A Calibration Approach for Accelerated Creep Testing for Electron Beam Melted Ti-6Al-4V Using the Wilshire–Cano–Stewart Model

Abstract

Abstract This study outlines a model calibration approach for an accelerated creep test called the dynamic negative stepped test (DNST) to enable the rapid screening of creep-resistant materials. In DNST, stress is stepped decreased based on the attainment of a sufficient minimum-creep-strain-rate (MCSR) at each stress level. Steps are repeated, torturing the material, until rupture occurs. The DNST is advantageous as a screening test for new alloys. Alloys and heats with superior creep resistance will be able to survive longer and with greater ductility than those with poor creep resistance. The calibration of a constitutive model to DNST data furnishes predictions of the conventional creep response being between 65 h and 6685 h from the relatively short (<130 h) DNST Data. In this study, DNSTs are performed on electron beam melted (EBM) Ti-6Al-4V at 650 °C with stepping through 150, 75, 60, and 50 MPa. Six build orientations are tested including 0 deg, 30 deg, 45 deg, 60 deg, 90 deg, and V (vertical) direction. The Wilshire–Cano–Stewart (WCS) model is employed to calibrate the experimental data. A systematic calibration approach is adopted. Each step is calibrated numerically. A unique set of minimum-creep-strain-rate (MCSR) and stress-rupture (SR) related material constants, i.e., the Wilshire and Sinh constants are obtained for each build direction. A nonhomogenous objective function is used to numerically optimize the strain trajectory and damage trajectory constants. To find the best-fit curve, the strain trajectory constants, and damage trajectory constants are numerically refined for each step. The WCS model shows a near-perfect prediction of the DNST data. Based on the calibrated constants, conventional creep curves are generated in order to determine which build orientations are likely to exhibit poor, moderate, and superior creep resistance. Predictions of MCSR and SR curves over a wide stress range are estimated outside the experimental range to investigate the extrapolation pedigree of the approach. This will allow the material designers to have more confidence in DNST-generated test data for predicting long-term creep response and structural lifetime.