New legislation has come into force, WLTP, the Worldwide Harmonized Light Vehicles Test Procedure, that ensures that the purchaser of a car is informed of its fuel consumption, or its range, in the case of an electric vehicle. This is where Petter Ekman’s research comes into the picture. At the Division of Applied Thermodynamics and Fluid Mechanics, he has looked into how car manufacturers can use virtual aerodynamic analyses to simplify and accelerate the certification of fuel consumption values for cars.
Complicated air turbulence“The fuel consumption is influenced by such factors as which wheel rim design the customer has chosen, which tyres and tyre patterns have been fitted, and whether the car is to have a towing hook. There are theoretically 300,000 different variants of a Volvo XC90, and even if only 200 of them are realistic and require certification, this still requires huge resources from the car manufacturer”, he says.
Certification can take place through wind tunnel testing, which takes a long time and is expensive; through road tests, which are less accurate; or through the use of computational fluid dynamics, CFD.
A complication with the computer calculations is how to handle air turbulence, the various-sized eddies that surround the vehicle. To be useful, a model must include turbulence in the calculations; the calculations must be accurate; and they must remain accurate even if changes are made to the car. Further, they must remain accurate when scaled up to industrial scale.
Results from two different calculations methods Photo credit Illustration Petter Ekman
The most accurate method is to solve for the turbulence by considering each eddy individually, but this requires so many calculations in a supercomputer that it is not possible in practice. In his thesis, Petter Ekman has investigated how much the calculations can be simplified without losing accuracy. “We have tried to find shortcuts by using models of the turbulence close to the surface, where the eddies are small. We then try to solve for the larger eddies a bit further away”, he explains.
In one of the articles included in the thesis, he has used a virtual generic car model called DrivAer developed by the motor industry. He shows that the time required for the calculations can be reduced by as much as 90% simply by using slightly longer time intervals when calculating how the eddies move.
Petter Ekman Photo credit Magnus Johansson“I show in the thesis that it is possible to do CFD calculations in which the accuracy is a few percent lower. The results are so reliable that we can even start to question the experimental results. The demands of the legislation are starting to be felt, so it’s important that car manufacturers can do the calculations both accurately and rapidly. CFD will be a great help here since the results from the calculations are reliable”, says Petter Ekman.
Timber trucks and goods vehicles
In other articles included in the thesis, he describes how fluid mechanics carried out by CFD can contribute to designs that significantly reduce fuel consumption for both timber trucks and smaller goods vehicles. These results are described in a link below.
The research has been carried out in collaboration with the Swedish motor industry and has been financed by, among other sources, the Swedish Energy Agency and the graduate school at the Department of Management and Engineering. The calculations have been carried out at the National Supercomputer Centre at LiU.
Petter Ekman defended his thesis on 5 May. After completing his thesis, he would like to continue and deepen the research, but is also attracted to working with calculation methods in industry.
Important factors for accurate scale-resolving simulations of automotive aerodynamics, Petter Ekman, Department of Management and Engineering, Linköping University 2020.
Supervisor: Professor Matts Karlsson
Translated by George Farrants
Turbulence in different time intervals. Photo credit Illustration Petter Ekman