Formula-1 racing cars could be made more aerodynamically efficient thanks to a new ‘virtual test bed’ being developed by British research engineers at Cambridge University. They have created one of the world’s most sophisticated systems for analysing the flow of fluids – including air – over surfaces, and have been working with Formula-1 teams to improve car design.
The work has important applications beyond the glamorous world of motor sport such as predicting the effects of explosions on oil rigs for safety assessments, or improving gas turbine design.
The research is being carried out in the Computational Fluid Dynamics (CFD) – Laboratory in the university’s Department of Engineering with funding and high-performance computing support from the Engineering and Physical Sciences Research Council, Hitachi, Silicon Graphics and several other sponsors.
The Cambridge researchers have cracked one of the big problems in the field by developing a suite of software that can take the initial computer-aided design (CAD) image of the component under examination, make it suitable for CFD analysis, and perform the flow calculations in the most efficient way.
Will Kellar of the CFD lab explains. ‘The extreme precision needed for CFD analysis is not usually provided in CAD programs, so the first part of the process is to patch up the deficiencies in the CAD format.’ To do this manually is highly labour-intensive, so the Cambridge team has written software that automates the procedure.
‘Once the CAD has been repaired you end up with a series of continuous surfaces in space which describe your object,’ says Kellar. ‘To be able to make meaningful calculations on the flow of fluid over these surfaces you need to break it down into discrete, manageable chunks.’ This is done by filling all the air spaces around the object with tiny tetrahedral-shaped cells.
A computerised model of a racing car could have several million individual cells over its surface.’You end up with a volume full of cells, each connected to each other and connected to the surface you are interested in,’ says Kellar. ‘This process is termed mesh generation. You are then in a position to be able to perform your flow calculations.’ A set of conditions is presented to the model – the speed of the air entering an inlet, for example, its temperature and pressure and how fast the air exits.
With this information the computer calculates the speed and direction of the flow within each individual cell. The researchers have recently been analysing how the drag of the wheel of a racing car is affected by the flow of air coming from the carÃs front wing. ‘It is possible for example to optimise the wheel drag by intelligent design of the front wing and other components,’ says Kellar.
By developing their own CAD repair tools and integrating these with their own programs for mesh generation and the actual mathematical analysis, the Cambridge researchers have opened up computational fluid dynamics to new areas of application. ‘What we are doing is generic,’ says Dr Mark Savill manager of the laboratory. ‘While Formula-1 is a serious application it is not the only one. This is very much an enabling technology which allows us to handle CAD descriptions direct from industry, do flow calculations on them and show them back as CAD in virtual reality.
Such numerical test beds can already provide data on flow properties for complex engineering flows that are difficult or impossible to measure experimentally, and this can help to minimise expensive testing. In future it will be possible to provide designers with powerful new tools to complement traditional wind tunnel testing.