fit, particularly on the inlet sector. Indeed, when the discharge pressure is low (0 – 6 bar) the pressure increases through inlet stages (1-7) up to P =20 bar.

That is due to strong interference fit which seals the cavities and a very small variation ( 0,2 %) of uncompressible liquid volume causes high pressure variation .On the other hand ,the slippage is located on the discharge stages ( P= 40 bar ).

Pump characteristics

Pum Name: Pump PCM Oilfield 100 TP 600

Flow rate = 108 m 3 /d

(680 bpd)

Head = 600 m (2000 ft); N= 500 rpm

Experimental conditions

Liquid: water, oil and Gas: air Flow rate (m3/d) = 10 to 60, Pressure = 0 to 40 bar GVF = 0 to 0.9, Oil viscosity (cPo) = 1200 (20°C) 600 (30°C)

Measured parameters

Pressure

= 21 sensors, Temperature = 10 sensors,

Flow rate: liquid, gas. Pump velocity (rpm)

Table 1: Test conditions

40 35 30 25 20 15 10 5 0 1 2 3 4 5 6
40
35
30
25
20
15
10
5
0
1
2
3
4
5
6
7 8
9
10
11
12
13
Stages
P= 40 bar
P = 0 - 6 bar
Pressure bar

Fig. 5 Pressure distribution in liquid flow, N= 300 rpm

The temperature developed in liquid flow (oil) is presented in Figures 6 and 7. Measured fluid temperature, at different pump velocities (Fig. 6: N= 300 and 100 rpm, P= 40 bar) confirms that the temperature increases with the velocity (see Eq. 3). However, the mean temperature ratio is half of the velocity ratio, due to viscous coefficient (f) variation with Re number of slippage liquid film. Taking into account both effects, velocities and viscous-friction coefficient, the experimental results (see Fig. 6) confirm the analytical formula (Eq. 3). Developed temperature, when the discharge pressure is P = 40 bar and 0 to 6 bar is shown in the Fig. 7. Using specific notations, the parameters (see Eq. 3) become marked: suffix a, for P = 40 bar and suffix b, for P= 0 to 6 bar. Then, the ratio corresponding to the Eq. 3 becomes:

T

a

T

b

P

a

P

b

(Eq. 5)

with both identical friction coefficients fa = fb, and velocities Na = Nb = 300 rpm.

SPE 95272

5

According to Figures 5 and 7 the ratio of mean pressure gradient and temperature are (Eq. 3 and 5):

P

a

P

b

= 1.65

T

a

T

b

= 1.54

This confirms the relationship between the pressure gradient and friction-viscous temperature (Eq.3). Moreover, temperature curves symmetry (see Fig. 7) is similar to pressure trend (see Fig. 5). Therefore, even in liquid flow the pressure gradient is the primary cause of friction-viscous torque and related temperature. The temperature is then a measure of viscous torque, reflecting the stator reliability.

12 10 8 6 4 2 0 1 2 3 4 5 6 7 8
12
10
8
6
4
2
0
1
2 3
4 5
6
7 8
9 10
11
12 13
Stages
T
N= 300 rpm
T mean
t
N= 100 rpm
t
mean
Temperature °C

Fig. 6. Temperature in liquid flow (oil) for various pump velocities (N= 300, 100 rpm), P= 40 bar.

12 10 8 6 4 2 0 1 2 3 4 5 6 7 8
12
10
8
6
4
2
0
1
2 3
4 5
6
7 8
9
10
11
12 13
Stages
P= 40 bar
P= 0-6 bar
P= 40 bar mean
P= 0-6 bar mean
Temperature °C

Fig. 7. Temperature in liquid flow (oil) for discharge pressure P = 40 bar and 0 to 6 bar, N = 300 rpm.

