The Aero Testing Process pt.1

Introduction

This blog post delves into the merits and challenges of aero testing on an outdoor velodrome. Nestled in the cycling haven of Cambridge, New Zealand, we are lucky to have both indoor and outdoor velodromes on our doorstep. So over the summer we tasked our aero engineer to investigate the aerodynamics of various track cycling handlebars. What learning can we take away from the process and how accurate will the results be from testing in outdoor conditions? 

Over several sessions with different athletes, various concepts were tested in the search for performance gains. The main purpose of the testing was to refine the aero testing process rather than the results themselves. This way we could establish a repeatable and documented process for aero testing in the future. 

Why conduct aero testing?

In a data-driven industry like track cycling, field testing stands as a pivotal aspect of research and development. It is a chance to validate theoretical development and ensure new products are robust and well suited to real world application. However the reliability of results hinges greatly on how rigorous the testing protocol is. I wanted to be confident in setting up the test parameters correctly so we could trust the results. Data on exactly how handlebars affect the aerodynamic drag of a cyclist is a valuable resource for athletes selecting handlebar types, in addition to driving our own future innovation.  

How to test aerodynamics?

To measure the aerodynamic drag on a rider, we use the term CdA. This quantifies the drag coefficient (Cd) and frontal area (A) of an object to tell us its aero efficiency. The larger the CdA the slower you will go, or the more power you need to hold the same speed.

World class CdA values in track cycling can be as low as 0.15 and are typically in the 0.20-0.30 range for the competitive amateur.

We used the Notio Aerometer which has a built in pitot-static tube that measures dynamic and static pressure. It uses this measurement to compute the all important CdA value.

• That narrower handlebars improve aerodynamics
• That longer reach improves aerodynamics
• That the Velobike Skat Grip C is faster than the other grip options currently available (Grip A and Grip B)

Velobike Skat handlebars with Grip C

What did we learn about the process?

Over several sessions, we had different athletes complete test runs on various handlebar setups. Using different athletes on different days increased the variables in the results, however accuracy wasn’t the intention of the test. As mentioned above, this was more about establishing and refining a repeatable process to measure how aero a system is. Using various athletes on different days meant we could build upon our learning to implement on future occasions when accuracy is more important. Given this, there were some aspects of the process that worked well and some that were more challenging;

• Testing on an outdoor velodrome had the advantage of being very accessible. It’s a pressure free environment with almost limitless track time so made for a great place to begin the hunt for useful data.  

• Windy days caused a constant change in wind direction when riding around an outdoor track. This made maintaining a certain speed during the test runs almost impossible, resulting in inconsistent CdA values. A calm day typically yielded CdA results that varied by 0.003 on repeat tests, compared to 0.006 with moderate wind. Testing on a day with anything above a gentle breeze (around 20 km/h) was therefore not worthwhile.

• Over the course of a 2 hour session riders began to fatigue and alter their position. We saw this in repeat baseline tests that showed a slightly higher CdA at the end of the session. Monitoring this helped eliminate the error due to a fatiguing rider.  We used Leomo motion sensors, but video footage or perceptual cues like the tail of a helmet staying in contact with a rider’s back could also work. Alternatively, bringing the rider up to speed behind a motorbike would reduce the power needed to accelerate and lessen fatigue.

• The Notio aero sensor and attached software takes care of the CdA calculation. But there are input variables like mechanical efficiency and rolling resistance that aren’t measured and need to be manually input. Spending some time researching these from reliable online sources is needed to be confident in the computed CdA value. Below is an example of rolling resistance values for the Vittoria Corsa G+ 2.0 tyre (bicyclerollingresistance.com).

Rolling resistance test results (bicyclerollingresistance.com) 

Here is a screenshot of the bike profile setup page of the Notio app. Total weight, tire circumference and rolling resistance are all important inputs to the CdA calculation.

Notio Aerometer bike profile setup page

• Lastly, these sessions are time consuming and can include upwards of 6 race-pace intervals. Consideration towards the rider’s personal training programme and relaying useful insights back to them was always appreciated by the athlete.

Joel Douglas carrying out an aero test run (Te Awamutu, NZ, 2023)

What were the testing results?

We found some interesting aero savings for the athletes we tested. They are specific to each rider and can’t be assumed to hold true across the board. Nonetheless, they show the magnitude of savings that simple equipment changes can make. Below are three results from different athletes:

A 180mm stem (compared to a 140mm length stem) saved 18 watts at 45 kph.

This session looked at the effects of stem length and position. We tested a longer stem, and then a higher handlebar placement with the longer stem. The results were consistent with the rule of thumb of a longer reach being faster. Here's data from one athlete where a higher handlebar increased CdA, so the fastest setup was running the 180 mm stem at the original stack height. 

Changing handlebars saved 36 watts at 40 kph

We changed two variables to achieve this massive saving. Firstly, the handlebars were narrower (330 mm compared to 360 mm). Secondly, the narrower handlebars had hood support which enabled the rider to adopt a horizontal forearm position in the hoods, rather than the drops. This explains the huge CdA saving; not all of it was from the bar width, but also from a positional change.

   

360 mm handlebars (top) and the Velobike bunch bar 330 mm (bottom).

For this session we upped the test run distance to 2.0 km from 1.5 km. This produced more consistent CdA values between repeat tests but had the obvious disadvantage of the athlete needing to exert more energy. The graph below shows the power output needed to achieve different speeds using the two handlebars.

Skat Endurance Handlebars with Grip C was 10 watts faster than Grip B at 45 kph

The Skat Endurance Handlebar has three grip options (at the time of testing) - A, B and C. They are a small component so have minimal contributions towards frontal area. However, the grip type (shape) determines much of how a rider positions their upper body; and rider position has much more impact on frontal area Grip C allows the rider to reach further forward and bring their hands closer together to emulate a pursuit position. Grip B has more similarities to a road bike shifter hood, which fits the rider's hands with a slightly reduced reach and a wider stance relative to grip C. This could explain the 10 watt saving that we recorded with this athlete.

Grip B (left) and grip C (right)

Discussion: what are the limitations and learnings to implement for next time?

For the purpose of ironing out the aero testing process, outdoor velodromes proved to be a practical and accessible resource. The sessions are a big time commitment, especially when testing multiple variables. So testing on an outdoor velodrome is a good place to start. You can dial in the process without the pressure and expense of an indoor velodrome. In terms of the results, we were limited by the weather, which played a major factor in the consistency. But by monitoring this there were certainly data take-aways which helped the athletes refine their position and equipment. The next step towards more precise results would be to eliminate some environmental factors by testing on an indoor velodrome (more details on this to come in a future blog). 

We also saw that results are very personal and subject to physiology and ability. The application of real world aero testing is therefore helpful for individuals looking to improve their aerodynamics. For a more general study into analysing the aero performance of certain components or positions, we would need a larger sample pool to detect trends. 

Overall, the old saying ‘garbage in - garbage out’ held true. Accurate input values and a regimented test protocol produced the cleanest results. We are excited to continue the deep dive into track cycling aerodynamics and will be sharing more insights as they develop. In the meantime, let us know in the comments below what aero data you would most like to see.

Written by Dan Gardner as part of an Engineering internship with Velobike Innovation