Synthesis of N-best Heat Recovery Networks with Consideration of Dynamic Control Performance

Recently, graph-theoretic methods have increasingly been employed to generate near-best (n-best) heat recovery networks, aiming to maximize energy recovery efficiency. The exploration of these n-best networks has proven pivotal for making informed decisions. Nevertheless, existing studies in this domain have not attempted to study the favourability of these generated networks based on their respective dynamic control performance. This performance metric reflects the network’s ability to maintain target temperature even under disturbances. The network topologies play important role in both economic (i.e., total annual cost (TAC)) and dynamic control aspects. To address this gap, this work introduces a hybrid approach. First, all combinatorically feasible heat recovery networks are generated using P-HENS. Thereafter, each network undergoes dynamic control performance evaluation through Aspen Plus simulations. The final step involves optimization of the network structures based on fuzzy method which avoids over-prioritization. To illustrate the efficacy of the proposed methodology, it is applied to solve a 5-stream problem. Results showed that Network A with the least TAC ($122,249) is not necessarily associated with the greatest dynamic performance (with failure rate of 15 %). Network C which offers the balance performance (with TAC of $122,666 and failure rate of 0 %) is chosen.