Multiphase flow. According to thermal process described in the previous section, the pressure distribution in multiphase flow is the main cause of temperature increase. The slippage flow rate compensates the gas volume of few compressed cavities and, therefore, the pressure distribution is limited to discharge stages (see Fig. 3). Indeed, the lack of slippage flow rate results in steepness pressure distribution,

even when the gas content is low (see Fig. 8). However, as the analytical approach shows (see Appendix), the pressure

distribution depends on both flow rate (or N), and GVF. Then,

when the flow rate is high (N= 300 rpm) compressed large gas

volume requires slippage compensation and the pressure is limited to discharge stages 10 to 13, and vice-versa , for low flow rate (N= 100 rpm) the pressure is distributed to stages 8 to 13 (See Fig. 8).

45 40 35 30 25 20 15 10 5 0 8 9 10 11 12
45
40
35
30
25
20
15
10
5
0
8
9
10
11
12
13
Stages
GVF=0.1 N=3OO rpm
GVF=0.9 N=300 rpm
GVF=0.1 N=100 rpm
GVF=0.9 N=100 rpm
Pressure bar

Fig. 8 .Discharge stages pressure distribution in multiphase flow .Experimental data for P= 40 bar, N= 300 and 100 rpm , GVF= α = 0,1 and 0,9.

Our new approach used for modeling the pressure distribution in multiphase flow (see Apendix) is compared with experimental data (see Fig.3) and demonstrates that pressure gradient can be predicted applying simple formula (Eq. 4). Based on previous pressure gradient analysis one can correlate with temperature experimental data. Figure 9 shows the temperature along the pump for gas content GVF= 0.5 and 0.9, P= 40 bar and fixed flow rate (N= 300 rpm). As the pressure

distribution is concentrated on discharge stages 10-13 (Figs. 3, and 8), the temperature curves reflect the:

̇ Thermo-hydraulic process on discharge stages 10 to 13; due to the pressur gradient the temperature is the result of gas polytropic compression ( Eq. 1 )

̇ Thermo-mechanic process on the inlet stages 1 to 9, where the temperature is due to the viscous friction multiphase lubrication (Eq. 3). Therefore, the relationship between the pressure gradient and the temperature increase due to polytropic compression (see

Eq. 1) on the discharge stages 10 to 13 and pressure variation of 40 bar, becomes (Fig. 9):

- For GVF = 0.9, the temperature variation (Eq. 1) is:

0,286

P

13

=

T

13

P

10

T

10

= 2.9

which corresponds to measured temperatures on the discharge stages: GVF=0.9 , T13 / T10 = 20 °C / 7°C .

6

SPE 95272

13

- For GVF= 0.5 the temperature variation (Eq. 1) becomes:

0,25

P

13

=

T

13

P

10

T

10

= 2.5

wich corresponds to measured temperature on the discharge stages 10 to 13: GVF=0.5 , T13 / T10 = 10 °C / 4 °C.

In multiphase flow, taking into account the liquid capacity to absorb heat flow the exponent (Eq.1) is a function of GVF. Nevertheless, when the gas content is high enough the liquid heat dissipation is not significant and the gas polytropic coefficient can be used, without any multiphase mixture consideration. Accordingly, this confirms the previous approach of thermo- hydraulic process i.e. the polytropic gas compression and temperature increase are related to pressure gradient.

20 18 16 14 12 10 8 6 4 2 0 1 3 5 7
20
18
16
14
12
10
8
6
4
2
0
1
3
5
7
911
13
Stages
GVF = 0,9
GVF = 0,5
Temperature °C

Fig. 9. Temperature in multiphase flow. GVF= α = 0.9 and = 0.5, P= 40 bar, N= 300 rpm.

