论文部分内容阅读
与确定疲劳可靠性曲线的原理一致 ,研究线性均值曲线拟合S—N数据误差的统计分布。根据误差数据的统计特点 ,指出可能的假设分布有三参数Weibull、正态、极大值和极小值 4种分布。提出良好假设的确定方法 ,包括 4个原则 ,即假设分布形状与误差数据真实分布形状的一致性、预测的总体拟合效果、与疲劳失效机制的一致性和尾部预测的安全性。 16Mn钢焊接接头试验S—N数据的分析 ,验证方法的有效性 ,说明应用一般推断方法 ,4种分布都可合理地作为数据的统计模型 ,考虑 4原则后 ,极大值和正态分布是良好假设分布。
In keeping with the principle of determining the fatigue reliability curve, the statistical distribution of the linear mean-curve fitting S-N data error is studied. According to the statistical characteristics of the error data, it is pointed out that there are four distributions of possible hypothesis distributions: Weibull, normal, maxima and minima. The method of determining good hypotheses is proposed, including four principles, that is, the consistency of the distribution shape of the assumed distribution and the true distribution of the error data, the overall fitting effect of the prediction, the consistency with the fatigue failure mechanism and the safety of the tail prediction. 16Mn steel welded joint test S-N data analysis to verify the effectiveness of the method, indicating the use of general inference methods, four kinds of distribution can be reasonably as a statistical model of the data, consider the four principles, the maximum and normal distribution is Good hypothetical distribution.