【摘要】 基于轴流压气机试验台，对带可调进口导叶的双级压气机在不同进口导叶角度和不同转速下进行了性能测试，得到了压气机稳定工作范围和失速边界。考虑到传统经验模态分解（EMD）方法易产生模态分量频率混叠，采用变分模态分解（VMD）方法对压气机失速信号进行特征识别提取。针对VMD方法在压气机失速特征提取中过于依赖模态数K与惩罚因子的现象，采用基于仿生学优化的改进算法：引入蛙跳算法构建自适应参数优化机制，以平均包络熵为优化目标实现K和的动态优化，并构建仿真信号验证参数优化VMD方法的优势。最后，将改进算法应用于压气机失速工况信号解析。结果表明，相较于EMD方法，参数优化VMD方法能更精准分离出表征失速特征的主导模态分量。更多还原

【Abstract】 Based on the axial flow compressor test bench, performance tests on a two-stage compressor with adjustable inlet guide vanes were conducted under different inlet guide vane angles and different rotational speeds, and the stable working range and stall boundary of the compressor were obtained. Considering that the traditional empirical mode decomposition(EMD) method is prone to frequency aliasing of modal components, the variational mode decomposition(VMD) method has been s adopted to conduct feature recognition and extraction of the compressor stall signal. In response to the phenomenon that VMD overly relies on the number of modes K and the penalty factor α in the extraction of compressor stall characteristics, an improved algorithm based on bionics optimization was adopted. By introducing the leapfrog algorithm to construct an adaptive parameter optimization mechanism, the dynamic optimization of K and α was achieved with the average envelope entropy as the optimization objective, and the advantages of the simulation signal verification parameter optimization VMD method were constructed.Finally, the improved algorithm was applied to the signal analysis of the compressor stall working condition. The results show that compared with the EMD method, the parameter-optimized VMD method can separate the dominant modal components representing the stall characteristics more accurately.更多还原