Abstract: In view of the difficulties in the accurate separation of each mode and in the accurate identification of its modal damping parameter of multiple modal signals, a method of parameter optimized variational mode decomposition based on cuckoo search (CS) algorithm named CS-VMD method and methods of modal damping parameter identification named envelope integral method (EIM) were proposed. The multiple modal components were accurately separated from the multi-modal vibration attenuation signals in the time domain by using CS-VMD method. And modal frequency and damping ratio of each mode were identified through EIM, and their identified results were compared with the theoretical values (or testing values) and the results identified by half power bandwidth method (HPB). The modal decomposition and modal parameter identification of a displacement simulation signal and the frequency testing signal of a compressor guide vane demonstrated that CS-VMD method can realize the correct decomposition of multiple modal signals, and the modal frequency identification errors of EIM are both below 1.0%. For the displacement simulation signal, the maximum error of the identified modal damping ratio through EIM is smaller than 2.5%. And for the frequency testing signal of the compressor guide vane, the maximum difference of the identified modal damping ratio between EIM and HPB method is 9.098%, and the accuracy of the modal damping identification of EIM is higher than that of HPB method.

Key words: Vibration attenuation signal;Signal separation;Damping parameter identification;Cuckoo search algorithm;Variational mode decomposition

摘要： 针对多模态信号中各模态难以准确分离和模态阻尼参数难以准确识别的问题，提出了布谷鸟搜索（CS）算法参数优化的变分模态分解方法（CS-VMD）和模态阻尼参数辨识的包络线积分法（EIM）。使用CS-VMD方法将多模态时域振动衰减信号中的多模态分量准确分离开来，利用EIM辨识各模态的模态频率和阻尼比，并与理论值（或测量值）以及半功率带宽法（HPB）辨识值进行对比。位移仿真信号与压气机导向叶片测频信号模态分解及模态参数辨识表明，CS-VMD方法可实现对多模态信号的正确分解，EIM辨识的模态频率误差均小于1.0%；对于位移仿真信号，EIM辨识的模态阻尼比最大误差小于2.5%；对于压气机导向叶片测频信号，使用EIM和HPB方法辨识的模态阻尼比最大差别为9.098%，EIM的模态阻尼辨识精度比HPB方法高。

关键词: 振动衰减信号;信号分离;阻尼参数辨识;布谷鸟搜索算法;变分模态分解