I took some time out to get some data on the drag and lift forces at differend angles and speeds from zero upwards of 250 km/h windspeed. The testbed of my experimend was a heavily modified version of the example car(Ps. it is broken at the moment). For the drag component I used a horizontal sledsystem that was fixed to the global coordinate system. For the lift a vertical one, The Collision triagles had a liftCoef of 100 at all time and gravity set to zero. After a long night I had a good sample size. I found alot of usefull things and also some weird behavior. The forces were messured with the node grabber, sounds tedious and it was. My first messurements were the effects of different surface areas to forces on the car in the horizontal sledsystem. And then converting the forces to forces per m^2 Example Testsurface 16m^2 at windspeeds of 150km/h, force messured F=3456N 3456N/16m^2 = 216,8N/m^2 All the data is converted this way as my first objective was to look how the game realises the conversion from the jbeam liftCoef to the game(this data is not present in the spreadsheet). No, to small differentes were found of the forces per m^2 with smaller or bigger test areas. Surface areas tested: 1m^2, 2m^2, 4m^2, 16m^2, 36m^2. Now to the more interesting things. During testing I found out that the testbed moved alot at no windspeeds and the lift force would taper of after 25 Km/h at low AoA and than come back at 50 to 100Km/h. On the 0° AoA test the results were very confusing. The airfoil had at 0Km/h a 10.81N/m^2 Force that strengthend to 20.44N/m^2 but then droped to 8.125N/m^2 at 100Km/h and recovert only at 250Km/h with a messured force of 21.25N/m^2. That could be a result of errors in the calculation of small numbers that will be present in this situation. What does that mean for the aero for cars and planes, loss of traction and speed as the surface must be higher angled to have the same effect and the drag would increase. For planes higher AoA to fly even at high speeds. The efficiency of the arfoil at high angle of attack was compromised as observed in the first graph. At all time the lift force was smaller than the force created in the 30° AoA test. Both drag and lift forces rise exponential, but the lift forces are in magnitudes stronger. What did I learn, well alot. Specially the unexpected behavior at low speeds and low AoA. Too really understand how the game calculates the forces I will need to make another test with a set of different surfaces that have other lift coefficents and then compare them.
If you need or want help, send me a message, I am a Dutch mechanical engineering student (tu Eindhoven) and I am also investigating this a bit, but not so much as you did
Great. I am an mechanical engineering student in Dresden Germany. My objective would be to reprogramm a lift and drag system that works on martix like for the performance of an engine Code: "aero":[ ["AoA", "Cl", "Cd"] [-20, -1.2010, 0.06783], [-15, -0.9549, 0.02863], [-10, -0.6670, 0.01504], [-5, -0.3273, 0.01071], [0, 0.0970 , 0.00829], [5, 0.5189 , 0.00828], [10, 1.0252, 0.01357], [15, 1.1395 , 0.04827], [20, 1.1650 , 0.08496], ], This are real values for a B-29 Airfoil. Before I can do this I need to invest some time in the old system
Why using a matrix? Why not using formulas and game data? I will send you a better explanation when I am back from uni
Sure, I am looking for the best of both worlds. Good performance and realistic aerodynamics. Don´t know how it will turn out
First of all, thank you for the analysis of the aerodynamics. It is very interesting to me when i get independent testing of the engine's features. Please note, that current aerodynamics only simulate local interaction with air. Each triangle individually. Having said that, we are still researching more advanced aerodynamic simulation methods that could be implemented in a real time setting. What would interest you concerning new aerodynamic "features"?
Thank you for writing , my biggest gribe with the aerodynamics would be that there is no stall behavior, not that I found one. Goosah said that at the moment the forces are based on a tilted plaid, but it seems that the triangles can still produce lift at 45 degrees. In Reality it would be able to to this till 20 degrees. What I want to say is that the Angle of attack should play some sort effect in calculating lift, ether by having a stall speed that is defined hardcoded or as an value you can set in jbeam. That way an airfoil could be simulated in an basic way without huge calculations
I am interested in seeing more interaction with the air under the car. From under the car aerodynamic elements to blow overs, where air getting under the car and lifts it into the air.
