PO191382 DOI: 10.

2118/191382-PA Date: 23-April-19 Stage: Page: 373 Total Pages: 12

Optimizing Jet-Pump Production

in the Presence of Gas
Andrew L. Kurkjian, GR Energy Services

Summary
A hydraulic jet pump with both gas and liquid phases at the intake is modeled analytically. A complete system model results when this
is combined with models of the surface pump, tubing, casing, and the pressure/temperature behavior of oil and gas. This model is used
to size jet pumps before installation and to optimize production. Production optimization includes first estimating the pump-intake pres-

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023
sure and then determining the drive frequency for the surface pump that minimizes the pump-intake pressure. There is an optimal fre-
quency at which the intake pressure is minimized and production is maximized.

Introduction
Hydraulic jet pumps are being used increasingly in horizontal wells drilled into unconventional (i.e., shale) oil reservoirs. This is pri-
marily because hydraulic jet pumps are highly reliable and can be resized quickly without a rig. By comparison, other methods of artifi-
cial lift in the same wells have reduced run times attributed to either premature failure or the need to resize at a significant cost. In this
paper, we focus on another advantage of hydraulic jet pumps: their tolerance to gas.
Unlike other artificial-lift methods, the modeling of hydraulic jet pumps in the presence of both gas and liquid at the intake is tracta-
ble analytically. This fact is exploited to enable the prediction and control of downhole conditions from the surface.
In this paper, we focus on how gas affects a downhole jet pump and apply this to the sizing of the jet pump and then to the optimiza-
tion of production. In this latter case, when only “daily production” values of oil, water, and gas are available, the methodology is lim-
ited to days when the production is stable.

Model Development
Cunningham (1995) developed a model for a hydraulic jet pump where there are both gas and liquid phases at the pump intake. This
same model was developed independently by Noronha et al. (1998). Cunningham had previously treated a liquid-only case at the intake
(Cunningham 1957) and a gas-only case at the intake (Cunningham 1974). In Cunningham (1974), a jet pump is used as a gas compres-
sor, and this is the extreme case in which there is only gas at the intake. These cases are all included in Cunningham (2008). The model
for two-phase flow at the intake supersedes the earlier model of Petrie et al. (1984, 1987).
Missing from the Cunningham (1995) model are the behaviors of oil and gas as functions of temperature and pressure, specifically
the calculation of oil/gas rates and pressure gradients at intake conditions. Also missing is the behavior of the surface pump that powers
the downhole jet pump. In addition, a model is required for the tubing connecting the two pumps and the casing annulus that connects
the jet-pump discharge to the surface facilities. Fig. 1 shows a simplified diagram of a surface centrifugal pump system and wellhead.
Fig. 2 shows a simplified diagram of a jet pump, indicating the positions within the pump for modeling purposes. (Not shown is the
standing valve below the pump.)

Power fluid
(water)

Power fluid (water)

Produced oil,
VFD Thrust
Motor Centrifugal pump water, and gas
chamber

Fig. 1—A centrifugal surface-pump system comprises a variable-frequency drive (VFD) providing electrical power to a three-phase
induction motor that turns the shaft within a centrifugal pump. The thrust chamber prevents axial forces on the motor. Power fluid
(water) is pumped down the tubing, and produced fluid (including the power fluid) returns up the casing. The power-fluid flow rate
and pressure are measured at the discharge of the pump, as is the pump-intake pressure and motor vibration. All measurements
are made in real time and monitored remotely.

Copyright V
C 2019 Society of Petroleum Engineers

Original SPE manuscript received for review 28 July 2017. Revised manuscript received for review 20 December 2017. Paper (SPE 191382) peer approved 22 May 2018.

May 2019 SPE Production & Operations 373

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030

 PO191382 DOI: 10.2118/191382-PA Date: 23-April-19 Stage: Page: 374 Total Pages: 12

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023
4

Fig. 2—A diagram of a jet pump showing the jet exit (Position 0), jet entrance (Position 1), pump intake (Position 2), throat exit (Posi-
tion 3), and discharge (Position 4). Note that the power fluid and the reservoir fluid mix in the throat and exit together to the surface.

The model assumes steady-state flow such that derivatives with respect to time are zero. This assumption applies because reservoirs
are slow to respond to changes. However, a gas slug that emerges from the horizontal section would result in a sudden upset of the
system. There would be a transient response during which the horizontal jet pump stabilizes. The pump response during this period is
not modeled in this paper.

Surface Pump. In the case of a centrifugal pump on the surface, the pressure created is a function of the flow rate. It is characterized
by six coefficients, the number of stages in the pump, the pressure gradient of the power fluid (G1), the pump-suction pressure, and the
drive frequency. At 60-Hz operation, the relationship between discharge pressure and flow rate is given by
X
6
Ppf ¼ Ps þ Ns G1 hi Qi1
1 : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð1Þ
i¼1

The affinity laws allow the calculation of the power-fluid pressure and flow rate as a function of frequency using
 
f
Q1 ð f Þ ¼ Q1 ð60Þ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð2Þ
60
 2
f
Ppf ð f Þ ¼ Ppf ð60Þ: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð3Þ
60
For a given drive frequency, Eqs. 1 to 3 define a pump curve describing the reduction of discharge pressure with increasing power-fluid
flow rate. For a centrifugal pump, the discharge pressure declines slowly with increasing flow rate and follows a smooth curve.
For a positive-displacement pump, the pump exhibits a rapid linear decline in discharge pressure with increasing flow rate. The
flow-rate variation with frequency follows Eq. 2, but there is no equation governing the pressure. The pressure adjusts to whatever is
required for the flow rate.

