U.S.-Tunisia Cooperative Research: Combined Cellular Automata and Multigrid Methods for Structural Topology Design

Virginia Polytechnic Institute And State University, Blacksburg VA

Investigators

Abstract

0353186 Gurdal Description: This project supports collaborative research between Dr. Zafar Gurdal, Department of Engineering Science and Mechanics, Virginia Institute of Technology and State University, Blacksburg, Virginia and Dr. Hichem Smaoui, Department of Civil Engineering, Ecole National d'Ingenieurs de Tunis (ENIT), Tunis, Tunisia. The research will develop a numerical design optimization methodology based on a combination of the cellular automata (CA) paradigm and multigrid methods (MG) for the design of large complex structures, and to investigate the new methodology in the context of design of large complex structures. Unlike the traditional approach, the CA does not require repetitive system level analyses, as it is based completely on local problem solving for both analysis and design. Multigrid methods are used to improve its efficiency. This innovative approach offers a new paradigm to design and analysis with great potential. This approach will be massively parallel and much more efficient than present design optimization algorithms. The PIs plan to apply this methodology to the topology optimization of 2-D elastic continua with highly refined grids, and to the topology and fiber orientation optimization of composite plates. Scope: Developing automated methods for engineering design is of immense significance, and will remain an important problem. The application of multi-grid approach in the design environment can benefit from the experience to be gained in this project. The results may change design paradigm for industry and thus have a major impact. The approach may be extended beyond structural design to other areas of multidisciplinary design optimization and to system identification, and would make many areas of optimization more efficient and allow them to be more widely applied, especially to nonlinear problems. The project would enhance teaching through the development of an associated course in Tunisia, and learning through the involvement of graduate students, with participation of a female Ph.D. student. A strong partnership is expected to develop between the research groups at the U.S. and Tunisian institutes. It is expected that the collaboration will be successful in completing the proposed work since the two researchers have already done preliminary work in the area. There will be participation of graduate students from Tunisia and VPI who will be working together. Widespread dissemination of results is expected.