Study optimizes gas lift in Gulf of Suez field

A.A. Abdel-Waly
King Saud University
Riyadh, Saudi Arabia

T.A. Darwish, A. Osman Salama
Cairo University
Giza, Egypt

M. El-Naggar
Gulf of Suez Petroleum Co.
Cairo, Egypt

A study using PVT data combined with fluid and multiphase flow correlations optimized gas lift in the Ramadan field, Nubia C, oil wells, in the Gulf of Suez.

Selection of appropriate correlations followed by multiphase flow calculations at various points of injection (POI) were the first steps in the study.

After determining the POI for each well from actual pressure and temperature surveys, the study constructed lift gas performance curves for each well. Actual and optimum operating conditions were compared to determine the optimal gas lift.

The study indicated a net 2,115 bo/d could be gained from implementing its recommendations. The actual net oil gained as a result of this optimization and injected gas reallocation was 2,024 bo/d.

Ramadan field

Ramadan oil field is 5 miles north of Gulf of Suez Petroleum Co.'s (Gupco) Ras Shukheir base. At the time of this study, 19 gas lift wells were producing about 17,000 bo/d.

The study was divided into three phases.

In the first phase, pressure-volume-temperature (PVT) data determined the best correlations for fluid property and multiphase flow for calculating pressure losses. Different PVT correlations were evaluated against reported PVT data of oil from the Nubia C reservoir.

The second phase evaluated, for each oil producing well, the best multiphase flow correlation to calculate the

pressure losses in the production tubing. Actual multipoint downhole pressure and temperature surveys from all wells were used as a comparison.

In the third phase, the selected fluid property correlations and the best vertical multiphase flow correlations were applied to determine the expected production rates for injected gas rates at various POIs.

After each well's POI was determined from actual pressure and temperature surveys, a gas lift performance curve for each well was constructed. A comparison between the actual and optimum operating conditions determined the most appropriate way to optimize the gas lift.

Fluid properties

A Fortran PVT program calculated the fluid properties at pressures and temperatures different from the laboratory analysis.

The fluid properties obtained were:

• Oil density, d_o
• Bubble point pressure, P_b
• Solution gas oil ratio, R_s
• Oil formation volume factor, B_o
• Oil viscosity, _o
• Gas compressibility factor, z-factor
• Gas formation volume factor, B_g
• Gas viscosity, _g.

Laboratory analysis of a bottom hole sample of crude oil collected from well Ramadan 5-17 on Jan. 1, 1984, provided the measured fluid properties. Fluid parameters input in the PVT program are API gravity, initial gas in solution, gas specific gravity, fluid temperature, and fluid pressure.

For each fluid property, measured laboratory data indicated the accuracy of each correlation. Fluid parameters were 32 API gravity, 0.945 gas specific gravity, 434 scf/st-tk bbl initial gas in solution, and 295 F. reservoir temperature.

Fluid property correlations evaluated include:

• Bubble point pressure-Standing,¹ Lasater,² Vasquez and Beggs,³ and Glas ⁴
• Oil density-Standing, and Vasquez and Beggs.
• Solution gas oil ratio-Standing, Lasater, and Vasquez and Beggs
• Oil formation volume factor-Standing, and Vasquez and Beggs
• Oil viscosity-Beggs and Robinson,⁵ and Vasquez and Beggs
• Z-factor-Hall and Yaborough⁶
• Gas formation volume factor-real gas equation of state⁷
• Gas viscosity: Lee, et al.⁸

To compare the calculated fluid properties with the measured values, statistical measures⁹ of error were determined with the following equations:

Percent error = ((Measured value - Calculated value)/Measure value) 3 100

Average percent error, Ew = (^EEi)/n

Average absolute error, |Ew| = (^|EEi|)/n

Standard deviation = (((Ewi - Ei)2)/(n-1))0.5

The n is the number of observations.

Multiphase flow

A computer program for production system analyses determined the vertical multiphase flow in the wells. Vertical multiphase flow correlations examined included:

• Hagedorn and Brown¹⁰
• Duns and Ros¹¹
• Orkiszewski¹²
• Beggs and Brill¹³
• Aziz¹⁴

All vertical multiphase flow correlations required the well profile, production test data, and surveys of bottom hole pressure and temperature.

After these data were input into a modeling computer program, the program calculated pressure-vs.-depth for the different vertical multiphase correlations. Pressure and temperature sonds (Amerada) obtained the actual measured data in wells under flowing conditions at various depths.

The flowing surveys typically recorded temperature and pressure at 50 ft depth stations above and below the gas lift valves. These actual pressure and temperature surveys determined the POI.

The calculated pressure data are determined from the pressure traverse with the previously mentioned multiphase flow correlations at the same flowing conditions and depth stations as the actually measured data.

