Beiranvand Choking Correlation
Beiranvand Choking Correlation
Beiranvand Choking Correlation
Summary cline with many deep grabens at the crest. The average reservoir
Flow-rate prediction of oil production wells is of prime impor- radius and area are 2125 km and 89 km2, respectively, with da-
tance to effectively confront high-water-cut and separator prob- tum of 2286 m in true vertical depth (TVD). The main producing
lems. (Semi-) empirical multiphase-flow correlations are proved reservoir is Burgan B sandstone, and the main production mecha-
quite useful for this purpose. This work presents new generalized nism is aquifer drive. Observations have shown an extensive
multiphase flow choke correlation, derived on the basis of actual regionally depleted aquifer, with no gas cap and considerable
production data from horizontal and vertical wells from an oil water breakthrough in some of the development wells. The initial
field in Iran. The newly established correlation predicts liquid pressure reported in year 2002 is 3,552 psia. It should be noted
flow rates as a function of flowing wellhead pressure, gas/liquid that initial pressure declined from virgin pressure of 3,715 psia,
ratio, surface wellhead choke size, and the newly incorporated pa- as a result of the regionally depleted aquifer. The reservoir tem-
rameters: basic sediment and water (BS&W) and temperature. To perature is 183 F (84 C) at 2286-m TVD. This field has heavy,
evaluate the influence of these two new parameters, a parameter- undersaturated oil that contains a moderate amount of gas (an
sensitivity analysis was performed and the results are shown. This average 400 to 500 scf/bbl). Fluid properties are tabulated in
proposed correlation exhibited an average error of roughly 2.89%, Table 1.
which is superior to those previous correlations in the literature The initial reservoir pressure has been determined from initial
that did not use these two newly incorporated parameters (BS&W drillstem-test pressures to be 3,715 psia at datum depth of 2286-m
and temperature). These new parameters can be added to the pre- TVD. This corresponds to a pressure of approximately 3,771 psia
vious correlations when the water cut and temperature become at the oil/water contact (OWC), midcase 2337-m TVD.
important in the production history of the wells. Some of the original-production-test results and pressure data
were checked and re-evaluated to verify the initial pressure at
datum depth and to establish whether a more conclusive OWC
Introduction could be interpreted from the pressure gradients observed in the
The phenomenon of multiphase flow (liquid and gas) happens in various wells. Unfortunately, no firm conclusions could be
the wellhead of the majority of the producing wells. As the regu- drawn with respect to the original OWC, although in general, the
lation of the flow rate becomes important during the production data appeared to support an OWC between 2342- and 2332-m
period in the producing wells, chokes are used for isolating the TVD. A summary of some of the more-reliable data points
underground reservoir from pressure variations in the surface that could be verified from original test reports is provided in
equipment, and also they are used for preventing or reducing the Table 2. A summary of the reservoir description is also tabu-
water production. Larger amounts of produced water from oil re- lated in Table 3.
covery result in increasing of operating costs and are a major In this paper, several production tests were taken and 182 tests
environmental concern for oil production (Jin and Wojtanowicz data were used to generate a new correlation for multiphase flow
2010). through the wellhead surface choke. The test field is one of the
Although numerous multiphase-flow correlations are included southern offshore fields of Iran. The history of production in this
in the literature (Al-Attar 2010), almost all of them are limited to field shows the lack of fast production in this reservoir because of
a special operational condition in which the correlations are water coning and eventual killing of the wells. Some parts of the
driven. As a result, the strength of those correlations for predict- reservoir have been separated by faults. At the beginning of dril-
ing the actual flow rate is restricted. ling and production, water coning of some wells dramatically
Gilbert (1954) developed the most popular multiphase flow increased. At that time, to control the water coning, wells were
surface choke correlation, but this correlation is valid for the criti- choked and some were killed/closed. At the beginning of this cen-
cal flow condition when the upstream pressure of the choke is tury, the field had been developed through the drilling of new hor-
70% or more higher than the downstream pressure (Ghareeb and izontal wells. Considering the fast water production of vertical
Shedid 2007). wells, horizontal-well technology is used to control water coning
Flow through the wellhead chokes is mainly divided into two and to decrease the costs associated with the drilling of individual
critical and subcritical conditions. Within this manuscript, the vertical wells in the sea. Fig. 1 shows a parameter-sensitivity
critical-flow condition refers to the state at which the flow rate analysis on field water cut performed at that time.
reaches a maximum amount independent of the downstream and
upstream pressure difference of the choke. Empirical correlations Background
are mainly used for critical-flow condition. In 1954, Gilbert proposed a correlation relating surface produc-
tion with the wellhead-choke size, wellhead (or tubing) pressure,
and gas/oil ratio (GOR), as follows:
Field Status
Almost 30 years ago, after initial reservoir exploration, 25 wells Pwh SB
had been drilled. The studied field is an elongated elliptical anti- QA ; A 0:1; B 1:89; C 0:546; . . . . . . 1
GORC
Copyright V
C 2012 Society of Petroleum Engineers where Q is the gross liquid rate (bbl/d); GOR is producing at
Original SPE manuscript received for review 14 August 2011. Revised manuscript received
standard conditions (Mscf/bbl); Pwh is wellhead (or tubing) pres-
for review 3 November 2011. Paper (SPE 158649) peer approved 15 January 2012. sure (psig); and S is bean size (1/64 in.).
