Slip Systems and Initiation of Plasticity in a Body-Centered-Cubic Titanium Alloy

• Published in: Metallurgical and materials transactions. A, Physical metallurgy and materials science Physical metallurgy and materials science, 2010-10, Vol.41 (10), p.2522-2531
• Main Authors: Lewis, Alexis C., Qidwai, Siddiq M., Geltmacher, Andrew B.
• Format: Article
• Language: English
• Subjects: Accuracy; Applied sciences; Characterization and Evaluation of Materials; Chemistry and Materials Science; Cost control; Data mining; Exact sciences and technology; Materials Science; Mechanical properties and methods of testing. Rheology. Fracture mechanics. Tribology; Metallic Materials; Metals. Metallurgy; Nanotechnology; Simulation; Structural Materials; Studies; Surfaces and Interfaces; Thin Films
• ISSN: 1073-5623; 1543-1940
• DOI: 10.1007/s11661-010-0284-5

To determine statistically relevant microstructure-yield correlations in three-dimensional (3-D) microstructures, large volumes comprised of many grains must be studied. With the aim of limiting computational loads without reducing the fidelity of the volume being simulated, this work investigates the use of reduced constitutive parameters, specifically the number of available slip systems, to analyze initial plastic flow in the microstructure. This is performed by embedding a 3-D reconstruction of a single-phase beta-Ti microstructure in a finite element (FE) computational model and subjecting it to a number of loading conditions. Three separate single-crystal plasticity formulations were used for each loading: reduced 12 slip systems (〈111〉{110} family), reduced 24 slip systems (〈111〉{110} + 〈111〉{112} families), and full 48 slip systems (〈111〉{110} + 〈111〉{112} + 〈111〉{123} families). The analysis results show that the 24-slip-system model accurately predicts the global stress-strain behavior and locations of initial yield under all loadings, with no more than 10 pct error in the spatial description of local state variables compared to the full 48-slip-system model. The 12-slip-system model generally follows the full model predictions and provides an even better cost improvement, but with errors in excess of 40 pct in local descriptions. Computational cost and data reduction are improved by 26 and 53 pct, respectively.