Due to high fuel costs and increasing pressure to minimise greenhouse gas emissions, commercial operators strive to reduce the fuel consumption of Heavy Goods Vehicles (HGVs). At motorway speeds, fuel consumption is highly dependent on aerodynamic drag, a considerable proportion of which is attributed to the significantly less well documented trailer underbody. Moreover, the underbody flow influences the topology of the near wake, which is also responsible for substantial drag generation. However, the underlying mechanisms of these interactions, and thus the role of the underbody flow in the development of the wake flow features, are not well understood. This thesis focuses on gaining further insight on HGV underbody and near wake aerodynamics to guide the effective design of drag reduction strategies.

Two 1/10th scale models of a generic HGV are tested in a water towing tank. The first model is a simplistic truck geometry with a smooth underside, whilst the second accurately represents the underbody detail and has rotating wheels. The models are towed along the tank, establishing correct ground conditions, at width-based Reynolds numbers between 5.96×10^5 -7.03×10^5 . Good optical access is enabled through the transparent walls and floor of the working section. Measurements are taken using Stereoscopic Particle Image Velocimetry (SPIV), which allows three-component velocity vectors to be measured in 3D space, and a force balance.

Significant differences are found between the mean flowfields of either model. Although the simple geometry already exhibits a weak tendency for inwards flow entrainment from the sides into the underbody, this is exacerbated for the geometrically more complex underside of the detailed model. In the vertical symmetry plane, the simple model displays a big recirculating region in the upper wake, the primary recirculation. This is much smaller for the detailed model, for which there is also evidence of another recirculating region of opposite sign in the lower wake, the secondary recirculation. Moreover, the lower part of the near wake behind the detailed model shows a distinct region of upwash. This is in contrast to the simple model where the streamlines are roughly horizontal. However, these flows are expected to be highly unsteady. Their unsteadiness may be assessed by evaluating turbulence properties. Whilst Reynolds stresses are generally found to be low, locally there are regions with low time-averaged velocities where the turbulent activity can be quite dominant, primarily in the wake. Instantaneously, no evidence of bi-stable wake dynamics is found for either model.

To determine the cause for the differences between the flowfields, the two most important ‘first order’ effects of the underbody flow, namely ground clearance and underbody roughness, are investigated. It is found that by decreasing the ground clearance of the simple model or adding simplistic roughness to its underside it is possible to change the wake flow to be more similar to the detailed model’s. Most notably, these modifications lead to the appearance of a small secondary recirculation close to the bottom of the model base, below the dominant primary recirculation. Adding roughness has the greater effect, such that this secondary recirculation is bigger, albeit still smaller than the detailed model’s. Overall, ground clearance and, particularly, underbody roughness appear to contribute to the differences between the flowfields of the simple and detailed models. Therefore, a realistic level of underbody roughness and so, an accurate simulation of the momentum loss of the underbody flow, are important in replicating the correct wake flow. Nevertheless, neither ground clearance or roughness effects fully replicate the complex flowfield around the detailed model. This suggests that the underbody geometric detail likely plays an important role in the development of the flow features.

The effectiveness of a range of drag reducing devices and other modifications to the underbody geometry of the detailed model is also assessed. Both mid and rear side-skirts reduce drag, albeit via different mechanisms. The former likely prevent the entrainment of high velocity flow into the underbody, whilst the latter are postulated to reduce the streamline curvature in the wake and thus to increase the base pressure. Tapering in the rear side-skirts provides additional benefits. Wheel covers also are effective drag reducing devices that likely prevent any detrimental interaction between the undisturbed free-stream flow and the rotating wheels. Furthermore, wheel flow deflectors, which emulate the effect of mudguards, are found to increase drag. Careful shaping can alleviate this. Finally, bluff bodies stowed in the trailer underbody appear to have a negligible impact on aerodynamic drag, probably due to two effects. Firstly, the trailer underbody flow has a low velocity to begin with, having been decelerated further upstream along the tractor underside. Secondly, there likely is a balance between the drag generated locally by these components and the aerodynamic benefits as they impede the entrainment of high velocity flow from the trailer sides into the underbody.