摘要
针对具有局部应力集中的机械零件对实际工程结构的损害问题,采用神经网络技术模拟基本随机变量和应力集中系数之间的关系,能够直接获得应力集中系数的显性表达式。在这个基础上,将神经网络技术获得的应力集中系数的显性表达式和可靠性理论相结合,解决应力集中情况下的机械零件结构可靠性问题。结果表明:模拟得到的应力集中系数显性表达式能够直接应用到机械零件的可靠性分析中,获得可靠性计算结果。这将为工程结构的可靠性分析提供了理论依据,而且提出了机械零件结构可靠性分析的新方法。
Aimed at the damage problem of the practical engineering structure brought by mechanical components with local stress concentrations,neural network（NN）technique has been adopted to simulate the relationship between the basic random variables and the stress concentration factors（SCF）so that an explicit expression of the stress concentration factors can be obtained directly.On the basis of this,the explicit expression of the stress concentration factors given by NN technique is combined with the reliability theory to solve the structural reliability problems of the mechanical components in the case of the stress concentration.The results show the explicit expression can be directly applied to perform the structure reliability analysis for mechanical components in order to obtain the reliability calculation result.This will provide theoretical foundation for reliability analysis of engineering structure,meanwhile a corresponding new method of reliability analysis for mechanical components has been presented.
出处
《机械设计与制造》
北大核心
2011年第7期260-262,共3页
Machinery Design & Manufacture
基金
陕西省自然科学基金资助项目(2007E218)
关键词
局部应力集中
神经网络
可靠性理论
应力集中系数
显性表达式
可靠性分析
理论依据
Local stress concentration
Neural network
Reliability theory
Stress concentration factors
Explicit expression
Reliability analysis
Theoretical foundation
