基于PSE-ANFIS的ZPW-2000A轨道电路故障诊断方法

Fault Diagnosis Method of ZPW-2000A Track Circuit Based on PSE-ANFIS

针对ZPW-2000A轨道电路故障的复杂性、随机性以及故障样本获取难的问题,提出一种基于功率谱熵(PSE)及自适应神经模糊推理系统(ANFIS)的ZPW-2000A轨道电路故障诊断方法。在移频信号(FSK)功率谱熵的理论基础上,利用ANFIS作为分类器进行故障模式识别,可简化轨道电路故障诊断的复杂性。运用LabVIEW平台对移频信号进行深入分析,对轨道电路故障的类型、位置进行识别。最终验证了该方法在轨道电路故障诊断方面效果良好,为实现轨道电路的在线故障诊断奠定了良好的基础。

For the complexity,randomization and difficulty for getting fault samples of ZPW-2000A track circuit,a method of fault diagnosis of ZPW-2000A track circuit based on the theory of power spectrum entropy(PSE)of frequency-shift signal(FSK)and the adaptive neuro-fuzzy inference system(ANFIS)was put forward.On the basis of the theory of PSE of FSK,ANFIS is used as classifier for fault pattern recognition,which can simplify the fault diagnosis of track circuit.LabVIEW platform is used to analyze FSK deeply,so as to recognize the type and location of track circuit failure.The results show that the method has a good effect on the fault diagnosis of track circuit and has laid a good foundation for realizing the online fault diagnosis of track circuit.