“ are rfactor2 physics broken” video

[Yzangard]

Yes it is something that has been brought up several times before, it is nice to see someone else realizing that probably when you reach the limit of the scope of the simulation there are other factors that are simply not ...factored in. In the end the result is the same, but one thing is scream around that something is broken, one thing is to realize that in real life other things can happen and this possibility is enough to change a human behaviour. In RF2 tires, and mechanics CAN'T get damaged from contact with the track surface, this is obviously false in the real world.

I know that this has been said several times, I myself have already mentioned it as a possibility, but the more I advance in my "research", the more I tell myself that we have underestimated the consequences a little too much.

The best proof in the end is to test with "normal" pressures and to realize that suddenly everything becomes much more "realistic", at least more in line with what one would expect.

Now, lowering the pressure comes with the advantages without the disadvantages, it seems normal that pilots in competitions use this to go faster, it would just be necessary to find a way to "penalize" in one way or another these low pressures.

I have several official documents that indicate the minimum pressure to be used (140 Kpa) and the recommended pressures (160 minimum to avoid accidents) and others that show the consequences if a lower pressure is used, so this is not an "interpretation" on my part, it is simply the competition regulations. Yes we accept pressures of 140 Kpa but it is more than warmly recommended to use a minimum pressure of 160 Kpa. I don't know if this is hot pressure or not.

https://www.gt-world-challenge-europe.com/documents/notice/903/Bulletin+03+Tyre+Pressure.pdf

Pirelli's website directly about P Zero :

 

[Comante]

Sure, there is a pletora of side effect that can happen to tires and are not modeled: Tires dislodging from rims, tire's shoulders puncture due excessive and prolonged deformation that can lead to heat buildup and material degradation, tire surface and body suffer localized or diffused damages due excessive heat or prolonged overheating, those can deflate or destroy the tire.
If RF2 could come up with a realistic system to at least simulated punctured tires, many problem will solve on his own (just to let other sprout to life) , at least before engaging a certain kind of driving one would know that is gambling with luck. It can go well for sometimes, but soon or later punition will come.

 

[Yzangard]

Sure, there is a pletora of side effect that can happen to tires and are not modeled: Tires dislodging from rims, tire's shoulders puncture due excessive and prolonged deformation that can lead to heat buildup and material degradation, tire surface and body suffer localized or diffused damages due excessive heat or prolonged overheating, those can deflate or destroy the tire.
If RF2 could come up with a realistic system to at least simulated punctured tires, many problem will solve on his own (just to let other sprout to life) , at least before engaging a certain kind of driving one would know that is gambling with luck. It can go well for sometimes, but soon or later punition will come.

This is exactly my point, I'm pretty sure tire model itself isn't bad, it just miss this kind of problem to arise hence the unnatural steering techniques spotted.

 

[Jack Dyson]

There is something that rF2 (or other sims actually, maybe better in ACC but not sure) don't simulate : tire stress under very low pressure. It seems to me that tires can't unseat from the rim in rF2, which should happen when driving with low pressures as we do. I'm not convinced there is a problem with tire model per see, in fact, the more I look for technical information on tires (in this case Pirelli P Zero DHD) the more I realize that what we think we know is shaped by what we see in reality, except that in reality it is impossible to drive as we see our "aliens" do for the good and simple reason that the tires would either be destroyed or stripped off. The forces at work are such that these tires would not hold, in reality the tire is not solidary with the rim, it is constrained but not solidary. In rF2, it is solidary, it can't unseat, so it "resists" those forces.

When I set tire pressure a bit higher (to fit Pirelli's recommendations), this steering behavior is much harder and tire is sliding way more. It seems to me that the problem may not be directly with tire model but more the consequences of low pressures that are possible to use in rF2. I'm pretty sure real counterparts would allow same driving style if it wouldn't mean tire destruction or unseat.

The curves we get from telemetry are not directly exploitable because they depend on so many factors that everything is possible and credible. Whether we accept it or not, Niels Heusinkveld was right on this point, the "ideal" curves obtained with the Pacejka model are only theoretical results in a perfectly controlled environment, which is not the case of a car on a circuit, not to mention the fact that the ideal temperatures will also vary according to the friction frequency, which is not modelled by any laboratory model to my knowledge. Therefore, judging the curves obtained is more than presumptuous.

