基于模糊聚类的城轨列车辅助逆变器故障诊断

A Method of Metro Vehicle Auxiliary Inverter Fault Diagnosis Based on Fuzzy Clustering Arithmetic

针对辅助逆变器结构复杂,易产生故障等问题,本文基于小波包频带能量分解的基本原理和模糊C均值聚类算法,研究了模糊C均值聚类算法在辅助逆变器故障诊断中的应用,并以MATLAB软件为仿真平台实现对故障信号的仿真,在仿真中设置电压频率变化、供电中断、脉冲暂态等几类故障,同时选用基于小波包频带能量分解的方式提取故障特征向量作为故障诊断的标准样本,通过计算待诊断样本与标准样本的贴近度,实现故障模式识别。仿真结果表明,模糊C均值聚类算法可以准确地进行故障分类。该研究为城轨列车辅助逆变系统的故障诊断提供了理论依据。

In this paper, the authors study the application of the fuzzy C-means clustering algorithm in the fault diagnosis of auxiliary inverter based on the basic principle of wavelet packet frequency band energy decomposition and the fuzzy C-means clustering algorithm in order to solve the problems of auxiliary inverter causing by its complicated structure. Using MATLAB software as the platform to simulate the fault signal, several fault types are set during the simulation, such as variations of voltage frequency, interruption of power supply, pulse transient and so on. Based on the decomposition of wavelet packet frequency band energy, the fault feature vectors can be extracted as the standard samples of fault diagnosis, then calculating the similarity degree of the samples to be diagnosed and the standard samples to realize the recognition of fault pattern. The simulation results show that the faults can be identified accurately based on FCM algorithm. This research provides theoretical basis for auxiliary inverter system fault diagnosis of metro vehicle.