Research on optimization of structural parameters for airfoil fin PCHE based on machine learning

The thermal hydraulic performance of the printed circuit heat exchanger (PCHE) is critical to the efficiency of the supercritical carbon dioxide (S-CO₂) Brayton cycle. Therefore, in this paper, four parameters (maximum thickness, maximum thickness location, transverse pitch and staggered pitch of the airfoil fin) were selected to carry out optimization research on the airfoil PCHE combined with CFD simulations, machine learning and optimization algorithms. Firstly, the simulation of the airfoil PCHE with different channel configurations was carried out to analyze the effects of four parameters on its performance. Then, the thermal hydraulic parameters were trained and predicted using the ANN model, from which correlations integrating the structural and arrangement parameters of the airfoil PCHE were obtained. Finally, the sequential quadratic programming (SQP) algorithm and non-dominated sorting genetic algorithm II (NSGA-II) were used to carry out the optimization studies of PCHE, respectively. The results showed that increasing the value of the maximum thickness location can improve the comprehensive performance of PCHE. After optimization, the performance coefficient η_h was improved by about 6.2 % compared to the basic structure when Re = 45000.