In this paper

a new infeasible path optimization method for complex modules is developed. It differs from the existing infeasible path optimization method in that it does not need the convergence of calculation for each round of iteration in complex modules. The problems of selecting the initial values of trial variables and scaling factors are also discussed and the algorithms for their selection are proposed separately. The new method has been successfully applied to the optimization of the flowrate of circulating water for the CO shift section in ammonia synthesis.