Wind tunnel talk 2012

[superdread]

Hmmmm maybe a dumb statement,,,wouldn't be my first,,,I thought the air temp was constant at a set humidity and density .

But you can't set the density of an air volume directly, you have to use other parameters to influence it, like pressure, humidity or temperature.

[Tommy Cookers]

even at low speeds big wind tunnels will need MW to circulate the air, presumably there is a corresponding heating effect
(and a heat exchanger ?)

so how does the basic circulation-related heat (+ moving ground plane) compare to any additional hot-plume modelling heat ?

(generating and using hot plumes may not be trivial either ?)

[mep]

so come on guys no input here ?
how is the exhaust system integrated into the model and how is the heat removed from the system again as it would be a
big factor to the windtunnel having the exhaust gas heating up the tunnel air in no time...do they have a big aircon or what?

A cold gas is used for the exhaust but the tunnel gets heated up by the friction of the belt and by external changes. That is why a air conditioning system is installed in big tunnels.

[strad]

This appeared in the most recent AutoWeek:
Inside the mind of Chip Ganassi

Ganassis's so called secret tunnel is like a rolling road wind tunnel, only full scale, cheaper and better.
It surfaced sometime in 2009 after a hiker snapped pictures of new infrastructure at the ends of the Laurel Hill Tunnel--a previously abandoned, roughly one mile Pennsylvania Turnpike bore under the border of Westmoreland and Somerset counties.
By then Ganassi had long since repaved the tunnel and equipped it with climate control, safety equipment and data collection systems. It allows limitless aerodynamic testing in controlled temperatures, free of wind, with a full size race car traveling at speed. It joins the best elements of real straight line testing and a conventional wind tunnel.
Ganassi first used the Laurel Hill Tunnel for testing in 2004 to develop the G-Force chassis for the Izod IndyCar series."It's not a secret," Ganassi says.
"It's not incumbent on me to publicize how we test, but it isn't a secret."
"Even if you rent a scale model rolling road, the models for cars in three different series gets expensive, and we were getting the most success out of our straight line program. My engineers told me,"We need a drag strip--let's buy an old drag strip or air strip and if there is still budget we'll build some walls and if there's more budget we'll put a roof on", I asked if a tunnel would work and they said"maybe better'. And I said "Good 'cause I know where there's a tunnel, so let's see if I can get that."
"You can work 365 without worrying about weather, noise or changing conditions." Ganassi continues "You build the real piece, not a model. Every team has a tunnel. Ours is just different, maybe more unique. I't the real thing because the heat is in the right places. The temperatures are real not calculated or simulated ..........

[Gee]

Hey, long time lurker, but something I felt I could contribute to a bit.

As just mentioned, a wind tunnel will tend to heat up over time, so most will have some form of temperature control. Mclaren use their huge ying (or yang) lake to help with their cooling for example. Very little work was done previously in wind tunnels with hot gases, after a couple of models went up in flames and most of this is now tackled with CFD (Source WT lecture courses/labs/job interviews). EDIT: Makes me wonder how well correlations can be made between WT and CFD with such an emphasis on hot gas blowing recently

I have never heard of the Laurel Hill Tunnel but sounds like a genius idea.

[olefud]

so come on guys no input here ?
how is the exhaust system integrated into the model and how is the heat removed from the system again as it would be a
big factor to the windtunnel having the exhaust gas heating up the tunnel air in no time...do they have a big aircon or what?

A cold gas is used for the exhaust but the tunnel gets heated up by the friction of the belt and by external changes. That is why a air conditioning system is installed in big tunnels.

A cold gas would have significantly differing aero characteristics vis-à-vis’ exhaust, I would think. For instance, the dynamic viscosity of air is 3.75 at 60degrees F vs. 9.5 at 1500 degrees F in the same units. This would feed back into laminar vs. turbulent flow and a number of other basic aero parameters that differ in kind rather than degree, i.e. discontinuities not subject to calibration.

Perhaps exhaust is either for real or an art practiced in modeling.

[skgoa]

so come on guys no input here ?

MGP has just made the step to 60 % in the midst of the 2012 season.Why would you do that ? They did not have to build a new tunnel for this as their tunnel is made for full scale testing since 2006...very mysterious .

Is you question (1) why would you do that at all or (2) why did they do this only now?
I don't have special insight into the team but my guesses are:
1) Because size (and wind speed) DOES matter a lot. That guy who claimed a 100% tunnel isn't a huge boon is talking out of his ass and clearly never worked in aero simulation (whatever kind) before. Because the air does behave quite differently under different parameters and you can't always predict in what ways. (E.g. the switch to a turbulent boundary layer, stalling, etc.) In this case, the most important one will probably be the relation of viscuous and inertial influences, aka the Reynolds number. You can't just slap a scaling factor on it, since these dependencies are non-linear. A switch from 50% to 60% is huge and it will definitely give them better results.
2) Probably because it either cost too much time or money to do it before or because they now realised that they do indeed need to make the step. Maybe they had to re-meassure a huge number of models to re-calibrate their wind tunnel and it just wasn't worth stoping all aero development in the mean time. They had been having big problems with correlating their wind tunnel data with the real world and only fairly recently got on top of that, so the switch might have seemed to be a big risk.

[mantikos]

so come on guys no input here ?

MGP has just made the step to 60 % in the midst of the 2012 season.Why would you do that ? They did not have to build a new tunnel for this as their tunnel is made for full scale testing since 2006...very mysterious .

