As a basic study on fatigue damage accumulation of structural members (for example; truck frame and plate spring of rolling stocks) subjected to service load, the authors have carried out fatigue tests under superposed stress wave on fifteen metalic materials (about 400 experimental points.)
In this paper, a few factors having influence on the predicted fatigue life have been discussed by comparing the experimental life N with the predicted life Nes with regard to these experimental points.
For the fatigue life prediction method, three kinds of stress cycle counting methods were used; the zero cross range pair mean method (ZRPM method), the range pair mean method (RNPM method) and the range pair method (RNPR method).
The modified Goodman's method was used to evaluate the effect of mean stress on stress amplitude. The lives were predicted by using the linear cumulative fatigue damage law.
The results obtained were as follows.
The most suitable cycle counting method was RNPM method, because of the scatter band of N/Nes being narrowest in three cycle counting methods. The mean stress effect on fatigue life amplitude was very complex, and the modified Goodman's method was not always adaptable for the prediction of the above experimental points.
In order to make narrow the scatter band of N/Nes, some discussions were made on the fatigue damage accumulation caused by the secondary stress amplitude cycling.