Probabilistic Material Modeling of Selective Laser Melted A357 Aluminum Alloy Subjected to Laser Shock Peening

Various processing parameters in Selective Laser Melting (SLM) such as scan speed, hatch distance, substrate temperature, etc., have significant impact on the residual stresses present in the print. Compressive residual stresses induced by Laser Shock Peening (LSP) enhances the fatigue life of various metallic components and their alloys. Considering the presence of tensile residual stresses on A357 Aluminum SLM specimens due to the dispersion of eutectic silicon particles, LSP is applied to induce compressive residual stresses. Efficient numerical simulation of LSP is achieved using the Single Explicit Analysis using Time-dependent Damping (SEATD) technique. Conventionally, the material model used in LSP simulation employs deterministic parameters for residual stress prediction. The residual stress distribution predicted by these deterministic parameters are prone to be inaccurate even for similar LSP configurations due to the intrinsic uncertainties associated with the material itself. Hence, a joint random field for the material model parameters for the high strain rate LSP process is developed based on a probabilistic approach known as Bayesian Inference. The working technique of Bayesian Inference process for material model calibration is demonstrated using a set of assumed residual stress data. The calibrated material model parameters are then used in simulating a “candy-bar” coupon subjected to different LSP patterns. Conventionally casted A357 specimens are processed with a similar technique with an aim of quantifying the difference in residual stress distribution as a result of varying manufacturing methodology. The results reveal that the finely calibrated material model parameters using Bayesian Inference predicts the assumed experimental residual stress field with a reasonable accuracy.