正态模糊因果图的推理算子及归一化算法研究

因果图理论是一种基于概率论的推理方法。本文在分析因果图理论发展和存在问题的基础上，将模糊数学引入因果图理论，即模糊因果图，从而可以克服因果图分析中概率难以精确赋值的缺点，将因果图理论应用扩大到了模糊领域。本文主要对事件概率为正态模糊数进行讨论，提出了因果图的模糊算子，得到了模糊条件概率的计算公式，提出了正态模糊数的归一化算法。最后，以核电站的一个子系统为例进行仿真实验，实验结果实际一致。研究表明：模糊因果图能有效地用于故障分析，比原来的因果图方法具有更大的灵活性和适应性。

Reasoning operator and normalizing algorithm for normal fuzzy causality diagram

Causality diagram is a probabilistic reasoning method. In this paper, fuzzy set theory is introduced to develop causality diagram methodology based on discussing the development and the restriction of conventional causality diagram. Fuzzy causality diagram can overcome the shortcoming that is difficult to assign probability of the event accurately in conventional causality diagram. By introducing fuzzy set theory, the application of causality diagram is extended to fuzzy field. Based on the normal fuzzy number, fuzzy operator and fuzzy conditional probability, normalization algorithm is proposed in this paper. A subsystem in nuclear plant is used as example to carry out simulation experiment. Simulation experiment result is coincident with practical fact. The research shows that fuzzy causality diagram is effective in fault analysis, and it is more flexible and adaptive than conventional method.