Improved infeasible path optimization for sequential modular simulators—II: the optimization algorithm

Recently it was shown that chemical processes modeled by steady-state simulators could be optimzied without repeatedly converging the process simulation. Instead, optimization and simulation of the flowsheet can be performed simultaneously, along an infeasible path, thus leading to much more efficient performance.In this two-part study, we describe several improvements to this infeasible-path approach. This second article deals with theoretical and computational improvements to the optimization algorithm. Because recycle convergence and optimization occur simultaneously, the consequences of algorithmic failure can be severe, since little useful information is recoverable. Here we consider and improve several factors that affect the efficiency and robustness of the successive quadratic programming (SQP) optimization algorithm.The improvements are demonstrated on several small problems as well as on three chemical process problems. The results show significant improvement in performance as predicted from the theory.