CAJ | 학술논문

建立了补水系统贝叶斯故障诊断网络。在结合补水系统结构特点、运行规程和专家经验的基础上构建了初始诊断贝叶斯网络，运用基于微粒群优化的贝叶斯网络学习算法学习故障数据集，进一步构建完整网络，并进行推理分析。所建网络能有效分析和更新系统中各节点故障概率，为故障诊断提供辅助决策。
건립료보수계통패협사고장진단망락。재결합보수계통결구특점、운행규정화전가경험적기출상구건료초시진단패협사망락，운용기우미립군우화적패협사망락학습산법학습고장수거집，진일보구건완정망락，병진행추리분석。소건망락능유효분석화경신계통중각절점고장개솔，위고장진단제공보조결책。
A fault diagnostic Bayesian network of reactor make-up system was constitu-ted .The system’s structure characters ,operation rules and experts’ experience were combined and an initial net was built .As the fault date sets were learned with the parti-cle swarm optimization based Bayesian network structure ,the structure of diagnostic net was completed and used to inference case .The built net can analyze diagnostic prob-ability of every node in the net and afford assistant decision to fault diagnosis .