Date of Original Version
Abstract or Table of Contents
Recently, embedded simplified process models have been shown to be very efficient for process simulation. When compared to the direct use of rigorous models, this approach has the potential to reduce the computational effort of process simulation by up to an order of magnitude or more. Application of this approach to process optimization should therefore lead to similar savings in computational effort as well as substantial improvement of the process. However, current simplified model embedding schemes applied to process optimization cannot, in general, converge to the optimum defined by the more rigorous process models. Consequently, they require an expensive rigorous model optimization starting from the solution of the simplified model optimum to guarantee convergence.In this paper we develop a framework that incorporates simplified models into an optimization algorithm and guarantees convergence to the rigorous model optimum. Here rigorous process models are evaluated only when necessary to insure progress toward the optimal solution. A theoretical justification of the algorithm is presented and several process examples are solved to demonstrate the effectiveness of this approach.