Aerodynamic effects of frontal elongation applied to a bluff body

Abstract

The EU has committed itself to reduce the greenhouse gas emissions with 20% by 2020 with respect to the 1990 level. This means that all industry sectors including the transport sector have to become more energy efficient. The long and short haul trucks are, with 4.95% of the total CO2 equivalent emissions, a major contributor in the transport sector. In order to achieve a reduction in emissions, the European Commission is proposing new regulations that allow for front elongation of trucks. This elongation might improve the aerodynamic efficiency and hence reduce the emissions and fuel costs. Reducing aerodynamic drag for tractor-trailer combinations has been researched for a long period of time. In 1965 Hoerner published the results of his experimental research on 3D bluff bodies and identified that the drag coefficient at the front was related to the total drag coefficient. In 1985 Cooper determined that an optimal front edge radius based on the Reynolds number and the frontal area of the vehicle exists. These two studies concerning bluff bodies formthe basis of numerous follow up studies and this thesis. This thesis analyses how the geometry of an elongated bluff body can be optimized in order to reduce the aerodynamic drag coefficient Typically the flow around a bluff body is characterized instable and separated flow. At the front of the truck the incoming flow velocity reduces to zero at the stagnation point. A high pressure region is formed at this location. Further downstream the flow reaches the front edge corners of the truck, where the flow is accelerated. Flow separation might occur at these front edge corners. At the rear a large low pressure region is formed which contributes to the total drag coefficient. In this thesis numerical simulations are performed in order to identify the relative importance of several front-end design parameters. The numerical calculations are performed by means of the steady-state RANS k ¡! turbulence model. The domain is refined near the surface model and its wake using two density boxes around the model with different cell dimensions and an inflation layer around the model. A wall model is used to approximate the flow behavior in the boundary layer. The wall model and the local refinements reduce the computational time while retaining a high level of accuracy of the numerical solution. The numerical simulation is validated using the data from an already existing wind tunnel experiment. This wind tunnel experiment has been performed on a 1:15 scale model in the Low Turbulence wind tunnel at the TU Delft. First the influence of the mesh cell size is determined by varying the cell sizes of the two density boxes surrounding the model. One density box is close to the model and has small cell dimensions, while the second density box is encapsulating a larger volume. This resulted in chosen cell dimensions with a ratio of 1:4. The discretization error is 4.1%, while the number of cells used is 4.47¤106. The second part is the validation itself. The drag coefficient of the model is 0.297 for the wind tunnel experiment and 0.300 for the simulation,while the base pressure is slightly lower for the numerical simulation. The pressure distribution at the base is not correctly captured in the simulation results. The final part of the validation is the comparison of the boundary layer development. The results of the simulation obtained in this thesis show great similarity with the results found in the wind tunnel experiment. To explore the real world effects the 1:15 scale results...