Highly scalable parallel computational models for large-scale RTM process modeling simulations, Part 1: Theoretical formulations and generic design

This article reports the recent development of a highly scalable parallel c omputational model for process modeling and manufacturing applications of l arge-scale composite structures with particular emphasis on resin transfer molding (RTM. Fundamental concepts and characteristic features of the propo sed scalable parallel algorithms are described and developed in technical d etail. The approaches for simulating process modeling and manufacturing app lications of composites includes: (I) the traditional, explicit control-vol ume finite-element (CV-FE) approach, and (2) a recently developed and new, implicit pure finite-element pure (pure FE) approach. SGI Power Challenge a nd SGI Origin 2000, which are symmetric multiprocessor (SMP) computing plat forms, are employed in this study. The issues of implementation and softwar e development of these manufacturing process simulations are parallel algor ithm development, data structures, and interprocessor communication strateg ies, with emphasis on performance and scalability on these symmetric multip rocessors. with the motivation for providing effective computational proced ures suitable for practical process modeling and manufacturing applications of large-scale composite structures and general finite-element simulations , the proposed developments not only provide a sound theoretical basis but also serve to be ideally portable to a wide range of parallel architectures . Whereas the theoretical formulations and generic design are described in Part I, the parallel formulation of the theory and implementation will be p resented in Part 2 and the techniques developed are applied to large-scale problems using Power Challenge and SGI Origin to demonstrate the effectiven ess and the practical applicability, which will be presented in Part 3 of t his work.