Actual production tests at the time of the flowing survey provided the production data of flow rate, produced

gas/liquid (GLR) ratio, and injected GLR ratio.

Well bore inclination angles were obtained from directional surveys taken while drilling. The actual tubing sizes in the well bore were used in the calculations.

Measured pressures compared with calculated pressures for different multiphase flow correlations indicated the best correlation for optimizing production.

The following statistical error measure compared the calculated pressure values with the measured values:

Absolute gradient error = (DPmeasured - DPcalculated)/DPmeasured

where: DP = P(i + 1) - Pi

Production optimization

At any given time in a well, two pressures always remain fixed and do not vary with the flow rate. One is the average reservoir pressure, Pr, and the other is the system outlet pressure.

The outlet pressure is usually the separator pressure, Psep, but if the well is controlled by a surface choke, as in the Ramadan field, the fixed outlet pressure is the wellhead pressure P_wh.

Data needed by the computer program to optimize the lift gas included well profile, production test data, bottom hole flowing pressure at the test rate, average reservoir pressure at mid-perforation or assumed datum, and surface lift gas pressure.

The well profile, the production test data, and the bottom hole flowing pressure were measured and defined as illustrated before. The average reservoir pressure was measured by pressure build-up or bottom hole pressure surveys.

The reservoir pressure is always measured at an arbitrary datum, such as the mid-perforation depth. The surface lift gas pressure was measured on the casing at surface.

These data were input into the same computer program that calculated production rate-vs.-gas injection rate using the best fluid properties and the vertical multiphase flow correlations. The aim was to determine the optimum well performance, which is the maximum production rate with minimum injected lift gas.

Well performance depends on production tubing size and injected lift gas rate. However, because the production tubing already is in all wells, its sizes were used in the optimization study. The only parameter left for optimization was the injected lift gas rate.

For the existing gas lift wells, the following steps determined the GLR and gas injection depth to produce maximum oil rate or maximum revenue:¹⁵

1. Define the best fluid properties and multiphase flow correlations.
   
2. Assume a value for GLR and different values of production rates.
   
3. For each production rate, calculate the pressure traverse, P_t, inside the tubing segment L.
4. Calculate the casing gas pressure, P_g, for the tubing segment h.
5. Compare the values of Pt and Pg. If P_t = P_g - DPvalve, set GLR = FGLR (formation GLR) for pressure traverse calculations until (SL) = Depth to mid-perforations.
6. Calculate the values of injection gas depth, h, and P_wf corresponding to the rate Q.
7. Change the value of GLR and repeat calculations from Step 2 to Step 6.
8. Plot Q-vs.-P_wf at different values of GLR, outflow curves.
9. Construct the inflow performance relationship (IPR) curve for the well, inflow curve.
10. Define the points of intersection between the outflow curves and the inflow curve, Q and GRL.
11. Construct the optimization curve, rate-vs.-injected gas, and determine the optimum production rate and the optimum injected gas.
12. At optimum production rate and optimum injection gas, determine the gas injection depth h.

Flowing pressure surveys determined the actual gas injection depth. The production rate and the injected lift gas rates were calculated at the optimum point of gas injection, and at the actual point of gas injection. A comparison is drawn between the two values to determine the optimum well operation.

Results

Figs. 1 and 2 show the reservoir fluid properties of the Nubia C formation. The best matches to measured values were obtained from the following correlations:

• Bubble point pressure (Fig. 1 [22166 bytes])-Glaso
• Oil density (Fig. 2a [38526 bytes]) - Standing
• Solution gas-oil ratio (Fig. 2b [38526 bytes])-Glaso
• Oil formation volume factor (Fig. 2c [38526 bytes])-Beggs and Robinson correlation for saturated oils, and Vasquez and Beggs for undersaturated oil.
• Oil viscosity (Fig. 2d [38526 bytes])-The calculated oil viscosity greatly overestimates the measured value for both correlations. However, the Vasquez and Beggs correlation is used for oil viscosity calculations because the Ramadan field, Nubia C reservoir, is undersaturated.
• Gas formation volume factor (Fig. 2e [38526 bytes])-The Z-factor was calculated with Hall and Yarborough correlation.
• Gas viscosity (Fig. 2f [38526 bytes])-Lee, et al.

• Beggs and Brill for nine wells
• Hagedorn and Brown for three wells
• Both Beggs and Brill, and Hagedorn and Brown for five wells.
• Duns and Ros for one well
• Orkiszewski for one well.

Fig. 3 [26668 bytes] illustrates the vertical multiphase results for Wells R3-3 and R6-40. Tables 1 and 2 present the calculated results for the two wells. Table 1 [40356 bytes] and Fig. 3a [26668 bytes] show that for well R3-3 above the POI, the Hagedorn and Brown method yields minimum error. Below POI, the slopes of the calculated pressure using all the correlations are the same and are close to the measured pressure curve.

