HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Feb 15, 2013, 12:27 PM
Registered User
yomgui's Avatar
France, RA, Cormaranche-en-Bugey
Joined Feb 2011
520 Posts
Discussion
Do vibrations increase drag ?

Hello,

This question came to me while testing some shapes in my homemade wind tunnel, and having results for drag far above theory.

The wind tunnel is not perfect and there is always more or less vibrations.

Can it have an effect on transition and separation leading to more drag ?

And what about in flight drag ?

Guillaume
yomgui is offline Find More Posts by yomgui
Reply With Quote
Sign up now
to remove ads between posts
Old Feb 15, 2013, 01:15 PM
The Great Filter
kcaldwel's Avatar
Joined Jan 2007
3,524 Posts
I suspect your problem isn't so much the vibration as it is turbulence. It is very difficult to get low levels of turbulence in a wind tunnel. If the small scale turbulence level is high, you will have completely turbulent boundary layers from the leading edge of your shapes. This will give higher drag.

Vibrations can cause separation, or keep flow attached. The frequency and amplitude make a difference:

http://ntrs.nasa.gov/archive/nasa/ca...1991002362.pdf

There are lots of papers on using acoustic frequency vibrations to control laminar separation bubble formation, etc.

Kevin
kcaldwel is offline Find More Posts by kcaldwel
Reply With Quote
Old Feb 15, 2013, 02:22 PM
Ascended Master
Sparky Paul's Avatar
Palmdale, CA
Joined Oct 2000
13,353 Posts
Amazing you should mention those acoustic studies, Kevin.
Our senior thesis for the Aero degree involved using a low-speed tunnel and playing various frequencies of noise into the wind stream, and measuring the lift and drag on a sensitive balance.
ISTR we got some useable data at 400 hz.
Sparky Paul is offline Find More Posts by Sparky Paul
RCG Plus Member
Old Feb 15, 2013, 03:06 PM
B for Bruce
BMatthews's Avatar
The 'Wack, BC, Canada
Joined Oct 2002
11,170 Posts
If you're serious about using your own wind tunnel to do this sort of work you should really find and read a copy of Soartech 9. It's the one which was the write up of the thesis study done by Michel Selig and his work on low speed airfoils. A rather bulky part of the thesis was all about efforts taken to minimize and measure the turbulence of the tunnel just due to the fact that at low Reynolds numbers the airflow in the tunnel has to be even smoother than for higher Rn work.
BMatthews is online now Find More Posts by BMatthews
Reply With Quote
Old Feb 15, 2013, 04:02 PM
Registered User
yomgui's Avatar
France, RA, Cormaranche-en-Bugey
Joined Feb 2011
520 Posts
Thanks all for the good links and comments !

Maybe I got obsessed by vibrations because I simply can see it, contrary to turbulence. Though obviously turbulence level should be quite high.

I built the wind tunnel for propeller testing but recently get involved in some drag studies. Very simple experiments about drag of various shapes, fuselage modifications, FPV camera or antennas, etc.

I've tested cylinder (antennas) and vertical flat plate with good correlation to published data (a bit higher drag coeff but I think I under estimate speed). But for streamlined shapes... I commonly find a drag coefficient higher by 50 to 100 % more than what one should expect from published data, drag built-up methods, etc.

Another thing that puzzle me is that in-flight data seems to match quite well with a 50 % to 100 % higher drag coefficient than theory (from propeller thrust at given speed and rpm, and steady flight speed, rate of climb).
yomgui is offline Find More Posts by yomgui
Reply With Quote
Old Feb 15, 2013, 05:38 PM
The Great Filter
kcaldwel's Avatar
Joined Jan 2007
3,524 Posts
There doesn't seem to be a very big discrepancy between XFOIL, good wind tunnel test data (Selig, etc.) and free air results in sailplanes. Interference drag, and the parasitic drag of things like control linkages, etc. are always hard to quantify exactly.

It would be interesting to know why you are seeing a 50 to 100% difference in drag. When you say "theory", are you sure you are capturing all the drag sources at the right Re? It is also very difficult to get good free-air results.

Kevin
kcaldwel is offline Find More Posts by kcaldwel
Reply With Quote
Old Feb 15, 2013, 08:56 PM
It's all fun till Crunch!
Capt. Crunch's Avatar
InSane Diego
Joined Jan 2002
1,940 Posts
Yes, any vibrations or flutter will degrade performance and ultimately lead to structural failure. Balance all rotating thingys and stiffen all flight surfaces so no movement is possible.
From some one that's been there.
Capt. Crunch is offline Find More Posts by Capt. Crunch
Reply With Quote
Old Feb 15, 2013, 09:36 PM
Registered User
Joined Jan 2008
1,148 Posts
I have read that some who fly the international class power free flight airplanes use side mounted engines to reduce drag. The idea being that a vertically mounted engine would vibrate up and down and effectively increase the wing thickness during the climb. I don't know if this is widely taken seriously.
Jim Thomerson is offline Find More Posts by Jim Thomerson
Reply With Quote
Old Feb 16, 2013, 01:29 AM
Registered User
yomgui's Avatar
France, RA, Cormaranche-en-Bugey
Joined Feb 2011
520 Posts
Quote:
Originally Posted by kcaldwel View Post
There doesn't seem to be a very big discrepancy between XFOIL, good wind tunnel test data (Selig, etc.) and free air results in sailplanes. Interference drag, and the parasitic drag of things like control linkages, etc. are always hard to quantify exactly.

