Development of generalized models for accurate prediction and optimization of thermal-hydraulic performance in chevron-type plate heat exchangers

Plate heat exchangers (PHEs) are widely used in diverse industrial applications due to their compact structure and high thermal efficiency. However, accurately predicting their performance remains challenging because of the complex flow behavior within cross-corrugated channels. This study presents a comprehensive thermal-hydraulic performance database generated through high-fidelity Large Eddy Simulations to support the development of robust and generalized predictive models. Correlations for the friction factor and Nusselt number were established and validated against multiple experimental datasets from different sources, demonstrating strong generalization across a broad range of design parameters and operating conditions. In addition, surrogate models based on five machine learning algorithms were developed to enable rapid performance prediction. A Voting ensemble model was introduced, combining the interpretability of empirical correlations with the predictive power of machine learning for enhanced accuracy and efficiency. Furthermore, the NSGA-II genetic algorithm was employed for multi-objective optimization, identifying optimal PHE design parameters that balance heat transfer and flow resistance. This study proposes a generalized dataset–model–optimization framework that can be extended to various heat exchanger types, offering valuable insights for both theoretical research and engineering applications.