Optimizing gas-lift production of oil wells: piecewise linear formulation and computational analysis

Because the internal pressure of a reservoir can recover only a fraction of its oil reserve, system operators apply artificial lifting techniques to force the flow of oil to the surface. A favored technique is gas-lifting which consists of injecting compressed gas into the production tubing to reduce the weight of the oil column and, thereby, sustain a continuous flow of oil. The optimal allocation of a limited amount of compressed gas to wells gives rise to the gas-lift optimization problem, a mixed-integer nonlinear programming problem, whose decision variables determine which wells should produce and the rate of gas injected into the active ones. This paper investigates a piecewise linear formulation of the problem, presenting some properties of the polyhedron associated with the space of feasible solutions, and delivering families of valid inequalities that are shown to improve the performance of optimization software.