It would be interesting to know why you are seeing a 50 to 100% difference in drag. When you say "theory", are you sure you are capturing all the drag sources at the right Re? It is also very difficult to get good free-air results.

Kevin
I use the drag built-up method found in Aircraft design: a conceptual approach by Raymer. It gives a very good answer finding CD min for the case of the Jibe2 tested by M.Scherrer (http://sailplane-matscherrer.blogspo...nd-tunnel.html).

For my wind tunnel tests of a streamlined shape I also use this method and compared it with published data for streamlined bodies... and I typically find values around 180 % of those theoretical values. So this could be the result of high turbulence in the wind tunnel.

When comparing the built-up to in-flight data, I've taken into account the antennas, the cameras, which I know the drag from my wind tunnel tests. I have no idea how to estimate the interference drag... could be up to 20 % or more of the overall CD min...?

The fuselage, wings and stab shapes are not perfect (motor mount in front, control surface leakage, thick trailing edge, air intakes, etc...), also drag can be majored by the propeller's thrust, so does CL max by the way.

In-flight data are taken from an automatic selection of the eagle tree data (those with small variations in course, altitude, speed and rpm for a few seconds), from a flight in a very calm weather, speed from GPS, checked by video, flying perpendicular to wind (less than 5 km/h). Drag is then seen as thrust, found with UIUC data (I can see a very very good correlation between theory and in-flight data for power, so I suppose the same for thrust). Of course the margin of error is wide.

The shape of the curve CL / CD follows very well the theory found by following the methods from Aircraft design: a conceptual approach (computing 3-D values from 2-D).

All is very fine, except I have to enter a huge factor to reach the actual seen CD min, what I don't need to reach the jibe 2 wind tunnel data.

All the difficulties of predicting drag of a draggy plane ??

BTW: I have no very noticeable vibrations in-flight, only in the wind tunnel.
yomgui is offline Find More Posts by yomgui
Last edited by yomgui; Feb 16, 2013 at 01:36 AM.
Reply With Quote
Old Feb 16, 2013, 10:13 AM
The Great Filter
kcaldwel's Avatar
Joined Jan 2007
3,524 Posts
Quote:
Originally Posted by yomgui View Post

For my wind tunnel tests of a streamlined shape I also use this method and compared it with published data for streamlined bodies... and I typically find values around 180 % of those theoretical values. So this could be the result of high turbulence in the wind tunnel.
180% seems much more than I would expect just from a high turbulence level. It seems there must be something else going on as well. Interference drag might account for a chunk of it - it is very hard to estimate, and varies considerably with the shape and number of intersections - but there still seems to be a sizable discrepancy.

When you say "published data for streamlined bodies", is the data for the correct Re?

I suppose the easiest way to get a reasonable Cd0 would be to see what the terminal velocity is in a vertical dive. It would be nice to do it without a propeller though, which would mean towing or other means of getting aloft. That would still only give you Cd0 at a pretty high Re for the model, and the lower speed Cd0 could be higher due to Re effects.

Kevin
kcaldwel is offline Find More Posts by kcaldwel
Reply With Quote
Old Feb 16, 2013, 02:46 PM
Registered User
yomgui's Avatar
France, RA, Cormaranche-en-Bugey
Joined Feb 2011
520 Posts
Thanks Kevin, well the published data are for Re = 10^7, so nothing to do with our models, but since the drag built-up method was fitting well with the data, I assumed it would do the same for Re = 5*10^5. That's maybe a mistake !

I've found the data in Aerodynamics, aeronautics and flight mechanics by McCormick (I've found it difficult to find other data). It gives CD = 0.65 or 0.66 (according to frontal area) for a body of L/D = 6.7, the built-up method gives CD = 0.67 at Re = 10^7, so quite the same. At Re = 5*10^5, it gives 0.123 and I find 0.209 (more or less a lot !), 70 % more in fact...

I've also tested a symmetrical piece of airfoil (15 cm span, 18 cm chord, around 10 % thickness), and compared the results with XFLR5, 3D VLM, using various Ncrit.
Wind tunnel tests gave CD0 = 0.0232 (according to wing area).
Ncrit = 9 gives CD0 = 0.0112
Ncrit = 1 gives CD0 = 0.0123
Ncrit = 1 and 10 % chord forced transition gives CD0 = 0.0154.

