Multi-objective optimization of a fuel-cell air compressor considering aerodynamic and exergy performance

Two-stage air compressor is widely used in hydrogen fuel cell systems. Performance improvement is limited when structural optimization ignores exergy efficiency. A multi-objective optimization method for two-stage air compressors is proposed, based on bayesian physics informed neural network and grey wolf optimizer, with pressure ratio, isentropic efficiency and exergy efficiency as optimization objectives. Five structural parameters of the impeller are selected as design variables, including tip clearance, impeller outlet width, impeller diameter, blade number and blade outlet installation angle. The entropy weight method is used to calculate the optimal parameter. Optimization results show that the pressure ratio increases from 1.39 to 1.44, isentropic efficiency increases from 89.55% to 90.50%, and exergy efficiency increases from 88.63% to 90.10%. Furthermore, the optimized impeller design greatly reduces irreversible exergy losses caused by turbulent dissipation, viscous friction and wall friction. The proposed method realizes synergistic improvement of both energy quantity and quality.