【摘要】 针对轮轨故障噪声信号非平稳性特征,提出一种基于经验模式分解(Empirical Mode Decomposition,EMD)与神经网络的轮轨故障诊断方法。该方法首先对轮轨噪声信号进行经验模式分解,信号分解为若干个基本模式分量(Intrinsic Mode Function,IMF)之和,再选取若干个包含主要故障信息的IMF分量,提取各分量的能量与峭度特征,对各分量的峭度特征综合得到多尺度峭度特征,然后将各分量能量特征与多尺度峭度特征作为神经网络的输入来识别轮轨故障的类型。对车轮扁疤、钢轨波浪磨耗和正常状态的分析结果表明,以EMD方法提取特征参数的神经网络诊断方法比以小波包方法提取特征参数的神经网络诊断方法具有更高的故障识别率。该方法能够对轮轨故障类型进行准确、有效地分类识别。更多还原

【Abstract】 Aiming at the non-stationary characteristics of wheel /rail fault noise signals,a wheel /rail fault diagnosis method based on empirical mode decomposition( EMD) and neural network was put forward. Frist of all,wheel /rail noise signals were decomposed into several intrinsic mode functions( IMFs),then some IMFs containing the main fault information were selected. The energy and kurtosis features of these chosen IMFs were extracted,and the kurtosis features of these IMFs were integrated into a muti-scale kurtosis feature. Finally,the energy features of these IMFs and the mutiscale kurtosis feature were taken as inputs of a neural network to identify the fault pattern of a wheel /rail system. The analysis results of wheel flats,rail wavy wears and their normal states showed that the neural network diagnosis method based on EMD method to extract feature parameters has a higher fault recognition rate than that based on the wavelet packet method,this method can be used to classify and identify wheel /rail fault patterns accurately and effectively.更多还原