齿轮的概率寿命预测与弯曲疲劳试验  

Probabilistic Life Prediction and Bending Fatigue Test for Gear

作　　者：李铭[1] 谢里阳[1] 张宇[1] 白鑫[1] 胡杰鑫 

机构地区：[1]东北大学机械工程与自动化学院,沈阳110004

出　　处：《机械工程学报》2017年第18期169-175,共7页Journal of Mechanical Engineering

基　　金：国家科技支撑计划(2014BAF08B01);国家自然科学基金重点(51335003)资助项目

摘　　要：精密长寿命齿轮的制造成本高、周期长,针对这种齿轮的疲劳试验更是费时费力。为了简单、准确地得到齿轮的寿命信息,基于最小次序统计量的概念建立齿轮的概率寿命预测模型,模型建立不同齿数的齿轮之间的概率寿命关系,其将特定齿轮的寿命数据作为输入变量,可以转化得到其他齿轮的概率寿命。在恒定应力水平下进行齿轮弯曲疲劳试验,在试验过程中说明了齿轮失效的监测与判据方法,同时将齿轮的失效机理分析与断口分析相结合,解释了齿轮的失效模式,最后借助试验数据详细阐述了模型的使用方法。利用随机截尾数据统计方法对模型进行验证,证明模型具有良好的预测能力和处理小样本数据的优势。Manufacture of the precision and long life gears has a characteristic of high cost and long cycle, furthermore, fatigue test for the gears is time-consuming and laborious. In order to get the life information of gears accurately and conveniently, the concept of minimum order statistics is used to establish a model for predicting the probabilistic life of gears. Then the probabilistic life relationship of gears with different number of teeth is established. The life data of specific gear are taken as input variable for the model, after translating by model the probabilistic life of other gears can be obtained. At the same time, a gear bending fatigue test is carried out at a constant stress level, and in the course of the test, the method of monitoring and evaluating for gear failure is explained. Next, two kinds of analysis of failure mechanism and fracture for gear are combined to show the failure mode, and test data are employed to show the instructions of the model. The model is finally validated by a statistical method of random censoring data, and the advantages of model with good prediction ability and dealing with small amount of data are proved.

关 键 词：齿轮疲劳试验 概率寿命 最小次序统计量 可靠性模型 随机截尾数据 

分 类 号：TB302[一般工业技术—材料科学与工程]