A simplified economic model predictive control for a gas lifted oil production network based on modifier adaptation

Abstract

We propose a novel simplified economic model predictive controller (EMPC) for supervisory control of an offshore petroleum production network composed of two wells, a riser, and a separator. The proposed controller uses a gradient-and-modifier-based methodology to mimic a nonlinear EMPC (it considers a nonlinear process model and cost function), while only solving quadratic programming. We alter the modifier adaptation to depend on an estimator and employ slack variables to improve feasibility. Tips for appropriate design, to improve stability and robustness, are presented. Since the tested facility is subject to multiple disturbances, we also use a regulatory controller to stabilize the system. Finally, we compare our method to a linear EMPC and self-optimizing control in various scenarios considering disturbances, noise, and model inconsistencies. The proposed controller always tends towards the economic optimal point, presenting similar nominal performance as the others advanced controllers and better robustness.