OPTIMIZATION OF PSA SYSTEMS - STUDIES ON CYCLIC STEADY-STATE CONVERGENCE

Two types of trace separation systems are used as case studies to inve stigate different methods to accelerate cyclic steady state convergenc e. The Aitken, Muller and the more rigorous Broyden based updating sch emes seem to speed up the convergence rate considerably, at least for systems for which the method of direct substitution converges slowly. A damped Newton-based approach is also investigated, but since this me thod requires an updated Jacobian at each iteration step, the integrat ion and sensitivity calculation time is increased correspondingly. For the two systems considered here, the total calculation time is in fac t longer than in the direct substitution case. In the context of optim ization, the key factors to success seem to be how the test on the CSS convergence is performed, how the acceleration methods are initiated and how close the initial guess is to the optimal value. Each of the i nvestigated acceleration methods are preoptimized according to these a spects and the Broyden method definitely shows the best result. The Ai tken and Muller methods are rather simple and they do not seem reliabl e enough to be applied during an optimization process. (C) 1997 Elsevi er Science Ltd.