基于CNN-GRU-Attention的道岔故障诊断算法研究

Research on Turnout Fault Diagnosis Algorithm Based on CNN GRU Attention

道岔是关系列车运行安全的铁路信号基础设备之一。通过分析道岔运行过程的功率数据,可以有效判断道岔的运行状况。为实现对道岔故障自动、高效、准确的诊断,研究并提出了一种基于深度学习的故障诊断方法。首先利用卷积神经网络提取数据空间性特征,再调用门控循环单元网络提取时间性特征,再引入注意力机制对特征进行权重分配,最后使用Softmax分类器进行分类。在对比实验中用多种指标评定该方法的性能,结果表明,所提方法相较于基础方法和另外2种现有方法在诊断性能上有着显著的优势。

Turnout is one of the railway signal infrastructures that affects the safety of trains.By analyzing the power data of the turnout operation process,the operation status of the turnout can be effectively judged.In order to achieve automatic,efficient and accurate diagnosis of turnout faults,a fault diagnosis method based on deep learning is studied and proposed.The study first utilizes convolutional neural networks to extract spatial features from data,then calls on gated recurrent unit networks to extract temporal features,introduces attention mechanisms for allocating weights to features,and finally uses Softmax classifiers for classification.In comparative experiments,multiple indicators are used to evaluate the performance of this method,and the results show that this method has significant advantages in diagnostic performance compared to the basic methods and two other existing methods.