Heat Exchanger Network Synthesis without stream splits using parallelized and simplified simulated Annealing and Particle Swarm Optimization

Abstract

Much attention has been paid to heat exchanger network (HEN) synthesis and optimization by using meta-heuristic approaches. In general, Simulated Annealing (SA) is able to provide good solutions, but with large computational efforts. In the present work, a two-level no-split HEN synthesis hybrid method is presented. SA is used for topology optimization, while continuous heat load variables are handled with Particle Swarm Optimization (PSO). SA is simplified and only one type of move is used (single heat exchanger addition), along with group optimizations to improve PSO performance. A parallel processing technique is also presented in order to improve local search performance. The method is tested in 4 medium and large scale benchmark case studies and the no-splits results are compared to literature solutions with and without splits. The solutions presented have lower Total Annual Costs (TAC) when compared to other no-split HENs, and even to some HENs with splits. The proposed method is able to present near-optimal solutions by more efficiently exploring the search space and using simple moves for local searches.