基于压缩小波变换与增强的发动机故障特征提取

Engine Failure Feature Extraction Based on Synchrosqueezed Wavelet Transform and Enhancement

针对发动机强噪声条件下故障信号信噪比低、提取困难的问题,提出基于压缩小波变换与同步增强的故障特征提取方法。首先,对测取的发动机信号进行压缩小波分解,得到不同尺度下对应的瞬时频率,并进行解调及消噪处理,提取微弱故障特征;其次,采用循环特征同步增强方法,强化故障特征表示,最终对故障特征进行诊断识别。仿真及实例分析表明,该方法能有效提取发动机故障特征及诊断故障。

Considering the low signal-noise ratio and extraction difficulty of fault signal under strong engine noise,the paper puts forward failure feature extraction method based on synchrosqueezed wavelet transform and enhancement. Firstly,it obtains instantaneous frequency at different scales by synchrosqueezed wavelet decomposition on engine signal and extracts weak failure feature by demodulation and de-noising. Then,it enhances failure feature representation with the method of circulation feature synchronous enhancement,and diagnoses and recognizes the failure feature. The simulation and analysis through real case show that this method can effectively extract engine feature and fault diagnosis.