Tubing. The pressure at the bottom of the tubing is related to the pressure at the top of the tubing by

P1 ¼ Ppf þ DG1  DPf r ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð4Þ

where the pressure drop caused by friction is calculated in this case using the Hazen-Williams (Hazen and Williams 1920) equation
(ASPE 2008),
 
L 2:916Q1 1:852 4:8655
DPf r ¼ 0:2083  G1 Dt : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð5Þ
100 F
A significant pressure drop from the discharge of the surface pump to the wellhead must also be accounted for if it exists.

Nozzle. In the absence of friction, the Bernoulli equation states that the energy per unit mass is conserved across the nozzle. Referring
to the positions within the pump shown in Fig. 2, the Bernoulli equation is
P1 v21 P0 v20
þ ¼ þ : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð6Þ
q1 2 q0 2

374 May 2019 SPE Production & Operations

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030

 PO191382 DOI: 10.2118/191382-PA Date: 23-April-19 Stage: Page: 375 Total Pages: 12

Ignoring the compressibility of water and the entrance velocity of the water (relative to the exit velocity), we arrive at the
“nozzle equation”:
 
Q1 2
P0 ¼ P1  ð1 þ Kn ÞG1 : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð7Þ
857An

A nozzle coefficient, Kn , has been included to account for a small loss of energy caused by friction.
Assuming the power-fluid pressure and flow rate are measured in real time at the surface, the nozzle-exit pressure, also called the
“jet pressure,” can be calculated using Eq. 7. Note that this pressure is not explicitly dependent on the reservoir production rate or
intake pressure.
If the reservoir-production rate or intake pressure changes, the power-fluid flow rate and pressure will change, and therefore the jet
pressure will change. Viewed in reverse, when changes are seen in the power-fluid pressure or flow rate at surface, these indicate
changes in the reservoir inflow and pressure at the intake. In this sense, the power-fluid flow rate and pressure behave as a downhole
sensor because they react to downhole conditions.
For a centrifugal pump at surface, the pump curve is relatively flat. This means that when intake conditions change, the power-fluid

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023
flow rate will change more significantly than the discharge pressure. For a positive-displacement pump at surface, the reverse is true.
The discharge pressure will change more significantly than the flow rate.
Because we can control the frequency of the surface drive, it follows that we also have control of the jet pressure. For the case of a
surface centrifugal pump, the dependence of the jet pressure with drive frequency results from the affinity laws as
 2
f
P0 ð f Þ ¼ ðP0  DG1 Þ þ DG1 ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð8Þ
f

where f  is the frequency the drive is currently generating and P0 is the associated jet pressure. Eq. 8 is used to determine what the jet
pressure will be if the frequency is changed. Because DG1 is larger than P0 in most applications, we see that increasing the frequency
reduces the jet pressure.
For a positive-displacement pump, the affinity law for pressure does not apply, and there is no simple equation to calculate the jet
pressure as a function of frequency.

Intake. The jet-pump model assumes that at the pump intake, the oil phase is in equilibrium with any free gas that might be present.
This means that gas is neither dissolving in the oil nor evolving from the oil as it enters the pump. Once inside the pump, the gas phase
will change rate and pressure gradient (i.e., density) as the pressure drops and then rises within the pump. It is assumed that the gas tem-
perature does not change within the pump. This is justified by the short residence time within the pump. It is also assumed that gas does
not dissolve into or evolve from the oil within the pump, again because of the short residence time. This implies that the reservoir-
liquid rate and density entering the pump remain unchanged through the pump. It also implies that the gas and oil are no longer in equi-
librium as they exit the pump. Equilibrium will be restored in the casing above the pump.
For  API values greater than 30, pressures less than approximately 6,000 psig, and modest downhole temperatures typical of onshore
shale plays, the solution gas/oil ratio (Rs) can be calculated using (Kartoatmodjo and Schmidt 1991)
 
ð11:289ÞAPI
Rs ¼ ð0:0315Þc0:7587
g 10 T þ 460 ðP þ 14:7Þ1:0937 : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð9Þ

The oil volume factor is given by (Kartoatmodjo and Schmidt 1991)

 1:5
1:5
B0 ¼ 0:98496 þ 0:0001 R0:755
s c0:25
g co þ 0:45T : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð10Þ

The gas volume factor is given by

  
14:7 T þ 460
Bg ¼ Z; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð11Þ
P þ 14:7 60 þ 460

where Z is the compressibility factor, as calculated in Appendix A. For a given pressure and temperature, this factor is the ratio of the
gas volume to the volume of an ideal gas.
The oil density is given by

350co þ 0:0764cg Rs
qo ¼ : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð12Þ
5:615Bo

The reservoir-liquid rate at the intake, but also at all points within the pump is given by

Q2 ¼ Qw þ Qo Bo ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð13Þ

where the oil volume factor is calculated at the intake pressure, consistent with the equilibrium assumption at the intake.
The pressure gradient of the reservoir-liquid phase at all points within the pump is given by
1 qo Qo Bo þ qw Qw
G2 ¼ ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð14Þ
144 Qo Bo þ Qw
where the oil volume factor is calculated at the intake pressure.

May 2019 SPE Production & Operations 375

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030

 PO191382 DOI: 10.2118/191382-PA Date: 23-April-19 Stage: Page: 376 Total Pages: 12

The gas rate at the intake to the pump is given by

ðPGOR  Rs ÞQo Bg
Qg2 ¼ ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð15Þ
5:615
where the gas volume factor and the solution gas-oil ratio (GOR) are calculated at the intake pressure, consistent with the equilibrium
assumption at the intake.
The gas pressure gradient is given by
0:0764 cg
Gg2 ¼ ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð16Þ
144 Bg

where the gas volume factor is calculated at the intake pressure.

