Outerra forum
User mods, screenshots & videos => Aircraft => Flight Model Development => Topic started by: bomber on March 21, 2015, 03:36:16 pm
-
Following on from completing the weight ballance file, we now need to work on co-efficients for the airfoils...
this is the primary tool for determining the Co-efficients
http://www.mh-aerotools.de/airfoils/javafoil.htm
ok right click 'save file as' the following working out spreadsheet for the Ask-13 which we'll modify.
http://target4today.com/_posted_files_/Cesna_172/Airfoils-Cesna_172.xls
-
Now using the javafoil app you can define your airfoil co-ords within the geometry tab...
(http://target4today.com/_posted_files_/Cesna_172/javafoil-1.JPG)
or
you can do a search on this web site
http://airfoiltools.com/airfoiltools/search/index?m%5Bgrp%5D=naca4d&m%5Bsort%5D=1&MAirfoilSearchForm%5BtextSearch%5D=2412&MAirfoilSearchForm%5BmaxThickness%5D=&MAirfoilSearchForm%5BminThickness%5D=&MAirfoilSearchForm%5BmaxCamber%5D=&MAirfoilSearchForm%5BminCamber%5D=&MAirfoilSearchForm%5Bgrp%5D=naca4d&MAirfoilSearchForm%5Bsort%5D=1&yt0=Search
and copy n past the airfoil data (far right link) into the co-ords window of javafoil and then 'update view'
(http://target4today.com/_posted_files_/Cesna_172/javafoil-2.JPG)
Next we need to work out a reynolds number to calculate the co-efficients at..
cord = 1.66m
cruise speed = 226km/p = 62.7m/s
Reynolds number = 7285740 (use the spreadsheet)
Within the polar tab we need to add the reynolds number and the range of AoA we wish to work in -30 to 30 degs and analyse it.
(http://target4today.com/_posted_files_/Cesna_172/javafoil-3.JPG)
Going into the reynolds number tab we do a copy n paste the data to the spreedsheet.
(http://target4today.com/_posted_files_/Cesna_172/javafoil-4.JPG)
The above is the basic procedure for getting the co-efficient of lift for an airfoil shape without any control surface deflection
But we need to perform the same procedure but with control surface deflection...
Assumption - I can't find any angles of deflection for alierons so we'll take a punt at 15 down, 30 up..... flaps are full deflection at 40 degs.
Within the modify tab we adjust the flap deflection (don't get hung up on the word flap) first to -30 [press enter], go back to the polar tab, analyse, copy n paste data to the spreadsheet....
(http://target4today.com/_posted_files_/Cesna_172/javafoil-5.JPG)
Then back to the modify tab, undo, 15 [press enter] go back to the polar tab, analyse, copy n paste data to the spreadsheet....
(http://target4today.com/_posted_files_/Cesna_172/javafoil-6.JPG)
HOWEVER
we don't work in co-efficients of lift but instead co-efficients of normal.... so within the spreedsheet we take the Cl's and convert them creating a new table to be used within the JSBsim files...
I've done that and updated the spreedsheet...
http://target4today.com/_posted_files_/Cesna_172/Airfoils-Cesna_172.xls
-
So once you've created all that data for wings, h-stabs and v-stabs you copy n paste the tables from -180 degs to 180 degs into the co-efficients file
http://target4today.com/_posted_files_/Cesna_172/co-efficients.xml
And that's the aerodynamics part done with... and now the engine and prop work begins.
Regards
Simon
(p.s. they have me back under a tube train next week, so study the above during the week and we'll start on the engine and prop when I get back from London)
-
One little question ... it takes the airfoil in what part of the wing ? ... i mean, its not constant along the lenght of the wing, so is it a too big problem ? ... also, what about leading edges and non-basic flaps ? :
(http://www.allstar.fiu.edu/aero/images/fig24.gif)
... its not criticism, i just have no idea about that stuff.
-
I think I understand the first part of the question, the example is for a cesna 172, which does have the same airfoil along the length of its wing...
If you have a plane with differing airfoils along the wing you just have to perform the task for each airfoil.
Now the flaps issue is a bridge we have to cross when we get to it.... I've not done one, but it's only a change in airfoil shape per notch of flap, so should end up being a bit of rinse and repeat, copy n paste.