The temperature due to viscous friction in multiphase flow, with GVF = 0.9 and 0.5, is illustrated by the inlet stages 1 to 9 behavior (See Fig.9). Thus, according to previous thermo- mechanical process (see Eq.3), for constant velocity N = 300 rpm, temperatures ratio is ( Eq.6 ):

TPf

c

cc

TPf

d

dd

(Eq.6)

where suffixes are: c, for GVF= 0.9 and d, for GVF= 0.5. Assuming that the rotor centrifugal acceleration results in a liquid film (with some gas bubbles), the interference fit is oil lubricated and the friction coefficients are equivalent. As for the pressure gradient , the experimental data show that for GVF = 0.9 the pressure is distributed on 3 stages (see Fig. 8) , whereas for GVF= 0.5 the pressure is on 4 stages . Then, the pressure gradient ratio is 4/3 = 1.33, and the temperature ratio corresponds to the measured ones (Fig. 9), i.e. Tc Td = 8 °C 6 °C = 1.33 .

measured ones (Fig. 9), i.e. Tc Td = 8 °C 6 °C = 1.33 . Therefore,

Therefore, the previous approach of thermo- mechanical process is confirmed, i.e. the pressure gradient is the cause of viscous friction torque, whereas the temperature is the measure of this mechanical interaction. Furthermore, observed thermal behavior depends on pump flow rate; Fig. 10 shows the temperature when flow rates vary (N= 100 and 300 rpm), for P = 40 bar and GVF= 0.9. Indeed, when gas flow rate is low (N= 100 rpm) the slippage overcomes compressed volume and that limits the temperature increase and improves the pressure distribution (stages 9 to 13). As we pointed out above, throughout the discharge stages the gas is compressed and the polytropic approach is still confirmed in both cases, high (N= 300 rpm) and low flow rates (N= 100 rpm). Then, the gas content is GVF= 0.9 and the temperature variation (between the stages 9-10 and 13)

is

temperature ratio is due to viscous friction process (Eqs. 3,7):

ratio is due to viscous friction process (Eqs. 3,7): TT 13 9 − 10 3 ≈

TT

13

9

10

3

. However, through the inlet stages the

T

fPN × ×

fPN

T

hhh

h

eee

e

(Eq.7)

where suffixes are: e for N= 300 rpm, and h for N= 100 rpm. At high gas flow rate the pressure is limited to discharge stages (10-13) and as the gas decreases the pressure distribution is extended, then the pressure gradient ratio is Pe/Ph = 1.6. Rotating velocities ratio is Ne/Nh = 3. As for viscous friction coefficient ratio, fluid film lubrication effect (Re number and oil- gas entrained) can be approximated and fe/fh =1/2. Therefore the temperature ratio is Te/Th = 2.4, which corresponds to measured mean values on the inlet stages (Fig. 10).

20 18 16 14 12 10 8 6 4 2 0 11 1 3 5
20
18
16
14
12
10
8
6
4
2
0
11
1
3
5
7
9
Stages
N = 300 rpm
N = 100 rpm
Temperature °C

Fig. 10. Temperature in multiphase flow for various flows rates, N = 300 and 100 rpm, P= 40 bar, GVF = 0.9.

Conclusions:

1. A new approach of the PCP’s behavior and reliability in two-phase flow is described:

̇ Thermo-hydraulic process: where the gas compression, due to pressure gradient, causes temperature to increase, ̇ Thermo-mechanic process: where the pressure gradient between the cavities induces strained stator

SPE 95272

7

effect that increases frictional (viscous) rotor-stator torque and related temperature,

̇ Pressure distribution in two-phase flow that depends on gas content (flow rate, GVF) and the compensation of compressed volume due to the slippage flow.

2. Experimental tests carried out on PCP’s confirm the

physical description; calculations correlate well with measured

values of pressure distribution and related temperature increase.

3. The pressure distribution is the cause of temperature

increase in both two-phase and liquid flows. Moreover, the temperature is a measure of gas compression effect and rotor- stator frictional (viscous) torque. Accordingly, that is equivalent of a criterion of PCP’s reliability and failure risk.

4. The temperature due to rotor-stator frictional (viscous)

torque is a function of pressure gradient, rotational velocity and frictional coefficient .The temperature due to gas compression depends on pressure distribution and gas flow rate (GVF, rotational velocity). In order to improve the PCP’s operational life, usually one reduces the rotational velocity, delivered pressure and rotor- stator fit. Therefore, the analytical formulas enable to optimize the pump performance in both liquid and two-phase flows.

