SPE 120628 Performance Analysis of ESP Systems in High-GLR Wells: From Lab Experiments to Practical Field Applications V.A. Elichev, R.A. Khabibullin, V.A. Krasnov, and K.V. Litvinenko, SPE, Rosneft, and M.G. Prado, SPE, University of Tulsa Copyright 2009, Society of Petroleum Engineers This paper was prepared for presentation at the 2009 SPE Production and Operations Symposium held in Oklahoma City, Oklahoma, USA, 4–8 April 2009. 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 This paper presents the results of field evaluation by Rosneft of correlations developed by the Tulsa University Artificial Lift Projects (TUALP) to predict the performance of the natural separation process and of Electrical Submersible Pumps (ESP) operating under multiphase flow conditions as well as Russian State University of Oil and Gas (RSUOG) correlations for the performance of rotary gas separators. The first two set of correlations were developed using data acquired with TUALP experimental facilities but have not been tested against field data. The correlations developed by the Russian State University were also accomplished using lab data and a field test validation was required. During the years of 2006 and 2007 Rosneft conduct several field tests with the purpose of evaluating the performance of several gas handling technologies for ESP installations. The valuable data from those tests enabled Rosneft to verify the accuracy and validity of TUALP and RSUOG correlations for natural separation efficiency, rotary gas separator efficiency and for ESP multiphase performance under real field operational conditions with real crude and natural gas. The TUALP-Marquez-Prado correlation was selected for natural separation efficiency, the correlations for rotary gas separators developed by RSUOG were used and the TUALP-Duran correlation was chosen for pump multiphase head degradation. An excellent match was obtained between the field test data and the correlations predictions with exception of the following cases: • Natural separation for installations with a horizontal ESP; • Multiphase flow head degradation in installations using gas handling devices. The paper presents a brief description of field data, a review of correlations for natural separation efficiency and pump head degradation; and a comparison between the predicted and measured performance. Introduction Electrical Submersible Pumping is a very important artificial lift method. The basic principle is the use of a down hole centrifugal pump that pressurizes the production stream reducing bottom hole flowing pressure. When the pressure decreases below the bubble point pressure, gas evolves out of solution from the liquid phase. If the amount of free gas present at the pump intake that is dragged into the pump is too high, significant operational problems can occur. Correct design of pumping system requires: • Prediction of how much free gas is dragged by the liquid phase into the pump • Prediction of the multiphase performance of the pump Many of Rosneft oilfields have a high bubble point pressure (higher then 1000 psi). It is estimated that up to 15% of oil production in the near future will be produce from those oilfields. ESPs are the most attractive artificial lift method for Rosneft, with more than 8000 wells currently producing by this method. It is very important for Rosneft to be able to predict correctly the performance of wells producing with ESPs under multiphase flow conditions. Tulsa University Artificial Lift Projects (TUALP) has been conducting research in the area of natural separation and multiphase performance of the ESPs for many years. TUALP experimental facilities have allowed researchers to investigate the performance of artificial lift equipment in near field operational conditions. The results from those experiments result in several models and correlations available to be used by the industry. Rosnef is currently one of the supporting members of TUALP as well as it companion project, The Tulsa University Fluid Flow Projects (TUFFP). Rosneft has also worked in collaboration with Russian State University of Oil and Gas (RSUOG) in developing correlations for the performance of rotary 2 SPE 120628 gas separators. This University also has an experimental facility that was used successfully to test many rotary gas separators under close to real operational conditions. In order to validate the correlations and models developed by TUALP and by RSUOG, several field tests were conducted at Purneftegas in North-Western part of Siberia. Most of Purneftegas wells have high GOR, so sizing problem for multiphase conditions is very important for those wells. Correlations for natural separation The