基于多模态融合的汽车充电桩故障诊断研究

Research on Fault Diagnosis of Automobile Charging Pile Based on Multi-Modal Fusion

摘要: 随着电动汽车的普及，充电桩故障诊断变得日益重要。本文提出一种基于多模态融合的充电桩故障诊断方法以识别和预测两类常见故障。第一类故障源于设备部件损坏或外部环境导致的绝缘故障，表现为电气信号的短时突变；第二类故障由散热部件损毁或长时间高温工作引发，属于累积性故障。采用支持向量机(SVM)处理第一类故障，利用长短期记忆神经网络(LSTM)分析第二类故障。通过多模态融合提高了故障诊断的准确性和鲁棒性。实验结果表明，所提方法在充电桩故障诊断中具有较高的准确率，为实际应用提供了有益参考。

Abstract: With the popularization of electric vehicles, the diagnosis of charging station faults has become in-creasingly important. This paper proposes a multimodal fusion-based method for charging station fault diagnosis to identify and predict two common types of faults. The first type of fault is caused by equipment component damage or insulation faults caused by external environments, manifested as short-term fluctuations in electrical signals. The second type of fault is caused by the destruction of heat dissipation components or long-term high-temperature work, which belongs to cumulative faults. Support vector machine (SVM) is used to handle the first type of fault, and long short-term memory neural network (LSTM) is used to analyze the second type of fault. The accuracy and ro-bustness of fault diagnosis are improved by multimodal fusion. Experimental results show that the proposed method has high accuracy in charging station fault diagnosis, providing useful reference for practical applications.