基于贝叶斯疑似度的有向二分图模型的故障元件诊断

针对目前已有的故障诊断方法存在的计算复杂度高、建模困难、诊断误差较大等不足,分析了故障与征兆之间的非确定性关系,建立了概率加权的有向二分图故障诊断模型,依据统计概率对模型进行初始化,结合开关量时序属性对征兆信息进行完备化处理,通过计算贝叶斯疑似度对电力系统中的故障进行诊断,输出故障诊断结果.;Aiming at the disadvantages, including high computational complexity, difficulty of modeling and error of diagnosis which exist in the current fault diagnosis, the study analyzes the relationship between the fault and symptom, establishes a fault diagnosis model based on weighed bipartite directed graph. The model is initialized according to the statistical probability, the symptom is completed based on switch temporal logic, the Bayesian suspected degree of fault is calculated, and finally the results of fault diagnosis are put out