SPE-171342-MS

Improved Gas Lift Valve Performance using a Modified Design for GLV
Seat
F. Elldakli and M. Y. Soliman, Texas Tech University; M. Shahri, Halliburton; H.W. Winkler and T. Gamadi,
Texas Tech University

Copyright 2014, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Artificial Lift Conference & Exhibition-North America held in Houston, Texas, USA, 6 8 October 2014.
This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Each gas lift valve (GLV) is a variable orifice until a fully open port area is attained (under maximum stem
travel). As the ball (stem) moves away from the ball/seat contact area, the area open to flow increases until
the flow area upstream to the port area equals or exceeds the fully open port area.
Laboratory gas dynamic throughput testing indicates that each injection-operated GLV often does not
open fully in actual operation, mainly because of the bellows stacking phenomena. As a result, the stem
forms a restriction upstream to the flow path. Therefore, actual flow through the GLV can be less than
expected. This paper addresses such issues and recommends a simple but effective solution. A modified
design for the GLV seat was created to help reduce the required stem travel to generate a flow area equal
to the port area.
Theoretical calculations confirm the actual gas dynamic measurements and show that the minimum
stem travel for the modified design improves from 5 to 58% compared to using a conventional
sharp-edged seat. This improvement should have a significant impact on GLV performance. The modified
seats for all different ports sizes were manufactured and tested using a benchmark valve test. The
experiments showed that for the same stem travel, the new design has a larger flowing area than that of
the sharp-edged seat. This paper details the new design, theoretical calculations, and experimental results.

Introduction
Generally, a gas lift is a simple, flexible, and reliable artificial lift system with the ability to cover a wide
range of production rates. Gas lift systems are a closed rotative system empowered by high-pressured gas.
Therefore, the surface facilities required to perform this application consist of a compressing unit and
supplemental source of lean gas. A GLV, however, is a backpressure regulator (Winkler 1987). The entire
process is used to reduc the wellbore fluid pressure gradient by supplementing gas through an external
source to withdraw more liquid from the reservoir under higher drawdown. Many parameters affect the
gas lift system design, such as a change in the wellhead and bottomhole pressures (BHPs), produced fluid
type, and productivity index of the reservoir. As these parameters change, the gas injection pressure
changes.

SPE-171342-MS

A GLV basically regulates the pressure on its upstream side to its downstream. While the upstream
pressure is higher than the dom-charged pressure, the GLV remains open. Therefore, calibrating each
GLV to achieve the best performance at the wellbore is vital to the artificial lift cycle of each well.
Because a GLV consists of many moveable mechanical compartments, achieving synergy between all of
those compartments should result in the best performance.
The objective of this study was to optimize the GLV performance by measuring dynamic gas
throughput performance of each GLV using a modified seat design. Before 1940, differential GLVs were
common. In those valves, the operation of the valve dependes on the differential pressure between the
injection gas in the casing and the fluid in the tubing. King (1940) invented the first GLV with a
gas-charged bellows assembly. With that, gas lifting of lower BHP wells became possible by means of a
controlled pressure change of the surface injection.
Overall, a gas lift is a forgiving method of enhanced production, in other words, even a poor gas lift
design can increase production. To achieve a higher ramp in fluid production rate using gas lift, however,
a more sophisticated design of each compartment of the system is required.
Because a gas lift is a system of multiple compartments, a comprehensive redesign may be nessary to
increase production rate. The architectural design of each GLV is as important as the depth of installation
and number of GLVs used in each installation. Some faulty designs may result in installing dozens of
GLVs to unload well that does not require even one to be installed. Therefore, it is very important to
understand the system and not rush its development.

Statement of the problem

The mechanics of a GLV are exclusively based on pressure balance across the valve itself. Current GLVs
are all based on King (1940) and consist of a dome section, which is charged with gas (usually nitrogen)
at a certain pressure. On one end, a dome seal is present for pressure charge and discharging purposes.
The dome section is attached to the bellows assembly. The bellows assembly is helical in shape and acts
like a spring. The bellows are attached to the stem, which ends at the ball. All of these mentioned sections
move as a single unit in each GLV. When the GLV is closed, the ball is seated onto a sized port area. As
a rule, in each GLV, the ball is 1/16-in. larger in diameter than each port size. A check valve on the
downstream side of the port prevents the backflow from either the tubing or casing to interfere with one
another.
The basic components that comprise a GLV are shown in Fig. 1. The loading element can be a
nitrogen-charged bellows, a spring, or a combination of both. Fig. 2 shows schematics of a benchmark
valve. Winkler (1987) originally came up with this design to measure the gas throughput capacity of each
GLV. Winkler and Camp (1987) implemented such a tool in field applications, with successful results. A
detailed description of this valve can be found in API RP11V2 (2001), ISO 17078-2 (2007), and Shahri
(2011).
As gas is injected the casing pressure rises, and applies on the bellows area. A fraction of the injection
pressure acts on the ball pushing the stem down. This amount depends on the ball surface area. In static
mode, the only pressure acting on the bellows is the dome-charged pressure, which is reffered to as closing
pressure. As the injection gas pressure ramps up, the force on the bellows increases, and when it passes
the bellows set charge pressure, the GLV initially begins to open. Note that the GLV opening mechanism
is gradual. The stem travel in the actual GLV system is based on the difference between the opening and
closing forces and the bellows-assembly load rate. The equivalent port area for a partially open valve is
defined by the lateral surface area of the frustum of a right circular cone. The frustum area is generated
between the ball surface and the valve seat as the valve stem moves away from the seat. Therefore, a GLV
is a variable orifice until maximum stem travel or a fully open port area is attained. Then it becomes
equivalent to an orifice (pipe flow).

