These simulations illustrate the connection between how a material deforms when stressed and the microstructure of its constituent particles. The dark lines in the simulation on the right show channels of damage formed within a single material crystal at the 2-micron scale. These channels can tell us how and where deformations are likely to occur within a material consisting of many crystals, shown with the dark shaded lines in the simulation on the left.
This modeling effort aims to establish a predictive connection between the manufacturing process and the final performance of parts.
The key is bridging the gaps between simulations at various length and time scales - from the microstructure of a single material crystal to the properties of a chunk of material - to provide information that's useful for engineers.
This model incorporates a time scale of seconds and length scales of hundreds of micrometers.
Connecting additive manufacturing processing with measured materials properties and extrapolating to performance is an essential element of the internally funded strategic initiative to accelerate the certificaiton of additively manufactured metals. Lawrence Livermore already has a robust multiscale modeling effort, spanning electronic structure calculations to macroscale performance calculations. (Marian et al.: 2004; Arsenlis et al.: 2006; Barton et al.: 2008; Barton et al.: 2011a; Barton et al.: 2011b)
This work leverages previous Laboratory Directed Research and Development investments, ongoing work in the Advanced Scientific Computing program, and jointly funded collaborations with Los Alamos National Laboratory. (Lebensohn et al.: 2008)
The emphasis is on directly connecting part performance to microstructural features induced by processing. We are pursuing two main thrusts:
Materials such as titanium alloys have complex deformation behavior, (Addessio et al.: 2003; Field et al.: 2005; Barton et al.: 2009) with strong anisotropy at the crystal scale that can induce both macroscale anisotropy and crystal scale initiation of damage. Other processing-induced structural changes such as ordering of alloying elements in the crystal lattice and distribution of impurities between grain interiors and boundaries also influence overall part performance.
Lawrence Livermore is delivering broadly useful scientific improvements that are needed both in crystal mechanics based modeling capabilities and in microstructurally sensitive homogenization methods. Given crystal mechanics capabilities in ALE3D and its parallel scalability, the Laboratory is uniquely positioned to assess the impact of complex microstructures on performance (Marinak et al.: 2005), with facilities for including more coupled physical effects than in more narrowly focused approaches (Lebensohn et al.: 2008).
We are elucidating damage initiation mechanisms at low macroscopic triaxialities and their connection to microstructure, building on an existing collaboration with researchers at the University of Illinois (Kweon et al.: 2010) and on experience in modeling Ti6Al4V microstructures (Addessio et al.: 2003; Field et al.: 2005; Barton et al.: 2009) Homogenization strategies for performance calculations are built from existing capabilities for including microstructural information (Barton et al.: 2011b) The process–properties–performance component of the project will actively engage the experimental and other modeling components of the project to enhance relevance and impact.
Group Leader, Materials Modeling & Simulation - Computational Engineering Division