根据风机齿轮箱故障信号的非平稳特性,提出一种基于总体平均经验模式分解(EEMD)和BP神经网络的风机齿轮箱故障诊断方法。首先,对原始信号进行小波去噪。然后,对故障信号进行EEMD分解,将其分解为多个固有模态函数(Intrinsic Mode Function,简称IMF)之和,选取若干含有主要故障信息的IMF分量做进一步分析。最后,从各IMF分量中提取故障信号能量特征参数,将归一化后的能量特征参数作为BP神经网络输入参数进行故障诊断。实测结果表明:该方法故障诊断准确率达到了99%左右。可以准确、有效的对风机齿轮箱进行故障诊断。

According to the non-stationary characteristics of wind turbine gearbox fault signal, propose a method of wind turbine gearbox fault diagnosis based on ensemble empirical mode decomposition (EEMD) and BP neural network.Firstly,use wavelet to denoise the original signal.Then, use EEMD to decompose the fault signal,which is decomposed into sum of several intrinsic mode functions (Intrinsic Mode Function,referred to as IMF),and select some main fault information for further analysis of IMF components.Finally,extract the energy feature parameters of fault signal from the IMF components, use the normalized energy characteristic parameters as the input parameters of the BP neural network for fault diagnosis of gearbox.Experimental results show that: the method of fault diagnosis accuracy rate has reached about 99%.This method can diagnose fault for wind turbine gearbox accurately and effectively.