摘要
针对ZPW-2000轨道电路分路不良故障,结合车载TCR设备提供的具体数据,根据不同位置故障对TCR感应电压幅值的影响,采用基于模态经验分解(EMD)、模糊熵的方法提取分路不良故障诊断所需的特征参量,通过改进粒子群优化支持向量机的混合算法实现轨道电路分路不良故障诊断。同时与SVM、PSO-SVM、GA-SVM算法对比,进一步验证本文所提方法的有效性,为分路不良故障诊断提供了新的快速、准确诊断方法,为轨道电路分路不良故障诊断提供了新思路。
Aiming at shunt malfunction of ZPW - 2000 Track Circuit, combining with the specific data provided by TCR equipment, based on the influence of faults in different locations on the induced voltage, the Empirical Mode Decomposition and Fuzzy Entropy were used to extract the parameters that diagnosis of shunt malfunction needed, the fault diagnosis of shunt malfunction of track circuit was achieved through the hybrid algorithm in which the improved PSO optimizes support vector machine. The proposed mothod is verified by comparion with SVM, PSO - SVM, GA - SVMalgorithm. The proposed method provides a new way for shunt malfunction,it is a rapid and accurate method
出处
《系统仿真技术》
2016年第4期291-296,306,共7页
System Simulation Technology