86 replies to this topic

[lokiman]

Breaking: http://www.autosport...rt.php/id/91924

[mstar]

This was on the cards before the Spanish GP. Shame i liked Wirth's approach i just wished his car was faster so all his hard work got recognised.

[Iridescent]

Well, could have seen it coming as there were some rumours circulating about them ditching the CFD-only approach. They NEED to gain performance in the next few races, so best of luck.

[oetzi]

This was on the cards before the Spanish GP. Shame i liked Wirth's approach i just wished his car was faster so all his hard work got recognised.

All his hard work making a slow, unreliable car?

Not saying he didn't do as well as he could, but he obviously didn't do as well as he needed to.

[lokiman]

Yes, it is a great shame, but I also trust Pat Symond's assessment of what the problems were. I'd love to see the CFD-only approach prove itself yet.

[highdownforce]

Well, could have seen it coming as there were some rumours circulating about them ditching the CFD-only approach. They NEED to gain performance in the next few races, so best of luck.

Until the summer break at least, all upgrades would have already being designed by Wirth, so any (lack of) development would be his credit.

[Les]

Yeah I agree with Oetzi. Maybe he worked hard and its genuinely sad for him but the name of the game is to produce a fast car and he failed in this. I hope for the drivers (in particular the talented Timo Glock) that they get their hands on something more competitive.

[lokiman]

Still think that the Virgin is the best-looking car on the grid by a country mile, but if it ain't quick...........

[Mandzipop]

Totally shocked, I thought the point was to be green and not use a wind tunnel, obviously their position has changed since HRT started to become more competitive and Team Lotus are starting to mix it with the midfield teams. Money and reputation talks.

[TURU]

Finally.

[oetzi]

Totally shocked, I thought the point was to be green and not use a wind tunnel, obviously their position has changed since HRT started to become more competitive and Team Lotus are starting to mix it with the midfield teams. Money and reputation talks.

It's more 'who wants to pay money to be seen as the only kid in class who doesn't get it?'

[Les]

Hey lokiman, beauty is in the eye of the beholder but I personally think its ugly. Slow and ugly!

[Iridescent]

Until the summer break at least, all upgrades would have already being designed by Wirth, so any (lack of) development would be his credit.

Pressure was on 100% all the way back to China really, but once the Turkish upgrade did not deliver the parting was almost inevitable. So I bet they had a B plan ever since and the wind tunnel work would now certainly help them with the basics at least. So imo they're going to improve somewhat, no miracles. As I said before wish them good luck, they're nice chaps.

[lokiman]

Totally shocked, I thought the point was to be green and not use a wind tunnel, obviously their position has changed since HRT started to become more competitive and Team Lotus are starting to mix it with the midfield teams. Money and reputation talks.

I guess, if CFD wasn't working............I remember Mike Gascoyne making a good case before last season about how the CFD route was too risky, because the technology was still relatively young. Maybe he was right.

[swerved]

Totally shocked, I thought the point was to be green and not use a wind tunnel, obviously their position has changed since HRT started to become more competitive and Team Lotus are starting to mix it with the midfield teams. Money and reputation talks.

I think being fast takes precedence over being green, i'm only surprised he's lasted this long tbh, I think Nick Wirth talks some great ideas, sadly, not all of them materialise.

[Fastcake]

Well it was a real shame Wirth's all-CFD method didn't work out. It had potential to really reduce aerodynamic development costs, and was very interesting to follow Virgin's attempts to match the midfield this year. I guess in 5 years time someone may try out this method again, when the technology has had time to advance more. Out of interest, what will happen with Wirth Research's shareholding in the team?

For the future a new tech team under Symonds then, interesting if they can affect this years car at all. At least by making this decision now means the MVR-03 can be more heavily designed by the new team, Timo at least deserves to be back up there.

[oetzi]

Got to wish all the teams good luck. It obviously wasn't working, and it was the only way to keep the money coming in.

[oetzi]

Well it was a real shame Wirth's all-CFD method didn't work out. It had potential to really reduce aerodynamic development costs

I really disagree with that. You can spend as much as you want on programmers, processors and verification methods.

and was very interesting to follow Virgin's attempts to match the midfield this year. I guess in 5 years time someone may try out this method again, when the technology has had time to advance more. Out of interest, what will happen with Wirth Research's shareholding in the team?

