Experimental investigations on parameter identification and failure predictions of titanium alloy by Gurson–Tvergaard–Needleman model

Abstract

This study focused on the experimental investigations on parameter identification and failure prediction of Ti-6Al-4 V by Gurson–Tvergaard–Needleman (GTN) model. Two approaches, i.e., direct (by image processing of SEM observations) and indirect measurements (by the reduction of effective Young’s modulus or FE trail-and-error), were used to determine the GTN parameters. Through digital image processed SEM micrographs of uniaxial tensile specimens at different deformation stages, the developments of void volume fraction (VVF) were obtained and compared with that from indirect measurements. The three nucleation parameters can then be analytically derived from this experimentally obtained VVF developments. The critical VVF at void coalescence was directly measured through a SEM observation of the section surface around the fracture region of uniaxial tensile specimen, and this value is also consistent with that from indirect measurement. However, as the matrix of a fracture surface is characterized with an obvious varying grayscale due to its dimple morphology, the VVF at final fracture is suggested to be determined from FE trial-and-error rather than a direct measurement of the fracture surface to avoid error induced by subjective judgements. Consistency of the determined GTN parameters was verified through failure predictions on notched tension (R = 6, 20, 100), notched Arcan tests (R = 1, 3, 5), and cracked Arcan tests (SECS, DECS, and CCS), which proved that GTN model with parameters from this experimental investigation cannot only predict the macroscopic mechanical behavior of ductile metals, but also well match the entire damage evolution, from void nucleation to final fracture.