基于预警系统的反应堆非能动设备故障的研究

Research on Faults of Reactor Passive Equipment Based on Early Warning System

在已有的核电智能诊断方法研究的基础上,为提高核电故障诊断的及时性,提出了以专家系统置信规则库技术为基础的基于时间周期性触发的预警系统。在运行过程中,预警参数超过阈值,系统进入预警状态,应用预警系统中的专家系统置信规则库,完成故障的定位与识别。采用故障机理模型与核电模型相结合的方式,通过报警时刻核电运行参数的研究,构建置信规则库。以此为基础进一步构建预警参数集合及预警阈值集。预警系统的研究对象为反应堆非能动设备故障,通过在核电模型中人为引入故障,预警系统能够在故障发生时进入预警状态,故障识别结果与引入的故障一致,验证了预警系统的有效性和可靠性。

Based on research on the intelligent diagnosis method of nuclear power plant,in order to improve the timeliness of nuclear power plant fault diagnosis,a time periodic triggering early warning system based on expert system belief rule base is proposed.In the process of running,the off-limit of early warning parameters make the system into the early warning state,and through the belief rule base,location and identification of fault is completed.Based on the fault mechanism model,both with the nuclear power model,the belief rule base is constructed by studying the operation parameters when alarm occur.In this way,the set of early warning parameters and the set of its threshold value will be built.The object studied about is the fault of the passive devices of the reactor.Faults artificially inserted into the nuclear power model indicate that the early warning system can enter the early warning state at the time of the fault.The diagnosis result is consistent with the fault insert.Proves that the early warning system is effective and reliable.