论文部分内容阅读
概率阈值法存在定位分支生成时间点不精确,建立的离散动态事件树具有较大时间不确定性等问题。针对概率阈值法存在的不足提出相应的改进措施——精确概率阈值法。通过分析动态水箱发现:在概率阈值为0.99的条件下,精确概率阈值法的总节点数比概率阈值法少25%,事故节点数仅少2.7%,计算时间缩短了37.7%,证明精确概率阈值法对事故节点的分析效率更高;精确概率阈值法得到的总事故概率比概率阈值法小33.6%,对于水箱干涸事故其留给操作人员采取措施干预事故发展的时间窗口更长;并且精确概率阈值法建立的离散动态事件树显示出更多的演化路径,反映了更多的事故信息,利于全面掌握动态水箱的事故演化过程,更适合研究系统的运行动态特性。
Probability thresholding method has some problems, such as inaccurate production time of positioning branch and uncertainty of the time-discrete discrete event tree. Aiming at the shortcoming of probability thresholding method, some corresponding improvement measures are put forward - the accurate probability thresholding method. By analyzing the dynamic water tank, it is found that the total number of nodes in the accurate probability thresholding method is 25% less than the probability thresholding method, the number of accident nodes is only 2.7% and the calculating time is shortened by 37.7% when the probability threshold is 0.99. Proved the exact probability threshold The probability of total accidents obtained by the accurate probability threshold method is 33.6% less than that of the probability threshold method, and the time window left for the operators to take measures to intervene in the accident development is longer for the tank dry-out accidents; and the exact probability The discrete dynamic event tree established by the threshold method shows more evolutionary paths and reflects more accident information, which helps to fully grasp the dynamic evolution process of dynamic water tank and is more suitable for studying the dynamic characteristics of the system.