For the future a new tech team under Symonds then, interesting if they can affect this years car at all. At least by making this decision now means the MVR-03 can be more heavily designed by the new team, Timo at least deserves to be back up there.

The rest, I agree, is interesting.

[bigginge]

The thing is, it was always obvious the CFD only approach was going to fail. All the other teams are not spending millions wind tunnel testing for fun - they all run extensive CFD simulations and are fully aware of what it's capable of. Over time I have no doubt that the balance will shift from Wind Tunnel testing to CFD, and when it does the front running teams will be at the forefront.

Wirth has no doubt done well out of the deal - his company have been paid for the work, and he's got a nice new supercomputer to play with. The critical mistake was assuming that success in a lower level (and more importantly closed wheel) series with the CFD approach does not equal success in F1.

[Watkins74]

Virgin is in F1 this year? Damn I never even noticed them.

[oetzi]

Over time I have no doubt that the balance will shift from Wind Tunnel testing to CFD, and when it does the front running teams will be at the forefront.

But it was interesting to see quite how far a full shift this is at the moment. Despite the very limited testing.

[Fastcake]

I really disagree with that. You can spend as much as you want on programmers, processors and verification methods.

You are quite right there, although a lot of that is a one time cost, compared to hiring out wind tunnels, and I had believed that it is more cost-efficient to run CFD compared to a comparative amount of wind tunnel testing.

[oetzi]

Virgin is in F1 this year? Damn I never even noticed them.

Wonder why they changed things?

[Dunder]

I strongly suspect that Virgin (initially conceived as Manor Grand Prix) would not have been allowed to enter F1 without Mosley doing a favour for his old friend Wirth.

Notwithstanding they applied to enter F1 as a budget-capped formula, their approach hasn't worked and they are barely ahead of (the widely ridiculed and ll but broke) HRT. Something had to give.

[oetzi]

You are quite right there, although a lot of that is a one time cost, compared to hiring out wind tunnels, and I had believed that it is more cost-efficient to run CFD compared to a comparative amount of wind tunnel testing.

Only if it works!

And the cost of constantly increasing processor power and refining models is ongoing.

[Rob]

Processor technology is lagging about 50 years behind the mathematics of CFD. Until the technology improves significantly, CFD simulations will be all about compromises in order to get the simulation time down. Unfortunately, with compromises errors can creep in.

[midgrid]

I feel really sorry for Wirth, as this is his third attempt to be successful in F1 and it hasn't worked out. Still, at least his company has been successful with the Acura prototypes.

[Mandzipop]

Got to wish all the teams good luck. It obviously wasn't working, and it was the only way to keep the money coming in.

Keep the money coming in? I'd like to know where the money has suddenly come from to start hiring the big guns and using the new facilities. They are on a very tight budget but having Pat Symmonds and Sam Michael onboard and on top of that using a wind tunnel, on top of that working with Mac/Merc must exceed their current budget. Money has been offered from somewhere to afford this. And it has to be serious money. Not an extra £5m, it has to be closer to £15m at least. That is if they are wanting to use the facilities for this year's car.

[engel]

never had any faith in Wirth, certainly not after the whole Simtek debacle, which was also supposed to be a cost effective foray into F1 design, 17 years ago.

[Fastcake]

Keep the money coming in? I'd like to know where the money has suddenly come from to start hiring the big guns and using the new facilities. They are on a very tight budget but having Pat Symmonds and Sam Michael onboard and on top of that using a wind tunnel, on top of that working with Mac/Merc must exceed their current budget. Money has been offered from somewhere to afford this. And it has to be serious money. Not an extra £5m, it has to be closer to £15m at least. That is if they are wanting to use the facilities for this year's car.

The money is coming in from Marussia, apparently they are investing a fair amount of money into the Virgin team - a lot more than Branson is. If Marussia want to make the team their big promotional vehicle, it certainly all adds up that they are paying for the new tech team.

[oetzi]

Keep the money coming in? I'd like to know where the money has suddenly come from to start hiring the big guns and using the new facilities. They are on a very tight budget but having Pat Symmonds and Sam Michael onboard and on top of that using a wind tunnel, on top of that working with Mac/Merc must exceed their current budget. Money has been offered from somewhere to afford this. And it has to be serious money. Not an extra £5m, it has to be closer to £15m at least. That is if they are wanting to use the facilities for this year's car.

