AUTOMATED FATIGUE DATA FITTING USING STRUCTURAL SHIFT DETECTION

A set of data fitting algorithms are presented for determining the cyc lic material properties used in the local strain approach to fatigue l ife prediction. These properties are obtained by fitting power-law cur ves to the elastic and plastic components of uniaxial strain-life fati gue data obtained under constant amplitude testing. It is well known t hat these power-law relationships may be expected to be valid only wit hin a limited range of strain amplitude; it is not uncommon however, t o have significant amounts of test data outside of this range. The alg orithms presented here address this problem by applying a structural s hift detection scheme to the elastic and plastic component life data s ets to identify the limits of the valid range; The curve fits obtained are then thefts of only the data lying within this range. The fitting approach is illustrated for two sets of SAE 1045 steel data. In the e xamples shown, proper fitting of the curves is seen to improve the qua lity of life prediction by as much as a factor three.