Abstract: Aiming at the problem that choice blindness for interval uncertainty parameters existed in reliability analysis of mechanical components,a global sensitivity and reliability analysis method was proposed based on partial elasticity theory.The unknown or highly nonlinear limit-state function was obtained by using response surface methodology,and a surrogate model that quadratic polynomial limit-state function containing interval uncertainty parameters was constructed.Based on functional relation and combined with partial elasticity theory,the corresponding rule of addition and multiplication were established to identify the interaction effect between parameters,and the calculation method of sensitivity index was derived.The smaller influence direction of interval uncertainty parameters change on the reliability was determined according to the sensitivity index,and then the choice reliability domain range for parameter values was .The feasibility and practicability of proposed approach was verified by taking the reliability sensitivity analysis of rolling bearing as the example.

Key words: interval uncertainty parameters, response surface methodology, reliability, partial elasticity theory, global sensitivity analysis

摘要： 针对机械零部件可靠性分析时存在区间不确定性参数盲目性选择问题,提出一种基于偏弹性理论的全局敏感度与可靠性分析方法。通过用响应面法获取未知或高度非线性的极限状态函数,构建了包括区间不确定性参数的二阶多项式极限状态函数代理模型。以该函数关系为基础,结合偏弹性理论,建立了相应的加法法则和乘法法则来辨识参数间交互作用的来源,并推导了敏感度指数的计算方法。根据敏感度指数确定出区间不确定性参数的改变对可靠性影响较小的方向,进而指导参数选择时的可靠域取值范围。以滚动轴承的可靠性敏感度分析为例,验证了该方法的可行性和实用性。

关键词: 区间不确定性参数, 响应面法, 可靠性, 偏弹性理论, 全局敏感度分析