Flow analysis between two bluff bodies in a close distance platooning configuration

Abstract

In the European Union greenhouse gas emissions of heavy duty vehicles make up 30% of the total caused by road transport. By placing vehicles in a platoon configuration the aerodynamic drag can be heavily reduced. The effect of platooning has been studied on both American and European type vehicles. However many of these studies only disclose the drag reductions and do not fully explain the flow behaviour between the models which in the end is what is causing the reduction in drag. Next to this studies done on European type vehicles mostly used time-averaged simulations so a next step is understanding the effect of unsteady flow on the vehicles. In this study the flow field between two vehicles in a platooning configuration is analysed numerically and experimentally for varying intervehicle distance and front- and rear drag reduction devices. The GETS model was used in this study which is a simplified model of a European heavy duty vehicle. For the numerical analysis Exa PowerFLOW is used which is a transient CFD package based on the LBM. Turbulence is modelled using VLES together with a RNG of the κ - ϵ equations. A study of the mesh size showed the sensitivity on the drag coefficient of the single model.
The experimental analysis was performed in the OJF of the TU Delft at a Reynolds number of 3.9 x 105 based on the square root of the model. Next to balance measurements CVV measurements in the gap between the two models were taken in order to visualize the flow field.
The single model showed a toroidal recirculation region in which a vortex ring can be seen. The drag force fluctuates at a Strouhal number of 0.073, side and lift force at 0.058 and 0.102 and 0.130 and 0.160. These fluctuations are caused by pressure fluctuations in the wake where the highest magnitudes are seen in the shear layer. The models with a smaller front radius saw a rise in drag with flow separation occurring for the smallest radius for the numerical results and for both smaller radii in the experiment. The addition of a boat tail lowered the drag of the models. With increasing tail angles the drag reduction also increases which is caused by the increase in base pressure. The tails also decrease the force fluctuations due to decreased pressure fluctuations in the shear layer.
At the closest spacing of 0.10 times the vehicle length all tested configurations benefit from the platoon. The flow between the two models is made up of a toroidal recirculation region much smaller in size compared to the single model. In general the flow takes an S-shaped path between the two models. From the underbody of the leading model it either stagnates on the front of the trailing model or moves into the gap where it ends up in the upper of lower vortex or stagnates on the front or rear of one of the models. When a sharper front edge radius is applied to the trailing model more flow is deflected into the gap leading to lower pressure vortices and a higher stagnation pressure. A small vertical misalignment of the trailing model, which was seen during the experimental campaign, leads to a higher stagnation pressure on the bottom of the trailing model, also here more flow is deflected into the gap. At this spacing the side-force fluctuations are a bit higher due to the vortices that leave the domain passing over the rounded edges of the model.
At the middle spacing of 0.45 times the vehicle length not all configurations benefit from the platoon. The trailing models with the baseline frontal radius saw an increase in drag due to the lack of thrust generated by the rounded edges. The other models did see a drag decrease. At this distance the wake of the leading model is quite similar to that of the single model. Flow enters from the top, bottom and sides of the model and ends up in the recirculation region or it stagnates on the rear of the model before it leaves the wake and stagnates on the front of the trailing model or flows over the rounded edges where it accelerates and leaves the gap. The effect of a sharper radius applied to the trailing model has little effect on the flow field. When a tail is applied to the leading model the recirculation region has been reduced as was seen for the single models. Due to the upwash the stagnation pressure on the bottom front has been increased but the drag has decreased compared to the platoon without a tail due to the increased suction of the rounded edges. At this distance the force fluctuations are much higher compared to the single model. The magnitude of drag and side force can go up to 2 and 6.5 times the values of the single model depending on the configuration. This is due to the stronger vortices from the leading model which pass over the trailing model.
For the last vehicle spacing of 0.91 s/L all the configurations benefit from the platoon again. For the baseline models the gains are only a few percent but for the trailing model with a sharper front radius the gains are higher than those at the middle intervehicle distance. This is due to the reduced stagnation pressure and almost undisturbed suction coming from the rounded edges. At this distance any effect on the wake of the leading vehicle has vanished. The unsteady forces still show an increased amplitude compared to the single model however compared to the middle intervehicle distance they are much lower.
The comparison between the results from the numerical and experimental analysis is quite good. The drag and flow field results are very similar. For the pressure and the unsteady forces this is less the case. The mismatch seen in the pressure comparison can be caused by alignment errors that were observed during the experiment. Next to this while reconstructing the flow field from the experimental measurements the resulting coordinates were outside the expected range. A manual coordinate shift was applied which may have resulted in additional errors. For the unsteady forces it is assumed that the ground plate and its interaction with the model as well as additional vibrations are the cause of the much higher fluctuations seen from the experimental data.