基于改进灰色模型的疲劳寿命预测及可靠性评估  

Fatigue life prediction and reliability evaluation based on modified grey model

作　　者：张爱菊[1] 张金豹[2] 张改丽[1] 

机构地区：[1]中国北方车辆研究所,北京100072 [2]哈尔滨工业大学机电工程学院,黑龙江哈尔滨150001

出　　处：《哈尔滨轴承》2015年第4期37-42,共6页Journal of Harbin Bearing

摘　　要：针对试验件中对数疲劳寿命均值随应力水平具有线性关系,但方差却具有波动性的状况,引入改进灰色模型对两者进行仿真分析。采用粒子群优化算法对模型多个参数进行估计。结果显示对数疲劳寿命均值与试验数据吻合良好,具有很好的线性关系;波动方差的拟合曲线顺滑,不仅清晰地定性显示其发展趋势,还能够定量地对均值和方差进行预测,尤其是对疲劳极限以下的寿命均值和方差的预测,避免了经验公式的使用。利用改进灰色模型预测得到的疲劳寿命均值和方差,结合疲劳累积损伤模型对块谱作用下的试验件可靠性进行了评估;通过中位秩计算试验件的失效概率,验证了可靠性仿真结果的正确性。The modified grey model was introduced to simulate the mean value of the logarithm fatigue life with linear relationship corresponding to the load level, and the variance with fluctuation. The multiple parameters of the model was estimated with particle swarm optimization. The result shows that the mean value fits well with the experiment data and has an excellent linear relationship; the fitting curve of the fluctuation variance is smoothed and demonstrates the growing trend clearly as well as quantificationally predicts the mean value and the variance, expecially for the ones under the fatigue limit, which avoids the employ of the empirical formula. With the mean value and the variance predicted by the modified grey model, the reliability was evluated under the block spectrum combining with the fatigue accumulated damage model. The simulation result of the reliability is verified with the failure probability computed by the median rank.

关 键 词：改进灰色模型 粒子群优化 寿命预测 疲劳累积损伤模型 可靠性评估 

分 类 号：TG115.57[金属学及工艺—物理冶金]