The “intake equation” governs the relationship between the intake pressure and the jet pressure. It is derived from the integral form
of the Bernoulli equation (Holland and Bragg 2002),
ð ð

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023
1
dP þ vdv ¼ 0; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð17Þ
q
which is required because density is now a function of pressure. (Eq. 17 ignores the pressure dependence of the gas-compressibility
factor.) The result is
2 32
P Q
    Q2 þ 2 g2
P2 Qg2 P2 Qg2 Gg2 66 P0 77
P2 þ ln ¼ P0 þ ð1 þ Ken Þ G2 þ 4857ðAt  An Þ5 : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð18Þ
Q2 Po Q2

In a jet-pump-sizing context, it is customary to specify the “desired” oil, water, and gas rates at the surface and to prescribe the asso-
ciated intake pressure. Because the intake pressure has been specified, the oil and gas gradients and rates can be calculated using Eqs. 9
through 16.

Estimation of Pump-Intake Pressure Using Daily Production Data. Once the well is producing, the actual production will be
known, as will the actual power-fluid flow rate and pressure. Although the power-fluid flow rate and pressure are measured in real time
and sampled every 5 or 10 minutes, the actual production is an accumulated value typically over the period of 1 day. Because of this,
we must restrict ourselves to days when the production is stable. If the measured power-fluid flow rate and pressure are observed to be
stable throughout the day, then the jet pressure will be stable and the production will also be stable. In this way, the daily production
rates are representative of actual real-time rates and can be associated with the real-time power-fluid flow rate and pressure.
Eqs. 9 through 16 cannot be recalculated using the actual production rates because the pump-intake pressure is not known. In this
case, a comprehensive set of possible intake pressures can be tried one after another until the one is found that solves Eq. 18.
The result is a virtual downhole-intake-pressure gauge. However, this can only be applied on days when production is stable. The
result is an average pressure for that day. This contrasts with the jet pressure, which can be provided in real time because the power-
fluid flow rate and pressure are measured in real time. An example of a virtual downhole-intake-pressure gauge can be found in
Appendix B.

Throat. The throat is modeled as a cylindrical tube where two fluids enter and one mixed fluid exits. The forces at either end (i.e., pres-
sure multiplied by area) must balance the difference in momentum at either end (i.e., mass-flow rate multiplied by velocity). The result
is a quadratic equation for the pressure at the throat exit,

P23 þ bP3 þ c ¼ 0; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð19Þ

where
 
P2 Qg2
2ðG 2 Q 2 þ G g2 Q g2 Þ Q 2 þ
Q1 þ Q2 2G1 Q21 P0
b ¼ ð2 þ Kt ÞðG1 Q1 þ G2 Q2 þ Gg2 Qg2 Þ 2
 P0   ; . . . . . . . . . . . . . ð20Þ
ð857At Þ 8572 At An 8572 At ðAt  An Þ

and

P2 Qg2
c ¼ ð2 þ Kt ÞðG1 Q1 þ G2 Q2 þ Gg2 Qg2 Þ : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð21Þ
ð857At Þ2

Diffuser. The integral form of the Bernoulli equation is applied across the diffuser to obtain

0 1
P2 Qg2 2
  Q þ Q þ
P2 Qg2 P4 Q1 G1 þ Q2 G2 þ Qg2 Gg2 B
B
1 2
P3 C C : . . . . . . . . . . . . . . . . . . . . . . . . ð22Þ
P4 þ ln ¼ P3 þ ð1  Kd Þ @ A
Q1 þ Q2 P3 Q1 þ Q2 857At

Casing. The pressure and temperature vary from the discharge to the surface. Starting at the known wellhead pressure and temperature,
we calculate the pressure gradient of the oil, water, and gas mixture using Eqs. 9 through 16 evaluated at the wellhead pressure and

376 May 2019 SPE Production & Operations

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030

 PO191382 DOI: 10.2118/191382-PA Date: 23-April-19 Stage: Page: 377 Total Pages: 12

temperature. We now introduce a small pressure increment. This pressure increment divided by the pressure gradient results in an asso-
ciated depth increment. The temperature at this depth increment is calculated assuming a linear increase from the wellhead temperature
to the fluid temperature at the jet pump. After this, the process is repeated: Another pressure increment is taken and the pressure gradi-
ent at the new pressure and temperature conditions is calculated, which leads to a new depth and temperature. In this way, the static
pressure at the depth of the jet-pump discharge is determined.
This model assumes that the power-fluid and reservoir-fluid temperatures are both equalized to a known value at the depth of the jet
pump. A more accurate model would include the warming of the power fluid as it progresses down the tubing and the mixing of two dif-
ferent fluid temperatures in the pump and then along the casing annulus to the surface. Holmes and Swift (1970) developed a model for
the circulation of drilling fluid in a formation with a linear temperature gradient. This model was not applied in this paper.
A small additional pressure is required to account for friction along the casing/tubing annulus. This can also be calculated incremen-
tally. Alternatively, an approximate value can be calculated using (Petrie et al. 1987)
2 3
6 ðDc þ DT Þ0:21 7
6 7
DPfr ¼ Ll0:21 G0:79 Q1:79 62:02ð106 Þ  0:1 7; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð23Þ
4 D 5

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023
c
ðDc þ DT Þ2 ðDc  DT Þ3
Dc  DT
where G and Q are mixture values averaged over depth.

Pump Efficiency. Following Cunningham (1995), the efficiency of the jet pump is the sum of liquid and gas terms,
 
P4
Q2 ðP4  P2 Þ þ P2 Qg2 ln
P2
g ¼ gL þ gg ¼ : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð24Þ
Q1 ½P1  P4 

The numerator is the hydraulic power output of the jet pump, while the denominator is the hydraulic power supplied by the power fluid.
(This assumes the nozzle is not withdrawn from the throat.) The presence of gas in the pump results in the second term in the numerator
of Eq. 24, which increases the pump efficiency. The fact that the jet-pump efficiency is improved by the presence of gas is explored in
greater detail by Cunningham (1995).

