模糊神经网络在摆式列车作动系统故障诊断中的应用

Use of Fuzzy Neural Network in Fault Diagnosis of Tilting Driving System in Tilting Train

针对摆式列车倾摆作动系统故障所具有的模糊性 ,提出了一种基于模糊神经网络的故障诊断方法 ,将模糊理论与神经网络融合在一起 ,实现了对故障的模糊诊断。给出了模糊神经网络的结构。该网络由两部分组成 ,前一部分为模糊量化部分 ,后一部分为神经网络部分。根据输入特征量的实际情况 ,确定了各模糊子集上的隶属度函数。提出了一种阈值向量故障判别方法 ,使故障判别更具灵活性和合理性。根据相关文献构造了倾摆作动系统的构造判别表 ,形成故障诊断理论样本。以摆式列车倾摆作动系统为对象 ,就倾摆作动系统中几种典型的故障模式 ,采用模糊神经网络方法作了识别。仿真结果表明 ,该方法是行之有效的 。

With the view of the fuzzy property used in fault diagnosis of the tilting driving system, the paper puts forward a fault diagnosis based on a fuzzy neural network (FNN), which combines fuzzy theory with neural network to implement fault fuzzy diagnosis. The structure of FNN is given, consisting of the fuzzy quantification section and the neural network section. The membership function is developed according to the actual situation. Meanwhile a fault discriminating method for threshold vector is employed to make fault identification even more flexible and reasonable. The table of fault identification is constructed on the basis of the related references, which are summarized in the theoretical sample of fault diagnosis of tilting driving system. It is applied to the fault diagnosis of tilting driving system. Several types of typical faults in tilting driving system are analyzed and identified with the fuzzy neural network. The simulation shows that it is useful and effective. The theoretical foundation of using the fault diagnosis in tilting driving system in real tilting train is provided.