natural separation process has been investigated by many researchers in Russia (Lyapkov) and the US (Alhanati, Serrano, Harun, Marquez, etc.). Many correlations and mechanistic models have been developed to describe and quantify this process. Alhanati developed a simple theoretical model to predict the natural separation efficiency of ESP systems incorporating a rotary gas separator. The two main assumptions in the model are: a uniform void fraction from the motor section up to the gas separator's gas outlet ports, and a no-slip condition between the gas and liquid phases within the control volume immediately preceding the gas separator's intake ports. The model is strictly limited to the vertical configuration only. Serrano conducted an experimental study of natural separation in an ESP system as a function of operating conditions and well inclination angle. Using a water and air fluid system, Serrano gathered 81 data points covering a maximum void fraction of 20 percent, liquid rate of about 2000 B/D, pressure of 150 psi, and inclination angles of 30 and 60 degree from horizontal. This data enabled the extension of Alhanati’s model. Harun et al. developed a simple model to predict the natural separation efficiency in vertical pumped wells. The model was developed based on the combined phase momentum equations and a general slip closure relationship applied to a single control volume in front of the pump intake ports. The model assumes no slip between the liquid and gas phases at the pump intake ports and a uniform void fraction across the annulus. Empirical correlations were developed to calculate the drag coefficient for all void fractions; however the model was limited to the vertical configuration only. Marquez and Prado presented a model to predict natural separation efficiency, based on the drift-flux model approach. The model considers the effect of the slip velocity in the radial direction, variable neglected in previous simplified models. The empirical data allowed the authors to obtain a correlation for the slip effects in the radial direction. Good agreement of this model with the experimental data showed the important effect that the slip velocity in the radial direction has in the prediction of natural separation Currently, commercial software allows using only Alhanati’s correlation developed in the early nineties with only lab work validation. After a survey of the most recent theoretical work, the Marquez mechanistic model resulted as the best option to be used, since it describes the physics of natural separation better than other models and correlations as can be seen in Fig. 1. Correlations for multiphase performance of ESP Modeling the pump performance under two-phase flow conditions is a challenging and complex problem. The homogeneous model commonly used in the petroleum industry is only applicable at very high liquid flow rates or very low gas flow rates, when there is sufficient liquid drag on the small bubbles to ensure a homogeneous mixture inside the impeller. When the drag force is not sufficient to overcome the pressure force exerted by the rotating liquid centrifugal field the gas bubbles start to slip in relation to the liquid in the radial direction. This slippage effect increases the local gas void fraction reducing the mixture density and causing a reduction in the head as compared to the predictions from the homogeneous mode. If the slippage is too big than the bubbles start to coalesce forming bigger bubbles that tend to stay stagnant at the impeller intake inducing an unstable behavior known as surging. Extensive work has been conducted both in the nuclear as well as in the petroleum industry to obtain experimental data on the performance of ESPs. Mathematical models have not been successful in predicting the phenomena due to the necessity of local closure models for turbulence, bubble size, bubble shape, drag coefficient that requires measurement of local variables in a rotating impeller. Therefore most of the available predictive methods are based on correlations developed based on laboratory experimental data. Sun and Prado developed a hybrid model where a rigorous mathematical formulation was used to describe the two phase flow inside impellers and diffuser. The model considers pump geometry effects but can only be closed and solved with an experimental correlation obtained at TUALP facilities. TUALP - Duran presented a very interesting and simplified method based on an empirical correlation that was based on experimental data obtained by TULAP. Duran correlation was selected to model multiphase pump performance in this work due to the lack of pump impeller and diffuser geometric information to use Sun and