A fast oil test suggested that transition might indeed be around 10 % chord, maybe 20 %, maybe there was not enough oil, or too viscous oil... anyway that's 50 % to 100 % higher CD0 than theory...

I'm afraid I can't find drag at terminal velocity without a windmilling propeller !
yomgui is offline Find More Posts by yomgui
Reply With Quote
Old Feb 16, 2013, 06:36 PM
The Great Filter
kcaldwel's Avatar
Joined Jan 2007
3,524 Posts
Not much of an airfoil if the transition is at 10 to 20%, especially at low Re! A good low Re airfoil shape has a transition point very far back, sometimes 75% to 100%, and there would be a large difference in drag between Ncrit = 1 and 9. With the transition so far forward on your example, you aren't seeing the effect of the turbulent flow.

XFoil usually under-predicts the drag a bit, and I'm not sure howXFLR5 would cope with such a low aspect ratio "wing". XFLR5 doesn't model the flow near the tips of a wing well, and that is all you have. I'm not sure I'd trust the XFLR5 results very much for an aspect ratio below 5.

At Re = 10^5, you are right in the region for the critical Re. There is a big step in the drag at around that Re. I think this could explain part of the problem.

How are you measuring the drag forces at that Re? The forces are very small, and difficult to measure accurately.

I've attached a chart for an airfoil across the critical Re. It is from a German publication, so Ca is the lift coefficient, and Cw is the drag coefficient.

Kevin
kcaldwel is offline Find More Posts by kcaldwel
Reply With Quote
Old Feb 17, 2013, 11:38 AM
Registered User
yomgui's Avatar
France, RA, Cormaranche-en-Bugey
Joined Feb 2011
520 Posts
OK, I think I understand that I should not trust XFLR5 here, since it's very far from a glider wing.

I can't be sure about transition, I guess that maybe I hadn't enough oil... or too viscous one...

I was using this small wing in order to compare the drag of foam vs tape covering (rough vs smooth), showing no difference at all BTW.

I measure drag using the "force sensor" (not sure it's the good word) from a chinese precision scale. The wing is fixed to a mount via a carbon tube, drag of the wing is then supposed to be the overall drag minus the mount + carbon tube drag alone previously measured (so not taking account of any interference between wing and mount). Drag at max speed (~18 to 19 m/s) is around 16 or 18 g (I can't recall exactly, and don't have the numbers under my eyes right now), for either the streamlined body and the small wing, accuracy is not great, measurement varying by 1 or 2 g at worst at max speed.

Maybe the difference I'm seeing is also for a big part due to the shapes not being optimal, same for the plane BTW.

Pehaps if I build a more "rational" plane I won't see such a difference. Then I guess it's hard to predict the drag of a "non-optimal or close to it" plane, thinking of all the FPV gear, fast and easy building, motor mount, etc.
yomgui is offline Find More Posts by yomgui
Reply With Quote
Old Feb 17, 2013, 01:56 PM
Registered User
ShoeDLG's Avatar
Germany, BW, Stuttgart
Joined Mar 2012
754 Posts
One challenge in making accurate drag measurements is that the lift of the thing you are measuring is often many times greater than the drag. If you're not VERY careful to isolate the drag component of the total force, your answer will be contaminated by "crosstalk" from the lift. Measuring even a small fraction of the lift as drag will cause big errors.
ShoeDLG is online now Find More Posts by ShoeDLG
Last edited by ShoeDLG; Feb 17, 2013 at 02:16 PM.
Reply With Quote
Old Feb 18, 2013, 02:11 PM
Registered User
yomgui's Avatar
France, RA, Cormaranche-en-Bugey
Joined Feb 2011
520 Posts
Yes, that's true, I have the drag force directly applied on the sensor, so the lift component should only slightly interfere as long as it's not in the sensor axis.

In fact, for the moment I think I just haven't test a good and nice shape that can be compared with other sources. Maybe I should start with a sphere and a flat plate !
yomgui is offline Find More Posts by yomgui
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion How do I stop Motor vibration? actioncircus Power Systems 4 Feb 13, 2013 07:52 AM
How do you calculate the drag of a simple object Mat Modeling Science 11 Nov 25, 2004 01:52 AM
Logo 10 - How to increas flight time? test3 Electric Heli Talk 5 Jul 11, 2003 01:28 AM
Anyone fly near Bandimere Speedway in Denver? We will be doing some drag racing there Duane Electric Plane Talk 0 Jul 28, 2002 09:34 PM
How do you characterize an aircraft for sleekness and low drag? KOMET 44 Power Systems 3 Dec 17, 2001 06:02 PM