航空高频自润滑衬垫材料寿命预测  

Life prediction of self-lubricating liner for aviation high frequency

作　　者：郝秀红[1,2] 王树强[1,2] 潘登 HAO Xiuhong;WANG Shuqiang;PAN Deng(School of Mechanical Engineering,Yanshan University,Qinhuangdao,Hebei 066004,China;Aviation Key Laboratory of Science and Technology on Generic Technology of Self-lubricating Spherical Plain Bearing,Yanshan University,Qinhuangdao,Hebei 066004,China)

机构地区：[1]燕山大学机械工程学院,河北秦皇岛066004 [2]燕山大学自润滑关节轴承共性技术航空科技重点实验室,河北秦皇岛066004

出　　处：《燕山大学学报》2021年第1期17-24,共8页Journal of Yanshan University

基　　金：国家自然科学基金资助项目(51605418);河北省自然科学基金资助项目(E2016203206)。

摘　　要：针对航空高频自润滑衬垫材料寿命试验周期长、样本量少问题,提出了一种基于蒙特卡洛模拟的短时小样本寿命预测方法。基于自润滑衬垫材料稳定磨损期短时数据,采用支持向量回归算法完成了磨损量退化轨迹拟合,得到了磨损量第一和第二特征参数向量及其分布模型,并基于蒙特卡洛模拟和逆变换法进行了样本量扩充。最后,基于自润滑衬垫材料失效判据得到了扩充样本的服役寿命及其分布模型,给出了其相应的可靠性寿命指标。结果表明:短时小样本寿命预测方法具有较高的预测精度,且最小样本量可低至3,这为大幅度缩短航空自润滑衬垫材料的寿命试验周期提供了理论支持。A new life prediction method with the short time and small sample is proposed based on the monte carlo simulation by considering the problem of the long service life and small sample size of the aviation high frequency self-lubricating liner.A support vector machine is adopted to complete the degradation trajectory of the wear based on the short time data of the stable wear period.The vectors and the distribution models of the first and second characteristic parameters are obtained.The sample size is expanded based on monte carlo simulation and the inverse transformation method.The service life and the corresponding distribution model of the extended sample are calculated,and the corresponding reliability life index are given based on the failure criterion of self-lubricating liner.The results show that the life prediction method with the short time and small sample has a high prediction accuracy,and the minimum sample size can be as low as 3.These results provide the theoretical support for shortening the life test cycle of aviation self-lubricating liner.

关 键 词：蒙特卡洛模拟 支持向量回归 自润滑衬垫 短时小样本 寿命预测 

分 类 号：TH133.3[机械工程—机械制造及自动化]