基于轨迹平行测量的发动机异响诊断方法

Diagnosis Method of Exception Sound Using Trajectory Parallel Measure in Engine Faults

为了利用机器的声 (异响 )信号进行故障识别 ,提出了一种异响信号的故障诊断思路 .首先利用连续小波消噪技术对声信号进行消噪和故障源识别 ,然后将混沌技术中的轨迹平行测量方法应用于异响信号的非平稳特征提取与识别 ,提取出时间序列动力学轨道运行特征指标 .经过对发动机的故障声信号进行实际应用后 ,表明该方法具有良好的识别效果 。

An idea for diagnosis exception sound is presented. The continuous Wavelet Transform (CWT) is used to purify the original sound signal and identify the source of faults. Meanwhile, the feature parameter of non-stationary for sound signal is extracted applying the trajectory parallel measure of chaos technique. At last, the character index of dynamics orbit in the observed time series is selected. Fault diagnosis for engine by using sound signal is illustrated. The result shows that this technique provides a new tool for diagnostic of sound signal.