Compressor design is always a prevalent research issue as nowadays the compressor demands higher performance. Designing afresh a new compressor of higher aerodynamic performance will cost a tremendous amount of time and expense. Consequently, the wiser method of acquiring a higher performance compressor is to optimize a current compressor, discussed in this paper. First, a detailed numerical simulation of al compressor is done. A parametric rotor model is created and the variations of lean and swept parameters are limited. Then, an optimization calculation is carried out based on the combination of an artificial neural network and a genetic algorithm. The optimization iteration is started at three specified working points – a choke point, a design point and a near stall point – to acquire a comprehensive improved solution. A novel penalty function considering both pressure ratio and efficiency is constructed to find the optimal model. Finally, a detailed review is made of the original and the optimized compressor. The computational results show that the isentropic efficiency is successfully increased and the stall margin is remarkably improved. Ultimately, a detailed vortex structure analysis with increasing outlet pressure is discussed. This helps the performance increment mechanism to be understood further.

压缩机设计一直是一个普遍的研究问题，因为如今压缩机要求更高的性能。重新设计具有更高空气动力学性能的新型压缩机将花费大量时间和费用。因此，本文讨论的获取更高性能压缩机的更明智的方法是优化当前压缩机。首先，对压气机进行了详细的数值模拟。创建了参数化转子模型，并限制了稀薄和清扫参数的变化。然后，基于人工神经网络和遗传算法的组合进行优化计算。优化迭代从三个指定的工作点开始（扼流点，设计点和接近失速点）开始，以获取全面的改进解决方案。构建了同时考虑压力比和效率的惩罚函数，以找到最优模型。最后，对原始和优化的压缩机进行了详细的审查。计算结果表明，等熵效率提高，失速裕度得到明显提高。最终，讨论了随着出口压力增加而进行的详细涡旋结构分析。这有助于进一步了解性能提升机制。讨论了随着出口压力增加而进行的详细涡旋结构分析。这有助于进一步了解性能提升机制。讨论了随着出口压力增加而进行的详细涡旋结构分析。这有助于进一步了解性能提升机制。