Estimation of mode I quasi-static fracture of notched aluminum–lithium AW2099-T83 alloy using local approaches and machine learning

Aluminum–lithium (Al–Li) alloys offer superior performance under different conditions that involve mechanical loading, high temperature and corrosive environment. Therefore, Al–Li alloys are being used in the defense, aerospace, and aircraft industries, specifically in structural parts such as fuselage, empennage, and wings. By modifying the chemical composition, a third generation of Al–Li alloys has been introduced to overcome the mechanical and thermal shortcomings of previous Al–Li generations. As structural parts usually contain notches, it is of paramount importance to study the strength of the newly introduced alloys in the presence of such geometrical discontinuities to select the suitable alloy for a particular application. The aim of this work is to analyze the strength of U and V-notched specimens machined from extruded AW2099-T83 Al–Li alloys under quasi-static loading. The fracture stress of specimens with various notch radii and angles were estimated using strain energy density (SED) and the theory of critical distances (TCD) methods. A support vector machine (SVM) regression model was also implemented to assess the applicability of estimating fracture stress using machine learning approaches. The results show an average absolute discrepancy of 6.6 %, 7.8 % and 2.8 % for SED, TCD and SVM methods, respectively.