Infeasible path optimization with sequential modular simulators

Abstract

Abstract Infeasible path optimization algorithms have been effective with equation‐solving or simultaneous modular process simulators. Using these simulators, successive quadratic programming is applied to large nonlinear programs. However, the algorithms are not directly applicable to commercial simulators of the sequential modular type. Here an infeasible path algorithm is presented which can easily be integrated with most commercial process simulators. The optimization problem remains small (10 or 20 variables) and the user merely replaces the convergence block with an “optimization” block. A simple process example demonstrates the operation, effectiveness, and potential of this algorithm.