-
If you have a plane with differing airfoils along the wing you just have to perform the task for each airfoil.
...so a angled wing of a lets say Boing 737 plane would need almost each meter of the lenght to have a new data-set, whyle at physics there is some parameter saying how much of a lenght of the wing uses this or that set ?
-
The twisting of the wing along its length is covered in the front end.. The change in airfoil shape is covered here within the co-efficients section and is a matter of dividing each wing into 8 sections and determining the airfoil shape change along its length.
-
... its not criticism, i just have no idea about that stuff.
Challenge EVERYTHING.....
The assumptions and compromises we are forced into making need to be transparant, understood and agreed... that way when someone takes an existing flight model for modification into a similair plane there's a level of confidence as to the output.
-
Challenge EVERYTHING.....
... i will. :D
... i did just thought initially that the calculator maybe could support multiple closed splines, doe i then thought about our lessons from hydrodynamics ( i dont know how its called specifically in technical Eng, in svk its literally taken as liquid-dynamics, cause the pricnipe is in both l and g flow-substances states and were then framed both as liquids for this purpose ) and laminars-Reynolds stuff. Must be interesting math behind the calculator ...
-
just a simple engineer here who thinks along the lines of .... a force acting at a point.
I looked at your flap issues and hypothesised if the cessna was equipped with double slotted flap with leading edge slat.
Imported the 2412 data into Autocad and then modifed the wing to match a leading edge slate of -5 degs and double slots of 15 degs with an increase in cord length of 12% and then exported the point data into javafoil.
(http://target4today.com/_posted_files_/Cesna_172/javafoil-7.JPG)
I've updated the autocad and excel spreadsheet with an extra tab showing the increase in Cl for this hypothesis.
(http://target4today.com/_posted_files_/Cesna_172/javafoil-8.JPG)
regards
Simon
-
Ok have a look at the Co-efficients of normal force for the wing..
(http://target4today.com/_posted_files_/Cesna_172/javafoil-9.JPG)
I've red circled the 2 areas that don't quite look correct....
Now if you look at the sinusoidal tab in the spreadsheet you can see what the co-efficients would look like if the effects of an airfoil never existed
The flat plate tab shows how even though it's a flat plate the air sticks to it and creates a virtual airfoil, it's not got a great AoA before the airflow breaks away.
The important thing to notice is that once the airfoil has stalled the co-efficient falls to the sinusoidal value, not below it.
So within the areas circled you must copy n paste the sinusoidal values, to get something looking like this.
(http://target4today.com/_posted_files_/Cesna_172/javafoil-10.JPG)
-
A question you're likely to be faced with (if there's no data) is just how much control surface deflection should you have, is too much a waste?
And the answer is yes... look at this first image, it's the above image only I've circled the areas of interest... only if you're either above or below the sinusoidal line are you getting bang for bucks..
(http://target4today.com/_posted_files_/Cesna_172/javafoil-11.JPG)
If I say to myself how about having -40 degree upwards deflection, you can see in this image that the co-efficient doesn't go above the sinusoidal line, so as a pilot at the extreme AoA of 30 degree, with your wing well stalled out, you'll have no ability to unstall it and get it back generating useful lift..
(http://target4today.com/_posted_files_/Cesna_172/javafoil-12.JPG)
-
... i noticed, there is quite a hole between the wing and the flap where airflow is partially allowed to get trough without being much obstructed. Doe i found this little digital visualization an noticed how small that effect actually is in a presure profile :
http://cfd2012.com/types-of-flaps.html (http://cfd2012.com/types-of-flaps.html)
Also nice to see how much a difference are all those profiles.
-
yes a nice set of images... enlightening.
(http://target4today.com/_posted_files_/Cesna_172/javafoil-13.JPG)
This is the first test on the cessna using flightgear as a wind tunnel... we know that the lift 2 drag ratio for this plane is 9:1, yet the max we get at present is 6.36:1 at 2 degs.... so we need some adjustment here.. but how much, do we lower the drag or increase the lift, or even both ?
What we know is that the cesna 172 stalls at 57mph (flaps up)...and by looking at the graph above the max Cl is at 13 degs.... so if I put the plane back in the wind tunnel at 13 degs with a wind of 57mph the lift generated by the wing has to be equal to the weight of the plane, if not we adjust the modifier until it is.
