Abstract
Gas lift is one of the most frequently used artificial lift methods in petroleum industries. The principle of this method is to maintain low well bottomhole pressure by injecting high pressure gas into the annulus and tubing, in order to enable the reservoir fluid to be delivered to the surface. Dual continuous gas lift is one type of gas lift installation, where oil is produced from two different oil formations in a single well system. Based on the field experiences, excessive gas usage frequently occurs in oil wells, which use dual gas lift well installation. Since the excessive injected gas could produce lower oil or liquid production or could create instability of the flow, the proper rate of gas injection has to be predetermined. The relationship between gas injection rate and liquid production rate is represented by gas lift performance curve (GLPC), which could be obtained implicitly from a two-parameter family ordinary nonlinear differential equation, which represents steady flow equation along the tubing, satisfying wellhead pressure and bottomhole pressure as boundary conditions. In a dual gas lift system, GLPCs are defined either for short-string or long-string, which represents oil production performance for each string, respectively. During the mature state of gas lift wells, the production instability usually occurs, and it becomes an important problem in a dual gas lift operation. The instability problems come from hydrodynamic instability that results in cyclic variations in wellhead pressure, oil production, and gas injection rate. Operating a well under this condition should be avoided because it has several disadvantages. The instability problems could be predicted, and the proper gas injection rate for each string is determined; therefore the stability conditions could be maintained. Gas allocation optimization for a dual gas lift system could be considered as maximization of a nonlinear function, which represents the total oil production for both strings. In this work, the gas injection rates are determined for each string, which yields maximum total oil production rate for the well, in the domain to satisfy the production stability criteria. Here, a numerical scheme using genetic algorithm is developed to find the optimum gas injection rate. This is a new approach since in the former approaches, the GLPCs are usually estimated using the curve fitting method. The results of optimization gas injection allocation using genetic algorithms are presented for given field data.
