bellancacf From United States of America, joined May 2011, 115 posts, RR: 0 Posted (10 months 2 days 2 hours ago) and read 2444 times:
There seem to be plenty of interactive airfoil analysis programs online that show upper and lower airfoil surface pressure profiles as a function of alpha, but I was hoping to find one that would show the vector for the force resulting from placing an airfoil into a free stream: I'd like to be able to vary alpha and the airfoil. Extra points for showing the normal to the chord line so that there's a meaningful visual reference as the incidence and the direction of the resultant change.
Is there a program out there that displays actual data obtained with physical airfoil models?
And out of curiosity -- not having searched for the answer -- how does one go about collecting such data?
tdscanuck From Canada, joined Jan 2006, 12709 posts, RR: 81 Reply 1, posted (10 months 2 days 1 hour ago) and read 2439 times:
Quoting bellancacf (Thread starter): I was hoping to find one that would show the vector for the force resulting from placing an airfoil into a free stream:
I'm not sure if there's one that actually draws the force vector, but there are several that give you the Lift/Drag coefficients so you can back out what the force vector must be. NASA has a bunch of fun free stuff: http://www.grc.nasa.gov/WWW/k-12/freesoftware_page.htm
Quoting bellancacf (Thread starter): Is there a program out there that displays actual data obtained with physical airfoil models?
Those would be more databases than programs. University of Illinois maintains a wonderful site that has 1500+ airfoils and links to a bunch of solvers that can give you the theoretical data (which is very good for many of them). For well established famlies like the NACA airfoils the actual physical data is also available: http://naca.larc.nasa.gov/search.jsp...976&qs=N%3D4294868711%2B4294448803
bellancacf From United States of America, joined May 2011, 115 posts, RR: 0 Reply 2, posted (10 months 2 days ago) and read 2432 times:
re tdscanuck@1: Wind tunnels. Lots and lots of wind tunnels.
Thanks for that info, Tom. Maybe I'll stumble across something in Python and, oh, Qt that someone doesn't mind sharing. That would be a good motivator to learn something about Python and graphics.
But what's been wrinkling my brow is whether it's done with levers and weights, for instance, or force transducers or what and in what geometry. I'd be lucky to come up with what the Wright brothers did in their shop if I tried to do it myself.
Thanks for that. I've downloaded the files and will take a look. I may be able to exhume my measurements of Bellanca's airfoil for the C.F -- a strange, extraordinary, undercambered, reflexed beast -- and feed that in.
BMI727 From United States of America, joined Feb 2009, 14315 posts, RR: 26 Reply 5, posted (10 months 2 days ago) and read 2417 times:
Quoting bellancacf (Reply 4): I may be able to exhume my measurements of Bellanca's airfoil for the C.F -- a strange, extraordinary, undercambered, reflexed beast -- and feed that in.
If you know the section you can use the UIUC airfoil database to download the nondimensionalised coordinates and feed that into the program.
Why do Aerospace Engineering students have to turn things in on time?
tdscanuck From Canada, joined Jan 2006, 12709 posts, RR: 81 Reply 6, posted (10 months 1 day 23 hours ago) and read 2411 times:
Quoting bellancacf (Reply 2): But what's been wrinkling my brow is whether it's done with levers and weights, for instance, or force transducers or what and in what geometry.
Sorry, I didn't realize you were going for the technical details.
For 2D airfoils, you can get excellent drag data without any force measurement at all; you just run a solid airfoil all the way across the tunnel and then use a pressure rake (a big array of pitot tubes that looks like a garden rake) to measure the momentum deficit in the wake across the tunnel.