I broadly agree with some of the arguments made here - they are interesting and ultimately I think, correct. What I wanted to suggest is that perhaps such a microscopic approach to generating grip is counterproductive in simulation technology anyway.

We could implement mathematical constraints based on real world performance on a good dynamic model like rF2.

Let me explain by giving one out of a million possible examples: A simulator depends on parameters a, b and c where these are given constants: output = simulator(a,b,c,input) where input, symbolically, is the input of the user. It will generally be the case that output, given a,b and c plus a certain driving style around a given track produces data like real_output = car(a,b,c, input). In order that the simulation is effective and realistic we would need to make sure that we choose (a,b,c) such that ||output-real_output|| is minimal over some range of (a,b,c).

Now, in one scenario what S397 could do is to "lock" these range values in the setup to begin with, e.g. tyre pressure, cambers or whatever a,b and c are, do not drop below or go over certain values etc. That would then produce a better experience and use the benefits of their physics engine in a more visible way. Should a user wish to evade these realism recommendations - they can change an option in the game which is visible to servers, like damage immunity. Racing leagues can then have this option to 1) keep setups free 2) make sure the models raced are realistic simulations.

It is just an opinion, and it would be good to hear other opinions as well. I am hoping lots of us contributing might get down to a set of well accepted solutions to problems where guys like @Lazza can then take them to S397 for consideration.

Cheers!

 

[Comante]

Setup variations can already be limited online. It just take a bit of effort to create a rfmod (I think).

 

[Yzangard]

I broadly agree with some of the arguments made here - they are interesting and ultimately I think, correct. What I wanted to suggest is that perhaps such a microscopic approach to generating grip is counterproductive in simulation technology anyway.

We could implement mathematical constraints based on real world performance on a good dynamic model like rF2.

Let me explain by giving one out of a million possible examples: A simulator depends on parameters a, b and c where these are given constants: output = simulator(a,b,c,input) where input, symbolically, is the input of the user. It will generally be the case that output, given a,b and c plus a certain driving style around a given track produces data like real_output = car(a,b,c, input). In order that the simulation is effective and realistic we would need to make sure that we choose (a,b,c) such that ||output-real_output|| is minimal over some range of (a,b,c).

Now, in one scenario what S397 could do is to "lock" these range values in the setup to begin with, e.g. tyre pressure, cambers or whatever a,b and c are, do not drop below or go over certain values etc. That would then produce a better experience and use the benefits of their physics engine in a more visible way. Should a user wish to evade these realism recommendations - they can change an option in the game which is visible to servers, like damage immunity. Racing leagues can then have this option to 1) keep setups free 2) make sure the models raced are realistic simulations.

It is just an opinion, and it would be good to hear other opinions as well. I am hoping lots of us contributing might get down to a set of well accepted solutions to problems where guys like @Lazza can then take them to S397 for consideration.

Cheers!

Limiting would only be a workaround, not a real fix. I am limiting myself to "regular" values as I don't care about competition right now but of course, when competition system will be live, there must be a solution to find.

 

[mantasisg]

There is something that rF2 (or other sims actually, maybe better in ACC but not sure) don't simulate : tire stress under very low pressure. It seems to me that tires can't unseat from the rim in rF2, which should happen when driving with low pressures as we do. I'm not convinced there is a problem with tire model per see, in fact, the more I look for technical information on tires (in this case Pirelli P Zero DHD) the more I realize that what we think we know is shaped by what we see in reality, except that in reality it is impossible to drive as we see our "aliens" do for the good and simple reason that the tires would either be destroyed or stripped off. The forces at work are such that these tires would not hold, in reality the tire is not solidary with the rim, it is constrained but not solidary. In rF2, it is solidary, it can't unseat, so it "resists" those forces.

When I set tire pressure a bit higher (to fit Pirelli's recommendations), this steering behavior is much harder and tire is sliding way more. It seems to me that the problem may not be directly with tire model but more the consequences of low pressures that are possible to use in rF2. I'm pretty sure real counterparts would allow same driving style if it wouldn't mean tire destruction or unseat.

