Process control strategies for constrained nonlinear systems

Abstract: "Control and state variable constraints are often overlooked in the development of control laws. Although recent developments with DMC and QDMC have addressed this problem for linear systems, incorporation of state and control variable constraints for nonlinear control problems has seen relatively little development. In this paper, a nonlinear control strategy based on operator theory is extended to deal with control and state variable constraints. Here we show that Newton-type control algorithms can easily be generalized using an on-line optimization approach. In particular, a special form of a successive quadratic programming (SQP) strategy is used to handle control and state variable contraints.Moreover, the constrained approach can be shown to fit into the IMC structure proposed by Economou and Morari. Using the small gain theory of Zames, we show that constraints on state and control variables can replace the design of a nonlinear filter for noninvertible processes and thereby lead to stable controllers. The advantage to this approach, however, is that constraints are relatively straightforward to apply since the limits of their response are known in advance.After presenting a detailed description of the method, a number of scenarios are considered for the control of nonlinear reactor models. These models cannot in general be controlled using linear control strategies. With the nonlinear constrained approach, on the other hand, regulation and control can be done in a straightforward manner, and emphasis to more important parts of the system can be given with appropriate placement of constraints."