There's a lot of people with deep pockets to back people they have faith in. Less to back a technology trial.

Last year Virgin got a lot of airtime and goodwill on the back of their (unsuccessful) experiment, with many hoping it would succeed. This year it's a known failure.

[Scudetto]

never had any faith in Wirth, certainly not after the whole Simtek debacle, which was also supposed to be a cost effective foray into F1 design, 17 years ago.

Curiously, then, in 1994-95, Simtek bumbled around the back of the field for one full season and through Monaco of its second season. Looks like, now, Wirth goes no further with his second outing.

Maybe third time's the charm.

[Disgrace]

Absolutely the right decision. Wirth has never created a top F1 car. Creating a fuel tank too small was a massive warning sign. Aero is clearly their biggest problem though and this was the only way they could have dealt with it.

I guess that's it for development of the current car, then, hope they don't fall too far behind when it comes to circuits like Spa and Suzuka with regards to the 107% rule.

[glorius&victorius]

[midgrid]

I guess that's it for development of the current car, then, hope they don't fall too far behind when it comes to circuits like Spa and Suzuka with regards to the 107% rule.

The source article states that there are several upgrades to come that have already been signed off by Wirth Research.

[Iridescent]

I guess, if CFD wasn't working............I remember Mike Gascoyne making a good case before last season about how the CFD route was too risky, because the technology was still relatively young. Maybe he was right.

Well, I’m very pro- CFD and I’m sure it will be used increasingly in the future. I remember reading years ago about a front wing analysis by Sauber – you need a careful meshing strategy for that, because capturing wakes and vortices is critical. A refined volume mesh was necessary, but it was also costly in cell count. Hence they introduced a new meshing scheme, which helps to resolve the flow field with better control and a reduced cell count requirement for the same model. This improvement in meshing technique eventually prompted an overhaul of the model. And at that time the software was hardly more than 50% of today’s in terms of development.

The thing is that it’s still not that reliable to be used as a sole method. This is because, as you know, the nature of the flow around a F1 car is highly complex due to the presence of large body and tyre wakes, strong vortices, separated flow regions etc. And while CFD is really good for the early concept phase and accessing various separate components, it becomes complicated when you don’t have another tool to verify the data. And it’s also questionable what the capabilities of the used software are, though F1 teams generally use some pretty advanced ones.

For example, with all the deflectors and external devices integrated to the bodywork, understanding of the coupling and interaction between the front and rear-end is vital. Minute changes in geometrical details of these components can often have a global impact on the overall flow topology, therefore influencing the performance. Hence, a simplified sub-modelling approach can provide only very limited information, sometimes misleading, and would not be sufficient.

Also a careful choice of meshing techniques is crucial for correctly capturing all essential flow field phenomena without creating a mesh of prohibitive size. The flow around F1 cars exhibits strong gradients travelling over large distances, and the control of the volume mesh in terms of both resolution and quality matters much for the prediction to be accurate. A right estimation of viscous vortex flows is known to be a true challenge, and excessive numerical dissipation resulting from insufficient grid resolution is still a major concern.

And finally, you cannot have just local analysis as a local solution that gives good feedback may be disastrous for the rest of the car, so you need to consider the test for the whole project. And that, believe me, can give you a great deal of pain.

So yeah, in a few words Gascoyne was spot on. Newey made similar comments not once. And excuse me for the long post, but I'm kind of a nerd when it comes to CFD.

[lokiman]

Absolutely the right decision. Wirth has never created a top F1 car. Creating a fuel tank too small was a massive warning sign. Aero is clearly their biggest problem though and this was the only way they could have dealt with it.

Meh, he gets a free pass from me on the fuel tank problem. Not his fault that the FIA changed the fuel spec regs after they'd finalized the chassis. Yes, maybe he should have played it safer, but if it had gone his way, people would be saying that he'd played a blinder. At the end of the day, the car is just too slow, so the writing was on the wall.

[pingu666]

didnt wirth research also make alot of the car aswell ?

[lokiman]

So yeah, in a few words Gascoyne was spot on. Newey made similar comments not once. And excuse me for the long post, but I'm kind of a nerd when it comes to CFD.

Am I also right in thinking that the regulations currently limit the computing power the teams can use?