Mach Number. The mixture of gas in the liquid lowers the speed of sound. When the velocity of the mixture equals the speed of
sound, the Mach number (MN) is unity. The ability to pump additional liquid will stop once MN ¼ 1. When sizing a jet pump in an
application where there is free gas at the intake, this MN must be considered. The MN is calculated as (Cunningham 1995)
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2P2 Qg2 ðG2 Q2 þ Gg2 Qg2 Þ
MN ¼ : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð25Þ
857P0 ðAt  An Þ

Cavitation. Cavitation occurs when the jet pressure P0 is at or less than the vapor pressure of the liquid Pv . The equation for the intake
pressure at which cavitation occurs is given by (Cunningham 2008)
 2
Q2
P2 ¼ rG1 ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð26Þ
857ðAt  An Þ
where r is a constant factor, used because the vapor pressure is not known. A value of 1.35 is recommended (Cunningham 2008). Note
that cavitation is not dependent on free gas.

Jet-Pump-Sizing Calculations. Sizing a jet pump is the procedure for selecting a nozzle/throat combination. Each of these is character-
ized by a diameter, or the corresponding area. For each specific nozzle/throat choice, there is a unique solution for the power-fluid pres-
sure and flow rate that gives rise to the “desired” surface production at the prescribed intake pressure. The power-fluid pressure and flow
rate are the same whether the surface pump is a positive-displacement pump or a centrifugal pump. For a given application, the wellhead
pressure and temperature, liquid temperature at the jet pump, and SGs of the oil, water, and gas must be specified. The calculated power-
fluid pressure and flow rate must be within the range of the surface pump. For each nozzle/throat combination, there will be an associated
power-fluid flow rate and pressure. There can be multiple nozzle/throat combinations that fall within the range of a given surface pump.
The nozzle/throat combination can be chosen from among these as the one associated with the lowest drive frequency or the one with the
highest pump efficiency. Alternatively, it can be chosen as the one with the highest jet pressure or lowest MN.

Model-Validation Example
In some cases, a memory gauge is run with a jet pump. Once retrieved to the surface, the pump-intake-pressure data are accessed. This
measured intake pressure can then be used in the sizing calculations along with the actual daily production values to calculate the
power-fluid pressure and flow rate. Again, the actual power-fluid pressure and flow rate and the memory-gauge intake pressure should
be stable over the same 24-hour period when the “daily production” is tallied. If the model is calibrated, the calculated power-fluid pres-
sure and flow rate will agree with the actual power-fluid pressure and flow rate. If there is a significant discrepancy, then assuming the
data inputs are accurate, the various loss coefficients for the pump—as well as the friction in the tubing and casing—can be varied until
a match is achieved. Once calibrated, the model will be consistent with the data and considered to be validated.
Without a memory gauge and stable daily production, the model cannot be validated. In this case, the loss coefficients can be taken
from values determined in previous wells or from values found in the literature. Manufacturers of hydraulic jet pumps do not provide
loss coefficients.
Grupping et al. (1988) surveyed the literature and published a table of loss coefficients showing the range. Noronha et al. (1998)
used experimental laboratory data to determine the loss coefficients. The values were consistent with the range published by Grupping
et al. (1988). Similar to Grupping et al. (1988), Cunningham (2008) also published a table showing the range of the loss coefficients.

May 2019 SPE Production & Operations 377

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030

 PO191382 DOI: 10.2118/191382-PA Date: 23-April-19 Stage: Page: 378 Total Pages: 12

An example of validation is shown in Tables 1 and 2. Table 1 contains the input data to the model. Hourly production indicated
stable conditions: 444 B/D of oil, 1,512 B/D of water, and 264 Mcf/D of gas. This translates to a water cut of 0.77, a gas/liquid ratio
(GLR) of 135 scf/bbl, and a liquid rate of 1,956 B/D. A memory gauge recorded 1,858-psig intake pressure. The wellhead pressure was
280 psig. There was an E:10 nozzle/throat combination in the jet-pump assembly.

Symbol Value Description

QL 1,956 Desired liquid-production rate (B/D)
GLR 135 GLR (scf/bbl)
WC 0.77 Water cut
P2 1,858 Jet-pump-intake pressure (psig)
Pump stage for surface centrifugal
Sp 2,700
pump (B/D)

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023
Suction pressure for surface
Ps 40
centrifugal pump (psig)
API 40 Oil °API
γ1 1.10 Water SG
γg 0.70 Gas SG
WHP 280 Wellhead-casing pressure (psig)
WHT 140 Wellhead-casing temperature (°F)
T 160 Fluid temperature at jet pump (°F)
D 6,920 True vertical depth of pump (ft)
L 6,920 Measured depth of pump (ft)
Dt 2.441 ID of tubing (in.)
DT 2.875 OD of tubing (in.)
Dc 4.892 ID of casing (in.)

Table 1—Data inputs for the model-validation example. ID 5 inside

diameter; OD 5 outside diameter.