The curves we get from telemetry are not directly exploitable because they depend on so many factors that everything is possible and credible. Whether we accept it or not, Niels Heusinkveld was right on this point, the "ideal" curves obtained with the Pacejka model are only theoretical results in a perfectly controlled environment, which is not the case of a car on a circuit, not to mention the fact that the ideal temperatures will also vary according to the friction frequency, which is not modelled by any laboratory model to my knowledge. Therefore, judging the curves obtained is more than presumptuous.

Absolutely, I think all the same about it all.

Low tire pressures, or very low tire pressures is interesting subject. A lot of interesting questions could be raised.

Effect on tire wear ? Effect on tire carcass durability ? Effect on staying firmly attached to the rim, keeping air inside tightly sealed ? Effect on rolling resistance !!!? Rolling resistance hard vs deformable surfaces differences ? Effect on fuel consumption ? Effect on contact patch pressures distribution ? Effect on wet surface performance ? Not asking about resistance to punctures. Effect for tire stiffness for soaking up bumps ? Effect for tire stiffness on deforming laterally during high G cornering ? Driving on sidewalls at high G cornering because loss of rigidity ? Possibility of having lower than atmospheric pressures inside the tire (there should be no question about such thing, but there is) ? Effect on load sensitivity, especially for high downforce cars ? Relation to tire width and low pressures ? Relation to tire diameter and low pressures ?

There are so many things to get tested. Some of which are arguably ok. I have no idea how many of these things are wrong or right and if wrong by what pressure these effects starts separating from reality. It is possible to limit minimum pressures for individual cars, therefore it is not a critical issue if underinflated tire simulation is not at its best.

Some other sims might avoid having any trouble with all that, since they simply don't simulate the tires physically or for that particular matter, they just have straight forward curves for what approximately should work ok as optimum pressures and they just saves whole lot of trouble, rF2 obviously chose to have these mechanics to "just happen" by simulating the tire physically. The right way to do it IMO. But of course challenging. I suppose tire, all these bristles should start acting rather chaotic with very low pressures, and I just assume that simulating anything chaotic is particularly super complex.

 

[Jack Dyson]

Absolutely, I think all the same about it all.

Low tire pressures, or very low tire pressures is interesting subject. A lot of interesting questions could be raised.

Effect on tire wear ? Effect on tire carcass durability ? Effect on staying firmly attached to the rim, keeping air inside tightly sealed ? Effect on rolling resistance !!!? Rolling resistance hard vs deformable surfaces differences ? Effect on fuel consumption ? Effect on contact patch pressures distribution ? Effect on wet surface performance ? Not asking about resistance to punctures. Effect for tire stiffness for soaking up bumps ? Effect for tire stiffness on deforming laterally during high G cornering ? Driving on sidewalls at high G cornering because loss of rigidity ? Possibility of having lower than atmospheric pressures inside the tire (there should be no question about such thing, but there is) ? Effect on load sensitivity, especially for high downforce cars ? Relation to tire width and low pressures ? Relation to tire diameter and low pressures ?

There are so many things to get tested. Some of which are arguably ok. I have no idea how many of these things are wrong or right and if wrong by what pressure these effects starts separating from reality. It is possible to limit minimum pressures for individual cars, therefore it is not a critical issue if underinflated tire simulation is not at its best.

Some other sims might avoid having any trouble with all that, since they simply don't simulate the tires physically or for that particular matter, they just have straight forward curves for what approximately should work ok as optimum pressures and they just saves whole lot of trouble, rF2 obviously chose to have these mechanics to "just happen" by simulating the tire physically. The right way to do it IMO. But of course challenging. I suppose tire, all these bristles should start acting rather chaotic with very low pressures, and I just assume that simulating anything chaotic is particularly super complex.

It is a classic problem of scale in deterministic physics - you could in theory start at the nanometer scale to simulate the rubber, tarmac and load interaction - in fact this is what is done in theoretical physics when we develop models for things like perfect gases and material equations of state (and we can never get it right except in the most idealised cases).

However on a practical level perhaps having an intelligent approximation within certain limits of how the tyre operates would do fine. It has to behave correctly on a global scale (time, distance, forces) and then further research and refinement can be added in successive versions (like more refinement of the initial object model methods that develop various physical quantities). Furthermore - modelling transition across different physical scales (multi-scale theories) require massive computing power. Now from what I have heard this is exactly what S397 do - they run hours of "dry" simulations before a tyre model is ready for use in the game.

