Novel multidisciplinary design and multi-objective optimization of centrifugal compressor used for hydrogen fuel cells

One-dimensional (1D) design and optimization of the impeller plays a significant role in performance improvement of the centrifugal compressor. However, most of the concentration has been paid to three-dimensional (3D) optimization of blades, few attention was focused on main control parameters determining aerodynamic performance and their optimal combination. Thus, this study innovatively developed a multidisciplinary design method combined with empirical 1D loss models, statistical analysis, and multi-optimization theory. The preliminary design of 1D parameters was developed based on empirical loss models. Besides, the analysis of variance of signal to noise ratio (SNR) was applied to find the main control parameters according to their contributions. To maximize the total pressure ratio and isentropic efficiency, the multi-objective optimization based on grey relational grade (GRG) was used to find the optimal combination of 1D parameters. The results showed that the impeller outlet width and impeller outlet radius are the most sensitive parameters affecting compressor performance. The optimal combination of 1D parameters is obtained. Compared to the initial design, the optimal impeller can reduce consumed power of 2.99%, enhance the isentropic efficiency of 1.24% at design point, and obtain the maximum increment of isentropic efficiency of 2.16% at 50 g/s operating point at 70,000 rpm.