Acknowledgements:

The author wishes to acknowledge the management of PCM Pumps for their permission to publish this work and the permanent support of the research effort. Also, I would like to thank those individuals at CREMHyG and PCM testing rig, who have been involved in the experimental program.

Nomenclature:

P: pressure, differential pressure, T: temperature,

GVF = α

ρ : densities (mixture, gas ( G ) , liquid ( L )), k: gas polytropic coefficient, Y: stator deformation, p: stator stress, f : rotor-stator friction coefficient, N: pump rotational velocity, W: cavity volume, W: gas compressed volume of the cavity, S: total slippage flow rate, q: slippage flow rate, compensating gas compressed volume W, h: pressure drop of slippage flow, i: inlet section, d: discharge section.

gas void fraction,

References :

1. Moineau René (1930), A New Capsulism, Doctoral Thesis, The University of Paris.

2. Vetter G and Wincek M (1993) Performance Prediction of

Twin-Screw Pumps for Two-Phase Gas/Liquid Flow,

Pumping Machinery, 154, 331-340

3. Robello G and Saveth K. (1998) Progressing Cavity Pump

(PCP): New Performance Equations for Optimal Design.

SPE 39786.SPE Permian Basin Oil and gas Recovery Conference, Midland.Texas.U.S.

4. Gamboa J, Olivet A and Espin S. (2003): New Approach for

Modeling Progressive Cavity Pumps Performance, SPE 84137, SPE Annual Technical Conference and Exhibition, Denver, Colorado, USA.

5. Wietstock P and White W. (1993): Studies in Multiphase

Pumping with Screw Pumps. 6-th International Conference on Multiphase Production, pp. 389-414. 6. Toma P. et al. (1998): Field assessment of Multiphase Progressing Cavity Pumping, SPE Progressing Cavity Pump Workshop, Tulsa U.S. 7. Prang J. (1991): Rotary Screw Pumps for Multiphase Products, British Pump Manufacturers Association, 12th international Pump Technical Conference, London U.K. 8. Martin A., Kenyery F. and Tremante A. (1999):

Experimental Study of Two Phase Pumping in progressive cavity pumps, SPE 53967. SPE Latin American and Cararibbean Petroleum Engineering Conference, Caracas.

9. Olivet A, Gamboa J and Kenyery F. (2002): Experimental

Study of Two-Phase Pumping in a Progressing Cavity Pump Metal to Metal, SPE 77730, SPE Annual Technical Conference and Exhibition. San Antonio, USA.

SI Metric Conversion Factors

1 inch ( in ) = 2.54 cm

= 0.305 m

1 foot ( ft )

1 barrel (B) = 0.159

1 bar

1 psig

( 5/9) ( °F-32) = °C

m

10

5

=

=

Pa

6.89 K Pa

3

Appendix: Multiphase Pressure Distribution

We propose a schematic hydraulic flow to understand the

pressure distribution. The multiphase mixture of the cavity

(W) moves from inlet section ( i ) toward the discharge (d),

while the leakage (S) through the rotor-stator contact is slipping from the discharge high pressure Pd (Fig.11). Due to

pressure P variation, the compressed gas volume W is compensated by a part q of overall leakage flow S. Therefore, the pressure distribution P is a function of leakage flow S curve, which depends of gas compression volume W. Since W is a function of gas flow rate and pressure P, there is an iterative process. Fluid mixture momentum and mass conservation equations describe the slippage multiphase compressible flow. Assuming that the kinetic energy can be neglected and the slippage pressure drop controls the pump pressure distribution, there results the nonlinear integral equation 8:

(

1-

)

αρ

L

dP

+ ρ

G

bα

∫∫ 0

P dP = C

.

