Development of High Order LES solver for GPUs architectures

Turbomachinery plays a vital role in almost all industrial sectors. They are used in changing the state of working fluids or transporting them in pumps or compressors, extracting energy (gas turbines) and creating propulsion (aircraft jet engines). Developers have long used computers to accurately simulate internal or external aerodynamics and hydrodynamics, thermal and structural state, combustion, and other mission-critical behaviours thereby increasing performance and reduce the cost of production. Computer simulations especially computational fluid dynamics (CFD) can provide valuable insight and can quickly identify potential problems thereby enabling the designer to investigate a multitude of operating conditions and parameters for evaluating alternative solutions. This prevents large numbers of costly and time-consuming late-stage design changes to verify the design with the minimum number of physical tests. The aim of the project is to combine the capabilities of high performance computing architectures such as graphics processing units (GPUs) and advanced numerical methods (compact high-order numerical schemes) to develop and implement a large eddy simulation (LES) method in Open FOAM C++ based solver. The proposed method will be developed for GPUs to enhance the speed of calculation without sacrificing accuracy thereby improving performance of jet engines and reducing the time to market.