At low pressure there is massive tyre compliance (an extreme or singular condition where the tyre pressures approximate to zero) and there is no simple equation I know of capable of simulating what happens to the rubber while rotating (finite elements ?) - one might argue it is not even a tyre anymore.

In chapter 2 of Carrol Smith's "Tune to Win" he talks about "tyre compound limits" and exceeding these means the tyre essentially comes apart. So keeping a sophisticated tyre model might well be fine as long as one modified the simulator object model to account for manufacturer "compound limits".

In my view though, a blend of theory and empirical data is the best way (like in ACC perhaps) since it provides solutions that are accurate enough at a relevant scale of time, distance and force.

And In fact, you made an excellent point about rolling resistance ! I forgot about that : how is it that cars are faster on long straights with the miraculous flat tyres (e.g. Hockenheimring 66)? That could give more clue as to what is going on underneath.

Best to all.

 

[Claudio Souza]

have you managed to make it work?

 

[Comante]

If I remember correctly toe in /out , caster, and all tire deformation that happen "going straight" doesn't concur in drag generation.

 

[mantasisg]

And In fact, you made an excellent point about rolling resistance ! I forgot about that : how is it that cars are faster on long straights with the miraculous flat tyres (e.g. Hockenheimring 66)? That could give more clue as to what is going on underneath.

I frequently think about it, especially since I have had lots of work to do with top speeds for 1954 and 1967 car packs car physics that I have worked a lot.

I haven't heard of tires actually havign more topspeed with very low pressures. I'd like to test it. Perhaps particular tire has certain expansion due to low pressures, but I don't know... By the way, during my work on cars I found that expansion of tires radius at high speed can reduce top speeds due to rideheight sensitivity to drag.... everything is so complex... On the second thought at high speed gasses inside tire might also get compressed towards outside diameter, and might not reduce tire radius expansion due this reason... anyway just a random thought...

I actually keep on remembering rolling resistance, and that it is definitely very important component on tire dynamics, car dynamics, also in economy. But I keep on forgeting to do some reasonable testing to understand it. I would also find it helpful to find some reference of some cars inflation pressures relation to tire pressures. Thats what I am going to do now, I will look up if I'll be able to find cues, data.

There are some lines about rolling resistance in TGM documentation, but those I have seen are more related to bristles. If tire run on lower and lower pressures, logically more and more rolling resistance should come from whole tire deforming, instead of micro defermations and macro deformations of the tread.

Indeed. For cars at least for modern road tires I find "high-speed pressures" mentioned everywhere. To put it short it tires with too low pressures might not survive high speeds due to fatigue, deflection on the bottom one way, and deflection on the rest of the tire the other way due to expansion. I couldn't really find top speeds stuff, just found that coasting test is common. And these tests shows massive differences in coasting distances. Interestingly in the world of bicycles there is plenty of talk about impeding point of rolling resistance, which happens when rolling resistance starts developing more from bumps in road than from tire hysteresis and deflection, but I suppose it might be insignificant for cars as they are more inertic, and run much wider tires than bicycles anyway.

 

[Jack Dyson]

I frequently think about it, especially since I have had lots of work to do with top speeds for 1954 and 1967 car packs car physics that I have worked a lot.

I haven't heard of tires actually havign more topspeed with very low pressures. I'd like to test it. Perhaps particular tire has certain expansion due to low pressures, but I don't know... By the way, during my work on cars I found that expansion of tires radius at high speed can reduce top speeds due to rideheight sensitivity to drag.... everything is so complex... On the second thought at high speed gasses inside tire might also get compressed towards outside diameter, and might not reduce tire radius expansion due this reason... anyway just a random thought...

I actually keep on remembering rolling resistance, and that it is definitely very important component on tire dynamics, car dynamics, also in economy. But I keep on forgeting to do some reasonable testing to understand it. I would also find it helpful to find some reference of some cars inflation pressures relation to tire pressures. Thats what I am going to do now, I will look up if I'll be able to find cues, data.

There are some lines about rolling resistance in TGM documentation, but those I have seen are more related to bristles. If tire run on lower and lower pressures, logically more and more rolling resistance should come from whole tire deforming, instead of micro defermations and macro deformations of the tread.

