基于贝叶斯网络的柴油机故障诊断研究  

Study of Diesel Engine Fault Diagnosis Based on Bayesian Network

作　　者：仝兆景[1] 石秀华[1] 许晖[1] 王文斌[2] 程永强[1] 

机构地区：[1]西北工业大学航海学院 [2]中国人民解放军68302部队

出　　处：《计算机测量与控制》2013年第5期1118-1119,1125,共3页Computer Measurement &Control

摘　　要：针对传统方法在不确定和信息不完备条件下故障诊断的不足,运用贝叶斯网络构建了柴油机的故障诊断模型,该模型实现故障原因征兆的因果推理,结合专家知识和故障样本,完成了网络结构参数学习和优化,实现了特定工况下故障的高效、准确诊断;实验表明,该诊断方法稳定可靠,准确度较高,在大型复杂装备的故障诊断领域具有一定的实用性和推广价值。Considering the limitation of the traditional method in fault diagnosis with uncertain and incomplete information, Bayesian net work model is established facing on the typical diesel engine fault diagnosis. The model is optimized by the structure and parameter learning with the help of expert knowledge and fault samples, and it can diagnose the fault efficiently and exactly under the special working conditions. Experimental results show that the technique is reliable and precise, and it could be expected to be applied to the diagnosis of large complex equipment.

关 键 词：贝叶斯网 柴油机 诊断 故障检测 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]