Perf. Interval Gauge Depth Gauge Pressure Gauge Depth Datum Pressure
Well Date (mAHD) (mAHD) (psig) (mTVDsa) (psia)
General Comments
Size and shape Elongated elliptical anticline, with many deep 30 km in length, 10 km in width and thickness is
grabens at crest approximately 80 m in average
Depth 2,286 mTVD Reservoir datum
Main producing reservoir Burgan B sandstone
Geological setting Deposited in a shallow marine and estuarine
environment
Area and depth 89 km2 * 15-m-thick stacked target sands, Average for Burgan B reservoir
117-m oil column
Main production mechanism Aquifer drive Extensive regionally depleted aquifer. No gas cap,
and reservoir performance and excellent communication across the field
Initial pressure 3,552 psia (in 2002, in the beginning of Initial pressure declined from virgin pressure of
development program) 3,715 psia (in 1966) because of a regionally
depleted aquifer
Reservoir temperature 183 F (84 C) 2,286 mTVD
1.0
Base Case BSW Development
Optimistic Relperms
Pessimistic Relperms
0.8
BSW [FT3/FT3]
0.6
0.4
0.2
0.0
2002 2006 2010 2014 2018 2022 2026 2030
Time, Year
In 1960, Ros developed a very likely correlation to the Gilbert Results and Discussion
(1954) correlation, but with different constant and exponents, as Results from 182 production tests performed in the field were
follows: used to derive a new equation for oil-flow rate through choke. The
Pwh SB data used are tabulated in Table 6.
QA ; A 0:574; B 2:0; C 0:5: . . . . . . . 2 In previous equations, including Gilbert (1954), Ros (1960),
GORC Baxendell (1950), Achong (1961), and Mesallati et al. (2000), sig-
Another Gilbert-form correlation was presented by Achong in nificant errors were observed in measured data. To minimize the
1961: errors of equations, two new parametersBS&W and Tare
assumed to be effective parameters in the proposed equation.
Pwh SB Through use of these two parameters, a considerable decrease was
QA ; A 0:2618; B 1:88; C 0:65; . . . . . 3
GORC observed in error of the equations. This proposed correlation dis-
played an average error rate of roughly 2.89% which is more
Other Gilbert-type multiphase flow choke correlations by other
accurate than other relevant correlations. The proposed correla-
researchers were proposed for the critical condition. These corre-
tion exhibits more accuracy (only 2.89% average error) than the
lations have the same form as that of Gilbert (1954) but with dif-
existent correlations. The comparison between the equations is pre-
fering constants and variable exponents. They are tabulated in
sented in the Table 4. Fig. 2 depicts a good correspondence
Table 4.
between the predicted and measured data for the driven correla-
tion. To evaluate the effect of these two new parameters, a param-
Development of New Correlations eter-sensitivity analysis was performed and the obtained results
Data from 182 production tests were collected for various hori- are depicted in Figs. 3 and 4. In the mentioned analysis, the trend
zontal and vertical wells, including the liquid-flow rate, GOR, of Q vs. T and vs. BS&W was obtained while keeping the other
choke size, and wellhead pressure within the range of 183 to 9284 effective parameters constant. As seen in these figures, the first
(stb/d), 36 to 885 (scf/stb), 25.6 to 40 (1/64 in.), and 133 to 883 and second derivations of new parameters are shown clearly. The
(psi), respectively. The range of the data used also is tabulated in general trend of Q vs. T and vs. BS&W has been reached by keep-
Table 5. ing the other parameters constant. As shown in Figs. 3 and 4, pa-
The newly proposed correlation considered a parameter that rameters 1-BS&W/100, T/Tsc, BS&W, and T are considered. In
was not included in the previous correlations: BS&W. Tempera- Fig. 3, both trends are positive and the curves are ascendant, while
ture also is considered an effective parameter in this correlation. in Fig. 4, the Q-vs.-BS&W curve has a negative trend. As can be
These two parameters are in addition to other parameters that seen, increasing T leads to a considerable increase in BS&W com-
appeared in Gilbert (1954) and other correlations. To minimize pared to Q. Consequently, temperature is a significant parameter
error in field condition, this correlation is as follows: whose effect should not be neglected in deriving an accurate
D T E equation to calculate oil-flow rate.
