Simulator technology

[Jersey Tom]

Biting my tongue here...

Think you guys are still giving simulators WAY too much credit for what they can predict. You can simulate and model any damn thing you want. Is it going to give you a reliable answer? Some things yes. Some things, eh, so so. Some things.. hell no.

Arnaud Dufournier is a nice guy, at least.

[Jersey Tom]

As a follow-up...

I'd say we can all agree that grip and wear change depending on what kinda surface you drive on, and how much downforce you have (among many other things).

Are you going to be able to confidently say what kind of surface you're going to drive on at the next race? How much has the track degraded since the last race? Have any patches been made? Will there be rubber laid down from a previous race or series? Will there be rain that washes some of it off? For a street course like Monaco, how much dirt and grime from regular street traffic will there be? How long will that last?

How hot will the track surface be? That certainly has an effect on things, no? Will it be sunny? Cloudy? Can alter the track temp immensely. Will it be sunny for the first half of the race and cloudy afterward?

What is the air temperature going to be that day? Density of air might be 1.2250 at 15C and 1.1839 at 25C. That's a 3.5% difference... if you have a car with 6000 lbf (~27kN) of downforce, that can be a change of 200 lbf (~1 kN)!! Likewise with the change in drag you'll have a different aerodynamic moment My, and front-to-rear balance.

So if we return to this question...

I wonder if they can correctly simulate things such as tyre wear and graining though... Those are, after all, quite crucial parts of the race weekend, if not in terms of outright pace, in terms of the pitstop strategy etc

You tell me Raymond, do you think they can accurately simulate and predict those things?

As I'm a fan of saying... creating math models on historical data is one thing. Forward prediction is another.

[747heavy]

sure JT, but it is exactly the same with track testing.
You will never be able to predict race day weather conditions with 100% accuracy.
Thats pretty independent of the fact that you use a simulation or doing track testing.

If you test (lets say as an example) during the offical test sessions in LeMans, does that mean the weather and track conditions will be the same come Qualifying time 3 weeks later?

corrected some spelling mistakes

[Giblet]

What is more important is simulating and quantifying the change on the car I think, as the tracks conditions are changeable.

If a simulator can show performance from a change to the car, under the same track conditions, then the sim has served it's purpose.

[DaveW]

If a simulator can show performance from a change to the car, under the same track conditions, then the sim has served it's purpose.

Apologies Giblet, but you have made two fairly important assumptions, I think. First that the simulator model is representative of the real vehicle. Second that the driver will interact with the simulator in the same way that he would with a real vehicle. Neither is necessarily correct. As an example of the first, how long did it take McL to solve the unexpected (hence, I guess, unmodelled) aero problem it encountered a couple of years ago?

[Giblet]

@DaveW

Speaking of assumptions, you assumed that I was speaking of a 'driver in the loop' sim, which I was not.

The sim is only as good as the data you put into it, and it can never replace the real thing, but it is getting better at it.

I was in iRacing, and Bobby Labonte (the NASCAR driver) came in to our Pontiac Solstice practice and was turning laps. We asked him what he was doing there, and he said "Getting some road racing practice in for the Rolex 24 at Daytona.

The larger the thing you are trying to accomplish, the more a sim can help. For smaller changes and goals, the sim becomes less useful.

Thats just how I see it

[Jersey Tom]

sure, but it is exactly the same with track testing.
You will never be able to predict race day weather conditions with 100% accuracy.
Thats pretty independent of the fact that you use a sumilation or doing tack testing.

I certainly agree.

What is more important is simulating and quantifying the change on the car I think, as the tracks conditions are changeable.

If a simulator can show performance from a change to the car, under the same track conditions, then the sim has served it's purpose.

Also agree. More interested in relatives than absolutes, presuming your sim and input data are good enough to pick out the right directionality.

Point I've really been trying to get at.. is that computer simulations of ANY sort (FEA, CFD, Lap Sims) are not the end-all-be-all solution... and all have significant limitations that the end user has to be aware of.

[autogyro]

There has always been simulation in motor sport.
We used to sit in the car in the garage going brrm brrm while moving the controls and imagining going round a circuit while looking at a circuit diagram on the garage wall.
Then we would cycle round the track on the Morning of the race.
Many teams with sports cars would drive to the circuits and the drivers would learn the car on the way.
Today it is all CAD CFD and simulators. Soon the actual racing will be reduced to almost nothing, so they might just as well race a computer game.
The giants who drove racing cars have been gone for years.

[lolzi]

Yeah yeah, autogyro, nothing will ever be as it was in the good old days - it's getting a bit dull, hearing that and the "ancient layshaft gearbox" stuff or whatever it is you are always writing.
What has your post got to do with what the author of this thread asked?

[autogyro]

Driving Simulator is the thread heading.

[godlameroso]

Undoubtedly the heavy restrictions on testing has contributed to the current development and use of simulations. I for one believe that if simulation tools continue developing at the current rate, in about 5 years, the high end simulators will be an adequate replacement for actual track testing. It won't be perfect because there are some things that simply cannot be simulated with technology available to our species.
Furthermore, if we were able to manipulate things like gravity for simulation purposes, we probably wouldn't be driving vehicles with wheels in the first place.

Also one can't deny at least a relative usefulness to simulation, not to mention that perhaps the software models evolve faster than the hardware that runs it. Reason being that the hardware side is driven by a different market with evolution driven more by competiton between brands that, most likely doesn't revolve around building race cars. Whereas the software side is developed in house by racing teams that are relentlessly trying to develop their cars.

[marcush.]

long gone the days when Alpina `s mechanics drove the Coupes across europe on wheels and only the lead car(!),the orange one was trailored to Brunn...

[marcush.]

we have recently seen all teams uprgrading their Driving Simulator technology ,with some teams apparently much more advanced than others .
The value of this simulation technology to the development of the car is emphasized by the teams but it seems some old school drivers are not too keen on the technology currently available or even get nausea in these devices due to unrealistic feedback of the machine.

I found the following about one of the simulators avaialable todays-it seems to be the technology used by Wirth /Virgin .

http://www.cruden.com/media/pdf/Cruden% ... %20app.pdf

one of the downsides of this concept seems to be you caanot generate sustained high g-forces with this approach and some drivers claim to really need the sensation of real g-force as the main reference to their driving feel.

To create a machine to produce sustained quick changing g-forces in the region of 5 gs would be of course quite a task ...


or what about this:
http://www.kuka-entertainment.com/en/products/

http://www.kuka-robotics.com/NR/rdonlyr ... oaster.pdf

I feel that could be well suited to found the base for sustained g-force sim.

[Jersey Tom]

Or how about a legit simulation package, worthy of doing engineering work.

http://prattmiller.com/software/
http://www.stackpoleengineering.com/products.aspx

Just sayin'. I'm a broken record when it comes to this topic though.

[marcush.]

something tells me this is what mercedes is planning to have for their simulator..
http://www.kuka-entertainment.com/en/pr ... institute/
http://www.dlr.de/desktopdefault.aspx/t ... ead-27431/


DLR has very close connection to Daimler Benz ...