The effect of roughness upon turbulent supersonic flows
- Started
- 1st October 2010
- Ending
- 1st October 2013
- Research Team
- Christopher Tyson
- Investigators
- Neil Sandham
Mean density contours showing the sudden changes in density caused by the presence of the surface.
Motivation
Improving our understanding over rough surfaces is important for being able to design increasingly more efficient air vehicles and has therefore been an active area for research for many years. One of the key issues has always been the vast range in different rough surface shapes encountered with the materials used in the industry. As such, creating robust prediction models able to account for all the different surface topographies is still a challenge. This is especially true for supersonic and hypersonic air flows, where not only are there many different surface roughness to contend with, but also the effect that increasing Mach number has in changing the relationship between the roughness and the flow. Therefore, the use of numerical simulations to investigate these effects has become widespread with the increase in computational power that has become available.
Method
Due to the turbulent nature of the flows in question, the Navier-Stokes equations are solved using a Direct Numerical Simulation (DNS). In contrast to other codes which use models to account for the behaviour near the surface wall, the DNS code is able to fully capture all the details associated with the flow at all locations. As a result, however, these codes are quite demanding in terms of computational power. Couple this with the need to run the simulations for a long time in order to average out any fluctuations associated with the turbulent features of the flow and it becomes clear that making use of advanced high powered computing clusters is necessary. Simulations will typically use around 128-384 cores simultaneously running in excess of a total time of 200 hours, which means that Iridis3 and HECToR are ideal resources.
Results
Some results collect so far illustrate that it is possible for the surface roughness to cause strong compressibility features in the flow. These features are able to dramatically change the properties of the flow, causing sudden changes in temperature, pressure and density. This in turn alters how the air flows over the surface. However, this is not true for all surface roughness, with some simulations showing the surface generates weak compressibility effects. Therefore, more simulations on more surfaces are currently being considered to further investigate these effects.
Categories
Physical Systems and Engineering simulation: CFD, Turbulence
Algorithms and computational methods: Finite differences
Visualisation and data handling software: ParaView
Programming languages and libraries: Fortran, Matlab, Python