基于信息熵的转子动力特征分析与诊断方法研究

Rotor dynamic feature analysis and diagnosis method based on information entropy

针对转子系统运行过程中的非线性、耦合性、非平稳突变性等复杂特征,提出基于信息熵特征提取方法,将转子系统动力分析与参数识别方法有机结合,分析转子典型故障——裂纹、碰摩及裂纹碰摩耦合等不同状态下响应参量的信息熵特征及变化规律,并构建小波神经网络模型实现转子系统故障状态诊断。通过对转子系统实测信号的分析诊断,验证理论方法的有效性。

Aiming at complex features in rotor system running process,such as non-linearity,coupling,un-steady abrupt states etc.,a new method based on information entropy feature extracting algorithm was proposed to integrate the rotor system power analysis with the parameter recognition method.Information entropy features of response parameters and its change rules in typical rotor fault states were analyzed,such as the faults of crack,rub-impact,crack and rub-impact coupling etc.,and a wavelet neural network model was established to realize the recognition and diagnosis of these different fault states.The validity of the proposed method was demonstrated by theoretical analyzing and verifying experiment on the rotor system.