David J. Benson - US grants

An important failure mechanism for ductile metals is the growth and coalescence of microscopic voids. In this work an explicit Eulerian code will be developed to study at the micromechanical level the dynamic failure of solids by void growth and coalescence. The highly sophisticated and accurate interface tracking methods developed for hypervelocity impact and penetration calculations will be adopted for evolving new surfaces. The accuracy of the Eulerian calculations will result in greater insight into this important failure mechanism.

This proposal aims to gain fundamental understanding of the
deformation mechanisms that operate in nanostructured metals and
alloys, in particular in those produced by severe plastic deformation
(SPD) methods. Based on this knowledge it further aims to develop full
capability to manufacture these materials in high quality bulk forms.

The extremely attractive (and rare) combination of mechanical
properties (high strength, ductility, fatigue resistance) and
manufacturability of these materials leads to a new class of high
performance alloys for structural uses. It is understood that this
combination of properties is due to the formation of nano-scale grain
sizes in these materials, but the mechanisms responsible for the high
strength combined with high ductility are not well understood. This
presents a fundamental obstacle to the optimization of these
materials, or to predictions of the performance of these materials in
applications.

An integrated approach with strong emphasis on manufacturing is
proposed. On the theoretical side, deformation mechanisms will be
simulated with crystal-plasticity aggregate models and with detailed
models of the grains and grain boundaries. The experimental program
covers a wide a range of strain rates and temperatures, texture
development, and in situ transmission electron microscopy and atomic
force microscopy to directly verify deformation mechanisms. The
experimental results will provide validation to the theoretical
modeling and manufacturing process simulations.

Finally, simulations of the manufacturing processes will enable
process parameter optimization. A complete, miniature, yet fully
scalable, manufacturing facility will be designed and implemented.

The significant impacts of the proposed research are made possible by
the acquired fundamental understanding of the deformation mechanisms,
and include advances in manufacturing techniques to produce these
highly desirable materials in bulk. The miniature manufacturing
facility will become a source of significant quantities of
nano-structured alloys. Finally, the proposal will provide students at
the UCSD and at local K-12 schools with interdisciplinary education in
a cutting-edge area of research.

This collaborative research is a partnership between the University of California, San Diego, the University of Texas, Austin and Alcoa (Grant Opportunities for Academic Liaison with Industry "GOALI" partner). This research will develop finite elements for virtual sheet metal stamping using NURBS (Non-Uniform Rational B-Splines), the same functions used for representing the geometry in CAD-CAM packages. Current finite element methods approximate the geometry of the sheet metal and the dies with linear interpolation and therefore don't represent curves well. NURBS, however, are used in CAD-CAM because they can represent complicated curves exactly. The easiest advantage to appreciate is a NURBS finite element formulation has the generality in its kinematics to be completely compatible with the CAD-designed die unlike the polynomials used in current finite elements. NURBS are also very well behaved numerically, allowing length to thickness ratios of 1000:1 to be modeled with solid elements, eliminating the zero normal stress assumption of traditional shell formulations. Being able to model the sheet with solid elements allows more general kinematics through the thickness than the linear displacement field in standard shell element formulations, enhancing the accuracy. Introducing NURBS in finite element analysis will therefore increase its accuracy.

This research is directly applicable to all engineering applications requiring the accurate analysis of linear (e.g., vibration) and nonlinear structural response. It initiates collaborative research between two universities and industry. It provides graduate students the rare opportunity to work with industry as part of their research, and will provide under represented groups of undergraduate students an opportunity to participate in research via the McNair Program at UCSD. Additionally, they will gain firsthand experience with industry through the collaboration with Alcoa.

The research objective of this collaborative research award is to develop methods that allow engineers to perform stress analysis using their computer aided design (CAD) models without the time consuming mesh generation required by traditional finite element analysis. The research approach combines isogeometric analysis with boundary element technology. Isogeometric analysis uses the same basis functions that are used in the CAD models for engineering analysis. Examples of these basis functions include non-uniform rational B-splines (NURBS) and T-Splines. Traditional numerical methods, such as finite element analysis, use piecewise continuous polynomials of low degree. A mesh generation step is therefore required to convert the CAD representation of the geometry to one that the numerical analysis can use. Using CAD basis functions directly allows us to eliminate this step. The current limitation with CAD basis functions is they only represent surfaces. To sidestep this limitation, we will use boundary element technology, which works directly with surface representations.

If successful, the results of this research will allow engineers to rapidly perform structural analyses with their CAD models during the preliminary design process. The high cost of doing detailed analysis during the design phase limits the scope of the design space that engineers are able to explore, and makes truly innovative design expensive. By reducing the time and cost of analysis, design engineers will be able to produce designs that are more optimal in terms of weight, performance, and cost. Automatic design optimization will also be facilitated since the numerical analysis will work directly with the design space parameterized by the CAD models, allowing optimization programs to work with less human intervention and correction.