Node dynamic adaptive non-structural model for efficient synthesis of heat exchanger networks

Abstract

Balancing the trade-off between structural diversity (solution space) and optimization efficiency is an urgent challenge in the modeling and global optimization of heat exchanger network synthesis problems. This paper presents a node dynamic adaptive non-structural model without stream splitting to enhance optimization efficiency under the premise of ensuring an adequate solution space. The node-based non-structural stream matching mechanism can be used to make the solution space scalable and achieve the efficacy of multi-stage stage-wise superstructure with a finite number of nodes. Two node dynamic adaptive strategies (uniformly distributed and random) are established to periodically adjust the number of nodes in process streams and the node distribution of existing heat exchanger units, thereby preventing a decrease in computational efficiency due to an excess number of preset nodes while still satisfying the solution space requirements. The node dynamic adaptive non-structural model is optimized using the random walk algorithm with compulsive evolution. The proposed method was applied to four well-known heat exchanger network case studies. High-quality solutions were obtained, and they are more economical than most of the optimal results previously reported in the literature. These results demonstrate the enhanced computational efficiency and applicability of the proposed synthesis method based on a node dynamic adaptive non-structural model for efficiently solving large-scale heat exchanger network synthesis problems.