It is a very good point about rotating compressible fluids and you can get a basic solution for the air density inside the tyre (away from the sidewall) by transforming to a frame of reference that is a perfect gas in a gravitational field of strength R*omega**2 at some atmospheric pressure P which is the pressure of the tyre itself: the result is then an exponential density increase from the center of the wheel to the outside of the tyre, which can be cut off at the inner radius of the tyre to get a first order approximation of what the rotating gas might do inside the tyre - yes this stuff is complex in every way in fact - according to Carroll "The subject [of tyre physics] is complex and imperfectly understood".

Two useful books for your travels :

"The Unified Physics of Tire and Rubber Friction" by Kummer

"The Physics of Tire Traction" by Hays

Do you have email ?

 

[Jack Dyson]

If I remember correctly toe in /out , caster, and all tire deformation that happen "going straight" doesn't concur in drag generation.

Caster and toe might be second order effects (generally toe heats the tyres up and they use it in cold weather to bring the tyres into their operating zone) and impulsive deformation of the tyre will probably not do much but if you change the size of the contact patch it definitely will. If you take a mountain bike, and really pump it up : it rides well and quite quickly on a road. If you deflate the wheels it takes a lot more power to turn the crank to reach a given speed - which ironically might be related to the tyre compliance itself.

 

[Yzangard]

It is a very good point about rotating compressible fluids - yes this stuff is complex in every way in fact - according to Carroll "The subject [of tyre physics] is complex and imperfectly understood".

Two useful books for your travels :

"The Unified Physics of Tire and Rubber Friction" by Kummer

"The Physics of Tire Traction" by Hays

Do you have email ?

Dunno if he does but I do have email if you can provide book samples

 

[Jack Dyson]

Dunno if he does but I do have email if you can provide book samples

Certainly - can you PM me on this forum at all (I'm quite new so I don't know) ?

 

[Yzangard]

Certainly - can you PM me on this forum at all (I'm quite new so I don't know) ?

done

 

[Raintyre]

None of you will be eager to reduce your own street car tyre pressure to half of the recommended because your instinct KNOWS exactly what will happen to straight speed, cornering speed, tyre temperature, fuel consumption, etcccccccccccc, if you do that...

Even if we are talking about hypercomplex things and hypersophisticated physics models there are some basic things that are simple.... It's a pity the basic things are forgotten and missing.....

 

[Bruno Gil]

None of you will be eager to reduce your own street car tyre pressure to half of the recommended because your instinct KNOWS exactly what will happen to straight speed, cornering speed, tyre temperature, fuel consumption, etcccccccccccc, if you do that...

Even if we are talking about hypercomplex things and hypersophisticated physics models there are some basic things that are simple.... It's a pity the basic things are forgotten and missing.....

Two completely different scenarios. Besides the differences in the tires themselves, on the road you're usually trying to use as little power as possible to get to your destination, so any little bit more of drag for example is very noticeable. Same effect on fuel consumption. On a track you're using all of your power, so it makes sense that the little bit of extra drag due to toes, rolling resistance, etc, is very minor in comparison. And from my testing toes do seem to produce more drag( even if just a little, you can see it in the straights with a delta timer).

Even tho I am also of the opinion that something is wrong with low pressures behaviour (I'm tending to go with yzangards hypothesis atm), if those effects you mention are so obvious, why does f1/ pirelli for example need to set a minimum pressure regulation? Why did the drivers all complain about the tires when that minimum pressure was raised the round after the blowouts in Silverstone in 2020?
Unfortunately it isn't as simple as you say

 

[FERNANDO CABRERA]

I like the tire puncture or blowup if you're pressure is set below a recommended point, the lower the tire pressure the higher the probability of a puncture.
That way people must use aero grip to stay on track, the gain reducing tire pressure woud not be worth it in the long run... But some stategies could come where tire are changes before n number of laps to gamble with the possibility of a blown tire.

 

[marti]

Ok im not sure where to post anything on this forum it's very difficult to navigate.
So sorry if I'm in the wrong place.

I need to know if there is a way to get this game to run properly with a simucube 2 pro.
I have tried for a week now and it's not working with my simucube 2 pro, feels like shit.
Anyone out there who can help with getting the game dialled in with the simucube, it would be
much appreciate. Thanks