Prado model. Correlations for gas separator performance Artificial separation was investigated by Alhanati and Harun of Tulsa University in the late 1990’s– early 2000’s. Three gas separators were tested and a mechanistic model was proposed. Similar research was carried out by some leading Western SPE 120628 3 pump manufactures. It has been elicited that artificial separation efficiency is affected by the following parameters: • Two phase flow performance of the separator inducer – a component of a gas separator that creates additional pressure inside the separator to remove extra gas into the annular space • Separator GLR at the intake • Separator load or liquid flow rate • Separator rotational speed The inducer performance depends on the construction type of the separator. The rest of the factors may be seen as external to the separator. It is therefore expedient in bench tests to vary the fluid rate, the gas rate, and the rotor frequency to obtain correlations between the separation efficiency and those parameters. To perform the testing, the experimental facility at I.M. Gubkin Russian State University for Oil and Gas was used (see Fig. 4). The tests used the water-surfactant-air fine mixture made by the ejector. The separators were run parallel with a 12-stage ESP 5-125. The facility was equipped with a frequency modulator capable of varying current frequency from 20 Hz to 90 Hz. During the experiments, data for separation efficiency with different GOR values at the intake, fluid rates and motor rotation frequency were obtained. The testing resulted in determining the characteristics of 20 Russian- and American-made gas separators. After processing the test results, a number of regularities were induced: 1. The separation efficiency in the main working zone is linearly dependent on the rate and tends to become non-linear at minimum and maximum rate values 2. The dependence of separation efficiency on intake GOR is described by a convex function 3. For the frequency, the correlation is close to linear. Depending on separator configuration, the function can be either ascending or descending In analyzing the obtained data, it was found that efficiency in all the separators changes depending upon the fluid rate, mixture GOR at the pump intake and the rotation frequency of the motor shaft by some correlations. It was, however, observed that individual measurements can be grossly different from the immediate values. It was therefore decided to carry out a mathematical processing of the bench test results to obtain an alteration function of separation efficiency as dependent on intake GOR, fluid rate, and rotation frequency. The class of parametric functions was defined to that end, which describes the behavior of efficiency, followed by solving an optimization problem of finding parameter sets for each separator. The problem boiled down to finding a surface in a 4D space with the maximum deviation of the real values no higher than 5%, with the minimum number of parameters. To solve this problem, a genetic algorithm was applied which resulted in finding a 12-parameter function: f( Q , β , F ) = θ 1 β 4 + θ 2 β 3 + θ 3 β 2 Q + θ 4 β 2 + θ 5 β Q + θ 6 β F + θ 7 Q 2 + θ 8 Q F + θ 9 β + θ10 Q + θ11 F + θ12 As a result, each separator can be characterized by only 12 parameters (Figure 3). Field testing Field measurements of separation efficiency were performed under Rosneft’s New Technologies System on oil fields operated by Purneftegaz. A total of 78 complex measurements were made at 41 wells. There were wells with conventional ESP and wells with specially designed ESP (gas handlers, axial-multiphase pumps). The submersible equipment was tested, with all parameters of this equipment being registered. A movable multiphase flow meter was used to determine the gas liquid ratio and phase flow rates at the surface. The readings of the multiphase flow rate meter were checked in the process of testing against standard OOO “RN-Purneftegas” phase flow rate meters (mass measuring units ASMA and AGZU “Sputnik”) (typical wiring diagram of measuring units is shown in fig.4). Electronic pressure gages-thermometers installed at the flowing APMa were used to measure temperature and pressure at the well head as well as the dynamic level of the liquid in the annulus. The equipment supplied with the submersible telemetric instruments (pressure and temperature transducers at the pump intake, for some pumps pressure and temperature transducers at the pump outlet) was used to control well parameters. It permits to evaluate the stability of ESP operation in time, pressure increment developed by ESP, as well as to specify the calculations of pressure distribution in the tubing