When testing aerodynamic performance, there are three main tools that are used:
With each of these methods, the test subject (in our case a wheel) is tested in a controlled environment, either physical or virtual, under specific, controlled wind conditions. Wind speed can be scaled mathematically, to show drag (and other forces such as sideforce) at a range of speeds. However, the wind angle, known as the yaw angle, remains fixed, with the resulting drag value being specific to that angle.
In real world riding conditions, of course, the wind direction is rarely if ever stable and constant, so we need to be able to translate test results from fixed yaw angles to the more variable conditions encountered whilst riding. To do this, we need a dataset for "average" wind conditions, especially wind direction / yaw angle.
There are several such datasets already in use across the cycling industry. The simplest takes the average recorded wind conditions from UK Met Office weather stations across the UK (12 in total). However, these are fixed locations and the majority of readings are taken at a substantial height above the ground. Whilst wind profiles vary according to terrain, in all cases the elevation of a bicycle and/or bicycle component is such that wind speeds will differ from weather station readings due to the impact of the wind gradient.
The majority of the existing datasets have been determined by data collection from an instrumented bike. These bikes all use a single instrument (analogue wind vane) mounted out in front of the rider. This does provide data at a more appropriate ground elevation, however it does not take into account the airflow interaction with the bike or rider. This is particularly relevant when the item being tested is more likely to be impacted by "dirty" airflow, e.g. the rear wheel.
As we set out to develop the next generation of wheelsets, we wanted to fully understand the yaw angles that would be impacting both front and rear wheels individually, so that we could design each rim profile to perform at its best in those conditions.
As part of our technical partnership with the Sports Engineering Department, we launched the #thinkwider project to analyse wind conditions at multiple points on the bike, specifically the front and rear wheels. In order to do so, we needed to fit smaller sensors to the bike that would be able to sit at the exact points we were interested in. Whereas previous studies used larger, more cumbersome wind sensors (anemometers), we were able to use smaller ultrasonic sensors that were originally designed for use in sailing. These wind sensors measure the disturbance in four ultrasonic beams to give readings of both wind speed and wind direction.
For data gathering, a sensor was fitted to 3D-printed brackets mounted onto the wheel axles. Data from each sensor was gathered using smartphones and then uploaded for analysis following each ride. During testing, riders were asked to ride "as normal", across a range of wind conditions, ride types and ride distances. Test rides took place over a 12-month period, capturing different seasons.
Having analysed over 25,000 data points, we could see that there was a meaningful and consistent difference between yaw angles observed at the front wheel and at the rear wheel:
Fig 1: Observed front wheel yaw angle
Fig 2: Observed rear wheel yaw angle
The results showed that, on average, the yaw angle at the front wheel was 1.5 degrees higher than that at the rear wheel.
During one of our wind tunnel test sessions, we also used some of our test runs to further calibrate the analysis and validate the readings we had been seeing in our real world testing. The sensors were fitted to the test bike, held in the wind tunnel’s static test rig. The rider, bike and test rig were then put through a standardised test run, with yaw angle increased from 0 degrees to 20 degrees at 5 degree increments, allowing airflow to settle at each test point.
Fig 3: Observed yaw angles in wind tunnel testing
Results from the wind tunnel testing were very similar to real-world riding, showing an average difference of 1.3 degrees between front & rear wheels.
The adoption of new technologies in road cycling has accelerated rapidly over the past few years. As recently as 10 years ago, the vast majority of frames and wheels were designed and optimised for a maximum tyre width of 23mm. This setup would be run at >100psi, with either a clincher or tubular tyre, and with braking provided at the rim.
However, it has become increasingly apparent that there are a number of advantages to running a wider tyre:
The introduction and adoption of tubeless tyre technology has only furthered the benefits of a wider tyre (and hence the demand with riders). Eliminating the inner tube from the system entirely removes the risk of a pinch flat where a tube could be “pinched” between the rim edge and road when run at a lower pressure.
However, whilst the benefits of a wider tyre have become established, wheel designers were for a time still constrained by rim design. A traditional rim brake caliper would only allow up to a c.28mm rim width before the rim would no longer fit between the brake pads. This created an issue, as in order for the wheel/tyre system to maintain its aerodynamic benefits, the rim width must measure c.105% of the tyre width. The “Rule of 105” limited tyre widths to c.25mm.
The final piece of the puzzle has been the shift to disc brakes in road cycling. Whilst the UCI have only recently approved disc brakes for racing, riders have been increasingly moving to disc brake setups over the past few years, so much so that many of the leading bike manufacturers are now releasing disc brake-only frame designs.
Moving the braking force from the rim to the hub has allowed for far greater flexibility and innovation in rim design. Firstly, the rim width is no longer constrained to the c.28mm clearance of a brake caliper. The rim edge can also be more aggressively contoured as there is no longer a requirement for a flat (or close to flat) contact area for the brake pad to exert a frictional force against. Finally, as no frictional force is exerted at the rim, there is no longer a need for heat-resistant properties at the outer edge of the rim. This is a benefit as the heat- resistant resins that are used in a carbon fibre brake surface are more rigid and therefore less resistant to impact.
The results from the NTU yaw angle study showed that front and rear rim profiles should be considered individually, so as to best suit the wind conditions at that point of the bike. We also wanted a rim profile that would perform aerodynamically with a wider, 28mm tyre.
For the front wheel, in order to perform well at higher yaw angles, our #thinkwider rim profile uses a more "blunt", U-shaped design. Prior testing has shown the more rounded shape allows airflow to remain attached for longer around the rim, reducing turbulence and therefore drag. In order to maintain this performance with a 28mm tyre, our CFD simulations showed an outer rim width of 32mm was required. As we have observed, when fitting a tyre to a wider internal rim channel it will often measure up wider than its stated width. Testing a range of commonly-used 28mm tyres on a 22.5mm internal rim width showed that they would measure up to between 28.5 & just over 30mm in actual width. This ties in with the "Rule of 105" for the front rim. An additional attraction of this shape is that the more gentle curvature also benefits handling stability in a crosswind.
With the rear wheel, the focus was more on performance at lower yaw angles. Here, a "sharper" V-shaped design has been shown to provide the optimum aerodynamic performance. Furthermore, with the reduced yaw angles, the importance of the "Rule of 105" is lessened, allowing for a slightly narrower rim width, in this case 30mm. This reduced width allows for a lighter rim, meaning the rim depth can be increased without a weight penalty. Finally, prior generations of wheels that used this V-shape profile have suffered from less stable handling in crosswinds. Whilst the wider shape of #thinkwider does not remove this, the importance of handling for a rear wheel is vastly reduced as the rear wheel is not free to turn on its axis.
Our #thinkwider rim profiles provide a new common design approach for our next generation of disc brake wheelsets. The differential front/rear profiles provide:
1. Optimum aerodynamic performance when paired with a 28mm tyre
2. Enhanced front wheel stability in crosswind conditions
3. Maximised rear wheel aerodynamics from greater rim depth