Characterization and modelling of quasi-static and fatigue notch effect in AZ31B-H24 rolled sheet
Loading...
Date
2018-10-25
Authors
Qian, Lin Feng Victor
Journal Title
Journal ISSN
Volume Title
Publisher
University of Waterloo
Abstract
The transportation sector releases large amounts of harmful greenhouse gases into Earth’s atmosphere. Vehicle weight reduction through the use of magnesium alloys is an effective way to reduce carbon dioxide emission and improve fuel economy. Magnesium is an attractive candidate for use as a structural material. In cars, notched structural members are common. Understanding of the fatigue behaviour of notched magnesium members is needed.
The monotonic and fatigue properties of AZ31B-H24 rolled sheet were characterized. It exhibits strong tensile-compressive yield asymmetry and weak anisotropy. At varying strain levels under strain-controlled loading, the response exhibits several characteristics, including yield asymmetry, mean stress evolution, cyclic strain hardening. Cyclic ratcheting is observed in stress-controlled tests with tensile mean stresses.
A finite-element model was created for the notched specimen. Three material models from Abaqus and LS-DYNA were compared. LS-DYNA material model 233 (MAT_233) was found to perform well in predicting residual strain at the notch root in quasi-static tension-unload as well as cyclic response under stress-controlled loading.
Smith-Watson-Topper (SWT) and Jahed-Varvani (JV) fatigue-life model parameters were obtained from fully-reversed strain-controlled tests on smooth specimens. The fatigue strength at 10 million cycles was estimated using a Weibull function to be about 80 MPa. A ‘direct-fit’ (df) method is proposed for SWT and JV. Fatigue life predictions using SWT, JV, SWT-df, and JV-df were compared against experimental stress-controlled tests. The direct-fit variants were found to be significantly better predictors of life under strain-controlled loading and marginally better under stress-controlled loading.
SWT with Miner’s rule is able to predict reasonably the fatigue lives for variable-amplitude loading (VAL) strain-controlled tests. However, all of the models examined underestimate the fatigue life of notched specimens in VAL load-controlled tests.