The need to improve the aerodynamic performance of a class 8 truck has become a higher priority for manufacturers with the constant increase in fuel costs and implementation of stricter governmental regulations on emissions. To meet these requirements, the ability to predict the aerodynamic performance and hence evaluate the impact of new designs and evaluate add-on devices is of critical importance.
Research conducted by engineers from Volvo Trucks North America and Exa Corp. highlighted the accuracy of simulation with a blind correlation to wind tunnel experiments and provided confidence in the use of simulation for heavy truck aerodynamics.
Experimental work for this study was conducted at the National Research Council (NRC) of Canada wind tunnel located at the Upland Campus (Ottawa, Canada). The 9 x 9-m (30 x 30-ft) wind tunnel is operated by the Aerodynamics Laboratory of the Institute for Aerospace Research (IAR), which is part of the NRC. Forces on models mounted in the tunnel are measured using a six-component pyramidal external balance mounted underneath the tunnel floor. A large turntable is used to allow a range of yaw angles to be considered.
In addition to being able to conduct experiments on full-scale trucks, the wind tunnel is used to evaluate scale models, including tractor-trailer combinations. In this configuration, a moving belt system can be used to mimic a moving ground condition beneath the vehicle. In addition, the tractor wheels can be driven at the same speed as the belt. The boundary layer forming upstream in the wind tunnel is removed using a boundary layer suction system.
For the simulation setup, the wind tunnel working section is modeled including the chamfers between the floor, wall, and ceiling. For the production Volvo VNL tractor-trailer combination all the mount plates, mount cradles, and air-bearings are included.
The half-scale model is a representation of the production truck, which is fully detailed including all underhood and underbody components. The half-scale model uses a simplified representation of the underhood detail, and baffles are used to include the remaining restriction. Further differences include the sidefairing details, wheel details, fuel tank, and exhaust configurations.
The final setup for the test model in the wind tunnel is a fixed floor condition with stationary wheels. For the simulation setup, the moving belt and rotating wheel boundary conditions are replaced with a standard friction floor and fixed wheel conditions.
In addition to the wind tunnel tests conducted in simulation, tests were also performed in an “open road” environment to investigate the influence of the wind tunnel. The setup for these cases removes the wind tunnel walls and introduces a complete moving ground as well as rotating wheel conditions on all five axles.
The aim of this research was to correlate numerical simulation results to NRC wind tunnel experimental results as well as to gain an understanding of differences between the wind tunnel environment and an “open road” condition. The differences that were investigated include the effects of model scale and floor conditions in the wind tunnel as well as the effects of ground conditions and trailer length in the open road configuration.
It was found that when comparing the simulation results for the production VNL with experiment, the differences were less than 2% for a range of yaw angles. The accuracy was confirmed when considering the wind averaged drag coefficient for the production truck in the wind tunnel. Further confirmation that simulation was able to accurately predict the drag performance in the wind tunnel environment was shown with the test model for a variety of belt and wheel conditions. The accuracy again was within 2% of the experiment.
Beyond the direct comparison with the experiments, simulations were used to investigate the effects of tunnel blockage on the production truck, the effects of trailer length and ground conditions in an open road configuration, and the effects of stationary belt, static wheels, and model scale using a test model. It was shown that the blockage effects are significant for a full-scale tractor-trailer combination. It was also shown that when a standard 53-ft (16-m) trailer is considered, the shape of the drag vs. yaw curve changes significantly. All of these effects are important to understand when considering wind-tunnel data and CFD analysis results for different truck designs.
Overall, this study provides confidence that simulation is accurate in terms of absolute drag and can be used as a predictive tool early in a design cycle. It also allows for optimization of forms and shapes, and it provides a streamlined process of developing new concepts based on analysis of the flow field around the previous design iteration. In addition, this study provides valuable information to understand the effects of the wind tunnel and helps to improve the techniques used to test and analyze full and half scale models.