对汽车钢圈疲劳试验寿命的模糊预测  

Fuzzy Prediction of the Fatigue Test Life for a Steel Wheel

作　　者：郑战光 王佳祥 杨兴 汪兆亮 

机构地区：[1]广西大学机械工程学院广西制造系统与先进制造技术重点实验室,广西南宁530004

出　　处：《机械设计与制造》2017年第10期85-87,91,共4页Machinery Design & Manufacture

基　　金：广西自然科学基金(2012GXNSFBA053145);广西制造系统与先进制造技术重点实验室项目(14-045-15S04);玉林市科技攻关项目(玉市科攻1421001)

摘　　要：汽车钢圈的疲劳台架实验时间长,费用高,数据量少,分散性大,给钢圈的设计与疲劳寿命评测带来了巨大的挑战。为了克服疲劳试验的天生缺陷,首先采用K-S检验对钢圈疲劳试验寿命数据的分布类型进行判断;再依据模糊贝叶斯理论进一步确定钢圈的疲劳试验寿命符合对数正态分布形式;最后采用对数正态概率分布模型对钢圈疲劳寿命进行预测。结果显示对数正态分布的似然函数值为0.448,较其它分布要大,采用对数正态分布描述汽车钢圈的疲劳台架实验数据更加合理,进而预测钢圈的疲劳寿命为6万个循环左右。这对于帮助研究人员合理进行汽车钢圈的设计与疲劳寿命评测具有一定的参考意义。Fatigue test of steel wheel has such traits as long time and high cost, little data and large scatter. These bring a challenge to the design and the fatigue life evaluation of the wheel. In order to overcome these defects, firstly, K-S test is used to found a distribution of fatigue life of a wheel; secondly, bayes method is used to determine a probability distribution of fatigue life and the distribution is log-normal; finally, the fatigue life of a wheel can be predicted by log-normal distribution model. The results demonstrated that the likelihood function value of log-normal distribution is 0.448 and log-normal distribution can be better described than others, at the same time, the fatigue life of wheel is forecasted about 60000 cycles. There is a certain significance to help the researchers in improving the design and the fatigue life evaluation of a wheel.

关 键 词：疲劳寿命 K-S检验 贝叶斯方法 对数正态分布 

分 类 号：TH16[机械工程—机械制造及自动化] TG405[金属学及工艺—焊接]