Symbols and Values G:11 F:11 F:10 E:10 E+:10 E+:11 F+:10 F+:12 F+:11 G:12
Q1 PF rate (B/D) 3,592 2,656 2,657 2,130 2,391 2,419 3,171 3,152 3,092 –
Ppf PF pressure (psig) 3,401 3,837 3,541 4,027 3,729 4,120 3,403 3,909 3,529 –
f Frequency (Hz) 73.4 71.3 68.9 70.7 69.3 72.5 70.7 74.3 71.2 –
η Efficiency 0.29 0.30 0.30 0.31 0.31 0.30 0.27 0.28 0.30 –
H Hydraulic power (hp) 205 171 158 144 150 168 181 207 183 –
MN MN 0.17 0.13 0.22 0.18 0.20 0.12 0.28 0.10 0.15 –
– Cavitation (yes/no) No No No No No No No No No –
Qg2 Gas rate (B/D) 65.6 65.6 65.6 65.6 65.6 65.6 65.6 65.6 65.6 –
P0 Jet pressure (psig) 1,373 1,515 1,216 1,344 1,285 1,542 1,053 1,628 1,457 –
P4 Discharge pressure (psig) 3,359 3,272 3,272 3,214 3,244 3,247 3,321 3,320 3,314 –
M Flow ratio 0.581 0.785 0.785 0.979 0.873 0.862 0.658 0.662 0.675 –
N Pressure ratio 0.579 0.411 0.450 0.354 0.404 0.364 0.525 0.443 0.494 –
P1 Nozzle press (psig) 5,953 6,707 6,412 7,041 6,674 7,059 6,108 6,621 6,261 –

Table 2—Ten different nozzle/throat combinations were evaluated. The resulting power-fluid (PF) pressure and flow-rate points for each
combination were calculated, along with the associated drive frequency for the surface horizontal pump. Other calculated values reflect the
downhole conditions.

The sizing results in Table 2 show a calculated power-fluid flow rate of 2,130 B/D and a calculated power-fluid pressure of 4,027 psig.
The actual power-fluid flow rate was 2,110 B/D, and the power-fluid pressure was 4,070 psig. (Both real-time quantities were stable over
the period where accumulated production was measured.) Although not an exact match, this is within 1% and considered good. In this
example, the nozzle-loss coefficient was 0.12, the intake loss was 0.0, the throat loss was 0.1, and the diffuser loss was 0.1. The tubing-
friction parameter was 80. These are reasonable values and consistent with published studies (Noronha et al. 1998; Cunningham 2008).
Table 2 shows that all the combinations have similar pump efficiencies and drive frequencies. (The G:12 combination is blank, indi-
cating that the power-fluid pressure and flow rate are outside the range of the surface pump.) However, the F:11, Eþ:11, and Fþ:12
combinations have somewhat higher jet pressures and lower MNs than the others.
In Fig. 3, we plot the values for power-fluid pressure and flow rate for each combination in Table 2 on the pump curves for the sur-
face centrifugal pump. All the combinations plot near the pump curve corresponding to the 70-Hz operation, indicating that resizing
does not result in a significant change.

378 May 2019 SPE Production & Operations

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030

 PO191382 DOI: 10.2118/191382-PA Date: 23-April-19 Stage: Page: 379 Total Pages: 12

5,000
4,500

Power-Fluid Pressure (psig)

4,000
3,500
3,000
2,500
2,000
1,500
1,000 Min
BE
Max 70 Hz
500 60 Hz
50 Hz
0 40 Hz
0 1,000 2,000 3,000 4,000 5,000 6,000

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023
Power-Fluid Flow Rate (B/D)

Fig. 3—The surface-centrifugal-pump curves for 40, 50, 60, and 70 Hz are plotted, along with the minimum, maximum, and “best-
efficiency” (BE) lines. For each jet-pump combination in Table 2, the associated power-fluid pressure and flow rate are plotted as
a point.

Production-Optimization Example
An example of production optimization is shown in Tables 3 and 4 and Figs. 4 and 5. The power-fluid flow rate and pressure are taken
during a 24-hour day when the values are stable. The daily production for that same day is associated with the power-fluid pressure and
flow rate. Table 4 shows that the jet pressure is calculated to be 1,911 psig and the intake pressure is 2,276 psig. There are 108 B/D of
free gas at the intake, which expands to 129 B/D at the entrance to the throat.

Symbol Value Description

Q1 1,900 Power-fluid rate (B/D)
Ppf 3,555 Power-fluid pressure (psig)
f 60 Drive frequency (Hz)
Qo 450 Oil-production rate (STB/D)
Qw 522 Water-production rate (B/D)
Qg 404 Gas-production rate (Mcf/D)
API 40 Oil °API
γ1 1.10 Water SG
γg 0.70 Gas SG
T 160 Fluid temperature at jet pump (°F)
D 6,920 True vertical depth of pump (ft)
L 6,920 Measured depth of pump (ft)
Dt 2.441 ID of tubing (in.)
2
Ab 0.0208 Nozzle area (in. )
2
At 0.066 Throat area (in. )

Table 3—Data inputs for the production-optimization example.

Symbol Value Description

P2 2,276 Pump-intake pressure (psig)
Qg2 108 Gas rate at pump intake (B/D)
Qg0 129 Gas rate at jet exit (B/D)
P0 1,911 Jet pressure (psig)
MN 0.196 MN
f 72 Optimal frequency (Hz)
Pump-intake pressure at optimal
P2(f ) 1,380
frequency (Hz)
MN(f ) 0.710 MN at optimal frequency (Hz)

Table 4—Production-optimization results.

The blue curve in Fig. 4 shows the dependence of the pump-intake pressure on drive frequency, assuming the use of a centrifugal
pump. We can see that the existing frequency of 60 Hz does not minimize the pump-intake pressure. Increasing the drive frequency
reduces the pump-intake pressure. At 74 Hz, the pump-intake pressure is minimized. Above 75 Hz, the pump-intake pressure increases dra-
matically (not shown) because of gas expansion resulting from low jet pressure. Fig. 5 shows that the gas rate at the jet exit increases with

May 2019 SPE Production & Operations 379

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030

 PO191382 DOI: 10.2118/191382-PA Date: 23-April-19 Stage: Page: 380 Total Pages: 12

frequency, whereas the gas rate at the intake starts to decrease at 75 Hz. The purple curve in Fig. 4 shows the MN vs. frequency. The MN
reaches a value of nearly unity at 74 Hz. When the MN reaches unity, the ability to pump additional liquid ceases (Cunningham 1995).
Table 4 indicates a recommended drive frequency of 72 Hz with an associated MN of 0.71 and a pump-intake pressure of 1,380 psig.

