Research On Aerodynamic Characteristics Of Flow Field Around Pickup Trucks

Wind tunnel test and numerical simulation were taken on a 1/5th pickup truck model in this dissertation. Research was done of the flow field around the model comprehensively. The model geometry has been purposefully made generic by discarding custom details such as side view mirrors and underbody components.Flow structure in the wake was the focus in the dissertation. In order to well understand the wake of pickup truck, the research compared pickup truck and SUV from the full-closing in the bed of pickup truck.Wind tunnel test included qualitative test such as ribbon flow test and smoke flow test and quantitative test such as PIV of pickup truck model and SUV model. Surface ribbon flow test of pickup truck show that there was reverse upstream flow near the cab base. Mesh ribbon flow test show that there was a pair of counter-rotating vortices in the near wake. The vortices became weak while the distance increased. Mesh ribbon flow test of SUV also show that there were two trace vortices in the wake. The velocity measurements were obtained at two planes of pickup truck: the symmetry plane and the vertical plane at y=80mm. However ,to SUV,two planes were: The symmetry plane and the horizontal plane at z=120mm. The average velocity was determined from 400 instantaneous velocities of the PIV test results.As external flow of road vehicle was three dimensions, viscous, turbulent, separate and unsteady, LES was adopted for numerical simulation because LES has better precision than RANS and fitted for unsteady flow simulation. QUICK scheme was chosen for spatial discretization scheme of governing equation because it has third order truncation error at convection term and has better precision than second order upwind scheme. Crank-Nicolson scheme was chosen for temporal discretization scheme because it has second order precision. The flow around body surface was typical separate flow. Because Wall Functions was unfit for resolving separate problem, the thesis adopted another method—Near-Wall Model. Model mesh was Tetre+Hexa+Prism mesh, which could make use of the strongpoint of the three kinds of mesh. The mesh was refined using local refinement algorithms in a parallelpiped region extending half vehicle-length upstream of the truck, one vehicle-length downstream, one vehicle-highness on the top and half vehicle-width on either side, in order to simulate the flow more accurately in the vicinity of the vehicle. In order to decrease the mesh magnitude of Hexa, mesh became few while away the vehicle because the fluctuation of the flow field was lessening. The road boundary set no slip wall to simulate the fixed floor of wind tunnel test. Simulation was carried at paralleled workstation of Institute of Automobile Aerodynamics of Jilin University. The total data was 280G for either model.Instantaneous flow fields, mean velocity vector fields and streamlines of the mean velocity were obtained from pickup truck and SUV model simulation data postprocessing. Mean velocity was determined from 300 instantaneous velocities.In order to validate the correctness of the simulation result, profiles of the mean velocity in the special planes were obtained, and compared the results with the PIV results from wind tunnel test. Mean velocity vector fields and streamlines in the symmetry plane of pickup truck at wake middle top region were compared with those result of PIV test, showed that there was very strong downwash behind the tailgate. The simulation results were consistent with the results of test, showed that the simulation results were credible. Pickup truck mean velocity profiles of the flow in the symmetry plane and vertical plane of y=80mm were compared with those of PIV test at the location of x=700mm,x=980mm,x=1100mm. In all cases velocity profiles predicted by CFD were seen to match the corresponding experimental profiles both qualitatively and quantitatively. It indicated that the CFD results were trusty. SUV mean velocity vector fields of the horizontal plane at z=120mm were compared with those result of PIV test, showed that the shear layer originated from the side of cab formed vortex in the near wake, the results were similar. SUV mean velocity profiles of the flow in the symmetry plane were compared with those of PIV test at the location of x=1100mm,x=1300mm. In all cases velocity profiles predicted by CFD were seen to match the corresponding experimental profiles both qualitatively and quantitatively. It indicated that the CFD results were credible. Compared the results of CFD with the results of ribbon flow test and smoke flow test, indicated that CFD results were trusty.Pressure distributions on the symmetry plane were obtained both pickup and SUV. The mean pressure on the inside surface of the tailgate was lower than on the outside surface suggesting that the force acting on the tailgate was in the forward direction, reducing aerodynamic drag. Numerical simulations were taken on the three conditions: tailgate off, tailgate down and 1/2 tonneau.The research indicted that: For pickup truck, there was a recirculation flow region over the bed. The cab shear layer did not interact directly with the tailgate, flowed above the top of the tailgate. There was a downwash flow in the symmetry plane behind the tailgate, was no reverse flow region in the symmetry plane, and the formation of two smaller recirculation flow regions on both sides of the symmetry plane. Mean flow fields in the near wake of the cab showed a weak pair of counter-rotating vortices behind the cab. In the cross-flow planes behind the tailgate, the mean flow fields show strong counter-rotating vortices behind the tailgate. These vortices dragged the air to the symmetry plane thus inducing strong downwash at the symmetry plane. The downwash generated between them promoted attached flow on the central portion behind the tailgate and brought higher pressure outside the tailgate. With the increase of x, the vortices in the wake became weak and disappear at last. Instantaneous flow fields in the cross-flow planes of the pickup truck near wake showed compact vortex structures located randomly in space. This suggested that the concept of a persistent pair of counter-rotating vortices in the near wake of a road vehicle was incorrect in agreement with Bearman observations in a car model. The mean pressure distribution on the tailgate showed a lower pressure coefficient on the inside surface compared to the outside surface suggesting that the tailgate reduces aerodynamic drag. Aerodynamic drag increased when tailgate was down or tailgate was off, however, aerodynamic drag decreased when the bed was 1/2 tonneau. These results were consistent with the results done by Kevin R. Cooper at National Research Council of Canada (NRC) wind tunnel, which confirmed further that CFD results were credible. Drag changes were due to a change in pressure at the vehicle near wake. Drag rise was corresponding with the rise of turbulent kinetic energy, which indicated that more energy was dispersed at the wake. For SUV, in the symmetry plane there were two shear layers, one originating form top of the model and another one originating from the underbody flow, and formed a big circulatory flow pattern at the near wake. The shear layers originating from body sides formed two counter-rotating recirculation flow regions in the wake flow. The length of the recirculation regions was in agreement with measurements in the symmetry plane.POD was a good method for the identifying of the coherent structure in the turbulent flow. POD was used for the analyzing of wake flow of pickup and SUV models. The result indicate that compared to SUV wake flow, pickup wake flow was more complex due to the interaction of the separated shear layer from the top and sides of the cab with the bed wall and tailgate. In order to show the near wake of pickup truck more modes were needed. The first two modes captured the most fluctuation energy. The first twenty modes dominated fluctuation velocity magnitude, they showed the different flow characteristic of coherent structure according to energetic distribution. The last eighty modes have little influence on the large scale structure, but they were corresponding with the small scale structure of the turbulence fluctuation which determined the dissipation of the whole flow field. POD modes not only ensured that the structures extracted by POD were prominent on"energy"in mathematic, but also depicted the typical structure in physics. Only needed a few modes could gain well results when reconstructed the flow field which showed that the reconstructed flow field captured the typical large scale structure of the original flow field.