PF SB 1 BS&W
Q A wh 100 TSC
; . . . . . . . . . . . . . . . 4
GORC Conclusion
where Pwh is wellhead pressure (psig); GOR is gas/liquid ratio On the basis of 182 actual data points, the available choke correla-
(scf/bbl); Q is gross-liquid-flow rate (bbl/d); S is choke size (1/64 tions were reviewed and one newly established correlation was
in.); T is temperature ( R); and BS&W includes free water, sedi- presented. The following conclusions are drawn:
ment, and emulsion, and is measured as a volume percentage of The proposed correlation has exhibited an average error of
the production stream. 2.89%, whereas current correlations such as Gilbert (1954) and
Average and absolute errors were computed using the equa- Ros (1960) present average errors of 60 and 160%, respectively.
tions as follows: The newly developed correlation considers a parameter that
was not included in the previous correlations: BS&W. This is in
Qtest Qcorrel addition to other parameters that appeared in Gilbert (1954) and
e . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Qtest other published correlations considered in this work, and is added
to minimize error in field condition.
And
Another effective parameter that was not included in previous
Qtest Qcorrel correlations is temperature. As can be seen, temperature has a
eabs : . . . . . . . . . . . . . . . . . . . . . . . .6
considerable effect on flow-rate prediction.
Q test
Empirical Coefficient
Pwh (psi) T ( F) Qo (B/D) BS&W (percent) GOR (scf/bbl) Choke Size (1/64 in.)
Pwh (psi) T ( F) Qo (B/D) BS&W (percent) GOR (scf/bbl) Choke Size (1/64 in.)
Pwh (psi) T ( F) Qo (B/D) BS&W (percent) GOR (scf/bbl) Choke Size (1/64 in.)
Q vs (T/TSC) Q vs T
Q vs (1-BS&W/100) Q vs BS&W
Fig. 3Flow rate vs. (T/Tsc) and vs. (1-BS&W/100) in log scale. Fig. 4Flow rate vs. T and vs. BS&W in log scale.
To analyze the effect of BS&W and T on flow rate, a parame- of Mining, Metallurgical, and Petroleum Engineers, Vol. 213,
ter-sensitivity analysis was performed. 202207. Dallas, Texas: Society of Petroleum Engineers.
The presented correlation undoubtedly can be used to control Ghareeb, M. and Shedid A. Shedid. 2007. A New Correlation for Calculat-
the water production and to reach the optimum production rate. ing Wellhead Production Considering Influences of Temperature,
GOR, and Water-Cut for Artificially Lifted Wells. Paper SPE 11101
presented at the International Petroleum Technology Conference,
Nomenclature Dubai, 46 December. http://dx.doi.org/10.2523/11101-MS.
A proportionality constant Gilbert, W.E. 1954. Flowing and gas-lift well performance. API Drilling
B bean or choke-size exponent & Production Practice 20 (1954): 126157.
C GOR exponent Jin, L. and Wojtanowicz, A.K. 2010. Coning Control and Recovery Improve-
D BS&W-term exponent ment Using In-situ Water Drainage/Injection in Bottom/Water/Drive Res-
e error ervoir. Paper SPE 129663 presented at the SPE Improved Oil Recovery
eabs absolute error Symposium, Tulsa, 2428 April. http://dx.doi.org/10.2118/129663-MS.
E temperature term exponent Mesallati, A., Biznati, M., and Mansouri, N. 2000. Multiphase-Flow Choke
F wellhead-pressure exponent Correlations for Offshore Bouri Oil Field (Correlations decoulement mul-
Pwh wellhead pressure, psig tiphase sur le gisement petrolier offshore de Bouri). Proc., International
Q gross-liquid-flow rate, bbl/d Gas Union 21st World Gas Conference, Nice, France, 69 June, 123.
Qcorrel calculated gross-liquid-flow rate from correlation, B/D Ros, N.C.J. 1960. An analysis of critical simultaneous gas-liquid flow
Qtest measured gross-liquid-flow rate, bbl/d through a restriction and its application to flow metering. Applied Sci-
S choke or bean size, 1/64 in. ence Research 9 (February): A-374.
T temperature, R
TSC standard condition temperature, R Mahmoud Safar Beiranvand is currently a graduate student in
reservoir engineering at the University of Tehran. His current
research focus includes water and aquifer management, fluid
flow in porous media, and production engineering and optimi-
References
zation. He holds a BSc degree in petroleum engineering from
Achong, I. 1961. Revised Bean Performance Formula for Lake Maracaibo the Petroleum University of Technology.
Wells. Internal report, Shell Oil Co., Houston.
Mohammad Babaei Khorzoughi is currently a graduate student
Al-Attar, H.H. 2010. New Correlations for Critical and Subcritical Two-
in mining engineering at the University of British Columbia. His
Phase Flow Through Surface Chokes in High-Rate Oil Wells. SPE current research interests are production and drilling engineer-
Proj Fac & Const 5 (1): 3137. SPE-120788-PA. http://dx.doi.org/ ing; reliability and maintenance engineering; and risk assess-
10.2118/120788-PA. ment in the oil, gas, and mining industries. He holds a BSc
Baxendell, P.B. 1958. Producing Wells on Casing FlowAn Analysis of degree in petroleum engineering from the Sharif University of
Flowing Pressure Gradients. In Transactions of the American Institute Technology.