A Hybrid Evolutionary Algorithm for Multi-Objective Heat Exchanger Network Retrofit for Multi-Period Processes

Abstract

With the growing awareness of the need to mitigate greenhouse gas (GHG) emissions, the world is transitioning towards more sustainable and energy-efficient alternatives. A large share of the Swiss process industry relies on multi-period production. Retrofitting such plants often has a high potential to improve energy efficiency and thus reducing the GHG emissions. To analyze such processes, in this work, an existing two-level algorithm using a genetic algorithm for topology optimization and a differential evolution for heat load optimization is extended to a multi-objective algorithm to consider GHG emissions besides total annual cost (TAC). The algorithm is applied to a chips production plant from industry. Comparing the results to the single-objective algorithm, GHG emissions can be further reduced by 50 %, causing an increase in TAC by 27 %. With the introduction of GHG emissions as the second objective, utility demand is included in both objectives leading to having a larger impact on the results than capital costs. However, it has been shown beneficial compared to solutions omitting this higher impact.