THERMAL ANALYSIS OF COMBUSTION CHAMBER WITH NEW GRAPHIC CARD

Simulation of heat transfer in combustion chamber like lots of computational fluid dynamics problems has long run time. To overcome this drawback, parallel processing is one solution. In 2006, newgeneration of graphic card introduced by NVIDIA company. NVIDIA corporation has manufactured GPU devices that can also be used for the processing of non-graphical data. This game card can be implemented for parallel processing. In order to employ a GPU for general-purpose computations, NVIDIA has introduced acomputing architecture called CUDA, which is an extension to C language. In the present work, simulation ofheat transfer in combustion chamber accelerated by these GPUs. For solving equations system, GeneralMinimum Residual Method (GMRES) as an iterative linear solver was implemented. It should be noted, this case study had a lot of application like rocket engine. Results show that GPU can simulate 6.7 times faster than CPU. Also our results show excellent agreement with those of available reports in the literature, while demonstrating exciting performance of the GPU simulations.

Keywords:

Heat transfer , combustion chamber , GMRES , CUDA , parallel processing , GPU