Modeling the Residual Stress Distribution and Experimental Characterization of Shot Peening on AZ31B Rolled Sheet

Abstract

The drive to reduce fossil fuel consumption due to its environmental impacts has generated renewed interest in employing magnesium (Mg), the lightest industrial metal, and its alloys in vehicle manufacturing. One of the qualifying metrics for structural application of Mg in transportation vehicles is its high durability. The low fatigue strength of these alloys has been an obstacle to using them in load-bearing components. Thus, methods for improving the fatigue properties of Mg alloys are of interest. Shot peening is a cold-working process employed to improve the fatigue properties of materials. The shot peening process induces compressive residual stress at the material’s surface and at a layer in the order of a few hundred micrometers deep, which improves the fatigue life by retarding the crack initiation as well as growth; however, the increased surface roughness has a detrimental effect on fatigue life. These competing effects of peening have created interest in finding the optimum peening intensity that will maximize fatigue life. Modeling reduces the cost of experimentally evaluating optimum peening conditions. However, modeling the shot peening of Mg alloys remains complicated due to the anisotropic and asymmetric properties of wrought Mg alloys, and the complex unloading behavior and rate-sensitivity behavior of these materials. To address these challenges, a comprehensive experimental and numerical-analytical study of shot peening on AZ31B-H24 rolled sheet was conducted and is reported in this thesis. First residual stress distributions through the depth of the material were measured. Among the methods for residual stress measurement, X-ray diffraction (XRD) has attracted researchers’ attention because: 1) it is a non-destructive method; 2) it can measure residual stress at the surface, and 3) the spatial resolution can be less than 0.3 mm. However, due to the low x-ray mass attenuation coefficient of Mg alloys, x-ray penetration in the material is significant which needs to be accounted for. The residual stresses in as-received and shot peened AZ31B-H24 rolled sheet samples were measured using the 2D-XRD method. The electro-polishing layer removal method was used to find the residual stress pattern through depth. Due to the high depth of penetration, a correction had to be made to account for the penetration depth. The results showed that the corrected residual stresses in a few tens of micrometers layer from the surface were different from the raw stresses. To better estimate the residual stress distribution in a few micrometers from the surface, the grazing-incidence x-ray diffraction (GIXD) method was applied to evaluate the stresses in the surface layer. This study also showed how small uncertainty in measuring the observed residual stress and in evaluating the depth of the polished area in layer removal leads to high uncertainty in the corrected residual stresses. The XRD results showed the creation of compressive residual stress through the depth as well as a good agreement between the XRD and hole-drilling and GIXD results. Modeling the shot peening process first requires an understanding of how Mg alloys behave at large strain values during loading-unloading. The tension-compression (TC) and compression-tension (CT) in the in-plane directions were obtained using an anti-buckling fixture. By comparing the compression part of the CT curves along the rolling direction (RD) with the ones using a cuboid sample, the negligible effect of using the anti-buckling fixture was shown. A novel fixture was designed to obtain the CT and TC curves in the through-thickness (normal direction: ND) of the rolled sheet, which is only 6.3 mm thick. FEM was employed to evaluate the consistent area for strain measurement using DIC in the designed setup. The CT and TC curves along ND were obtained using the new fixture. The results of the new fixture were verified by comparing the curves obtained by the new fixture in RD with those obtained by using the anti-buckling fixture. Different effects of shot peening on the AZ31B-H24 rolled sheet were characterized in this study by measuring the residual stress and micro-hardness distribution through the depth, followed by measuring surface roughness and texture evolution at the surface of samples shot peened under Almen intensities ranging from 0.05 mmN to 0.6 mmN. To obtain the optimum peening intensity, rotating bending fatigue tests were performed on peened samples at different intensities. It was found that increasing the peening intensity, increases the surface roughness and hardness at the surface layer. In addition, the depth of the maximum compressive stress and the depth of the induced compressive residual stress layer have a direct relation with the peening intensity. The material showed a high sensitivity to shot peening under different intensities, due to the over-peening effects in the peening on Mg alloys. Peening at the optimum intensity increases the fatigue strength moderately, from 130 MPa to 150 MPa. During investigations to find an accurate and a computationally efficient method for capturing the complex behavior of Mg alloys, it was found that stringent assumptions are needed to allow for a closed-form analytical solution when calculating residual stresses induced by shot peening. This limits the application of these models to idealized conditions. On the other hand, and because of the complex behaviors of Mg alloys, such as complex unloading behavior and rate-sensitivity, it is difficult to provide numerical solutions such as finite element that are capable of mimicking actual material’s behavior once it is released from an over-strain loading state. Moreover, modeling full coverage shot peening condition is time-consuming and computationally expensive. A single-shot finite element model was combined with an analytical model using actual loading-unloading material behavior to propose a hybrid FEM-analytical model for prediction of the residual stress distribution in shot peening. First, the shot peening process was divided into a loading phase, modeling the impact of a shot and substrate, and an unloading phase, modeling the rebounding of the shot. Finite element was employed to model a single shot impingement on a substrate using the actual loading properties of the substrate. Using the results of the loading phase, an analytical model was proposed to predict stresses due to the unloading phase, using the actual unloading behavior of the material. The proposed hybrid model accounts for the actual behavior of a material, actual elastic-plastic contact analysis, strain rate effect, and friction. The model was then verified by predicting residual stresses induced in a SAE1070 and an Al2024-T351 sheet. Results were compared with the available experimental results and showed close agreements. The application of the proposed hybrid numerical-analytical model was extended to use with an asymmetric and anisotropic material that also has complex unloading behavior, i.e., Mg alloys. First, the loading state of material under peening and the effects of the material’s asymmetry and anisotropy were discussed, then the numerical modeling of the loading step was provided. Finally, the actual unloading curves, measured using the designed fixture, of the material were used to estimate the residual stress profiles. The strain rate effect was also considered in the modeling. The results were matched closely with the XRD and hole-drilling experimental measurements.