论文部分内容阅读
车轮踏面擦伤的预示诊断对列车安全运行具有重要意义。在分析粗糙集和神经网络特点的基础上,结合预示诊断中多传感器、多特征的要求,提出了一种粗糙集与多个神经网络相结合的车轮踏面擦伤预示诊断方法。该方法采用时频域都具有高分辨率的小波分析从车轮振动信号中提取擦伤特征,利用粗糙集的数据约简确定神经网络的初始拓扑结构,通过网络训练建立故障特征与故障之间的映射关系,从而实现踏面擦伤的多传感器融合诊断。实验结果表明该方法具有良好的预示诊断性能。
The prediction of wheel tread abrasion is of great significance to the safe operation of the train. Based on the analysis of the characteristics of rough set and neural network, combined with the requirements of multi-sensor and multi-feature in predictive diagnosis, a diagnostic method of wheel tread flaw prediction based on rough set and multiple neural networks is proposed. In this method, wavelet analysis with high resolution in the time-frequency domain is used to extract the chafing characteristics from the wheel vibration signals. The rough set data reduction is used to determine the initial topology of the neural network. The network training is used to establish the relationship between the fault features and the fault Mapping relationship, in order to achieve tread scratching multi-sensor fusion diagnosis. Experimental results show that this method has a good predictive diagnostic performance.