High fidelity simulation of reacting flows using GPUs
The coupling of fluid flow turbulence and chemical reaction extends the range of spatial and temporal scales at which these physical phenomenon occur, thereby greatly increasing the computational requirements for numerical calculations. The project involves the implementation of chemical species transport and reaction calculations into a highly-optimized Direct Numerical Simulations (DNS) fluid dynamics code. Calculations will be performed and validated against a variety of well documented experimental and numerical test cases to verify the accuracy of the implementation. High-fidelity simulations of reacting flows are to be conducted using large parallel CPU and GPU systems, to evaluate and improve the computational performance of the numerical code. The highly-optimized code will allow simulations of greater detail to be performed, resulting in increased detail in the data providing additional insight into the physics, which many be difficult to measure experimentally. The simulation results obtained can also be used to develop and improve current modelling approaches for turbulent combustion. Since modelling approaches are more commonly used in the industry over DNS approaches, the study will not only have an impact in the field of research, but also in the industry.
Leader: Richard Sandberg
Staff: Mohsen Talei
Students: Dominic Ma
Optimisation of resources and infrastructure
combustion and emissions; computational fluid dynamics; fluid dynamics; numerical modelling; turbulence