Windtunnel modelling of vehicle aerodynamics : with emphasis on turbulent wind effects on commercial vehicle drag

摘要

Fuel represents a major proportion of road transport expenditure and it is likely that this proportion will increase. At typical road speeds approximately half of the total fuel used is consumed in overcoming aerodynamic drag, hence the determination and reduction of aerodynamic drag is of considerable importance. This is normally performed by scale testing in wind tunnels with relatively smooth flow. When modelling an atmospheric crosswind in the tunnel the relative air direction is generated by yawing the model at an angle to the oncoming flow. This procedure does not reproduce the inherent turbulence in atmospheric winds. A review of the literature showed a poor correlation between road and wind-tunnel results often attributed to the lack of tunnel turbulence. The work presented herein involves road and wind-tunnel tests to investigate these discrepancies and aims to improve the accuracy of wind-tunnel modelling for commercial vehicles. Wind-tunnel and on-road tests which determine drag coefficient reductions from aerodynamic devices fitted to commercial vehicles are described. Two series of road tests utilised pairs of commercial vehicles: International Harvester Australia low-forward-entry articulated vehicles with maximum road-legal size containers and Isuzu rigid (box-van) vehicles fitted with cuboid containers. Drag coefficient reductions were calculated from fuel meter readings in the trucks and measurements of yaw angle and relative velocity from an instrumented chase car. Tunnel testing was performed on scale vehicles in the Royal Melbourne Institute of Technology (RMIT) Industrial Wind Tunnel in relatively smooth flow (longitudinal intensity = 1.7%). Large differences between road and tunnel drag coefficients at high yaw angles were found. The on-road turbulent wind environment was measured utilising a vehcle instrumented with mast-mounted cross-wire and propeller-vane anemometers. Atmospheric mean wind speeds of 1 m/s to 9 m/s, aligned at various angles to .the road direction, were encountered and data were taken with the vehicle stationary and moving at 27.8 m/s (100 km/h). Longitudinal and lateral intensities and spectra were calculated thus providing new information on the wind environment for vehicles. A mathematical model of the turbulence intensities perceived by a moving vehicle was developed. This utilised atmospheric wind data obtained whilst the vehicle was stationary to predict moving vehicle data. Measured and predicted intensities for the moving vehicle were in good agreement for roads with no local roadside obstructions (eg. trees) thus validating the model, but the obstructions increased data scatter and augmented the lateral intensities by typically 30%) with little change in the longitudinal intensities. Peaks in the longitudinal and lateral spectra for the moving vehicle were at approximately 1.0 Hz and most of the energy was contained between 0.1 Hz and 10 Hz. Subsequent tunnel tests were performed using five levels of grid-generated turbulence and the mathematical model was used to predict the on-road data from tunnel tests. Better agreement was found at high yaw angles when the correct longitudinal intensities were used. However the scales of turbulence in the tunnel were too short for correct modelling. Flow visualisation studies over the model and full-size cab roofs indicated differences in flow patterns that were attributed to Reynolds number differences. The mathematical model and measurements described in this thesis showed that high yaw giigles are always accompanied by relatively high turbulence intensities and it was concluded that the modelling of turbulence characteristics for commercial vehicles is more important than other modelling parameters such as a moving ground. Most major vehicle aerodynamics tunnels have very low turbulence levels (longitudinal rms intensities commonly less than 0.5%) whereas measured on-road values of 2% to 5% are typical (with higher values of lateral intensities). It is therefore recommended that for vehicle aerodynamic generally more attention be paid to correctly modelling the intensities and scales of turbulence in wind tunnels and understanding the effects of typical turbulence characteristics on vehicle drag.