Underwood, O.; Madison, Jonathan D.; Martens, R.M.; Thompson, G.B.; Welsh, S.; Evans, J.
This study offers experimental observation of the effect of low strain conditions (ε < 10%) on abnormal grain growth (AGG) in Nickel-200. At such conditions, stored mechanical energy is low within the microstructure enabling one to observe the impact of increasing mechanical deformation on the early onset of AGG compared to a control, or nondeformed, equivalent sample. The onset of AGG was observed to occur at specific pairings of compressive strain and annealing temperature and an empirical relation describing the influence of thermal exposure and strain content was developed. The evolution of low-Σ coincident site lattice (CSL) boundaries and overall grain size distributions are quantified using electron backscatter diffraction preceding, at onset and during ensuing AGG, whereby possible mechanisms for AGG in the low strain regime are offered and discussed.
Using the kinetic Monte Carlo simulator, Stochastic Parallel PARticle Kinetic Simulator, from Sandia National Laboratories, a user routine has been developed to simulate mesoscale predictions of a grain structure near a moving heat source. Here, we demonstrate the use of this user routine to produce voxelized, synthetic, three-dimensional microstructures for electron-beam welding by comparing them with experimentally produced microstructures. When simulation input parameters are matched to experimental process parameters, qualitative and quantitative agreement for both grain size and grain morphology are achieved. The method is capable of simulating both single- and multipass welds. The simulations provide an opportunity for not only accelerated design but also the integration of simulation and experiments in design such that simulations can receive parameter bounds from experiments and, in turn, provide predictions of a resultant microstructure.
The goal of this project is to generate 3D microstructural data by destructive and non-destructive means and provide accompanying characterization and quantitative analysis of such data. This work is a continuing part of a larger effort to relate material performance variability to microstructural variability. That larger effort is called “Predicting Performance Margins” or PPM. In conjunction with that overarching initiative, the RoboMET.3D™ is a specific asset of Center 1800 and is an automated serialsectioning system for destructive analysis of microstructure, which is called upon to provide direct customer support to 1800 and non-1800 customers. To that end, data collection, 3d reconstruction and analysis of typical and atypical microstructures have been pursued for the purposes of qualitative and quantitative characterization with a goal toward linking microstructural defects and/or microstructural features with mechanical response. Material systems examined in FY15 include precipitation hardened 17-4 steel, laser-welds of 304L stainless steel, thermal spray coatings of 304L and geological samples of sandstone.