Transonic compressor, Combined casing treatment, Surrogate model, Multi-objective optimization
For modern high load compressors, an excellent stability-enhancing capability by casing treatment (CT) is desirable. However, it is very time consuming to accomplish effective CT design. In this study, a new combined CT structure composed of axial skewed slots and end-wall injection, was proposed to be installed in transonic axial compressors to improve the overall performance. Considering the high computation cost for CFD simulation of the flow field in transonic compressor, a Gaussian Process Regression (GPR) surrogate model combined with Latin hypercube sampling, was utilized to predict compressor performance. For optimization process, a multi-objective evolutionary algorithm (NSGA-Ⅱ) was adopted to obtain the Pareto-optimal front. The main geometric parameters of the slot and the mass-flow rate of injection were selected as design parameters, with the peak efficiency and pressure ratio being two objectives. The results indicated that the surrogate model works well in capturing the key features of the concerning target and accelerating the optimization process. The optimal scheme of the combined CT was found able to increase stall margin (SM) by 19.5% with low efficiency penalty, showing a better performance than the reference combined casing treatment (CCT) scheme. What’s more, the analysis results of entropy generation showed that the superior effect of optimized scheme (OPT) can be attributed to the improvement of exchange flow in slots and the decreased loss in the whole passage.