3,500 1
Pump-intake pressure
0.9
3,000 Operating point
Jet pressure 0.8
MN
2,500 0.7

Pressure (psig)
2,000 0.6

MN
0.5
1,500
0.4

1,000 0.3

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023
0.2
500
0.1
0 0
50 55 60 65 70 75
Frequency (Hz)

Fig. 4—A plot of intake pressure, jet pressure, and MN vs. drive frequency. The operating point of the system results in a higher
intake pressure than the minimum, indicating that production can be increased by increasing the drive frequency.

1,400

1,200
Gas at intake
1,000 Gas at jet
Gas Rate (B/D)

800

600

400

200

0
50 55 60 65 70 75
Frequency (Hz)

Fig. 5—As frequency increases, the intake and jet pressures reduce, resulting in gas expansion. The pressure drop from the intake
to the jet increases with frequency, causing the two curves to separate.

Increasing-Gas-Production Example
The behaviors of a jet pump with increasing amounts of produced gas are shown in Figs. 6 through 8. We use a pump-intake pressure
of 2,000 psi and increase the GLR from 0 to 300 scf/bbl. The other parameters are taken from Table 1.

75 0.8
0.7
70
0.6
Frequency (Hz)

Efficiency, MN

65 0.5 Frequency
0.4 Efficiency
60 0.3 MN

0.2
55
0.1
50 0
0 100 200 300 400 500 600
Produced-Gas Rate (Mcf/D)

Fig. 6—Behavior of drive frequency, jet-pump efficiency, and MN with increasing surface-gas production. Liquid production, water
cut, and intake pressure have been held constant.

In Fig. 6, we plot the drive frequency, jet-pump efficiency, and MN vs. the produced-gas rate. The drive frequency drops from a
maximum of 75 Hz to a minimum of approximately 50 Hz as the gas rate increases. The presence of gas makes it easier to lift the liquid.
For gas-production values lower than 200 Mcf/D, the gas is dissolved in the oil at this intake pressure and temperature, resulting in
MN ¼ 0. However, the MN rises to nearly 0.67 at 587 Mcf/D. This is close to a cutoff condition in the pump. The pump efficiency is
effectively constant. The beneficial effect of gas on efficiency is offset by associated changes in intake and discharge pressures. In this
example, the jet pump is able to lift 1,956 B/D of liquid with 586 Mcf/D of associated gas.

380 May 2019 SPE Production & Operations

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030

 PO191382 DOI: 10.2118/191382-PA Date: 23-April-19 Stage: Page: 381 Total Pages: 12

5,000 200
4,500

4,000 150

Pressure (psig)

Power (hp)
3,500
3,000 100
2,500
2,000 Power-fluid pressure 50

1,500 Hydraulic power

1,000 0
0 100 200 300 400 500 600
Produced-Gas Rate (Mcf/D)

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023
Fig. 7—Behavior of the power-fluid pressure and surface hydraulic power with increasing surface gas production. Liquid produc-
tion, water cut, and intake pressure have been held constant.

1,600 700

1,500 600
Pressure (psig)

500

Rate (B/D)
1,400
Jet pressure
400
1,300 Gas rate at intake
300
1,200
200
1,100 100

1,000 0
0 100 200 300 400 500 600
Produced-Gas Rate (Mcf/D)

Fig. 8—Behavior of jet pressure and gas rate at the intake, with increasing surface-gas production. Liquid production, water cut,
and intake pressure have been held constant.

In Fig. 7, we plot the power-fluid pressure and the hydraulic horsepower (from the surface pump) vs. the produced-gas rate. The gas
lift effect results in a significant reduction in surface pressure and power.
In Fig. 8, we plot the jet pressure and gas rate at the pump intake. The gas rate increases linearly once the oil is saturated. The jet
pressure declines, and because the pump-intake pressure is 2,000 psi in this example, the gas rate at the intake will continue to increase
at the jet, where it enters the throat.

Conclusions
Because surface production is not measured in real time, we must make use of accumulated values over some period. To validate the
model on a well or optimize production for the well, the production must be stable during this period. In this way, the daily production
values are representative of the actual flow rates. Stability is determined by examining the real-time measurements of the power-fluid
flow rate and pressure or the calculated real-time jet pressure. If these are stable, then the production is stable.
The system model can be calibrated using the pump-intake pressure from a memory gauge on a day when production is stable. This
calibration validates the model and involves the adjustment of friction-loss parameters.
The jet pressure within the pump can be calculated in real time from the power-fluid flow-rate/pressure measurements. Stable flow is
not required to do this.
The calculated pump-intake pressure does require stable flow over the period when production is accumulated. The result is an
average intake pressure for that period.
For a centrifugal pump at the surface, the variation in pump-intake pressure with drive frequency can be calculated. When there is
free gas entering the pump, this variation might show an optimal frequency at which the intake pressure is minimized and production is
maximized. The drive can be adjusted to this new frequency. After the change in frequency and after production has restabilized, the
process can be repeated.
The MN can be calculated. This indicates the proximity of the pumping conditions to optimal production and to severe production
reduction associated with excessive gas expansion.
A hydraulic jet pump is tolerant of free gas at the pump intake.

