【摘要】 材料的疲劳损伤是一非稳态过程 ,难以用传统的数学函数准确表达。摒弃了能量均匀耗散及线性累积损伤理论假设 ,采用神经网络技术准确描述其与材料性能、载荷应力间复杂的非线性映射关系 ,真实描述了材料疲劳损伤过程 ;建立了疲劳失效动态准则 ,综合考虑了材料疲劳性能及载荷的随机性 ,能够正确反映两者的个体性能及相互关系 ,并运用离散事件仿真原理 ,构造了材料疲劳可靠性分析的智能仿真系统 ,用于准确地进行疲劳寿命可靠性分析。经正火 3 5钢随机载荷疲劳实验验证 ,其可靠性分析精度较高更多还原

【Abstract】 The fatigue damage of materials is an unsteady process and is difficult to be expressed accurately by the use of traditional mathematical functions. This paper abandoned the theoretical hypothesis of energy even dissipation and linear accumulative damage and adopted the techniques of neural network to describe accurately the complicated non-linear mapping of its relationship with material properties and loading stress, thus really depicted the fatigue damage process of materials. It established the dynamic criterion of fatigue failure which considered comprehensi vely the fatigue properties of materials and the randomness of loading and could reflect correctly the individual characters and mutual relationship between them.Using simulation principle of discrete event to construct the intellectual simulation system for fatigue reliability analysis of materials so as to carry out accurately reliability analysis of fatigue life. Through fatigue testing of normalizing 35 steel under random loading that its accuracy of reliability analysis is verified fairly high.更多还原