For lift and moment measurements, you use some kind of force measuring device. The device is called the "balance" in wind tunnel parlance. You mount your test specimen to the balance, fire up the tunnel, and directly measure the forces and torques. Good balances have adjustable joints so you can vary angle of attack. Modern balances are all finely calibrated multi-axis load cells. Here's an example: http://www.mmtool.com/balances.html
There's *ton* of calibration and correction, especially on transsonic tunnels, to turn the raw balance data into actual coefficients.
bellancacf From United States of America, joined May 2011, 115 posts, RR: 0 Reply 7, posted (10 months 1 day 22 hours ago) and read 2404 times:
re tdscanuck@6:
Tom, thanks a lot for taking your time to answer in such detail. That mmtool balance reminds me of instances when, to get something done in the lab, I would have to chase deeper and deeper into a corporation, finally arriving at, almost always, a single person who knew how to do it and in fact usually had pretty much done it all by himself. (Males in the cases I remember, but over in electron microscopy, there was a good chance it'd be one devoted, skilled woman.) I used to wander out into the office saying, "Technology is so d**ned THIN! Wherever you look, it comes down to just one person." So blasted fragile, too, and so hard to re-establish or reinvent once the chain is broken.
Just to add to Tom's post. You can also get lift, pressure drag and moment coefficients by measuring the static pressure distribution about the airfoil and integrating the results with respect to the chord (for lift) and "height" (for pressure drag) of the airfoil. This method does not capture the skin friction component of drag. To get the total drag of a 2D airfoil (pressure and skin friction) you need to use a wake rake as mentioned by Tom.
The benefit of the static pressure method is that you can actually see the shape of the pressure distribution about the airfoil, and pick out such features as the location of separation. A balance once setup however, is generally much quicker if you are just after the force coefficients.
Quoting tdscanuck (Reply 6): There's *ton* of calibration and correction, especially on transsonic tunnels, to turn the raw balance data into actual coefficients.
A whole science in itself!
Regards, JetMech
JetMech split the back of his pants. He can feel the wind in his hair.
tdscanuck From Canada, joined Jan 2006, 12709 posts, RR: 81 Reply 10, posted (10 months 1 day 22 hours ago) and read 2399 times:
Quoting jetmech (Reply 8): Quoting tdscanuck (Reply 6):
There's *ton* of calibration and correction, especially on transsonic tunnels, to turn the raw balance data into actual coefficients.
A whole science in itself!
A few months ago I found a book on clearance at the MIT bookstore that was a collection of NACA papers on transsonic wind tunnel design, with an introduction by Von Karman (yes, *that* Von Karman). My head nearly exploded.
bellancacf From United States of America, joined May 2011, 115 posts, RR: 0 Reply 11, posted (10 months 1 day 22 hours ago) and read 2391 times:
re JetMech@8: You can also get lift, pressure drag and moment coefficients by measuring the static pressure distribution about the airfoil and integrating the results with respect to the chord (for lift) and "height" (for pressure drag) of the airfoil.
I'd been wondering about that. I realized I didn't know whether I had to integrate over the chord length (a lot of identical delta-lengths) or take the slope of the airfoil surface into account (secant(slope-sub-i) * delta-length). How does that work? And the drag calculation? Is this, at each station, the height above and normal to the chord line of the airfoil surface regardless of alpha?
Sorry to ask this kind of question, but it's this or get lost in stuff I'm almost certain not to understand . . . :-]
tdscanuck From Canada, joined Jan 2006, 12709 posts, RR: 81 Reply 12, posted (10 months 1 day 14 hours ago) and read 2338 times:
Quoting bellancacf (Reply 11): I realized I didn't know whether I had to integrate over the chord length (a lot of identical delta-lengths) or take the slope of the airfoil surface into account (secant(slope-sub-i) * delta-length). How does that work?
You do need to take the slope of the surface into account, although you can do that at two different places.
If you're going for lift directly, you can just take the delta length at each station projected onto the line parallel to the free-stream; that will give you the component of pressure force normal to the freestream (definition of lift). Adding up all those little components will give you lift. You can do the same thing for drag by projecting the delta length onto a line normal to the free stream.
However, you can be more general and just stay in vectors. The pressure force on each delta length acts normal to the surface at that point. That gives you a bunch of tiny little vectors that you can just directly add to get the total force on the airfoil, at which point you can resolve that force into lift/drag or any other axis you want.