The high error values above POI result from multipoint lift gas injection.

Table 2 [42464 bytes] and Fig. 3b [26668 bytes] show that for Well R6-40 the Beggs and Brill method yields the minimum error above the POI, while all correlations have the same slope below the POI and are close to the measured pressure curve.

Results for all wells indicate the following conclusions:

• For GLR < 400 scf/st-tk bbl, all correlations can be used for pressure drop calculations.
• For GLR 400 scf/st-tk bbl, Beggs and Brill is the best, but Hagedorn and Brown can also be used.

Fig. 4 [26557 bytes] and Table 3 [38248 bytes] show the calculated production rates-vs.-injected gas rates at actual POI and optimum POI for wells R3-3 and R6-40. For well R3-3, the actual fluid production point is 1,035 b/d and the injected gas is 1.18 MMscfd, while the optimum fluid production point is 1,203 b/d and the injected gas is 1.38 MMscfd.

It is obvious that the gain in the fluid rate is 168 b/d (138 bo/d) with 0.2 MMscfd more injected gas. For well

R6-40, the gain in fluid rate is 59 b/d (6 bo/d) with 0.98 MMscfd less injected gas.

Table 4 [58744 bytes] summarizes the results of the optimization study. From this table it is clear that the excess gas required to produce the wells at optimum rate is only 2.64 MMscfd injected gas, and the corresponding increase in oil production is 2,115 bo/d.

Table 5 [29952 bytes] shows the wells arranged in descending order of production rates. Well R6-32 is producing with the highest and well R6-40 with the lowest oil production rate per MMscf injected gas.

The injected gas required for optimum production, 2.64 MMscfd (Table 4 [58744 bytes]), can be obtained from wells R6-40 and R1-29.

By shutting in well R6-40, 2.37 MMscfd injected gas will be saved and 48 bo/d will be lost; and by reducing the injected gas for well R1-29 to save 0.27 MMscfd the oil production will be reduced by 44 bo/d.

The total oil lost to save the amount of gas needed for optimum production is 92 bo/d. The net oil gained as a result of this optimization and injected gas reallocation is 2,024 bo/d. The optimum average oil production rate per MMscf injected gas is 556 st-tk bbl. This is compared to 535 st-tk bbl/MMscf under current actual conditions.

Acknowledgment

The authors are grateful to Gulf of Suez Petroleum Co. for the data and support.

References

1. Standing. M.B., "A Pressure-Volume-Temperature Correlation for Mixtures of California Oil and Gases," Drilling and Production Practices, API, 1942.

2. Lasater, J.A., "Bubble Point Pressure Correlation," Trans. AIME, Vol. 213, 1938, pp. 379-81.

3. Vasquez. M., and Beggs. H.D., "Correlations for Fluid Physical Prediction," JPT, June 1980, pp. 968-70.

4. Glas , O.S., "Generalized Pressure-Volume-Temperature Correlations," The Foundation of Scientific and Industrial Research Technology, N-7034, Trondheim, Norway.

5. Beggs. H.D., and Robinson, J.R, "Estimating the Viscosity of Crude Oil Systems," JPT, September 1975, pp. 1140-41.

6. Hall, K.R., and Yarborough, L., "How to solve equation of state for z-factors," OGJ, Feb 18,1974, pp. 86-88.

7. Dake, L.P., Fundamentals of Reservoir Engineering.

8. Lee, A.L., Gonzalez, M.H., and Ealin, B.E, "The Viscosity of Natural Gases," JPT, 1966.

9. Murray, R.S., Theory and Problems of Statistics, McGraw-Hill Co., New York, 1972

10. Hagedorn, A.R., and Brown. K.E., "Experimental Study of Pressure Gradient Occurring During Continuous Two Phase Flow in Small Diameter Vertical Conduits," JPT, April 1965, p. 475.

11. Duns, H., and Ros, N.C.J., "Vertical Flow of Gas and Liquid Mixture in Wells," 6th World Petroleum Congress, Frankfurt, pp. 43l-55.

12. Orkiszewski, J., "Predicting Two-Phase Pressure Drops in Vertical Pipe," JPT, June, 1976, pp. 829-38.

13. Beggs, H.D., and Brill, J.P., "A Study of Two-Phase Flow in Inclined Pipes," JPT, May 1973, pp. 607-17.

14. Aziz, K., Govier, G.W., and Fogarasi, M., "Pressure Drop in Wells Producing Oil and Gas," Journal of Canadian Petroleum Technology, July-September 1972, pp. 38-48.

15. Brown, K.E., The Technology of Artificial Lift Method, Vol. 4, PennWell Publishing Co., Tulsa, 1984.

The Author