Nomenclature
A; B; C; D; E; F; G ¼ terms used in calculating the gas factor
An ; At ¼ areas of the nozzle and throat, respectively, in.2

API ¼ API value of the oil
Bo ; Bg ¼ oil and gas volume factors, respectively
D¼ true vertical depth of the jet pump, ft
Dt ; DT ¼ ID and OD of the tubing, respectively, in.
Dc ¼ ID of the casing, in.
f ¼ drive frequency (centrifugal pump on the surface), Hz
F¼ Hazen-Williams friction coefficient
Gg0 ; Gg2 ; Gg3 ; Gg4 ¼ pressure gradient of the gas at Positions 0, 2, 3, and 4, respectively, within the jet pump, psi/ft

May 2019 SPE Production & Operations 381

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030

 PO191382 DOI: 10.2118/191382-PA Date: 23-April-19 Stage: Page: 382 Total Pages: 12

G1 ¼ pressure gradient of the power fluid, psi/ft

G2 ¼ pressure gradient of the liquid at the jet-pump intake and throughout the jet pump, psi/ft
GLR ¼ produced GLR, scf/bbl
hi ; i ¼ 1,2,…6; centrifugal (surface) pump-head curve coefficients
H ¼ hydraulic horsepower of the surface centrifugal pump, hp
Kn ; Ken ; Kt ; Kd ¼ friction-loss coefficients of the nozzle, entrance (i.e., intake), throat, and diffuser, respectively
L ¼ measured depth of the jet pump (tubing length to jet pump), ft
MN ¼ Mach number
M ¼ flow ratio ¼ Q2 =Q1
N ¼ pressure ratio ¼ ðP4  P2 Þ=ðP1  P4 Þ
Ns ¼ number of pump stages (centrifugal pump on the surface)
PGOR ¼ producing GOR, scf/STB
Pi ; i ¼ 0,1,2,3,4; pressure at Positions 0, 1, 2, 3, and 4 within the jet pump
Ppf ¼ discharge (power-fluid) pressure of the surface pump, psig
Pr ¼ reduced pressure

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023
Ps ¼ suction (intake) pressure of the surface pump, psig
Pv ¼ vapor pressure of liquid, psig
Qg ¼ produced-gas volumetric rate at the surface, scf/D
Qg0 ; Qg2 ; Qg3 ; Qg4 ¼ volumetric rate of the gas at Positions 0, 2, 3, and 4, respectively, within the jet pump, B/D
QL ¼ produced-liquid volumetric rate at the surface, STB/D
Qo ; Q w ¼ produced-oil and produced-water volumetric rate, respectively, at the surface, STB/D
Q1 ¼ power-fluid volumetric rate, B/D
Q2 ¼ liquid volumetric rate at the jet-pump intake, B/D
Rs ¼ solution GOR, scf/STB
Sp ¼ pump stage used in the surface centrifugal pump, B/D
T ¼ temperature, F
Tr ¼ reduced temperature
v ¼ velocity
v0 ; v1 ¼ velocity of the power fluid at Positions 0 and 1, respectively
WC ¼ water cut
WHP ¼ wellhead pressure, psig
WHT ¼ wellhead temperature, F
Z ¼ gas-compressibility factor
cg ¼ SG of the gas
co ¼ SG of the oil
c1 ¼ SG of the power fluid
DPfr ¼ pressure loss caused by friction in tubing, psig
g; gL ; gg ¼ jet-pump efficiency, liquid efficiency, and gas efficiency, respectively
q ¼ density
qo ; qg ; qw ¼ oil, gas, and water densities, respectively, lbm/ft3
q0 ; q1 ¼ density of the power fluid at Positions 0 and 1
r ¼ cavitation parameter
l ¼ viscosity, cp

Acknowledgments
Wayne Richards, Mike Trice, Jeff Wilhelm, Jim Speer, Jason Cruz, and Bryce Wickliffe are recognized for their support of this work.
Peter Goode contributed technically to the author’s understanding of transient reservoir behavior. Jefferson Creek advised on the
pressure/volume/temperature calculations for gas and oil rates and gradients.

References
Americal Society of Plumbing Engineers (ASPE) 2008. Plumbing Engineering and Design Handbook of Tables, Rosemont, Illinois: ASPE.
Cunningham, R. G. 1957. Jet Pump Theory and Performance With Fluids of High Viscosity. Trans. ASME 79: 1807–1820.
Cunningham, R. G. 1974. Gas Compression With the Liquid Jet Pump. J. Fluids Eng. 96 (3): 203–215. https://doi.org/10.1115/1.3447143.
Cunningham, R. G. 1995. Liquid Jet Pumps for Two-Phase Flows. J. Fluids Eng. 117 (2): 309–316. https://doi.org/10.1115/1.2817147.
Cunningham, R. G. 2008. Jet Pump Theory. In Pump Handbook, fourth edition, ed. I. J. Karassik, J. P. Messina, P. Cooper, et al. Chap. 7. New York
City: McGraw Hill.
Fayazi, A., Arabloo, M., and Mohammadi, A. H. 2014. Efficient Estimation of Natural Gas Compressibility Factor Using a Rigorous Method. J. Nat.
Gas Sci. Eng. 16 (January): 8–17. https://doi.org/10.1016/j.jngse.2013.10.004.
Grupping, A. W., Coppes, J. L. R., and Groot, J. G. 1988. Fundamentals of Oilwell Jet Pumping (includes associated papers 17106 and 17113). SPE
Prod Eng 3 (1): 9–14. SPE-15670-PA. https://doi.org/10.2118/15670-PA.
Hazen, A. and Williams, S. G. 1920. Hydraulic Tables, New York City: John Wiley & Sons, Inc.
Holland, F. A. and Bragg, R. 2002. Fluid Flow for Chemical Engineers, second edition. Waltham, Massachusetts: Butterworth-Heinemann.
Holmes, C. S. and Swift, S. C. 1970. Calculation of Circulating Mud Temperatures. J Pet Technol 22 (6): 670–674. SPE-2318-PA. https://doi.org/
10.2118/2318-PA.
Kartoatmodjo, T. R. S. and Schmidt, Z. 1991. New Correlations for Crude Oil Physical Properties. SPE-23556-MS. Unsolicited.
Noronha, F. A. F., Franca, F. A., and Alhanati, F. J. S. 1998. Improved Two-Phase Model for Hydraulic Jet Pumps. SPE J. 3 (3): 227–231. SPE-50940-
PA. https://doi.org/10.2118/50940-PA.
Petrie, H. L., Wilson, P. M., and Smart, E. E. 1983. Jet Pumping Oil Wells. World Oil 197 (6): 6p.
Petrie, H. L., Wilson, P. M., and Smart, E. E. 1987. Hydraulic Pumping. In SPE Petroleum Engineering Handbook, Vol. IV, ed. J. D. Clegg, Chap. 6,
36–47. Richardson, Texas: Society of Petroleum Engineers.

