The previous four posts generally outlined the design methodology for individual wheels. I may not have stated it but you can see in the photos posted in the previous posts that the modeling only included a single wheel by itself. This was an intentional choice. Generally speaking a wheel that will be more aerodynamic in the front will be more aerodynamic in the front. The reason why we didn’t specifically optimize front and rear independently is that each rim shape that we start manufacturing is a high fixed cost for the mold, so doubling rim shape (front/rear individual) essentially doubles cost. Also since the rear wheel is in a lot of dirty air (really only the trailing edge of the rear wheel has significant drag effects), the assumption was made that if a particular rim depth is good enough for the front, it’s good enough for the rear.
Additionally you’ll notice that the frame/bike we setup here is pretty old school/plane jane. This was done very much intentionally (had nothing to do with ease of CAD’ing). We didn’t’ want to do a complex Aerodynamic frame-set since there is so much variation in Aero frames out there (cut-outs, airfoil shapes, etc). This basic frameset would give us a good baseline to analyze the wheel-set without worrying about further complex interactions.
If for whatever reason you’re interested in the frameset I used, check it out HERE
However, when we started looking into a disc wheel, there is no way getting around NOT analyzing a full bike model. Obviously 99% of the time, a disc wheel that we’ll be producing gets used as a rear wheel. So the decision was made to not even bother analyzing the wheel by itself. And since we were not going after the ultra-elite track cyclist market (you got it covered Mavic), the wheel would never be used as a front wheel.
Anyway since the disc wheel is essentially used only as a rear wheel we needed to analyze the whole wheel bike system. Again another assumption we made was to exclude the rider in the CFD model. There were a couple of reasons for this. First a persons body is too variable and constantly moving to accurately model in our CFD model. Second, we wanted to purely examine the rear wheel, and attempting to model legs would only dirty the air going to the rear wheel, further confusing results. So it was decided to model just the full bike and rear wheel without rider to get the best stratification of results for the different rim shapes while still maintaining the effects of the bike (which will always be present regardless of how erratically you might be pedaling)
There were fewer differences in the disc wheelset results as the Disc design is already naturally pretty aerodynamic. We also took a look at a few more parameters when analyzing the results. Due to limitations in computational power and time constraints we only executed a 0 degree AoA case and a 10 degree case. However the 10 degree case we did from both sides (wind coming from drive side, and non-drive side). We wanted to look at that to determine if there were significant differences or benefits of a particular design on drive side or non-drive side due to flow differences.
After running probably 5 different rim shapes (after the baseline straight disc, 90mm case, and sub-9 type cases), I came up with what I thought was a pretty good design that appeared to be lower drag over most cases than either the straight disc or Sub-9 disc. This is yet to be confirmed by wind tunnel results (we’re still a work in progress). But I’m pretty happy with the design because it’s backed up by making intuitive sense and is fairly simple. And if you’ve read any of my other blog posts on engineering you know that I always thing elegant and simple designs are the best.