The fatigue design process is addressed using a probabilistic methodology. The structural fatigue reliability prediction methodology is based upon a damage fraction concept. Reliability predictions are based on Palmgren-Miner's Rule and reliability definitions. From experimental test data, it has been shown that the number of cycles to failure is a random variable and can be represented by a lognormal probability density function. Thus, "Miner's Rule," or the damage fraction, is considered a function of a random variable. An expression for fatigue reliability utilizes an excepted design criteria for predicting structural integrity. Engineering analysis tools incorporating this criteria are developed for the purpose of comparing designs under both the constant and multiple amplitude cyclic stress model cases.;The reliability prediction methodology is applied to multiple amplitude sinusoidal cyclic stress models. It has been shown that the methodology is credible in comparison to results from the direct integration method, multiple convolution integration method, Fourier transform technique. and Monte Carlo simulation. Polynomial equations have been incorporated as an enabling technology to approximately describe the characteristic functions for the lognormal distribution and are utilized with the Fourier transform technique. Computer implementation of the four mathematical methods is provided in the form of Mathematica Version 3.0 computer programs. An assessment of these four methods was achieved through comparisons of computational accuracy and efficiency.
展开▼
机译：疲劳设计过程使用概率方法解决。结构疲劳可靠性预测方法基于损伤分数概念。可靠性预测基于Palmgren-Miner规则和可靠性定义。从实验测试数据中可以看出，失效的周期数是一个随机变量，可以用对数正态概率密度函数表示。因此，“矿工规则”或损害分数被认为是随机变量的函数。疲劳可靠性的表述利用例外的设计标准来预测结构完整性。开发了包含此标准的工程分析工具，目的是在恒定和多振幅循环应力模型情况下比较设计。可靠性预测方法应用于多振幅正弦循环应力模型。与直接积分法，多重卷积积分法，傅立叶变换技术的结果相比，该方法是可靠的。和蒙特卡洛模拟。多项式方程已作为一种使能技术被纳入，以近似描述对数正态分布的特征函数，并与傅立叶变换技术一起使用。这四个数学方法的计算机实现以Mathematica Version 3.0计算机程序的形式提供。通过比较计算精度和效率对这四种方法进行了评估。
展开▼