using hydrodynamic two phase flow pressure drop correlations. Separation measurements The total efficiency of gas separation for a well can be calculated through individual measurements of gas rate through the tubing and through the annular space. The efficiency will amount to 4 SPE 120628 ηtotal _ sep = СУ Qtubing СУ Qtubing + QcaСУsin g СУ where Qtubing is the rate of gas flowing out of the separator and through the ESP, reduced to standard conditions (m3 per day). This rate can be calculated using the tubing gas rate measured at the well head, considering that in the pumping conditions (given the pressure and temperature equal to those at the pump intake) part of the gas measured on the surface was dissolved in СУ oil as a free phase. Qca sin g is the rate of the gas travelling through the annulus, reduced to standard conditions (m3 per day). СУ СУ СУ СУ Given the total rate of the well, Qca sin g can be defined as Qtotal − Qtubing where Qca sin g is the total gas rate through the tubing and the annulus measured on the surface and reduced to standard conditions. To measure the gas, fluid, and water cut rates mobile multiphase flow meter was used. These measurements were confirmed by Purneftegaz’s standard measurement tools – the Sputnik meter station and the mass meter unit ASMA-T. The viability natural separation model was verified at wells equipped by ESPs with no separators. Different gas rates and percentages were produced at the intake by changing pump frequency by means of a frequency modulator. Out of the models analyzed, the Marquez mechanistic correlation appeared to be the best fit (Fig. 5). Separation efficiency for wells having separators was calculated as follows: • Natural separation was calculated using the Marquez model; • Rotary gas separation was calculated using the separator correlation; • The total efficiency can be determined as: η total _ sep = η natural _ sep + η artificial _ sep (1 − η natural _ sep ) Figure 6 shows that in most cases the estimates of the separation efficiency have degree of precision applicable to practical applications (20%). The individual points on the graph correspond to well measurements. A detailed analysis of outlier deviations showed that some equipment or well problems had been present during the tests. As a conclusion, the correlation of gas separation efficiency obtained from laboratory bench-test data, together with the mechanistic correlation to calculate natural separation, can be used in practical applications for selecting and analyzing ESPs equipped with a rotary gas separator. Pump multiphase performance measurements Measured head was found using pressure intake sensor and pump discharge pressure. Some pumps had pressure discharge sensor as well and in this case this information was used. For pumps without pressure discharge sensors a unified multiphase flow correlation, developed by TUFFP, was used for estimating discharge pressure from the reading of the wellhead pressure sensor. It was found on a previous study of natural flowing wells that TUFFP unified correlation has a very good performance against real data for Purneftegas conditions. Duran correlation was used for estimating pump head under two phase flow conditions. The comparison of the results of testing of standard electric centrifugal pumps with his theoretical calculations shows a good matching. (Fig.7). The calculations do not match the measurements for electric centrifugal pumps equipped with special devices, because Duran correlation was developed for conventional pumps. Two tests indicates bad match between measured and calculated head for conventional ESP. Analysis showed us, that those were old pumps, that could be deteriorated by wear due to long operational life. Conclusions The following has been concluded from the investigation results: The academic work conducted by The University of Tulsa Artificial Lift Projects (TUALP) and Russian State University of Oil and Gas (RSUOG) resulted in correlations and mechanistic models that are very useful in predicting the performance of ESP in high GLR wells. This work showed that those correlations and models can be used to successfully analyze the performance of wells operating under real conditions. As a consequence those models should also be used in de selection and design of ESP installations. Those correlations and mechanistic models provide a more fundamental engineering tool for this problem and have a much stronger basis than rules of thumb. In particular: Marquez (TUALP) mechanistic correlation for down hole natural separation has a very good performance against field data. Total separation efficiency derived from natural and artificial separations is a good