Cyclist Aerodynamic Drag Analysis through Large-Scale PIV

Abstract

The use of large-scale particle image velocimetry (PIV) is proposed for cycling aerodynamic study to advance the general understanding of the flow around the rider and the bike, leading to new strategies for cycling aerodynamic drag reduction in the future. The investigation concentrates on the measurement of the wake velocity and its relation to the aerodynamic drag of stationary models in wind tunnels and of transiting models in the field. In the first part of this work, PIV measurements are conducted in a wind tunnel to capture the wake flow topology of a full-scale cyclist model and determine the cyclist aerodynamic drag. In-house built seeding systems are employed to inject Helium-filled soap bubble (HFSB) tracers upstream of an elite time-trial cyclist replica. The obtained flow topology compares well among different experimental repetitions and with literature, demonstrating the robustness of the PIV measurement approach. The aerodynamic drag is obtained by a so-called PIV wake rake approach, which relies on the conservation of momentum in a control volume surrounding the model. Comparison of the PIV wake rake aerodynamic drag against that of a force balance demonstrates that a drag accuracy of the latter below 1% is possible. The PIV wake rake measurements are conducted in a plane downstream of the bike’s rear wheel to avoid shadows and optical blockage. At this distance from the athlete, however, investigation of the separated and reverse flow regions, that are the main driver of the aerodynamic drag, is not possible. In the second part of this dissertation, therefore, robotic volumetric PIV measurements are conducted to retrieve the velocity description close to the cyclist. The near-wake of the cyclist limbs is presented, which somehow resembles that of isolated bluff bodies, such as cylinders, featuring a recirculation region bounded by two shear layers. The size of the recirculation region, however, is not only governed by the width of the limb, but also by the coherent vortical structures emanating from these limbs near the limb junctions (e.g. elbows and knees). Moreover, interaction of the limbs with the wakes of the upstream body parts also plays a role in the local wake properties. In addition to the measurement of the cyclist’s near wake at typical race speed, also the cyclist Reynolds number effects are investigated to understand how to reduce the aerodynamic drag by dedicated skinsuits designs in the future. This is achieved repeating the robotic volumetric PIV measurements in a wide range of freestream velocity. While reductions of the wake width are observed on both lower leg and arm with increasing free-stream velocity, the wake of the upper leg follows an opposite trend increasing in size at higher velocity. These variations of wake width with increasing freestream speed are related to the behaviour of the local drag coefficient, indicating a drag crisis behaviour on both leg and arm. The distribution of the so-called critical velocity upon these body segments is discussed, as it determines the freestream speed where a minimum value for the drag occurs. The third, and last part of this work, is dedicated to the development of quantitative flow visualisation and drag determination of cyclists in the field. This so-called Ring-of-Fire system allows, among others, aerodynamic studies that are practically impossible in the wind tunnel, such as model accelerations and model curved-linear trajectories. A tomographic PIV wake rake is employed to measure the flow around a simplified transiting bluff body, a towed 10 cm sphere. These scaled experiments serve as a proof-of-concept of this novel measurement system. The aerodynamic drag is obtained invoking the control volume momentum balance in a frame of reference moving with the object. The expression for the time-average drag consists of three terms, a momentum, Reynolds stress and pressure term, which are individually evaluated at increasing distance downstream of the sphere. It is shown that the aerodynamic drag is most accurately evaluated when the contribution of the momentum term dominates the overall drag and that the PIV pressure evaluation can be avoided five sphere diameters into the wake. The latter largely simplifies the data reduction procedures of the Ring-of-Fire. Finally, the present system estimates the aerodynamic drag with an accuracy of 20 drag counts. This is evaluated from repeated model passages in a range of Reynolds numbers in which the model’s drag coefficient is constant. This resolution is comparable to other aerodynamic drag measurement field techniques. It is rather poor, instead, in comparison to force balance measurements in wind tunnels. In contrast to the latter drag measurement techniques, the Ring-of-Fire also provides information about the flow yielding advanced insights into cyclist aerodynamics in the future.