Multi-objective optimization of a double-faced type printed circuit heat exchanger

The zigzag flow channels in a printed circuit heat exchanger (PCHE) of the double-faced type have been optimized to enhance heat transfer performance and reduce friction loss by using three-dimensional Reynolds-averaged Navier–Stokes (RANS) analysis and a multi-objective evolutionary algorithm. The response surface approximation (RSA) model was applied in light of the surrogate fidelity of the approximate analysis. A shear stress transport turbulence model and high-resolution scheme were used for numerical analysis. The effectiveness and non-dimensional pressure drop of zigzag channels were employed as the two objective functions of the optimization, and three nondimensional variables, i.e., the ratios of the fillet radius, wavelength, and wave height to the hydraulic diameter of the channels, were selected as the design variables. The design points within the design space were selected using Latin hypercube sampling. Two objective functions were calculated at each design point through RANS analysis to construct RSA models. A Pareto-optimal front was obtained with the multi-objective evolutionary algorithm, and four optimal designs were selected on the Pareto-optimal front by using K-means clustering. The thermal and hydraulic characteristics of these designs were compared with those of a conventionally designed PCHE. An exergy analysis showed that both designs located at opposite extremes of the Pareto-optimal front have an energy savings advantage over the reference design. •Multi-objective optimization of a double-faced type PCHE with zigzag channels was performed using RANS analysis.•The effectiveness, Feff and non-dimensional pressure drop, Fp of the cold channel were selected as objective functions.•In optimum Design A, Feff decreased by 16.1%, although Fp also decreased by 76.9% relative to reference design.•In optimum Design D, both Feff and Fp increased by 13.1% and 22.6%, respectively, relative to reference design.•The Design A exhibited the best performance in terms of net exergy gain.