Nonlinear Model Predictive Control for Large Scale Systems

In the past decade the field of nonlinear model predictive control (NMPC) has witnessed steadily increasing attention from control practitioners. Its popularity comes from the fact that today’s processes need to be operated under much tighter performance specifications while at the same time more and more constraints, stemming for example from environmental and safety considerations, need to be satisfied. These increasing demands can only be met when process nonlinearities and constraints are explicitly considered in the controller design stage. Nonlinear predictive control, the extension of well established linear predictive control to the nonlinear world, appears to be a well suited approach for this kind of problems. One of the main difficulties that often permits NMPC from being applied in practice is the high online computational load: At each sampling instance a nonlinear constrained finite horizon optimal control problem needs to be solved numerically. In this paper we discuss how recent system theoretic results for NMPC, namely the use of the so-called quasi-infinite horizon approach to NMPC, can improve its applicability by allowing a reduction of the prediction horizon and thus the online computational load, without affecting closed loop performance and stability. With the use of a realistic process control example we demonstrate that even large scale problems can be solved using NMPC techniques if state of the art optimization techniques, that are properly adjusted for the receding horizon implementation, are combined with the quasi-infinite horizon technique.