Simulator technology

[dren]

So when simulators are upgraded, do teams rework the source code to run on a new hardware set-up or do they start from scratch? I would think they rework their modeling and add more complex computations when the hardware can do it.

I would think these machines have tons of processors with lots of cache and ram. Or are they just high end linked desktop machines?

[raymondu999]

I would think they somehow dump numbers from CFD to the simulators so that the simulators only need to run "lighter" calculations

[dren]

If that was the case, you could run the cars on a PS3. I think they do to some extent, but I don't know how much and where the benefits come from by keeping the calculations on the simulator. I'd think you'd want a lot of the tire modeling on the simulator.

[Jersey Tom]

Eh? rFactor Pro isn't a game. That's rFactor; which, confusingly enough, bears no relationship to rFactor Pro in terms of product lineage. They're entirely separate products

Still, why deviate from something that exists and you feel comfortable with (say from previous professional experience).. to something that's an unknown? Would have to do all the validation work.. and what do you gain from it? Perusing their website I don't see what advantage it offers over anything else. To me, the "driver in the loop" and "visualization" bits are nice-to-have rather than need-to-have.

Also, the claim on their website...

rFactor Pro is in use in F1 and NASCAR as the primary simulator for engineering development, driver training and telemetry visualization.

... screams of BS and puts me off a bit. Unless they mean that at some small number of organizations, this is used as their primary simulation tool, rather than the way it sounds of this being the primary tool used among all teams.

[Richard]

JT - I note that rFactor comment refers to visualisations, and might have been valid for the smaller teams without in-house CGI teams when rFactor was ground breaking at the time of release.

However, my conclusion from the previous thread sums up my view:

Driver in loop - useful educational tool for the driver, some feedback to designers about diver/car/track interaction, but time limited for exploring permutations.

Driver out of loop - able to undertake zillions of laps using zillions of permutations. Perfect for engineering development. Also useful for post-processing the data logs to inform the driver and team about the optimum set up and driving style.

[Tim.Wright]

...Also having read the website, it does seem it has (or had) some F1 applications. I would guess it could be with Toyota since they are now offerring it it commercially.

Anyway, I can guarantee any serious customer would ditch the standard vehicle model and run co-simulation with Carmaker, simpack etc as mentioned on their website. At the end of the day, this is the important part of the simulator. Everything else is just fluff. But they seem to be very good at the fluff.

As JT said, going down any other route for the vehicle model would mean repeating your validation work and would very likely leave you at a dead end when you realise its not adequate.

Tim

[Gatecrasher]

So when simulators are upgraded, do teams rework the source code to run on a new hardware set-up or do they start from scratch? I would think they rework their modeling and add more complex computations when the hardware can do it.

I would think these machines have tons of processors with lots of cache and ram. Or are they just high end linked desktop machines?

With some of the teams having been or in some cases still are sponsored by tech companies I am sure that the equipment is more than just a few desktops thrown together. Since testing is limited I am sure the IT budgets are not small.

I have 2 of these at work http://h10010.www1.hp.com/wwpc/us/en/sm ... 69399.html

It is amazing the computational horsepower you can get for a relatively small amount of cash.

[Richard]

Here's a relevant article by James Allen. OK, he's describing a mobile simulator for the public, but he does mention a few useful bits. He talks about the feel of teh tyres when they are cold compared to warm, and also that the seat belts tighten to simulate the feeling of braking!

http://www.jamesallenonf1.com/2011/08/a ... simulator/

[silente]

I pick this thread up again to ask you about a doubt i always have.

I don't know it F1 teams use RF pro for their simulator or they developed software on their own for this purpose. But anyway, Rf pro claims that one of its best features is that it can be linked to bigger toys like ADAMS or Matlab.

I know Adams quite a bit and i have some ideas on how it could be useful to let it interact with a driver in the loop simulation, but i would like to hear your opinion about that.

[Jersey Tom]

I know Adams quite a bit and i have some ideas on how it could be useful to let it interact with a driver in the loop simulation, but i would like to hear your opinion about that.

Well don't be shy. Let's hear them.

[silente]

Adams doesn't have very good or accurate lap time solver, for example. But on the other hand, you can build very datailed models with it and it is pretty good doing what the users tell it to do. You can ask to your vehicle model to both follow a certain line executing certain drivers commands or to follow a certain speed and line and find itself what to do in order to achieve this target.

It could be interesting to let a driver to drive a certain vechicle model to HIS limit and, if he is good enough or the model is driveable enough, to model limits and then to analyse in great detail with ADAMS what is happening to each subsystem...

any other ideas?

[Jersey Tom]

Why go through the bother of having the driver do a limit lap in a driver-in-the-loop simulation when you could collect data much easier at the track?

[silente]

to test new parts for example or their effect,before you go on the track you could evaluate which solutions are worth to be really tested.

[Jersey Tom]

And do you need a human driver to evaluate all of those changes and upgrades? (This is a matter of opinion).

[silente]

yes this is a matter of opinion.

But the question was different. I am not asking why to use driver in the loop, just why you could want to use a driver in the loop simulation together with another software like adams.

From driver in the loop software sites, like RF pro, they claims one of their features is that they can work with adams...but with which purpose they do it?