Quoting bellancacf (Reply 11): And the drag calculation? Is this, at each station, the height above and normal to the chord line of the airfoil surface regardless of alpha?
Height above will matter for moment calculations but not for drag; drag just cares about normal to the freestream (not the same as normal to the chordline). The chordline is often used as the reference zero for alpha measurement but it doesn't have to be. You need to know alpha to get the transformation from airfoil coordinates (what the pressure probes are measuring) to freestream coordinates.
Lift is calculated by plotting the non-dimensional static pressure (Cp) vs. the non-dimensional chordwise location (x/c) of the pressure tapping as below. This distance is non-dimensionalised by the chord length, so it varies from 0 to 1
The area inside the curve, which is found by integration, is the normal force coefficient (Cn), the normal force being that perpendicular to the chord line regardless of angle of attack. Pressure drag is calculated by plotting the non-dimensional static pressure (Cp) vs. the non-dimensional location above or below the chord line (y/c) of the pressure tapping as below. This distance is non-dimensionalised by the chord length, so it varies from 0 to some positive and negative non zero number depending on the section thickness
The area inside the curve, which is found by integration, is the tangential force coefficient (Ct), the tangential force being that parallel to the chord line regardless of angle of attack. My integration gives Cn as 0.4006 and Ct as -0.02732.
Once you have Cn and Ct, you resolve trigonometrically using the angle of attack (a) to get the lift (Cl) and pressure drag (Cdp) coefficients as below.
The pressure distributions shown were taken at an angle of attack of 4 degrees, thus using the above; I get Cl as 0.402 and Cdp as 0.001.
Anytime the pressure coefficient curve crosses over itself, such as the Cp vs. y/c example above, the alternate closed areas will make positive and negative contributions to either the normal or tangential force coefficient. Such crossing over occurs much more frequently for the Cp vs. y/c curve.
Regards, JetMech
JetMech split the back of his pants. He can feel the wind in his hair.
I thought it was an important subject! Do you know if it is possible to use adaptive wall wind tunnels for transonic testing? I get the felling it may be tricky achieving transonic velocities with the added complexity of the divergent / convergent shape of an adaptive wall.
Regards, JetMech
JetMech split the back of his pants. He can feel the wind in his hair.
tdscanuck From Canada, joined Jan 2006, 12709 posts, RR: 81 Reply 16, posted (9 months 4 weeks 1 day 12 hours ago) and read 2125 times:
Quoting jetmech (Reply 15): Do you know if it is possible to use adaptive wall wind tunnels for transonic testing? I get the felling it may be tricky achieving transonic velocities with the added complexity of the divergent / convergent shape of an adaptive wall.
Anything's possible but I think you'd need a really cool control system to keep the tunnel stable. And I don't even have the foggiest clue as to how to approach figuring out the tunnel corrections for a dynamically moving wall.
From what I understand, the inlets of many supersonic aircraft are designed to provide subsonic air to the face of the engine in the most efficient manner possible. The focus is thus on efficient deceleration. Concorde uses oblique shocks to slow down and compress the air prior to entering the engine. The trick is the fine adjustment of the intake ramps to achieve the desired shock pattern.
Adaptive wall wind tunnels - for 2 dimensional airfoil testing at least - is all about capturing a given streamline, and is focussed much less on acceleration or deceleration of the air velocity. The problem is, for a fixed angle of attack, adjustment of the walls to capture the desired streamline will as a side effect, change the velocity of the air as well, which then may go on to change the streamline pattern.
This may present additional problems for transonic testing; as changes in the adaptive wall could also change the locations where the flow field transitions from sub to supersonic and vice versa. Achieving and maintaining a desired test-point may involve much time consuming iteration.
Regards, JetMech
JetMech split the back of his pants. He can feel the wind in his hair.
Sponsor Message: 
From United States of America, joined May 2011, 115 posts, 
From Canada, joined Jan 2006, 12709 posts, 
From Australia, joined Mar 2006, 2586 posts, 
From Hong Kong, joined Dec 2006, 7725 posts, 
Sponsor Message: 
Printer friendly format
