摘要
通过对各种疲劳门槛值的理论模型的分析 ,认为结构钢的拉伸性能指标和循环加载条件是裂纹体疲劳门槛值的决定因素 .建立了用屈服强度、抗拉强度、断裂延性和循环加载应力比预测疲劳门槛值的人工神经网络模型 ,并用 1 0种结构钢的 60个样本对该模型进行了训练 .结果表明 ,人工神经网络模型可以很好地描述疲劳门槛值与结构钢拉伸性能指标及应力比之间复杂的定量关系 .应用所训练的人工神经网络模型预测了部分结构钢的疲劳门槛值 。
According to the analysis of some theoretical models for predicting fatigue threshold of a cracked body, it is held that the tensile properties of structural steels and the stress ratio under recycle loading are main dominant factors of the fatigue threshold. A neural network model for predicting thefatigue threshold is established, in which yield strength, tensile strength, fracture ductility of structural steels and stress ratio in recycle loading are included. The model is learnt by 60 samples of 10 structural steels, and the results show that the neural network model can well represent the complicated quantitative relationship between fatigue threshold and the tensile properties and the stress ratio. The learnt neural network model is used to predict the fatigue thresholds of some steels, and the predicted results tally well with those obtained by experiments or analysis.
出处
《西安石油学院学报（自然科学版）》
2003年第3期50-54,共5页
Journal of Xi'an Petroleum Institute（Natural Science Edition)
