AN IMPROVED TECHNIQUE FOR THE PREDICTION OF AXIAL FATIGUE LIFE FROM TENSILE DATA

This paper introduces an improved technique over existing methods for the prediction of axial fatigue life from uniaxial tensile data. The method uses multilinear regression analysis for initial prediction of the elastic and plastic strain amplitudes at 10(6) and 10(4) reversals to failure respectively from uniaxial tensile data. The fatigue strength coefficient is assumed to be equal to the true fracture stress of the material. The fatigue ductility coefficient is approximated to be equal to the logarithmic ductility of the material. The fatigue strength exponent and fatigue ductility exponent are then located on the strain-life curve by making use of the estimated elastic/plastic strain amplitudes, logarithmic ductility and the true fracture stress of the material. The validity of the method was investigated by examining 49 steels whose material properties cover a wide range of variables that might affect fatigue behaviour.