Abstract
Improving energy efficiency through retrofitting heat exchanger networks (HENs) is among the key strategies for industrial decarbonization. However, identifying the optimal HEN retrofit design is complex due to numerous constraints in operating requirements and site layout, particularly for multi-period processes, i.e. a single plant operated continuously in multiple periods. Existing literature on HEN optimization shows an increasing application of stochastic mathematical programming (MP) methods, especially a trend towards metaheuristics, due to their capability to solve large, complex problems. Simulated Annealing (SA), a metaheuristic algorithm, utilizes neighborhood moves that avoid major modifications of the existing HEN design, thereby providing practical retrofit designs. Despite its potential, its application to HEN retrofit problems remains limited. In particular, there is no existing literature on the application of SA to retrofit HEN for multi-period processes. This study therefore proposes an SA-based optimization framework for multi-period retrofit. The framework integrates SA for topological modification, a repair algorithm for continuous variable optimization, and a post-optimization step for enhancement of heat exchanger (HEX) area utilization across multiple periods. Using an industrial case study, the proposed framework generated a retrofit design that reduced the total annual cost (TAC) by 78% compared to the initial existing HEN design. The work demonstrates the first successful application of SA for retrofitting industrial processes in multi-period operation under multiple practical constraints.