摘要
考虑结构和材料等随机参数对零件动态可靠性的影响,避免轴系产生共振,利用Workbench软件建立扭矩轴参数化有限元模型.用模态分析法求解前六阶固有频率和临界转速,并与传递矩阵法比较,证明了结构和转速设计的合理性.结合谐响应分析,说明共振失效应考虑一阶固有频率.通过响应面设计和拉丁超立方抽样法完成对扭矩轴结构、材料参数的抽样,运用BP神经网络拟合一阶固有频率的功能函数,并求解随机参数的可靠性灵敏度.采用一次二阶矩法(FOSM)计算轴在特定转速下的可靠度,并用Monte-Carlo模拟法(MCS)进行了验证,说明转速设计较为可靠.通过灵敏度分析,明确了对扭矩轴动态可靠性影响最大的因素,为轴的稳健优化设计奠定基础.
The Workbench software was used to establish a parametric finite element model of the torque shaft.Through modal analysis and comparison with the transfer matrix method,the first six orders of natural frequency and corresponding limit speed were obtained to verify the rationality of the structure and speed design.The harmonic response analysis illustrated that resonance invalidation should consider the first natural frequency.In addition,the response surface design and Latin superpower square sampling methods were used to realize the sampling analysis of the structure and material parameters.The BP neural network technology was used to fit the functions of first order natural frequency,solving the reliability sensitivity of each random parameter.The first order second moment and Monte-Carlo simulation were used to calculate the reliability at a specific speed and to find out the biggest influencing factor on the dynamic reliability,thus laying a foundation for the robust optimization design of shaft.
作者
杨周
姜超
张义民
姜红猛
YANG Zhou;JIANG Chao;ZHANG Yi-min;JIANG Hong-meng(School of Mechanical Engineering&Automation,Northeastern University,Shenyang 110819,China;School of Mechanical Engineering,Shenyang University of Chemical Technology,Shenyang 110142,China)
出处
《东北大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2020年第2期217-222,共6页
Journal of Northeastern University(Natural Science)
基金
国家自然科学基金联合基金资助项目(U1710119)
关键词
扭矩轴
有限元
BP神经网络
可靠度
可靠性灵敏度
torque shaft
finite element
BP neural network
reliability
reliability sensitivity
