Return to the Table of ContentsObjective: The goal of this project is to investigate the algorithmic and implementation issues as well as the system software requirements pertaining to the direct-coupled, multi-disciplinary computational aero-science simulations on distributed memory (DM), multiple instruction stream, multiple data stream (MIMD) parallel architectures.
Approach: The design of future generations of civil transport aircraft that are competitive in the global marketplace requires multi-disciplinary analysis and design optimization capabilities involving the direct coupling of diverse physical disciplines that influence the operational characteristics of the aircraft. An immediate outcome of a such an approach would be the greatly increased computational requirements for the simulation, in comparison to what is needed for current single discipline simulations on conventional supercomputers. In the near future, it appears that the computational resources of the scale required for such multi-disciplinary analysis and/or design optimization tasks may only be fulfilled in a cost-effective manner by the use of highly parallel computer architectures. In order to effectively harness the tremendous computational power promised by such architectures, it is imperative to investigate the algorithmic and software issues involved in the development and implementation of concurrent, directly-coupled, multi-disciplinary simulations. This study takes a necessary preliminary step towards the development of this enormously complex capability by attempting to compute the unsteady aeroelastic response and flutter boundary of a wing in the transonic flow regime through the direct coupling of two disciplines, viz. fluid mechanics and structural dynamics on a DM-MIMD computer.
Accomplishment: A direct-coupled, fluid-structure interaction code capable of simulating the highly nonlinear aeroelastic response of a wing in the transonic flow regime was implemented on the 128 processor Intel iPSC/860 computer. The performance and the scalability of the implementation realized on the iPSC/860 was demonstrated by computing the transient aeroelastic response of a simple High Speed Civil Transport type strake-wing configuration. Also as a part of this study, the efficacy of various concurrent time integration schemes that are based on the partitioned analysis approach were investigated. The effort also helped in gaining a greater understanding of the system software requirements associated with such multi-disciplinary simulations on DM-MIMD computers. The algorithmic and implementation details as well as the results can be found in the following papers: AIAA-94-0095 and AIAA-94-1550.
Significance: This implementation for the first time exploits the functional parallelism in addition to the data parallelism present in multi-disciplinary computations on MIMD computers. It demonstrates the feasibility of carrying out complex, multi-disciplinary, computational aeroscience simulations efficiently on current generation of DM-MIMD computers.
Status/Plans: The future efforts will further explore the possibility of developing more robust and scalable concurrent algorithms for fluid-structure interaction problems, the incorporation of additional disciplines and the feasibility of using emerging parallel programming language standards for developing direct-coupled, multi-disciplinary CAS applications.
Point of Contact:
Return to the Table of Contents