基于DBSCAN_SOM的道岔故障诊断

雩风三日

基于DBSCAN_SOM的道岔故障诊断

$ o- V% A5 P* Z& U5 @$ K. N      针对铁路开关控制电路的常见故障现象，提出了一种基于密度的带噪声应用空间聚类（DBSCAN）和自组织特征图（SOM）相结合的开关故障诊断方法。基于微机系统记录的三相电流曲线，根据开关机的动作原理对电流曲线进行分段，初步诊断特征为计算。对于初始特征尺寸较大的问题，使用DBSCAN分离敏感特征。引入粒子群优化（PSO）方法调整SOM网络权重修改规则和避免了“死亡神经元”现象。设计了PSO-SOM故障分类器，并对测试样本进行分类。实验表明，该方法能够以较少的样本识别出开关控制电路的故障模式。比较一下与传统的SOM网络相比，诊断的准确性更高，可以满足诊断的要求。开关控制电路的常见故障。
* O, H7 q& H& g2 K# [. ~5 Z5 u' k# B

关键词：投票率；故障诊断；基于密度的含噪声应用程序空间聚类（DBSCAN）；粒子群优化（PSO）；自组织特征图（SOM）
, k# r3 N- E& Y
" }- E3 N0 K$ b& y" [           Aiming at the common fault phenomenon of railway switch control circuit, a switch fault diagnosis method is proposed based on the combination of Density-Based Spatial Clustering of Applications with Noise (DBSCAN) and Self-Organizing Feature Map (SOM). Based on the three-phase current curve recorded by the micro-computer system, the current curve is segmented according to the action principle of the switch machine, and the initial diagnostic features are calculated. For the problem that the initial feature dimension is high, the DBSCAN is used to separate the sensitive features. The Particle Swarm Optimization (PSO) method is introduced to adjust the SOM network weight modification rules and the phenomenon of “dead neurons” is avoided. The PSO-SOM fault classifier is designed, and the test samples are classified. Experiment shows that this method can recognize the failure modes of the switch control circuit with less samples. Compared with the traditional SOM network, the accuracy of diagnosis is higher, which can meet the diagnosis requirements of common faults of switch control circuits.
7 g. W" ~: j5 l! Y. z
Keywords：turnout; fault diagnosis; density-based spatial clustering of applications with noise (DBSCAN) ; particle swarm optimization (PSO); self-organizing feature map (SOM)

5 {$ B( L! p; t9 i7 c