-
Some more thought on this....
http://cessna.us/cessna-172/
This says that the zero lift drag is 0.0319, and the lift to drag of 7.5.
We have the zero lift at -2 degs with the drag being 0.039.... so only a small adjustment needed in drag modifier..
-
Oh more investigation has reveal this....
to generate a lift of 2300lb at 57mph (49.5kn)... the Cl needed is greater than any Cl prediction (in a reasonable reynolds number) by either javafoil or xfoil....and to top it all we know we have a twist in the wing of 1.5 deg at the root to -1.5 at the tip which make it even worse.
So is someone telling porkies here ?
-
So all I did was adjust the lift modifier to get the correct lift force at the correct speed, and the correct lift to drag of 7.5
(http://target4today.com/_posted_files_/Cesna_172/javafoil-14.JPG)
So anyone got an idea where all this extra lift that the cesna has over the ask-13 is coming from ?
Could it be fairy dust ?
Simon
-
ok adjusted the drag modifier to give a drag 0.0319 at zero lift...
POH says L/D is 9:1.... we get 8.28 at 10mph wind so I'm happy here
and we have a lift of 1.61 generating the required lbs force at the correct speed..
(http://target4today.com/_posted_files_/Cesna_172/javafoil-15.JPG)
-
This is the equivalent whole plane lift n drag for the Ask-13...
Anyone spot the huge elephant in this room ?
ASK-13
(http://target4today.com/_posted_files_/Cesna_172/javafoil-16.JPG)
-
Its just a tiny fox-hole on the hillside ... a proper trench position actually ... someone has a fine soldier on the loose.
-
http://www.aerospaceweb.org/question/aerodynamics/q0194.shtml
-
Nice article Uriah and validates the use of force normals, thank you very much.
However it doesn't tell the whole story with regards the 2 systems...
I've experimented with both, a year on one, a year on the other and then backwards and forth... and then it dawned on me...
When looking closely at the forces created by the airfoils, force normal system with it's reference to the airframe works best.
When looking closely at the drags, induced, skin and form then the lift/drag system with it reference to the wind frame works best.
However they're mutualy exclusive, you either use one of the other.... not both... even JSBsim manual says so...
And that's why I spent a few years going back and forth..... untill deciding to ignore it and use BOTH systems... because we have a computer not a piece of paper and slide rule.
Simon
-
Levi, I don't want to mess with the official 172, so can you knock up an otx, adding these files as per the ask_13...
don't forget to rename the Flight Damage Model file to the name of the plane...
Which reminds me, Devs can we please not have to do that...
For those that fly it whilst I'm down the pub tonight... it's a glide model to demonstrate the aerodynamics... the engine comes when we're happy with this...
Oh and I've not tested it, I like to live dangerously :)
http://target4today.com/_posted_files_/Cesna_172/Cesna_172.zip
-
Levi, I don't want to mess with the official 172, so can you knock up an otx, adding these files as per the ask_13...
don't forget to rename the Flight Damage Model file to the name of the plane...
Here you go: Download (http://www.mediafire.com/download/belyv2akq4ekazp/c172(New_FDM).Levi.otx)
The plane is standalone, and installs under my username.
FDM files are located within main package, as I usually do.
Oh and I've not tested it, I like to live dangerously :)
It works. Though, the steering on ground (when using rudder pedals) is inverted.
-
don't forget to rename the Flight Damage Model file to the name of the plane...
Which reminds me, Devs can we please not have to do that...
Oh, you can actually have any name for your FDM. You'll just have to write that name within the .objdef files, like this:
"physics" : "jsbsim/Flight-Damage-Model",
And the folder (c172) located in c172/FDM/aircraft, must have the same name as your FDM .xml file.
c172/FDM/aircraft/Flight-Damage-Model/Flight-Damage-Model.xml
But I still prefer having the FDM named the same as the aircraft...
-
Thanks Levi,
http://target4today.com/_posted_files_/Cesna_172/c172.xml
updated the fdm for the steering, for those interested the nose wheel had a -2 for steering and that's for tail draggers..
-
Here's the updated .OTX (link remains the same): c172(New_FDM).Levi.otx (http://www.mediafire.com/download/belyv2akq4ekazp/c172(New_FDM).Levi.otx)