基于贝叶斯理论的复杂系统快速诊断算法

针对系统结构复杂、维修保障难度大,多具有不确定性(不完整性)信息、多故障和关联故障等特点,融合贝叶斯理论和策略优化算法的优势,提出一种混合快速诊断算法.将诊断任务分解成若干单故障子任务,对子任务用故障树和贝叶斯网络分别建模,根据贝叶斯-蒙特卡洛算法获得先验概率,进行诊断推理分析,获得系统故障薄弱环节,同时运用综合策略优化算法,对故障原因进行诊断策略优化,以实现快速诊断为目的.以电控发动机复杂系统为例,对混合快速诊断算法进行应用研究,验证了该算法的正确性和有效性.

Research on rapid diagnosis algorithm for complex system based on Bayesian theory

According to the characteristics of complex system,which are complex structure,very difficult to maintain and support,uncertain and incomplete information,multiple faults and coherent faults,fused of Bayesian theory and strategy optimization algorithm,a hybrid rapid fault diagnosis method is put forward.In this method,system fault diagnosis task is divided into several single-fault subtasks;and the fault model is established respectively based on fault tree and Bayesian network;the prior probability is got ...