基于Bayesian网络的柴油发动机故障诊断分析系统的研发

The Diesel Engine Fault Diagnosis Based on Bayesian Network Analysis System Research and Development

根据电控柴油机故障源的多样性和不确定性,提出了先使用传感器数据技术检测发动机不同阶段的工作状态,然后结合经验法、专家指导意见等预算出故障源的可能发生的概率,最后使用Bayesian网络推断查找故障源的一种故障诊断方法。经某电控柴油机的实验结果表明,该系统结合了传感器诊断技术的实时性和Bayesian网络技术的判断决策能力,提高了故障诊断的正确率和实效性。

According to the diversity of electronically controlled diesel engine fault sources and uncertainty, this paper proposes to use the sensor data to detect the working state of the engine at different stages then combined with the empirical method, expert guidance, such as the probability of the budget for the source of the fault may occur, and finally the use of bayesian networks to infer a fault diagnosis method to find the source of the fault. An experimental results show that the electronic control diesel engine this system combined the sensor diagnosis technology of the real-time performance and decision-making ability of the bayesian network technology, the accuracy of fault diagnosis is improved and the effectiveness.