Nonlinear model predictive control application for gas-lift based oil production

Abstract

Abstract This work investigates the application of Nonlinear Model Predictive Control (NMPC) using Hammerstein Models in an offshore oil production well with gas-lift. The problem is especially challenging from the control point of view because it presents regions with gain sign inversion. This reversion of the gain direction makes this problem difficult to control with the usual Model Predictive Control (MPC) or even with the traditional Proportional-Integral-Derivative (PID) controllers. Thus, it is proposed to investigate the use of methodologies applying NMPC based on Hammerstein model combined with the Moving Horizon Estimator (MHE) approach. The oil flow rate is controlled by the gas-lift flow rate through the gas-lift choke valve. The plant was described by a mathematical model including the gas-lift well, the flowline and the production choke valve. Some control scenarios were tested and the nonlinear model predictive controller was capable to successfully stabilize the process.