Test The Limits

Why aerodynamic optimization needn’t be a drag

By Exa

August 18 2015

Regulatory requirements for emissions reduction are increasing the importance of vehicle aerodynamics. A further reduction in drag is key: it’s thought that most OEMs are targeting a drag coefficient (Cd) of 0.2, down from a best of around 0.24 currently, as achieved by Tesla with its all-electric Model S.


But that goal could also bring further conflicts with the demands for effective thermal management and attractive styling. It’s highly unlikely that such an ambitious shift will be attainable using traditional, trial-and-error development using wind tunnel testing, or even with the type of advanced computational fluid dynamics (CFD) digital simulation methods used up to now.


Existing CFD tools already enable many more design iterations to be explored than is possible with a physical test in the tunnel, saving valuable time and resources. However, the tools still rely on the experience and intuition of the engineer, who must know, working from a baseline design, which parameters to change in order to create the permutations to be tested. If the wrong avenues are explored, resulting in no aerodynamic improvement, then the process must begin again.


To achieve the competing development goals of the future, it’s thought that automakers will have to find ways to evaluate 10 times the number of designs as before, in half the time. Achieving this with the widely adopted CFD approach of recent years will clearly be a challenge, but one that could potentially be overcome by software that systematically explores all combinations of design change possibilities, to efficiently understand their potential and interactions, and then optimize a chosen design to arrive at the best possible solution.


Exa Corporation believes it could have an answer. Our PowerFLOW Optimization Solution simulates complex fluid-flow problems and then uses statistical analysis methods to systematically explore the potential of design alternatives. Unlike the existing, experience-based method, this new approach can be used even by new engineers and can also handle many more changes in parallel.


So how does it work? Once the stylists have provided areas that the aerodynamicists can work on, the engineers take a baseline geometry and parameterize a ‘design space’ of shape changes, in order to search for better design solutions within the constraints provided. The PowerFLOW Optimization Solution works within the cutting edge optimization tools modeFRONTIER and Isight to explore large numbers of design alternatives and generate a mathematical ‘response surface’ from a small number of simulations. Advanced algorithms then identify trends and evaluate trade-offs, providing a multidimensional analysis often reaching beyond the capability of human reasoning.


Crucially, high-performance visual flow analysis tools are available to help the engineers understand and explain the trends in the statistical analysis. The data can be visualized with full geometric detail and high-quality morphing, such that the aerodynamic and stylistic impacts of any design changes can be clearly seen on-screen.


The PowerFLOW Optimization Solution has already been demonstrated in more than 150 design optimization projects for aerodynamics, thermal packaging, aero-acoustics and brake cooling. Performance improvements of up to 13% have been achieved – far beyond what would have been possible by guessing the optimal design parameter combination. In aerodynamics alone, over 60 production projects have been carried out, yielding improvements of 1% to 6% in fuel economy.


So there we have it: a solution that can find bigger improvements from a more efficient approach to testing, earlier in the development cycle. It turns out that aerodynamic optimization needn’t be a drag, after all…


To download a new whitepaper on optimization by Ora Research's Bruce Jenkins, please click here