Nonlinear Model Predictive Controller for Electrical Submersible Pump Lifted Wells

Artificial lifting utilizing electrical submersible pump is widely used to increase oil production from wells. A suitable control system is required in order to increase efficiency of ESP system and to avoid damage to pump and to increase safety of ESP production. Automatic control of ESP unit is not a trivial task due to the high number of parameters and variable involved. Many methods have been proposed for control of ESP lifted wells. Most methods rely on a linear model of ESP lifted well and controller designed based on this linear approximation. Linear model and controllers can fail if the process undergoes large changes in operating conditions or huge disturbances are introduced to process. Other methods solve dynamic equations governing the ESP lifted well operation. Although these methods are highly accurate; however, they are computationally expensive, and they cannot be implemented on conventional control systems. In this paper, a nonlinear dynamic model is developed for ESP lifted well. The model is then utilized inside a Nonlinear Model Predictive Control (NMPC) System. The developed model and controller performance is then tested and assessed under various scenarios. The developed controller performance shows proper reference tracking and disturbance rejection properties while the process constraint are completely satisfied.