F1 simulator with 5g load - a discussion

[hardingfv32]

maybe the sim was inaccurate.

Exactly. Simulations aren't perfect. They do performance prediction, but that doesn't mean they nail it 100%. You don't just run some sims, throw the car on a 7-post, and arrive at a racetrack perfect. Or at least that would be the exception to the rule.

Could they be better than you imagine? Have you had the opportunity to use a simulation before your track testing?

Brian

[hardingfv32]

Even if they were perfect, you can only run them with some specified weather.
And you cannot predict how the weather will be... Neither how the track will be.

There is no need predict the weather or track conditions, as that is an constant unknown.

You can develop the simulation for different possible weather and track conditions. Then at the track you make your last minute setup decisions based on these simulations results.

Brian

[Pandamasque]

I don't think you need the acceleration magnitude to match the real car.

Accuracy of the direction of the acceleration I feel is more important. I would also put accuracy of rotational movements as infinitely more important for the driver than the lateral or long accelerations.

It depends if the purpose of the simulator is to develop the car or the driver.

[Tim.Wright]

To me its clear they are a vehicle development tool.

They won't help a driver because the feedback is too far removed from a real car. Besides learning a new track, theres nothing a driver would really get out of a simulator.

Tim

[hardingfv32]

To me its clear they are a vehicle development tool.

They won't help a driver because the feedback is too far removed from a real car. Besides learning a new track, theres nothing a driver would really get out of a simulator.

1) How is this a vehicle development tool if the driver is not able to give good feedback? Why is the driver even part of the simulator?

2) 'There is plenty to be learned by the best drivers pounding around in the simulator. The car's performance can be optimized into a condition that the leaves the driver unable to control the car. The simulator allows the driver to modify his skill set to better cope with previously un-drivable conditions. We are trying to modify the drivers preferences to better suit the car.'

Fact, we have no idea how good the simulators are or how well they have been correlated to actual track tests.

Brian

[Jersey Tom]

Have you had the opportunity to use a simulation before your track testing?

Brian

Yes.

[Lurk]

Even if they were perfect, you can only run them with some specified weather.
And you cannot predict how the weather will be... Neither how the track will be.

There is no need predict the weather or track conditions, as that is an constant unknown.

You can develop the simulation for different possible weather and track conditions. Then at the track you make your last minute setup decisions based on these simulations results.

Brian

You will have a good setup but not the perfect one, so you'll have to improve it. It is what I meant.
Otherwise the driver is a lucky guy, it happens sometime: a team my brother worked for did the pole and won the 2 race in an event at Laguna Seca with a stock setup. A restful race weekend for him.

[Mogo]

I must be really stupid, but surely stop the stupid testing ban with real cars on the track then we wouldn't need to go the extraordinary lengths being discussed here?

[Jersey Tom]

I see no need to go to these lengths regardless.

[hardingfv32]

I see no need to go to these lengths regardless.

You don not find it unsettling that many other engineers have convinced their team directors that the simulators are worth the big expense? We are not talking about one team.

Could they be developing systems beyond your knowledge base or imagination?

Brian

[wesley123]

No they cannot, or we differ in what we believe from that

Like said, it is an simulation. Thus you run a computer program based on the track and knowledge of the car. Weather and such, don't even try to simulate it, you cannot do that without an huge(and with huge I mean really huge) and then still it can differ. You can test things like tire wear, and all such things, but it is still an abstract simulation of what they know of the car as engineers. They are likely to be able to import CFD/Windtunnel data directly but apart from simulating tires, engine, gearbox aerodynamics and such. These are good enough for drivers to learn the track and the engineers to make an set up for the weekend. Drivers in the end still require changes of their cars in the weekend to suit the car it's liking. Afterall you are simulating reality, you arent running reality into a machine.

Just as an example take a look at rFactor, one of the few simulation games available. Look how simplified it's physics are. Real life probably has a zillion more parameters and the simulators used by the teams have a lot more parameters too.

Systems 'beyond our imagination' are yet something that are for our imagination. And if it will ever become real life what you are running? No, since then it wouldnt be simulation anymore. Current computer systems can barely correctly predict weather, let galone run reality in that

[Jersey Tom]

I see no need to go to these lengths regardless.

You don not find it unsettling that many other engineers have convinced their team directors that the simulators are worth the big expense? We are not talking about one team.

They haven't convinced their team directors to build simulators in a damn centrifuge.

Could they be developing systems beyond your knowledge base or imagination?

I'd say I'm quite familiar with what is on the leading edge of development. Honestly, there's no magic to vehicle simulation. I'd say everyone who works on it knows what's out there as far as fundamental capabilities and drawbacks of various solutions.

Just a question of how much emphasis you want to put on certain things than others.

[hardingfv32]

[quote="Jersey Tom"]They haven't convinced their team directors to build simulators in a damn centrifuge.

1) That is not the question I asked. You questioned the value and I asked how could you if many F1 personnel are backing up their opinion with serious money. Say they are just as knowledgeable as you, would not their expenditures be a convincing argument that they perceive value where you don't. And if many such people exist on several teams, that if fact there is value in the simulators?

2) Is a centrifuge required for the simulator if we are not concerned with driver strength training?
Can't the brain/body be tricked in regard to G forces?

Brian

[wesley123]

The main reason for these simulators is just the lack of testing. Although mainly McLAren and Ferrari started with simulators in the early 2000s it is just replacement for testing, no longer allowed on unlimited basis

[Jersey Tom]

We are talking about high G simulators here, are we not? That was this whole hypothetical thread? I just see no reason to pursue them. Huge investment, little return. Money best spent elsewhere.

Is there some value to driver-in-the-loop simulators? Sure, there's some. However, for as high profile and "flashy" as they are, there are much more valuable engineering tools that don't get shown off to the public - probably because they are boring. Boring and/or proprietary, where the real work happens.

With regard to Brian's early question to the effect of, "Well if everyone else is doing it, isn't it probably worth doing?" I absolutely do not subscribe to the "monkey see, monkey do" philosophy that is surprisingly prevalent in racing. While I certainly like to keep aware of the other approaches in industry, if I don't agree with them then that's that.

There are people in my profession who have made a name for themselves in my profession doing it for years or decades longer than I do, and have entirely different opinions on how to approach and solve problems. That's fine by me. They can do their thing, I'll do mine. Does that make me stubborn? Cocky? Sure.