Globally optimal reactor network synthesis via the combination of linear programming and stochastic optimization approach

Abstract

Reactor network synthesis based on superstructure optimization is often a complex non-linear programming problem (NLP), which is very difficult to solve by means of the traditional optimization approaches. To solve this problem, a double-level new optimization method which combines linear programming and stochastic optimization approach is proposed in this paper. In addition, a superstructure network that includes continuous stirred-tank reactor (CSTR) and plug flow reactor (PFR) which is approximated as a series of CSTRs is constructed. By the proposed method and the superstructure network, the NLP is divided into a linear programming in flow rate and reactor volume space, and a stochastic optimization problem in concentration space. We solve two cases to illustrate the feasibility of the proposed method. The results show that this new optimization method can reduce the scales and difficulties of the problem and give more suitable structure of the reactor network, as well as better reactor size than those reported in the literature.