Parallel computing of overset grids for aerodynamic problems with moving objects.

摘要

When a store is dropped from a military aircraft at high subsonic, transonic, or supersonic speeds, the aerodynamic forces and moments acting on the store can be sufficient to send the store back into contact with the aircraft. Therefore, store separation analysis is used to certify the safety of any proposed drop. Time accurate computational fluid dynamics (CFD) offers the option of calculating store separation trajectories from first principles.; In the Chimera grid scheme, a set of independent, overlapping, structured grids are used to decompose the domain of interest. This allows the use of efficient structured grid flow solvers and associated boundary conditions, and allows for grid motion without stretching or regridding. However, these advantages are gained in exchange for the requirement to establish communication links between the overlapping grids via a process referred to as “grid assembly.”; Relatively little work has been done to use parallel computing for time accurate, moving body problems. Thus, new techniques are presented for the parallel implementation of the assembly of overset, Chimera grids.; This work is based on the grid assembly function defined in the Beggar code, currently under development at Eglin Air Force Base, FL.; The parallel performance of each implementation is analyzed and equations are presented for estimating the parallel speedup. Each successive implementation attacks the weaknesses of the previous implementation in an effort to improve the parallel performance.; The first implementation achieves the solution of moving body problems on multiple processors with minimum code changes. The second implementation improves the parallel performance by hiding the execution time of the grid assembly function behind the execution time of the flow solver. The third implementation uses coarse grain data decomposition to reduce the execution time of the grid assembly function. The final implementation demonstrates the fine grain decomposition of the grid assembly through the fine grain decomposition of the hole cutting process. Shared memory techniques are used in the final implementation and appropriate dynamic load balancing algorithms are presented. (Abstract shortened by UMI.)