The gold standard for aerodynamic measurement is the wind tunnel. However it has several downsides: Cost for many is prohibitive; the flow is unrealistically steady; the cyclist does not tire as they would over the course of a 40km time trial.
Several companies have already tried to develop a device to measure real-time cycling aerodynamic drag, with mixed results. It is well known that Garmin have been developing a device for some years. One of their product leads told us it is an “8-degree of freedom problem”. Talking about ways of just evaluating aero devices Ray Maker commented in his DC Rainmaker blog “Holy @#$#@ this is hard”. Aerodynamics is difficult, but not impossible.
How do you measure the aerodynamic drag on a moving cyclist, on the road? We cannot directly measure the aerodynamic drag force on the cyclist and rider, but we can measure or estimate the other forces. In this context it is generally easier to talk about power rather than force (although of course they are directly related): overall there are 5 sources and sinks of power:
Rider power (source)
This comes directly from the power meter.
Kinetic power, due to acceleration or deceleration (sink or source)
This comes from the speed sensor: a function of rider and bike weight, and change in speed over a given time.
Potential power, due to climbing or descending (sink or source)
This comes from aerosensor’s high-precision barometric altimeter: potential energy is a function of rider and bike weight, and altitude. We are able to resolve changes in altitude of just 10cm.
Friction power losses, due to rolling resistance (sink).
As rolling resistance is typically small compared to the other contributions, we can assume a constant Crr (Coefficient of Rolling resistance) with little loss in accuracy. Rolling resistance power is just a function of Crr, rider and bike weight, and bike speed.
Aero power (sink)
What we are trying to measure!
Since the sum of all these 5 powers has to be zero at any time, by measuring or estimating 4 of them, we can calculate the 5th and obtain aerodynamic drag. Dividing the aerodynamic drag by dynamic pressure, which is measured directly by aerosensor, we get CdA.
The reason this is hard to do well is that there are many measurements going into this equation. The end number is only as accurate as the least accurate measurement. Drawing on my experience in experimental aerodynamics we spent over five years digging into each one and understanding how to extract the best accuracy from each. Here are some examples:
To further enhance accuracy, our Aerodynamic Cycling System has two other devices:
If you have bad data then the performance difference you are looking for is drowned out by noisy data. Not only is this confusing and frustrating, but it can lead to bad conclusions, ultimately causing you to reduce, not increase your speed.
With our system, we are able to achieve lap on lap repeatability better than +/-1% in the velodrome, or +/-1.5% over a 1km average on the road. Over the coming weeks we will be publishing case-studies showing how this data can be used to make you faster.
Our Aerodynamic Cycling System is being released in July. We’d love you to register your interest so we can keep you updated on our progress.