SPE-171342-MS

Figure 1Schematic of an injection pressure operated (IPO) GLV and

components.

Figure 2Schematic of a benchmark valve and components (from

Winkler 1987)

Flowing area is one of the most important parameters in each GLV. This area is generated by the stem
movement away from the seat. This movement depends on the bellows efficiency. During the gas lift
operation, the bellows is frequntally exposed to the injection pressure (opening force) and dome-charged
pressure (closing force). Therefore, the inner and outer convolutions of the bellows come in contact one
another. As a result, the bellows experiences stacking, and the GLV stem does not travel a sufficient
distance to create a flow area equal to the port area. Therefore, the GLVs ball and stem form a restriction
in the flow path and the actual throughput flow in the GLV becomes less than the theoretically calculated
value.

Modified Design
A modified design for the GLV that helps reduce the stem travel required to generate a flow area equal
to the port area was achieved. The workflow in such a process begins with calculating the minimum stem
travel for a modified seat and comparing the results with current geometry (sharp-edged seat) before
manufacturing the new seat/port assembly. The new seat design is easy to manufacture and is compatible
with other GLVs. The specifications for the modified GLV seat are controlled by changing the angle of
the taper.
Different angles are used with each port to create different port top diameters, as shown in Fig. 3. The
theoretical minimum stem travel can be calculated using Eq. 1 which is similar to the equation developed
by Kulkarni (2005). The stem travel in Eq.1 is function of the port bottom radius (rp), port top radius (rT),
and ball radius (rb).
(1)

The ball size for each port top diameter is used based on API recommendations (ball OD Port top
is based on the ball/seat contact area. Table 1 shows the amount of
diameter 1/16 in.). The ratio
travel required to achieve a flow passage equivelent to the port. The calculations are basically the same
for a sharp-edged seat.

SPE-171342-MS

Figure 3Comparison of sharp-edged seat with beveled seat (at different beveled angle).
Table 1GLV PORT/BALL AND STEM CHARACTERISTICS FOR GAS THROUGHPUT TESTS

Theoretical calculations show that the minimum stem travel for the modified design improves the stem
travel from 5 to 58% compared to using a conventional sharp-edged seat designs. This improvement
should have a significant impact on GLV performance.

Testing
In each gas lift design, GLVs should be tested to assure the proper amount of gas is passed to lift the
predicted volume of liquid. To quantify the GLV mechanics and behavior, a benchmark valve (Fig. 2) that

SPE-171342-MS

Figure 4 Variable beveled seat arrangements for gas throughput capacity measurements.

has the body of an actual GLV with a controlled

flowing area is used for these sets of experiments.
The banchmark valve is used to test the valve performance at different stem positions. The only physical difference between a true GLV and the benchmark valve is the existence of the bellows assembly.
The bechmark valve is not equipped with a bellows
or dome section, but the stem position in relation to
the valve seat is manually adjustable. This information could be applied to predict the stem travel for
actual GLVs when the volumetric gas rate and upstream and downstream pressures are known.
For the modified design, different sizes of seats
were tested to determine flow through values. A Figure 5Variable dimensionless flow area of variable beveled-angle
1-1/2-in. (IPO) Camco J-20 was used in all exper- seat and different ball sizes for -in. port.
iments. The benchmark valves with identical seats
were installed in an encapsulating tester, and then the stem travel was adjusted at six different positions
relative to the ball/seat contact area. Therefore, the ball/seat distances were based on the theoretical
minimum required to fully travel that the upsteam area would be identical to the downstream area.
Because a ball/stem was involved as well, the benchmark valve was set at two other positions greater than
the theoretical fully open area.
The testing system used in this paper was based on API RP11V2 (2001). The testing procedure was
based on a constant production pressure test (CPPT), in which the downstream production pressure used
was atmospheric pressure. The testing was performed at transient conditions; however, it could be
performed under steady state conditions as well. The pressure-time data were recorded using a very fast
data acquisition system (DAS) utilized by National Instruments (NI) (2014). This system is capable of
recording up to 12,000 samples per second per channel. In the experiments for this paper, the sample rate
was set at 100 samples per second per channel. Winkler (2010) shows that using 100 samples per second
works as well as using 12,000 samples per second. Data analysis is also accelerated. Shahri (2011) and
Shahri and Winkler (2011) published some discharge coefficient values as these tests were underway.
Their results show that as the dome-charged pressure changes, the liner stem travel of the ball changes.

SPE-171342-MS

Figure 6 Measurement of flow throughput capacity of variable beveled-angle seat for -in. port.

