Adaptive model predictive control applied to submersible pump lifted wells

Abstract

Electric Submersible Pumps (ESPs) are one of the most widespread oil artificial lifting technologies for deepwater exploration. In the operation of an ESP there is a large number of parameters that must be monitored and held within operational constraints in order to guarantee stable and optimal operation. Model predictive control (MPC) is one of the strategies able to guarantee stable and optimal operation with constraints handling. Previous literature has proposed the use of linear MPC based on system identification, however no adaptive strategy has been employed to overcome system nonlinearities, instead a single internal model has been used. Moreover, all previous works have considered relevant system variables measurements to be available. In this dissertation, the problem of losing measurements of the state variables due to the aggressive subsea environment is addressed. We show that a non-adaptive single linear model strategy lacks in quality for state estimation and a robust MPC is not possible under this configuration. Therefore, an adaptive MPC coupled with Kalman Filter is proposed and three adapting strategies are compared. Two scheduling strategies based on linear interpolation of a set of local models are proposed and compared to successive linearization. All strategies guarantee internal model accuracy and stability over the whole operational range. The proposed scheduling strategies presented a similar performance compared to the successive linearization strategy, avoiding the need of obtaining a local linear model at each sampling time by interpolating among a number of linear models previously obtained by identification instead. In addition, a parameter estimation strategy is proposed and coupled to the MPC scheme so that measurement and model structural uncertainties are overcome.