A Simulated Annealing Framework for Multi-Period Heat Exchanger Network Retrofit

Retrofitting heat exchanger network (HEN) is a complex task, especially for multi-period processes wherein the HEN retrofit design must handle temperature variations over time. For such complex problems, metaheuristics, a subset of stochastic mathematical programming (MP) methods, are gaining attention due to their capability to provide optimal or near-optimal solutions compared to deterministic MP methods. Simulated Annealing (SA), a metaheuristic algorithm, can be utilized to perform discrete modifications in the existing HEN design and evaluate the objective using the Metropolis criterion, thereby generating optimal or near-optimal HEN retrofit solutions. Previous literature had shown the application of SA in several HEN optimization problems, but none of the studies have explored its application in multi-period HEN retrofit problems. This work presents a simulated annealing (SA)-based optimization framework for solving multi-period HEN retrofit problems. The SA-based optimization framework has the following features: Reduction of complexity by utilizing SA for discrete structural changes for the existing HEN design and a repair algorithm to optimize the continuous variables.The modified methodology incorporates the addition of mixers in the list of moves and mixer fractions as another optimization variable in the repair algorithm to ensure flexibility in the multi-period HEN.A post-optimization step is proposed after attaining the modified and feasible HEN design to ensure that the heat exchanger (HEX) area utilization is maximized across all operating periods.