【摘要】 针对由于轨道交通车辆走行部故障数据缺乏,故障信号检测困难等因素造成列车走行部故障难以及时、准确地诊断。按照轨道交通车辆技术规范标准,采用多体动力学软件Simpack构建列车模型,通过计算机仿真列车正常运行,抗蛇形减震器故障以及一系悬挂弹簧失效时的三种工况的信号,基于结合小波变换和信息熵的小波熵理论(Wavelet Entropy Theory)对三种工况信号进行故障诊断分析。选取列车构架中部通道的三种工况信号,采取Tsallis小波能量熵,小波奇异熵,小波近似熵以及小波时间熵作为分类特征,通过SVM分类器进行仿真,最后经过仿真分类三种工况的准确率能到达92.25%,从而验证基于小波熵理论构造分类特征的可行性和可靠性。更多还原

【Abstract】 The train bogie fault is difficult to be diagnosed in time and accurately, because of the lack of actual data of bogie fault, the difficulty of fault signal detection and other factors. According to the technical standards of rail transit vehicles, the train model is constructed by using multi-body dynamics software Simpack. Through computer simulation, the signals of normal operation, anti-snake shock absorber failure and primary suspension spring failure are simulated. Based on wavelet entropy theory, which combines wavelet transform and information entropy, the signals of three working conditions are diagnosed and analyzed, three kinds of working condition signals in the middle channel of the train frame are selected. Tsallis wavelet energy entropy, wavelet singular entropy, wavelet approximate entropy and wavelet time entropy are used as the classification features. The SVM classifier is used for simulation, and the accuracy of the three working conditions can reach 92.25% after simulation. Thus, the feasibility and reliability of constructing classification features based on wavelet entropy theory are verified.更多还原