提出了一种利用矢量量化(VQ)算法对列车闸片温度进行状态监测的新方法.根据不同闸片温度序列设置模式库中的参考样本,进行数据训练和闸片温度数据识别与预处理,对高速列车闸片温度数据与各个参考样本进行相似性度量,最终根据VQ的思想求出欧氏距离,从而得到检测结果.结果表明:该方法误报漏报率低,响应时间快,有效消除列车闸片温度跳变的影响,能够满足高速列车闸片温度状态监测的实时性和准确度要求.
A method of condition monitoring of train brake pad temperature based on vector quantiza- tion (VQ)is presented. The data training was is based on different temperature sequences of brake pad, the reference samples in pattern library is set up, and then the data of brake pad temperature arei- dentified and preprocessed, the high-speed train brake pad temperature data are compared with refer- ence samples by similarity measurement. The Euclidean distance according to vector quantization pro- gramming is calculated, and the diagnosis results are obtained. Results show that the method has a low rate of false alarm and a fast response time. It also can eliminate the influence of noise and jump. The method can meet the requirements of high-speed train brake pad temperature monitoring in real-time and accuracy.