认知不确定多态系统模糊可靠性评估

Evaluation on Fuzzy Reliability of Multi-state System in Consideration of Epistemic Uncertainty

针对高可靠部件短时间内很难获得足够的性能分布信息,致使对部件的认知存在一定的不确定性问题,为降低其对部件性能分析的影响,实现对系统可靠性的准确估计,假定部件性能分布参数为模糊数,根据模糊集合分解定理,建立基于β-截集变量的性能分布模型,给出了部件模糊状态性能区间的定义和状态模糊概率的计算方法。对传统的通用生成函数方法进行改进,定义了α-截集通用生成函数及其运算法则。提出了考虑认知不确定的多态系统模糊可靠性评估方法,并以仿真实例进行验证说明。该方法不仅克服了性能分布信息缺少,无法准确地建立状态信息分析模型的不足,且方法简单、思路清晰,具有强通用性和工程应用性。

There is epistemic uncertainty on high-reliability components since it is hard to obtain adequate performance data of the components within a short time.The components′performance distribution parameter is assumed as a fuzzy number,according to the decomposition theorem of the fuzzy set,aperformance distribution model based onβ-level cut set is built,and a definition of the components′performance interval in the fuzzy state and a computational method of the fuzzy state probability are provided.The traditional universal generating function is improved,theα-level cut set universal generating function and its algorithm are defined.An assessment method of fuzzy reliability of multi-state system in consideration of epistemic uncertainty is proposed,and verification and illustration are conducted with simulation examples in order to reduce the influence on components′performance analysis and achieve precise estimation of system reliability.This method overcomes the shortcoming that a state information analysis model cannot be accurately built on account of the deficiency of performance distribution information.It can be widely used and has great engineering application value.