煤矿机械

针对极寒环境下矿用电动轮卡车核心部件故障频发、传统诊断方法适应性差等问题，设计一种基于多源信号融合与改进故障树分析-模糊贝叶斯网络（FTA-FBN）的故障诊断方案。首先，通过CAN总线和传感器采集动力电池、电动机、逆变器等关键部件的多维度信号，经数据预处理与特征融合后构建故障数据集；其次，引入极寒环境影响因子改进FTA，梳理故障因果链并转化为FBN结构，实现定性与定量诊断的结合；最后，通过实际故障案例验证该方案的有效性。结果表明，该方案对极寒环境下电动轮卡车核心部件故障诊断准确率达到96.8%，较传统方法提升12.3%，可有效缩短故障排查时间，为设备运维提供技术支撑。

In the view of the problems of frequent faults of core components of mining electric wheel truck and poor adaptability of traditional diagnosis methods in extremely cold environments, a fault diagnosis scheme based on multi-source signal fusion and improved Fault Tree Analysis-Fuzzy Bayesian Network(FTA-FBN) was designed. Firstly, the multi-dimensional signals of key components such as power batteries, motors and inverters were collected through CAN bus and sensors, and a fault dataset was constructed after data preprocessing and feature fusion. Secondly, the influence factor of extremely cold environment was introduced to improve the FTA, the fault causal chain was sorted out and transformed into a FBN structure, so as to realize the combination of qualitative and quantitative diagnosis. Finally, the effectiveness of this scheme was verified by actual fault cases. The results show that the accuracy of this method for fault diagnosis of core components of electric wheel truck in extremely cold environments reaches 96.8%, which is 12.3% higher than that of traditional methods. It can effectively shorten the fault troubleshooting time and provide technical support for equipment operation and maintenance.