Figure 7Variable dimensionless flow area of variable beveled-angle

seat and different ball sizes for 5/16-in. port.

Figure 8 Measurement of flow throughput capacity of variable beveled-angle seat for 5/16-in. port.

This causes the value of the discharge coefficient to

change. Not considering these changes can result in
overestimating the liquid production by up to 30%.
The calculation is implemented at a critical condition and orifice flow regime. To creat this situation, the calculated ratio of downstream pressure to
upstream pressure (Pdown / Pup) should be less or
equal to 0.528 (the critical value for nitrogen).
Knowing the injection pressure, the pressure
drop at a depletion time, and the capacity of working gas allows the volumetric flow rate to be calculated. The effect of temperature was included
throughout the test. Because the testing time was
short, the temperature changes was insignificiant Figure 9 Variable dimensionless flow area of variable beveled-angle
seat and different ball sizes for 3/8-in port.
but were incorporated into the test.
The volumetric calculations are based on the real
gas law. Some values must be measured, such as the pressure as the gas is venting from the system and
the corresponding temperature with time. Working gas pressure (upstream) was recorded with an analog

SPE-171342-MS

Figure 10 Measurement of flow throughput capacity of variable beveled-angle seat for 3/8-in. port.

dial gauge, as well as a high-speed digital DAS from NI. Gas temperature (upstream) was read using an
electronic laser thermometer. Because the length of each experiment was short, the temperature variations
were negligible. Therefore, the temperature measurements were not continuous and were measured at the
beginning and at the end of each test. The temperature variations were to be 0.5 F. Atmospheric
pressure was read using a barometer in mmHg and then was recalculated to psia. The basis of standard
pressure was set to 760 mmHg or 14.696 psia. On the downstream side, the initial temperature was
assumed to be equal to the atmospheric temperature; however, because of the gas cooling effect, that
temperature varied under experimental conditions. The gas compressibility factor was calculated at each
pressure and temperature based on available correlations. Some constant inputs were used in this analysis
that were dominantly depenedent on the location and testing facility such as the gas constant, ratio of
specific heats of the active gas, the capacity of the storage facility, and the specific gravity of the working
gas.

Results
The experimental results confirmed the theoretical predictions. Figs. 5, 7, and 9 compare the effective flow
area obtained when using the sharp-edged seat with that obtained using the new design for three different
port sizes: 1/4, 5/16 and 3/8 in. Figs. 6, 8, and 10 demonstrate the gas throughput capacity of each port
size respectively. At the same stem travel time, the new design provides a larger flow area, and as the port
top diameter increases, the flow area increases. This improvement should have a significant impact on the
GLV performance, which was measured to be between 5 to 30% more than the gas throughput capacity
of GLVs using a sharp-edged seat.

Conclusions
The following conclusions are a result of this work:

GLVs do not pop open as the injection pressure passes the initial opening pressure; and thus, the
static force balance equations used to calculate opening and closing pressures are not appropriate
for calculating dynamic flow performance.
A new GLV seat was successfully designed, manufactured and tested.
The port geometry and maximum stem travel effect on the volumetric gas throughput of each GLV
were improved.
The experiment results showed that for the same stem travel, GLVs using a beveled seat had a
larger open area to flow compared to those using a sharp-edged seat.

SPE-171342-MS

References
ISO 17078-2 Petroleum and Natural Gas Industries Drilling and Production Equipmentpart2:
Flow-Control Devices for Side-Pocket Mandrels, Annex H and Annex O. 2007. Geneva, Switzerland:
ISO.
King, W.R. 1940. Time and Volume Control for Gas Intermitter. US Patent No. 2,339,487.
Kulkarni, M.N. 2005. Gas Lift Valve Modeling with Orifice Effects. MS thesis, Texas Tech University,
Lubbock, Texas.
National Instruments. 2014. NI 9237, http://sine.ni.com/nips/cds/view/p/lang/en/nid/208791 (accessed
9 June 2014).
Recommended Practice 11V2, Gas-lift Valve Performance Testing, second edition. 2001. Washington
DC: API.
Shahri, M.A. and Winkler, H.W. 2011. Practical Method for Measurement of Injection-Gas Throughput of Each Gas-Lift Valve before Well Installation. Paper SPE 141055 presented at the SPE Production
and Operations Symposium, Oklahoma City, Oklahoma, USA, 2729 March. 10.2118/141055-MS.
Shahri, M.A. 2011. Simplified and Rapid Method for Determining Flow Characteristics of Every
Gas-Lift Valve. PhD dissertation, Texas Tech University, Lubbock, Texas.
Winkler, H.W. 1987. Petroleum Engineering Handbook, volume 4. Richardson, Texas: SPE.
Winkler, H.W. and Camp, G.F. 1987. Dynamic Performance Testing of Single-Element Unbalanced
Gas-Lift Valves. SPE Prod Eng 2(3): 183190. 10.2118/14348-PA.
Winkler, H.W. 2010. Field Testing Gas-Lift Valves Before Well Installation. Paper presented at the
Proceeding of South Western Petroleum Short Course (SWPSC), Lubbock, Texas, USA, 2122 April.