疲劳寿命呈现异方差特性,其标准差随弹性应变幅和塑性应变幅的减小而增大,因此在Manson-Coffin公式中引入标准正态变量斗和线性标准差σe、σp,将ε-N曲线参数表示为随机变量μ的函数,建立了低循环疲劳寿命的概率模型。在低循环疲劳试验数据的基础上应用异方差回归分析方法获得了该模型的参数,通过蒙特卡洛模拟验证了模型的精度。应用该模型进行等厚空心轮盘的低循环疲劳寿命可靠性分析,得到了轮盘中心孔危险点的疲劳寿命分布。由于没有事先假设疲劳性能参数的分布特性,参数均由试验数据分析得到,并且采用异方差回归分析能够充分利用数据信息,在提高分析精度的同时能够节约疲劳试件,因此该分析方法具有较好的工程应用前景。
The standard deviation of fatigue life increases with reducing of elasticity and plasticity strain components. A standard normal random variable μ and linear standard deviations σe and σp were introduced into Manson-Coffin equation,so the probabilistic model of fatigue life was established in which the f our parameters of Manson-Coffin equation were expressed as the functions of μ The parameters of the probabilistic model were obtained by the heteroscedastic regression analysis method based on the LCF test data,and the precision was tested by Monte-Carlo simulations for fatigue lives of several strain levels. The relaibility analysis of a disk was achieved by this model. This reliability analysis method was suit for engineering application for the variables' distributions was not assumed in adwance and the heteroscedastic regression analysis could not only inereas precision of analysis but also save test specimens.