PNNL

Abstract

This work proposes a computational modeling framework capable of predicting the behavior of materials when subject to solid phase processing (SPP). A macroscopic dislocation density based constitutive model with internal state variables correlated to microscopic dislocation densities is incorporated into a meshfree smoothed particle hydrodynamics (SPH) framework able to handle large deformations and high temperatures typical of what a material undergoes during SPP. Unlike phenomenological constitutive models that typically neglect material microstructural contributions to macroscale response, the dislocation density based constitutive model improves significantly the predictivity of our modeling framework in a wider range of conditions and allows the multiscale material responses to be simulated simultaneously. With dislocation density based model parameters calibrated to AA1100-O aluminum test data, the generality of the proposed framework is tested on two SPP techniques, namely friction extrusion (FE) and friction stir processing (FSP). Simulation results with different process parameters are systematically validated by experimental data for both cases. The validated models are then used to elucidate the inherent physics associated with SPP by revealing the distributions of temperature, strain, strain rate, and dislocation density, stick-slip contact conditions, and heat generation rates. Model-predicted high dislocation density regions in the processed AA1100-O also agree well with the dynamically recrystallized grain refinement zones observed from experiments.