Wind Tunnel Sim Work

This is a compilation of some aero-based demos and content I’ve developed.  Some of this was by and for me personally, some for work, and some as developed as testing environments used at work (but owned by me).  Likewise, the software used includes Unity, as well as Project Falcon (Autodesk Flow Design now), SimScale, and several custom solutions.


I’ve always loved aerodynamics.  Like game design and rendering it’s a beautiful combination of hard science and black magic; a stunning example of non-polynomial fact-finding.  You can reasonably calculate the effect of a change, but not the form change required to affect the result.  It favors those with an encyclopedic knowledge of history, a long list of dirty tricks and the willingness to think outside the box and around the rules.


The first is my first proposal for a Hololens (or similar) application that uses Unity to help calibrate wind tunnel and computational fluid dynamics (CFD) results.  Typically, the lift/downforce/drag/etc figures between testing types can differ drastically.  For example, CFD tends to perform poorly when developing more organic shapes – racecars designed exclusively in CFD with no wind tunnel testing are unnaturally smooth, slab-sides and relatively featureless.  Wind tunnels, on the other hand, encourage finer detail development, but often under-perform or under-estimate when measuring the effect of underbody aerodynamics and have a relatively small size limit for the model.

Furthermore, for when the vehicle is out of development and in the field, it is important to be able to quickly and easily evaluate the effect of certain parts.  One example is the Red Bull Racing Formula 1 Team – after heavy contact with another competitor during a race.  The floor and sidepod area of their car was damaged, and they needed to work out how much of an effect the damage was having on the car.  It’s plausible that similar situations may play out for aircraft operators, as well as high performance trains and boats, and I think it would be beneficial to be able to visualize the performance impact readily and easily.


This project actually utilized some of my earlier work; I modeled and simulated a Toyota TF110 Formula 1 racecar, and the wind tunnel seen in several of these pics.  Technically, the TF110 never raced (team quit F1), but that’s beside the point.  Using the handful of references available at the time I modeled, UV’d, and imported the car into the Autodesk Falcon simulation system.

A quick bit of history on the car:

While Toyota announced they would be withdrawing from Formula 1 at the conclusion of season 2009, development of the 2010 car, the TF110, continued until two complete cars were built. This was in the event of another team wishing to buy them and race them the following year, and in actuality, several parties came close. Hispania Racing Team had agreed to the use of them, though apparently broke their contractual agreements late in 2009. After that, a privateer intended to buy them for his F1 operation, Stefan GP. However, even after tricking Toyota into painting one of the cars Serbian Red, Toyota caught on that Stefan didn’t actually have the money to run a F1 operation.


In an embarrassing moment, the red TF110 was then driven by Toyota Motorsport’s John Howett crashed it into a wall. A shame for several reasons; firstly because it has the highest crash-to-distance-raced ratio of any car (infinite), and secondly because the performance of the TF110 would appear to be very promising. Featuring some of the most extreme rear aerodynamic concepts in recent years as well as what would turn out to be successful front aero solutions (high nose, fold-over endplates), in the hands of good drivers the car could be a consistent challenger for the lead.

Since then, numerous rule changes have made the car obsolete; and it would now be illegal to race in F1.



Another project that was later used at work as a testing environment for the (unfortunately, cancelled) general intelligence system was my ARX1 program.  This took the form of my old plane generation sim (think Spore or Kerbal Space Program, but with planes and missions).  By adjusting parameters like wing chord, angle, position, etc, a machine (and/or a player) could learn to ‘discover’ the laws of physics, from the basics of lift, mass, drag, power, and PID tuning, through to stability control theory and purposeful design principles.  In the case of the machine, this general intelligence could then be saturated, tested and trained with realistic data not possible using just government reporting data.

The project was divided in two parts – the editor and the world.  In the editor, it was possible to change most aspects of the vehicle, or lock them out to constrain parameters (for beginners / testing software).  This applied to both the airframe and the powerplant, and included several types of jets, piston engines, and electric motors.  The world was to be a fully-realized island nation, and players/machines would be free to build cargo/transport/passenger planes as the economy developed, and stealth/attacker/fighter planes in times of war.

Unfortunately, as the business intelligence project was cancelled, development on the ARX1 program just about ground to a halt.  I still work on it from time to time, and it’s great for testing out new designs and trying out fantastical ideas for planes.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s