approximation to solve practical problems, for example while selecting and analyzing ESPs. Duran (TUALP) correlation could be used for estimating conventional ESP multiphase performance SPE 120628 5 RSUOG correlations for rotary gas separator performance can be used in real practical applications References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Lyapkov P. D., Dunaev V. V. Results of ЭН-160-800 Pump Testing in the well with gas content in the fluid produced – Oil Economy, 2, 1960 Lyapkov P. D., Doroshuk N. F., Zlatkis A. D. Results of testing of submersible centrifugal pump operation in oil and gas-oil mixtures. – Tatar Oil, 4, 1962 Cirilo R.: Air – Water flow through electric submersible pumps, MS thesis, the University of Tulsa, Oklahoma? 1998 Lyapkov P. D. On water-air mixtures flows through the channels of the working parts of the centrifugal pump. – Chemical and Gas Engineering, 10, 1968 B. Hirth, J. Curfew, R. Waygood, P. Julstrom, J. Miller, S. Ossia, C. Ollre. Short Story: Case Study – Performance Evaluation of a Helico-Axial Multiphase Pump in a CO2 Flood. SPE ESP Workshop 2005 K. Boerger, A. Cooke Short Story: Production Of Highly Gassy Wells Utilizing A Helico Axial Multiphase Gas Handler SPE ESP Workshop 2005 Lea, J.F. “Effect of Gaseous Fluids on Submersible Pump Performance”, paper SPE 9218 presented at the 1980 SPE Annual Technique Conference and Exhibition, Dallas, TX. (Sep. 21-24). Beltur, R., 2003, ‘‘Experimental Investigation of Performance of Electrical Submersible Pumps in Two-Phase Flow Condition,’’ M.S. thesis, The University of Tulsa, Oklahoma. Pessoa, R., 2001, ‘‘Experimental Investigation of Two-Phase Flow Performance of Electrical Submersible Pump Stages,’’ M.S. thesis, The University of Tulsa, Oklahoma. Sachdeva, R., Doty, D. R., and Schmidt, Z., 1994, ‘‘Performance of Electric Submersible Pumps in Gassy Wells,’’ SPE Prod. Facilities., 9, pp. 55–60. Sun, D., and Prado, M. G., 2003, ‘‘Modeling Gas-Liquid Head Performance of Electrical Submersible Pumps,’’ Ph.D. dissertation, The University of Tulsa, Oklahoma. Romero, M. “An evaluation of an electrical submersible pumping system for high GOR wells”. Thesis for Master of Science, Tulsa University, 1999. Marquez, R.A. and Prado, M.G., “A new correlation for predicting natural separation efficiency”. Southwestern Petroleum Short Course. April 16-17, 2003. Lubbock Memorial Civic Center, Lubbock, Texas Duran, J; Prado, M.G.: “ESP Stages Air-Water Two-Phase Performance - Modeling and Experimental Data”, paper SPE presented at the 2004 SPE ESP Workshop. Drozdov A. N., Verbitzky V. S., Dengaev A. A., Litvinenko V. A., Khabibullin R. A. Litvinenko K. V., Elichev V. A. Results of Investigation of Submersible Centrifugal Gas Separator Operation in OOO “RN-Purneftegas” Wells with high gas content at the intake. Comparison of Bench and Field Tests. SPE 117415. Russian Federation Patent No. 2075656. Method of testing hydraulic machines and the corresponding electric motors, and a bench facility / Authors: A.N. Drozdov, L.A. Demyanova. – Moscow, F 04 D 13/10, F 04 F 5/54, F 04 B 51/00, submitted 14.03.1995, published 20.03.1997, Б. И.№ 8. Richard Márquez, “Modeling down hole natural separation”, а dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Discipline of Petroleum Engineering, The Graduate School. The University of Tulsa – 2004. Bedrin V.G., Khasanov M.M., Khabibullin R.A., Krasnov V.A., Pashali A.A., Litvinenko K.V., Elichev V.A., Prado M. Comparison of Technologies for Operation of Electric Centrifugal Pumps at High Gas Content at the Pump Intake on the Basis of Field Tests. SPE 117414. 6 SPE 120628 0,35 separation efficiency,% 0,3 0,25 0,2 0,15 0,1 0,05 0 0 0,2 0,4 0,6 0,8 Watercut,% Marquez correlation Lyapkov correlation Marquez mechanistic model Fig. 1. Natural separation efficiency against watercut Fig. 2. RSUOG bench test facility 1 1,2 SPE 120628 7 Fig 3. Correlation for gas separator efficiency Pad 261 Well # 255 Layout В коллектор PhaseTester ACMA-T 03-400-300 Задвижка закрыта во время замера Задвижка открыта во время замера Направление потока жидкости Штуцерная камера Обратный клапан Fig. 4. Typical connection diagram of measuring instruments to measure well operation 8 SPE 120628 30 Separation efficiency, deviation from actual data, % 25 20 15 10 5 0 Regim e 1 Regim e 2 Regim e 3 Regim e 4 Regim e 5 Regim e 6 Average deviation, % -5 -10 -15 -20 Serrano Marquez mechanistic model Marquez correlation Lyapkov correlation Fig. 5. Comparison of different natural separation models for various oil well regimes Fig 6. Comparison of measured against calculated separation efficiency SPE 120628 9 Fig 7. Comparison of measured against calculated dimensionless head