x

S

2

(

x P

,

,

α

). dx

(Eq. 8)

8

SPE 95272

where b and C are: system constants (inlet pressure, pressure drop coefficient, hydraulic diameter and flow section). This equation (8) shows that the pressure distribution P (left side) is primarily due to pressure drop of slippage flow (right side), which S flow rate varies while the compressed gas is compensated. Therefore, it is necessary to solve the relationship between the gas compressed – compensated volume (W) and local slippage flow rate (S).

i

S

q d Wo x L x Wx Pd Pi Px P Wi ∆Wx W Sd
q
d
Wo
x
L
x
Wx
Pd
Pi
Px
P
Wi
∆Wx
W
Sd
S
Sx
qd
q
Fig. 11.
qx
Multiphase pressure distribution: compressed-
compensated volume and slippage flow rate.

Mass conservation of pump flow rate, including compressible gas volume, gives the volume Wx that have to be compensated in the section x:

W

x

α

W

i

P ×− 1 i ⎜ ⎛ ⎟ ⎞ P ⎝ ⎠ x
P
×−
1
i
⎜ ⎛
⎟ ⎞
P
x

(Eq. 9)

Assuming linear approximation of pressure distribution, the compressed volume is a part of initial gas volume (αWi). Therefore, the slippage flow rate qx that compensates Wx in the section x, becomes:

q

x

=

W x

⎝ ⎜

x

L

=

α

×

1

qW d i

(Eq. 10)

Consider the total slippage flow rate consumed between the dischage d and the section x:

S

cx

=

q

x

.

dx

=

α

×

q

d

×

⎝ ⎜

1

x

2 L

⎠ ⎟

×

x

(Eq. 11)

The slippage flow in the section x is:

SSS=

x

dcx

(Eq. 12)

So, the flow through the section x is the difference between the discharge initial slippage flow rate Sd and consumed one S cx to compensate the compressed gas volume. Actually, in

multiphase flow the entire slippage flow Sd compensates few cavities Sd = Sc, and the flow rate has the form:

S (x)

x

=

S

d

×− 1

2

α⎛

⎜ ⎝

x

L

2L

1

x

(Eq.

13)

Now, one can turn to pressure drop expression hx (Eq. 8, right side) from the dischage d to x section, and the integral is:

h

x

CS

d

2

×

F

α

,

x

L

(Eq. 14)

where, F is a polynomial function of: α, x and L. For liquid flow (α=0), the pressure drop function is linear

(i.e., F = x /L), and the overall liquid pressure drop is:

h

0,L

=

C ×S

d

2 , x = L

Then, the pressure drop in multiphase flow becomes:

h

x

=

h

0,L

×α F,

x

L

⎟ ⎠

(Eq. 15)

It is now easy matter to get the pressure distribution (left side of the equation) and the integral becomes:

(

1

− α ρ

)(

P

Ld

P

x

)

+ αρ

G

1

2P

i

(

P

2

d

P

x

2

)

=

h

x

(Eq. 16)

The pressure distribution of PCP in multiphase flow is then given by simplified analytical solution:

P

x

P

d

,

L

=− 1F

x

α

(Eq. 17)

where, F denotes the pressure drop function depending of gas flow rate (α) and the coordinate along the pump (x /L). This result is still valid for liquid flow (α = 0 and F = x / L) that gives a linear pressure distribution:

P

x

=

1

x

P

d

L

(Eq. 18)

The new approach for modeling PCP’s internal slip in multiphase flow shows that pump pressure distribution can be analyzed with practical formula. Therefore it is not necessary to use complex numerical method.

It is interesting to notice that the simplicity of this result is

coherent with thermo-hydraulic-mechanic description. In order to illustrate the pressure distribution along the pump, in multiphase flow, the calculated non-dimensional pressure ratio (see Eqs.17,4) is compared with experimental data (See Fig. 3). The discharge stages are compensated by the slippage flow and pressure distribution is limited to 1/3 of the overall

stages.