Is you question (1) why would you do that at all or (2) why did they do this only now?
I don't have special insight into the team but my guesses are:
1) Because size (and wind speed) DOES matter a lot. That guy who claimed a 100% tunnel isn't a huge boon is talking out of his ass and clearly never worked in aero simulation (whatever kind) before. Because the air does behave quite differently under different parameters and you can't always predict in what ways. (E.g. the switch to a turbulent boundary layer, stalling, etc.) In this case, the most important one will probably be the relation of viscuous and inertial influences, aka the Reynolds number. You can't just slap a scaling factor on it, since these dependencies are non-linear. A switch from 50% to 60% is huge and it will definitely give them better results.
2) Probably because it either cost too much time or money to do it before or because they now realised that they do indeed need to make the step. Maybe they had to re-meassure a huge number of models to re-calibrate their wind tunnel and it just wasn't worth stoping all aero development in the mean time. They had been having big problems with correlating their wind tunnel data with the real world and only fairly recently got on top of that, so the switch might have seemed to be a big risk.

#2: You can test more using a smaller scale model than a larger one...you can do x number of days of 60% scale testing OR x+y days of 50% scale

[mantikos]

as we are back to the tunnel discussion with MGP now having 60%model size.
I had thought Honda err MGPs tunnel had 100% capability allready as they introduced their new tunnel in 2006? Interestingly
enough the phasing in of this tunnel correlates to the fall from grace of Honda .....may it be this tunnel was not trusted and the team kept the 50% modelsize to be able to correlate to their trusted old unit -which is ´now no longer available (rented by HRT) as you canonly use one Tunnel anyways....

The 100% capable tunnel can and is used to test 50 and 60% models...so it is the same tunnel. MGP's tunnel is unique that it is has an open section - forgot the techincal name for this. Willis commissened it and is a firm beliver in this type of open tunnel because it allows them to introduce some additional external elements for testing the affect on the airflow (read this in a Willis interview).

[marcush.]

a 10% bigger model will neever ever reduce your input -how could it? no more parts ,no more complexity -a few % added weight a few percent added material...but surely a lot fuss to accomodate everything needed inside the model itself methinks...sensors are not infinetely small .as are junctionboxes and connectors.

IK´d think those tunnels with exhaust flow modeling need a real air con unit to get rid of the waste heat..or you would already start with an open tunnel?

The scale factors i have hammered these concerns since long time .My point always was: you are looking into minute differences and small gains -so it seems rather difficult to separate gains from tunnel scale effects /errors -obviously mirrored perfectly by teams difficulties to correlate their tunnel and cfd developments to on track performance over long periods of time.

To me this all looks like modeling and "adjusting" the model via fudge factors gained in track testing .Over some period this may work very fine until you finally leave the "linear"zone of your model and suddenly NOTHING works as your calcs are just flawed and in reality you were always just iterating towards the outer rim of the envelope and finally dropping like a stone.

[Pierce89]

a 10% bigger model will neever ever reduce your input -how could it? no more parts ,no more complexity -a few % added weight a few percent added material...but surely a lot fuss to accomodate everything needed inside the model itself methinks...sensors are not infinetely small .as are junctionboxes and connectors.

IK´d think those tunnels with exhaust flow modeling need a real air con unit to get rid of the waste heat..or you would already start with an open tunnel?

The scale factors i have hammered these concerns since long time .My point always was: you are looking into minute differences and small gains -so it seems rather difficult to separate gains from tunnel scale effects /errors -obviously mirrored perfectly by teams difficulties to correlate their tunnel and cfd developments to on track performance over long periods of time.

To me this all looks like modeling and "adjusting" the model via fudge factors gained in track testing .Over some period this may work very fine until you finally leave the "linear"zone of your model and suddenly NOTHING works as your calcs are just flawed and in reality you were always just iterating towards the outer rim of the envelope and finally dropping like a stone.

I don't know about the actual wind tunnels, but I've read quotes about "fudging" the CFD conditions until they get closer to reality. So, to match the wind tunnel and track testing, they're inputting "wrong" boundary conditions etc. into their cfd runs.

[strad]

It seems to me that when Ganassi says:

The temperatures are real not calculated or simulated .

he's talking temperatures in this case but I would think it carries through to all aspects..The bigger your model and the closer to actual conditions, the less you're going to have to rely on guesstimations.

[Tommy Cookers]

even a 100% tunnel with negligible blockage will give results that need correlation with real-life data
(air moved at speed between walls over a stationary model is not ideal in flow quality eg turbulence)

correlation enhanced via improved flight instrumentation methods has been substantial development in recent years
(extending the applicability of tunnel data in supporting aircraft certification, saving time and money)

there is no substitute for correlation (so good real-life data are needed), apparently Ferrari has a pressing need for this
compared to this a 50% Re no is trivial ?(presumably F1 designs to a Re no)

information in the public domain suggests Re no sweeps of perhaps 2 - 20% are standard where viscosity is important
(and that major programs eg the FA18 E/F were deficient in this regard ?)

[Huntresa]

Mod Edit - this post and next couple of pages moved from F2012 thread.
_________________________________________________________________________________________________________

Great shots of wind tunnel testing:





The car is connected to the vertical rod holding it in place not on top of the tv camera pod like other wind tunnels but inside the cockpit

Is that Ferraris own Wind tunnel tho ? I mean are the shots zoomed or is it 100% ?

[beelsebob]

Is that Ferraris own Wind tunnel tho ? I mean are the shots zoomed or is it 100% ?

Why would the shots need to be zoomed in any way? A 60% model will look exactly like a 100% model (assuming it's accurate enough), with no context for the size of things.