摘要
针对多失效模式相关的机械零部件,利用Copula函数描述相关性方面的优势,提出了一种多失效模式相关机械零部件的可靠性建模方法。基于Vine Copula函数将复杂多失效相关的机械零件可靠性问题转换为对多个二维Copula函数进行分析,采用极大似然估计法和AIC信息准则识别最优二维Copula函数。机械零部件失效模式相对较少,获得失效样本就少。为了避免AIC值的变异性,基于改进参数的Bootstrap法提出了失效样本容量较小时最优二维Copula函数的识别。利用Matlab进行模拟检验了改进参数Bootstrap法可行性,进而运用Vine Copula模型来求解多失效模式相关机械零部件的可靠度。最后通过传动轴的算例与其它方法进行比较得出所述方法的合理性、有效性。
Aiming at mechanical parts for multi-failure modes,the advantages of using Copula functions to describe correlations,a reliability modeling method for multi-failure mode related mechanical parts is proposed.Based on the Vine Copula function,the reliability problem of mechanical parts related to complex and multi-failure is transformed into the analysis of multiple two-dimensional Copula functions,optimal two-dimensional Copula function is identified by using maximum likelihood estimation and AIC information criterion.Mechanical parts failure mode is relatively small,access to less failure samples.To avoid the variability of AIC values,based on improved parameters is proposed to identify the optimal two-dimensional Copula function when the failure sample size is small.The feasibility of the improved parameter Bootstrap method is verified by Matlab simulation,then the reliability of the multi-failure mode related mechanical parts is solved by using the Vine Copula model.Finally,through the comparison between the calculation example of drive shaft and other methods,the method is rationality and validity.
作者
胡启国
周松
HU QiGuo;ZHOU Song(School of Electro-mechanical&Automobile Engineering,Chongqing Jiaotong University,Chongqing 400074,China)
出处
《机械强度》
CAS
CSCD
北大核心
2019年第6期1365-1371,共7页
Journal of Mechanical Strength
基金
国家自然科学基金项目(51375519)
重庆市基础科学与前沿技术研究专项基金(cstc2015jcyjBX0133)资助~~
作者简介
胡启国,男,1967年12月生,重庆市人,汉族。重庆交通大学机电与车辆工程学院教授,博士,研究方向为机械可靠性分析及优化、机械系统动力学教学与研究;周松,男,1990年9月生,湖北广水市人,汉族。重庆交通大学机电与车辆工程学院硕士研究生,研究方向为机械系统可靠性分析及优化。
