随机冲击情况下考虑认知不确定的多态系统可靠性评估

摘要

Components may suffer from random shocks in some environments or other conditions.The performance degradation of components consists of normal performance degradation and random shocks.Since it is hard to obtain adequate performance data of high-reliability components within a short time,there are epistemic uncertainties on components and the system reliability cannot be accurately estimated.For the purpose of accurate estimation of system reliability,the distribution parameters of components'performance deterioration caused by random shocks were assumed as interval variables,the components'performance distribution model based on the interval variables was built,the definition of interval-continuous sequences of components'state performances was presented and a method to calculate the interval-valued state probability were provided.The traditionally universal generating function method was improved,the interval-valued universal generating function and its algorithm were defined,a method to assess multi-state system reliability in consideration of epistemic uncertainty under random shocks was proposed,and the verification and illustration were conducted by using simulations.The method not only overcomes the inaccuracy of the reliability analysis model,but also has strong versatility and engineering application value.%由于环境或其他原因，部件可能受到随机冲击，部件性能退化由正常性能退化和随机冲击两部分构成，对于该类高可靠部件短时间内很难得到足够的性能数据，致使对部件认知存在一定不确定性，无法准确估计系统的可靠性。为实现对系统可靠性的准确估计，假定冲击引起的部件性能损伤分布参数为区间变量，建立基于区间变量的部件性能分布模型，给出了部件状态性能区间连续序列定义和区间状态概率计算方法，对传统的通用生成函数方法进行改进，定义了区间通用生成函数及其运算法则，提出了随机冲击情况下考虑认知不确定的多态系统可靠性评估方法，并以仿真实例进行了验证说明。该方法不仅克服了性能分布信息缺少，无法准确建立状态性能分布模型的不足，且具有很强的通用性和工程应用价值。