Multi-Scale Microstructure-Informed Modeling of Residual Stresses and Distortions in Additive Friction Stir Deposition

The objective of this study is to develop a multi-scale framework for thermo-mechanical analysis of Ti-6Al-4V structures manufactured using the Additive Friction Stir Deposition (AFSD) technique. During the AFSD process, the structure undergoes repeated heating and cooling cycles as hot layers are deposited continuously over previously cooled material, leading to a large thermal gradient that can result in residual stresses and structural distortions. The proposed framework aims to predict such residual stresses and distortions efficiently and accurately. To capture the microstructural evolution during the process, Düsseldorf Advanced Material Simulation Kit (DAMASK), an open source Crystal Plasticity model, is used to determine the slip parameters needed to simulate the plastic deformation based on the experimental stress-strain data. Since crystal plasticity simulations are computationally expensive, a surrogate Forward Neural Network (FNN) is trained using DAMASK to predict the elasto-plastic stress-strain response of Ti-6Al-4V microstructure as a function of strain rate and temperature. The trained surrogate model is shown to significantly reduce the computational time for a low fidelity analysis to just 1 minute as compared to 71 hours when using the full physics-based DAMASK crystal plasticity simulations. The ANN architecture is then integrated with the macro scale model in Simulia Abaqus. Process simulation for AFSD is done through a sequential thermo-mechanical analysis. A transient heat transfer analysis is conducted first and the obtained temperature profiles, varying in both time and space, are used as predefined temperature field boundary conditions for the following thermo-mechanical analysis on the same mesh. The material properties such as Young’s Modulus, coefficient of thermal expansion, etc. are defined as a function of temperature. Framework capabilities are demonstrated using a ring stiffened panel example with Stainless Steel 304L substrate and Ti-6Al-4V stiffener.