Modelling and Extremum Seeking Control of Gas Lifted Oil Wells

This paper addresses the design of a perturbation-based extremum seeking control (ESC) method to maintain the oil production around the optimum point of the well-performance curve in gas lifted wells. The method uses periodic perturbation which is injected into the plant with intention to extract the information about the gradient of the well-performance curve. A simple nonlinear dynamic model is proposed and the essential dynamics of the Eikrem's model are captured, i.e., the transient behavior and the optimal steady state well- performance curve. Based on this simple model, a pre-compensation is developed which allows the application of the ESC scheme without reducing excessively the perturbation frequency. The control performance is evaluated via numerical simulations using an appropriate environment for modelling, simulation and optimization (EMSO) of process dynamic systems.