| Surface Mechanical Attrition Treatment (SMAT) has become a promising technique to produce nanostructured surface layers in bulk material. This work targets to numerically investigate and predict the material behavior under the random high strain rate impacts during SMAT.;A number of parameters need to be controlled and regulated during SMAT, such as ball size, impact energy, impact frequency, impact velocity, ball density and incident angle. A computational modeling of SMAT process is thus developed. AISI 316L stainless steel is chosen as the target material. First, the influences of ball parameters on the residual stress are systematically analyzed, and then the numerical solutions of the indent size are proposed and validated by experiments. In addition, the depth of plastic zone which is induced by impact with oblique angle is numerically evaluated. At the same time, a global random impact model and a local impact frequency model are developed to analyze the statistic characteristics of SMAT coverage.;A new realistic SMAT model is thus developed which considers the full coverage, random impact location and random impact oblique angle simultaneously. The energy partition during SMAT process is explored by using the new developed model. The components of impinging and rebounding velocities during SMAT are monitored. The stored energy and the fraction of plastic work converted into heat (β) are numerically evaluated.;Further investigations focus on the mechanism of highly improved strength of SMATed material. In developing a new material model, not only the varying strain, strain rate, temperature, but also the changing of microstructure such as dislocation density, grain size and/or space between the deformation twinning need to be considered. In this study, the micro-level variable, i.e., grain size, is introduced. A new continuum-based constitutive model involving the evolution of grain size is proposed and implemented by user-defined subroutines. The deformation behavior corresponding to the grain refinement during SMAT process is thus predicted. Finally, a numerical comparison between SMAT and SP is conducted by using the proposed constitutive equation and full coverage random impacts model. |
