Thermodynamic entropy to detect fatigue crack initiation using digital image correlation, and effect of overload spectrums

Abstract Current approaches to fatigue life prediction are mostly based on classical fracture mechanics and it is difficult to predict lifetime under different loading conditions, particularly under overload spectrums. A thermodynamic entropy-based approach is less sensitive to variability in loading conditions and provides a dependable measure of damage. Volumetric thermodynamic entropy is utilized in this study to measure damage. Digital Image Correlation-DIC is used to measure plastic deformation in close vicinity of the notch in AA7075-T6 specimens. The plastic deformations are converted to a corresponding thermodynamic entropy, using a modified Clausius-Duhem inequality. A comparison of thermodynamic entropy generation using global measurement consisting of load frame data and Linear Variable Differential Transformer-LVDT extensometer, and local measurement using DIC data is done. The entropic-endurance level for fatigue crack initiation regardless of loading spectrums is shown to scatter within a short interval and is represented by a Weibull distribution.