The general idea of testing.

[Belatti]

Ask Red Bull racing about "the idea of testing"... as they did 19 tests more than the rest of the teams this season...

[richard_leeds]

X - KERS is a good example of why they need to test. They keep failing in real life, so that shows how integration with other systems does result in unforeseen problems.

I always thought every F1 car a relatively untested prototype compared to most automotive engineering?

Anyway, you should switch to aerospace because they can't test the product that gets launched into space since many of them are single use only. What's more they don't even gets the bits back when it breaks.

Structural engineers don't even build prototypes, but you'd find their methods a little too crude. The lack of testing means they build in a serious amount of redundancy because even a small error can kill a lot of people. There's an irony - something that has the most potential to kill has the least testing and no prototypes.

[xpensive]

Quite obviously, I can fully understand the need for testing systems, in particular when said system might contain an entity not 100% known and certified, such as a dynamic sequence, like a turbo-engine's characteristic during revving-up.

What I react against however, is when people refuse to trust rather straightforward calculations, their own as well as others', for xample the critical speed of a simply supported compressor. If I have done this with Dunkerley's, Rayleigh-Ritz and a number of hocus-pocus computer softwares and came to the same result, I don't need to "test" it?

That's what engineering is about, but that's just me of course.

[Jersey Tom]

Depends on your definition of "straightforward."

Bear in mind, I'd say for a long time "build and test" was the way to go in engineering, fed by some lesser measure of predictive capability. Only in recent years has there been the explosion of computing power and the ability to do predictive work on increasingly difficult problems and challenging systems.

Even now, there are times where "build & test" is cheaper, faster, and gives you a more direct answer than doing simulations & predictions.

[xpensive]

Depends on your definition of "straightforward."
...
Even now, there are times where "build & test" is cheaper, faster, and gives you a more direct answer than doing simulations & predictions.

So very true JT, but there are instances where "testing" does not really apply, such as lifespan of a bearing, when there's simply no way you can hang around for 20 000 hours to make sure that it will last as for as long as xpected, is there?

[marcush.]

in theory you can calculate everything but for example a full blown thermal cosimulation with CFD is just only getting available these days..You can be happy to get a sim of a heatsoak situation but where is the value of calculations when you know in advance that your error is bigger than the increments you hope to find in improving your product?repeatability seems to be a big issue for racing not only when testing .
The drivers feedback is still a source of information you should extract to the last digit and interpret .

[WhiteBlue]

Testing in reality is never fully substituted by simulation as we know very well from the Virgin adventure. They have come back to the idea to stick their aero parts in a tunnel because it represents the reality better than heir computer models.

F1 has gone from almost no testing where everything new was bolted on a car and raced to orgies of testing in the noughties when the tyre war was fully on.

Right now it looks like a sensible compromise if you look at total testing. Nevertheless I would do less pre season testing and do one or two tests in the season.

[Robbobnob]

Also in a sport where driver preference is significant to the cars overall performance, just because a system is theoretically faster on paper, driver preference for the way he drives the car may be affected by the new system causing him to drive slower than theoretically possible.

A situation that comes to mind is the beginning of the exhaust blown diffuser last year with RBR, where Webber had an advantage over Vettel in terms of car pace, and it wasnt until a throttle mapping change where the performance gap changed.

Also at Japan (IIRC) this year, RBR were trying a different suspension and their relative performance to Mclaren had dropped, and when Vettel suggested a revert to the tried and trusted he was able to put the car on pole the next day.

So in recap, when the system is chaotic in nature with driver preference an important factor, engineering numbers dont always represent the whole picture.

[richard_leeds]

not a specific RB8 theme but I don´t get why teams do not equip the cars as rolling labs in FP1+2 ,along the lines of the pirelli test car and even more ...Is this a time constraint thing as your work allowance from Friday to saturday would not be enough to get the car back to race spec?

[raymondu999]

You mean littered with rakes pitots and whatnot? Or do you mean using Friday/Saturday as parts-test sessions?

[AnthonyG]

not a specific RB8 theme but I don´t get why teams do not equip the cars as rolling labs in FP1+2 ,along the lines of the pirelli test car and even more ...Is this a time constraint thing as your work allowance from Friday to saturday would not be enough to get the car back to race spec?

I think it just isn't allowed.

[raymondu999]

Why would it not be allowed? I think the cars don't need to be race-legal in FP. Remember McLaren in 2009 and 2010? They had pitot rakes and the like every other Friday.

[myurr]

Probably because each test interferes with the airflow so that it would ruin the results of all the tests behind it. I believe much of the aero rake data is also captured at constant speed so you miss out on important running whilst trundling around collecting that data.

McLaren in 2009 were probably trying to sort basic concepts and large scale changes rather than the more subtle fine tuning the F1 teams will be doing in season this year. If the teams thought they would learn something then they'd be doing it.

[marcush.]

Probably because each test interferes with the airflow so that it would ruin the results of all the tests behind it. I believe much of the aero rake data is also captured at constant speed so you miss out on important running whilst trundling around collecting that data.

McLaren in 2009 were probably trying to sort basic concepts and large scale changes rather than the more subtle fine tuning the F1 teams will be doing in season this year. If the teams thought they would learn something then they'd be doing it.

the issue with the rakes and other temporary appendages does strike me all the time.I feel it would make a lot more sense to have the tub and all mjor components designed with measuring airflows ,preuusres and deflection from the word go to make sure your test results are not influenced heavily by the test equipment.(Putting a rake into the wake will surely change the wake,doesn´t it?)
I feel all this Aero stuff is done in a pseudoscientific way in the same cotext i often see people do temperature measuring completely ignoring the limitations of their test setup .

[Lycoming]

It is difficult to gather reliable data from the car itself. Thats the (relative) advantage of the wind tunnel and CFD, but you have to test the actual car to know that those tools are producing useable results. Additionally, the way wind tunnels and CFD eliminates effects such as crosswinds, wake turbulence, etc. can be an advantage or a disadvantage depending on what you're trying to do.

I wouldn't go so far as to call it pseudoscience, but it is a bit of a black art to extract data from the running car. And don't forget, integrating all that test apparatus into the monocoque comes with its own disadvantages.

In any case sometimes its unavoidable; If you want to know about the flow in a region somewhat distant from the surface (ie. just behind the front wing) there's no way around sticking some large contraption out there, and incorporating that into the chassis will just mean you carry dead weight and drag into the race. You basically have to test it bit by bit, as the information from such a test is only good for the behavior of the front wing. And care must be taken to ensure downstream effects from the apparatus do not significantly effect the performance at the front.