Thermal Performance Prediction and Optimization of “Heat Exchangers” by Artificial Intelligence Techniques

With the rapid development of the artificial intelligence (AI) techniques used in thermal systems in recent decades, more and more AI algorithms are used in the optimal design of heat exchangers. Thermal performance prediction, optimal design, and cost minimization of heat exchangers are important targets for both designers and users. In this article, four common AI methods (artificial neural network, ANN; genetic algorithm, GA; particle swarm optimization, PSO; and Taguchi method) are applied to predict and optimize the thermal performances of three common types of heat exchangers (shell-and-tube heat exchangers, STHXS; finned tube heat exchangers, FTHXS; and primary surface recuperators). It can be concluded that ANNS can be applied to predict thermal performances of the heat exchangers, especially for engineers to model the complicated heat exchangers in engineering applications. The optimized results acquired by the GA and the improved PSO methods for the comprehensive performance of heat exchangers may be used in engineering application and comprehensive study, as long as the optimization objective functions, optimization variables, and restrictive conditions can be chosen appropriately. The Taguchi method can be applied and conducted quickly and inexpensively to study the effect of the design parameters on the comprehensive performances of heat exchangers based on the CFD or experiment methods.