基于贝叶斯网络的态势评估诊断模型

Situation Assessment/Diagnosis Model Based on Bayesian Networks for Hydropower Equipment

针对传统的水电设备诊断模型通用性差等问题,提出了基于贝叶斯网络的水电设备态势评估诊断模型.该态势评估模型根据功能分为三层结构:特征级、理解级、评估级.并将贝叶斯网络中的节点按照功能分为态势节点和事件节点,采用网络推理过程将传感器采集信息作为事件节点的证据用来更新态势节点概率,并反过来影响事件节点的概率.该诊断模型在水电设备调速系统的诊断应用中的准确率达到95.2%,证实了该模型的判决可信度.

Aiming at uncertainty and the poor generality of hydropower equipment fault diagnosis, a general situation assessment model based on Bayesian Networks is put forward. The model includes three levels, i.e., character level,understanding level, and assessment level. Nodes in Bayesian networks are divided into situation and event nodes according to their functions. During reasoning the information acquired by sensors are taken as the evidence of event node to update the probability of situation node and in turn, to influence the probability of event nodes. The diagnosis model in application of hydropower speed governor system shows that the veracity can be up to 95.2%, thus indicating the reliability provided by the situation assessment model.