Navigation

Multiscale Design for Desired Material Properties

This work addresses a materials-by-design problem of controlling a cold-working process so that desired property distributions are obtained in the final product.

We use a multiscale finite element simulator which predicts microstructure development under known processing conditions. The design problem allows prediction of the right process conditions for generating a desired microstructure. This is by far (at least on the basis of computational complexity alone, with over a billion degrees of freedom) a challenging problem. As an example, we demonstrate a tailored 'uniform elastic modulus' extruded product by optimizing extrusion die shape. The polycrystalline texture at all points in the exit cross section is controlled by modifying die shapes. The efficiency of the methodology is exemplified by achievement of optimal Youngs Modulus distribution in just four iterations.

im


  1. V. Sundararaghavan and N. Zabaras, "A multi-length scale continuum sensitivity analysis for the control of texture-dependent properties in deformation processing", International Journal of Plasticity, in preparation. [PDF]
  2. V. Sundararaghavan and N. Zabaras, "On the control of microstructural degrees of freedom in deformation processes", presented at the `Processing and Mechanical Response of Engineering Materials' symposium in the 2006 TMS Annual Meeting & Exhibition (T. R. Bieler et al., organizers), San Antonio, TX, March 12-16, 2006[PPT]

Process-Property-Structure maps

reduced

Property cross-plots, a standard approach for materials selection, as generalized by Ashby, have successfully demonstrated graphical quantification of property-performance relations for material selection. Is it possible to build
similar graphical representations for the selection of optimal microstructures (and processes) in polycrystalline materials? Although the space of microstructures is extremely high dimensional, our studies seem to indicate that visualization of structure-property maps is indeed possible.

Microstructure representation techniques that we have recently developed allow generation of detailed property-structure maps for materials allowing metallurgists to choose the right microstructure for the application and right processes for generating such microstructures. These representation schemes decompose the microstructure feature space into smaller dimensional spaces using proper orthogonal decomposition. In this space of three coefficients, all possible microstructure features (obtained from given process conditions) can be visualized as a plane, which we call the 'process plane' (see figure above) . Properties can be visualized as surfaces intersecting the process plane. 


1. V. Sundararaghavan and N. Zabaras, "Linear analysis of texture-property relationships using process-based representations of Rodrigues space" Acta Materialia, accepted, 2006.[PDF]
2. V. Sundararaghavan and N. Zabaras, "On the design and control of properties in polycrystalline materials using process-texture-property maps", presentation at the `Materials Processing and Manufacturing Division Symposium: Mechanics and Materials Modeling and Materials Design Methodologies, in the Honor of Dr. Craig Hartley's 40 years of Contributions to the Field of Mechanics and Materials Science Materials Science' symposium in the 2007 TMS Annual Meeting & Exhibition (Brent Adams and Hamid Garmestani, organizers), Orlando, FL, February 25 - March 1st, 2007[PPT]