Sequence model approximation approach for optimization and control of steady-state systems

Abstract

A new approach—a sequence model approximation method—is proposed for determining the optimum steady-state operating point of an industrial process. The method makes full use of first-order information of reality to construct an approximate model which can match real output and the required derivatives of reality at each iteration. In order to ensure that the real objective index descends, the Newton step and a one-dimensional minimization search technique are then used respectively. The approach's convergence and optimality conditions are investigated under mild assumptions. Simulation results show that the new approach requires fewer set-point changes, and exhibits a higher convergent rate than previous methods in this field.