CAJ | 학술논문

针对核动力装置系统参数密集且相互影响、故障征兆随故障发生的位置、程度以及警报阈值的不同而发生变化、难以获得确定的故障诊断规则的问题,在采用多层流模型描述核动力装置的运行特性的基础上,引入了影响因子来描述故障征兆之间的因果关系强度,进一步结合警报分析方法和贝叶斯理论进行不确定性故障推理. 建立了压水堆主冷却剂系统不确定性故障诊断模型,通过仿真证明了该方法可以合理有效地识别系统故障,解决了由于诊断规则的不确定性而可能造成的误诊和漏诊,可有效辅助操纵员进行深层次决策.
침대핵동력장치계통삼수밀집차상호영향、고장정조수고장발생적위치、정도이급경보역치적불동이발생변화、난이획득학정적고장진단규칙적문제,재채용다층류모형묘술핵동력장치적운행특성적기출상,인입료영향인자래묘술고장정조지간적인과관계강도,진일보결합경보분석방법화패협사이론진행불학정성고장추리. 건립료압수퇴주냉각제계통불학정성고장진단모형,통과방진증명료해방법가이합리유효지식별계통고장,해결료유우진단규칙적불학정성이가능조성적오진화루진,가유효보조조종원진행심층차결책.
Nuclear power plants have large significantly interacting parameters. The symptoms of a fault may show up well away from the fault′s location with missleading degree of severity, while escaping pre-set alarm thresholds. It is very difficult to diagnose faults by definite rules. In this paper, Multilevel Flow Models were used to describe the operational characteristics of nuclear power plants where impact factor was introduced to quantify the strength of causality between fault symptoms. An uncertain fault diagnosis approach based on alarm analysis and Bayes theory was explored. That is, an uncertain fault diagnosis model of the primary coolant system was developed. The sound-ness of the proposed approach was verified through simulations. The proposed approach can identify faults logically and effectively. It can solve misdiagnosed and undiagnosed problems missed by the misleading certainty of rules and can assist operators in their deep level decision-making.