基于LSSVM和PNN的车轮状态安全域估计及故障诊断

Safety Region Estimation and Fault Diagnosis of Wheels Based on Least Squares Support Vector Machine and Probabilistic Neural Networks

为了实时监测城轨车辆的车轮状态,及时发现车轮故障,提出一种基于最小二乘支持向量机和概率神经网络的车轮状态安全域估计及故障诊断方法。对仿真得到的钢轨振动信号进行经验模态分解,提取各本征模函数特征,采用LSSVM实现对车轮服役状态的安全域估计,采用PNN对正常车轮、扁疤车轮、不圆车轮这3种状态进行模式识别。实验结果表明,该方法能够准确有效地识别车轮工作状态与故障类型。

To implement the real-time monitoring of the wheel state of urban rail vehicle and discover the wheel fault in time, this paper proposes a method, in which the safety region of the wheels is estimated and the fault disgnosis is done, based on least squares support vector machine （ LSSVM ） and probabilistic neural networks （PNN）. Then, empirical mode decomposition （EMD） of the track vibration signals obtained by SIMPACK is used to obtain the Intrinsic mode functions and extracted the features of each IMF. And then, LSSVM is used to estimate the safety region of the wheels service status and PNN is used to recognize the three states of the wheel： normal wheel, flat wheel, out of round wheel. Experimental results show that the proposed method can be used to identify the wheel working state and fault pattern accurately and effectively.