Fatigue Test by Staircase Method with Small Samples

Abstract

By the conventional method of the fatigue test, the accuracy of the endurance limit cannot be estimatd. On the other hand, the statistical methods represented by the probit method, the staircase method and so on require a large number of specimens. In this report the fatigue test by staircase method with small samples is presented. The step size of the staircase method should be chosen as close to the standard deviation σ of the endurance limit as possible. When σ is unknown, the following procedure is recommended. The conventional fatigue test is performed in the range where S-log N curve can be replaced by the straight line with a slope. The experimental results are plotted on the S-log N diagram. The linear regression analysis is applied to find σ. The assumptions included in this procedure are discussed. The methods of computing the confidence interval of the mean and the lower tolerance limit of the endurance limit were investigated. The lower tolerance limit of the endurance limit is a statistical concept and the conventional term "lower limit of the endurance limit" is defined from the statistical point of view. Finally these methods were applied to some experimental examples.