382 May 2019 SPE Production & Operations

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030

 PO191382 DOI: 10.2118/191382-PA Date: 23-April-19 Stage: Page: 383 Total Pages: 12

Appendix A—Z-Factor Calculation

A closed-form expression for gas-factor calculation comes from Fayazi et al. (2014),
 4
Pr
Z ¼ A þ BPr þ ð1  AÞexpðCÞ  D ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ðA-1Þ
10
where
T þ 460
Tr ¼ ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ðA-2Þ
169:2 þ 349:5cg  74:0c2g
P þ 14:7
Pr ¼ ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ðA-3Þ
756:8  131:07cg  3:6c2g
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
A ¼ 0:101  0:36Tr þ 1:3868 Tr  0:919; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ðA-4Þ
0:04275

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023
B ¼ 0:021 þ ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ðA-5Þ
Tr  0:65
C ¼ Pr ðE þ FPr þ GP4r Þ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ðA-6Þ

D ¼ 0:122exp½11:3ðTr  1Þ ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ðA-7Þ

E ¼ 0:6222  0:224Tr ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ðA-8Þ

0:0657
F¼  0:037; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ðA-9Þ
Tr  0:85
G ¼ 0:32exp½19:53ðTr  1Þ: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ðA-10Þ

Appendix B—Virtual Downhole-Pressure Gauges

The real-time measurement of the power-fluid pressure and flow rate at the surface allows us to calculate the jet pressure in real time.
This is performed with Eq. 7, which requires Eqs. 4 and 5. This is the pressure at the exit of the nozzle, or Position 0 in Fig. 2. This cal-
culation constitutes a virtual downhole-pressure gauge, internal to the jet pump. The surface-drive frequency can be changed to raise or
lower this jet pressure. In this way, the jet pressure is under control from the surface.
The pump-intake pressure is higher than the jet pressure. The difference between the pump-intake pressure and the jet pressure is
determined by Eq. 18. Unlike the jet pressure, the pump-intake pressure depends on the surface-production rates of oil, water, and gas.
Because these are not measured in real time, we must rely on “daily production” values, which are accumulated values over a 24-hour
period. If the production is steady over this period, then using these values will lead to a representative value for the intake pressure,
and a virtual “daily” intake-pressure gauge is achieved. If the actual production varies considerably over the 24-hour period, then the
calculated intake pressure is some type of average intake pressure. If the real-time surface measurements of power-fluid pressure and
flow rate are steady during the 24-hour period, then this indicates that the reservoir production is steady and so the virtual intake pres-
sure is representative of what a downhole gauge would measure.
Fig. B-1 is an example of a well that had a downhole real-time pressure gauge at the pump intake. The data from the actual gauge
have been averaged each day. The daily values of virtual jet pressure and virtual intake pressure are also shown. The jet pressure is
lower than the intake pressure, and the virtual intake pressure trends closely with the actual gauge measurement. The nozzle-loss coeffi-
cient and the tubing-friction coefficient were adjusted to achieve this close match.
1,650

Real-time intake-pressure gauge

1,600 Virtual jet-pressure gauge
Virtual intake-pressure gauge

1,550
Pressure (psi)

1,500

1,450

1,400

1,350

1,300
June 2 June 3 June 4 June 5 June 6 June 7 June 8 June 9
Date

Fig. B-1—A well with a real-time intake-pressure gauge is compared with calculations of the jet pressure and intake pressure.

Unlike the virtual jet pressure, the virtual intake pressure is not in real time. It is computed from daily production values and gives
the pressure “yesterday.”

May 2019 SPE Production & Operations 383

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030

 PO191382 DOI: 10.2118/191382-PA Date: 23-April-19 Stage: Page: 384 Total Pages: 12

Andrew Kurkjian is a technical adviser at GR Energy Services in Sugar Land, Texas, USA. Before joining GR Energy Services,
he worked for Schlumberger in the areas of borehole geophysics and wireline-formation testing, for Global Artificial Lift and
Halliburton in the area of electrical submersible pumps, and under a wide range of contracts while leading OTNA Services.
Kurkjian has numerous publications and holds numerous patents. He holds a PhD degree in electrical engineering with a special-
ization in digital-signal processing from the Massachusetts Institute of Technology. Kurkjian is a member of SPE.

Downloaded from http://onepetro.org/po/article-pdf/34/02/373/2144426/spe-191382-pa.pdf by University of Engineering and Technology Lahore user on 04 February 2023

384 May 2019 SPE Production & Operations

ID: jaganm Time: 18:38 I Path: S:/PO##/Vol00000/180030/Comp/APPFile/SA-PO##180030