贝叶斯网络理论及其在设备故障诊断中的应用

在分析机电设备故障诊断技术中广泛存在的不确定性和复杂关联关系的基础上，指出目前的故障诊断方法在处理不确定性和关联性问题时存在的局限性，提出了应用基于概率理论和图论的贝叶斯网络作为设备故障诊断模型具有很好前景的观点。阐述了贝叶斯网络的提出与发展、模型数学描述及研究现状，讨论了贝叶斯网络在故障诊断领域应用的可能方式及其应用情况。指出贝叶斯网络技术在故障诊断领域中的应用将进一步得到推广，提出了将贝叶斯网络广泛应用到故障诊断领域中需要解决的关键技术。

Bayesian Network and Its Applications for Device Fault Diagnosis

In this paper uncertainty and complex coupling widely existed in device fault diagnosis are firsly analyzed, and the limitations of current methods to handle uncertainty and coupling are pointed out. The Bayesian network model based on probability theory and graph theory is set forth with its nice features. Then the history and mathematical basis of Bayesian network are introduced, and the current research tasks are discussed. The application modes are depicted with many examples for fault diagnosis. At last it is suggested that Bayesian network can be used more widely in fault diagnosis, and several important techniques for Bayesian network general applications in this field are provided.