基于振动分析的风力机齿轮箱故障诊断

Wind Turbine's Gearbox Fault Diagnosis Based on Vibration Analysis

为实时、准确、简易地诊断出风力机齿轮箱故障类型,提高风力机的稳定性,提出一种面向多故障的基于多尺度小波分析和希尔伯特变换的实时齿轮箱故障诊断方法。利用故障发生前期微弱的故障征兆,首先用小波降噪技术滤除齿轮箱振动信号中的噪声,然后对信号进行多尺度小波分解,通过分析高频重构信号,来判断是否将要产生故障;如果确定故障将要发生,再对高频重构信号进行希尔伯特变换,通过分析能量包络谱相应的波形参数值来判定预测故障的类型。利用试验数据对该方法进行分析验证,证明了该算法的简单和有效性。

A fault diagnosis method of gearbox based on multi-scale wavelet analysis and Hilbert transform,which can diagnose many kinds of fault immediately,was proposed in order to diagnose the wind turbine gearbox fault type easily,improve the stability of the wind turbine. Firstly,the wavelet denoise method was used to remove the noise of gearbox vibration signals, then the vibration signals of gearbox were decomposed by the multi-scale wavelet decomposition method. Through the analysis of high frequency reconstruction signal,whether the fault will occur can be predicted. If the fault is going to happen,Hilbert transform method is used to the high-frequency reconstruction signals. The predict fault type can be determined by the analysis of certain parameter values of energy envelope spectrum. The new experimental data was employed to show that the algorithm is simple and effective.