[oetzi]

CFD...can provide only...limited information, sometimes misleading

I've taken the liberty of editing your quote.

I think all this shows that comprehending aero is still an art, not a science, and complex concepts can only be proved in real life, not by models.

Maybe not for ever, but definitely for now.

Hopefully for a while yet, as models effective enough to accurately predict complex outcomes this would make life very dull in many ways.

ps anyone know if it's still true that no computer model can reliably predict the path of the eighth snooker ball in a sequence initiated by contact with the first? I know it was a while ago...

[engel]

Am I also right in thinking that the regulations currently limit the computing power the teams can use?

time not power

[Craven Morehead]

Well, no surprises really. The car is not at all competitive. This is motorsport, the whole point is to build a quick car. The fuel tank debacle this year was really something..

[Iridescent]

Am I also right in thinking that the regulations currently limit the computing power the teams can use?

You're quite right. It's really all about a bargain between physic testing i.e. only windtunnel now (since in-season testing is forbidden) and computional software use as part of the cost-cutting. The restrictions as such are on the number of people involved and on the hardware performance of the CFD computer systems so that smaller teams aren't disadvantaged. But the truth is that some margin is left on a year-to-year basis for development and since the big teams are investing more, they still have better equipment/ people/ resources. It's another reason I imagine it must have been very hard for Virgin.

@oetzi: nice edit... But still, it can be very useful for some stuff, do not totally undermine my work

[KateLM]

Meh, he gets a free pass from me on the fuel tank problem. Not his fault that the FIA changed the fuel spec regs after they'd finalized the chassis. Yes, maybe he should have played it safer, but if it had gone his way, people would be saying that he'd played a blinder. At the end of the day, the car is just too slow, so the writing was on the wall.

Lotus and even HRT got the fuel tank right despite the late change, I don't think Wirth has an excuse.

Anyway, its not really surprising that Wirth has gone, he's clearly been on borrowed time for the last few months. Being so close to HRT when they are barely keeping afloat just isn't acceptable. It does make me wonder though - the average user on this BB had a fairly good idea that an all-CFD approach wasn't going to work with current technology. Did Wirth know something that everyone else didn't, or did he just think he did?

[bigginge]

@oetzi: nice edit... But still, it can be very useful for some stuff, do not totally undermine my work

Actually, I'd like you to touch on the complexity of modelling the influence of the rotating wheels, internal flow, cooling, exhaust thrust etc etc

It will be a long time before we have the CFD to cope

[bigginge]

Did Wirth know something that everyone else didn't, or did he just think he did?

Or did Wirth need someone to bankroll his new supercomputer

[primer]

I don't think there's anything wrong with a CFD only approach, someone has to be the first to do it and pave the path. Perhaps Mother Russia Virgin do not have the resources to do so in the best way. But it is their unreliability which has been very painful, and IMO cost Wirth his job.

[oetzi]

it can be very useful for some stuff, do not totally undermine my work

If you know of any line of work which can provide anything more than limited information which is sometimes misleading, I'd be interested to hear about it

Small edit for polite smiley.

Edited by oetzi, 01 June 2011 - 23:04.

[PNSD]

I cant speak for F1, but for aerospace flows the limiting factor is CFD is data storage. Obviously the larger the mesh the greater the number of, well numbers you need to store. This limit's mesh size and density which will possibly effect results.

I think Iridescent summed it up well.

It must also be noted that wind tunnel technology is by no means standing still. Technology is getting better, costs are getting lower. Only in the last 5-8 years have F1 teams started using optical methods, techniques used in aerospace for many many years. Measurement techniques and instruments are getting more accurate and methods such as PIV compliments CFD perfectly.

People seem have a crazy idea that wind tunnels are dying, whilst its my belief that they are nowhere near their peak and still have alot to offer.

My experience with CFD has not been pleasant, so im more inclined toward experimental methods however.

It was expected that this would happen, Pat Symonds is a great engineer and should lead the team toward great improvements.

[jonnoj]

CFD is never going to produce a decent car, until it's used by a good car designer. All we've seen from Virgin is rubbish in = rubbish out. We know CFD has some value in F1 as all of the big teams are using it, in conjunction with wind tunnels.

There have also been reports that Virgin is seeking a technical partner, with McLaren among the options.

I do hope McLaren avoid any partnership with Virgin. Their previous partners have a habit of losing a lot of money.