Board index FlightGear Development Aircraft

JSBSim Piper J-3 Cub  Topic is solved

Questions and discussion about creating aircraft. Flight dynamics, 3d models, cockpits, systems, animation, textures.

Re: JSBSim Piper J-3 Cub

Postby david.megginson » Tue Sep 05, 2023 10:48 pm

Alant wrote in Tue Sep 05, 2023 9:11 am:If you read it, the report finds drag increases and lateral side force increases due to sideslip when large tyres are fitted. There is also a small increase in yaw moment due to sideslip. Roll moment due to sideslip is not affected.
The changes are presented using the same aerodynamic derivatives and coefficients that are used in the existing JSBSim model, so are simple to implement.

Two "fuselage" objects on each side would probably do it if you're modelling that level of detail.
David MEGGINSON (he/him)
I need more tea.
Scenery: FlightGear Americas Scenery
Fediverse: @david_megginson@mstdn.ca
User avatar
david.megginson
 
Posts: 561
Joined: Sat Dec 10, 2016 8:25 pm
Location: CYRO (Ottawa, Canada)
Pronouns: he/him
Version: next
OS: Linux Ubuntu

Re: JSBSim Piper J-3 Cub

Postby Alant » Sun Sep 10, 2023 11:12 pm

A quick report on my Datcom and OpenVSP investigation.
So far I made a comparison of the lift, drag and pitch moments of the Cub wing. As a reference I have used the wind tunnel tests in report https://ntrs.nasa.gov/citations/19930091301.
This includes a USA-35b wing section having an aspect ratio of 6 at a Reynolds number of 3470000. So I have set a simple Datcom and OpenVSP model of this airfoil at those conditions.
OpenVSP offers 2 methods. The simple one is a degenerative geometry in which the wing is represented as a thin plate having the same curvature as the airfoil . This is quite easy to set up.
The second method is a panel method where the wing surfaces are split into small panels and the effects of the air between one panel and its neighbours is calculated. This method requires more care when setting it up, and is slower to run.
Here are the results:-
Image
Now that I am happy with these results I can run the same process on the full aircraft. Fortunately I already have OpenVSP and Datcom models prepared, They will probably require some checking and tuning.
Alan
Alant
 
Posts: 1209
Joined: Wed Jun 23, 2010 6:58 am
Location: Portugal
Callsign: Tarnish99
Version: from Git
OS: Windows 10

Re: JSBSim Piper J-3 Cub

Postby TheEagle » Mon Sep 11, 2023 9:02 am

Well … all three (VSP Panel, VSP Degen and Datcom are failing miserably on the post-stall region … :(
Cessna 210 (Wiki)
My other aircraft: my wiki profile !
Other: FGTools (GitHub)
World tour: View on SkyVector
User avatar
TheEagle
 
Posts: 3391
Joined: Sat May 01, 2021 3:27 pm
Location: France
Pronouns: You, he
Callsign: F-EAGLE
IRC name: none
Version: Git next
OS: Ubuntu Studio 22.04

Re: JSBSim Piper J-3 Cub

Postby dany93 » Mon Sep 11, 2023 11:10 am

Hi,

Thank you Alant.

It's not a mere coincidence, I was on this issue too.

A few benchmark values:
(calculated with standard atmosphere ro = 1.225 kg/m3)
The stall speed musy be calculated with flaps for the PA18. In the FDM, thh flap CL is added to the "clean" wings. Addition +0.3875 for 50 deg flaps.
J3Cub 65 hp, 1200 lbs: stall speed 33 kts ==> CL = 1.82
PA18 95 hp, 1500 lbs: stall speed 37 kts ==> CL = 1.81 (1.42 clean if +0.39 flaps)
PA18 150 hp, 1750 lbs: stall speed 37 kts ==> CL = 2.12 (1.73 clean if +0.39 flaps)

=> CL = 1.82 clean would be enough for all configurations. Hovever, this value is not reached by the published measurement curves that I see for the USA35B (I see 1.37 - 1.38 for the NACA TR curves). And a high CL max value (calculations... possibly for a 2D profile) also means a high value at 0 AoA ("nose down" issue).

https://ntrs.nasa.gov/api/citations/199 ... 091301.pdf
https://www.abbottaerospace.com/downloa ... -sections/
https://digital.library.unt.edu/ark:/67 ... 091401.pdf
http://airfoiltools.com/polar/details?p ... il-1000000

Calculations by taking max CL= 1.38 for the wings (from NACA USA35B curves),
1200 lbs: Vstall = 37.9 kt (no flaps). J3Cub
CL = 1.38 + 0.39 = 1.77 (with flaps): PA18
1500 lbs: Vstall = 37.4 kt
1750 lbs: Vstall = 40.5 kt
________________________________________

To improve the possibility to reach the pitch angle at stall (3 point landing needs 11 to 13°):

I also tried by changing the CoG location in the FDM.
FDM, empty CG: 15.75 instead of 13.80
Code: Select all
        <location name="CG" unit="IN">
            <x>    15.75 </x> <!-- ini 13.80 -->
            <y>     0.00 </y>
            <z>   -23.23 </z>
        </location>

The empty CoG can be between 23 and 28% MAC. 13.8" is at 22%. 15.75" is at 25%.
https://scindeks-clanci.ceon.rs/data/pd ... 01031Q.pdf

I also smoothed the current CL curve close to the stall Aoa (14 to 16 deg). Weak effect, I think.

I combined the CoG change with the NACA 331 / 352 lift curve (max CL = 1.374 clean).

It partially improves the 3 point landing possibility, but partially only. It depends on the variant, a bit of throttle helps to get the pitch angle prior to touchdown.

I can upload one or two branches with this if you wish to try or pick some values, but no miracle. The 3 point landing is sometimes possible, bu tricky.
dany93
 
Posts: 861
Joined: Mon Sep 07, 2009 4:43 pm
Location: France (Paris region)
Version: 2020.4.0
OS: Linux Mint 20.3

Re: JSBSim Piper J-3 Cub

Postby Ysop » Mon Sep 11, 2023 11:28 am

@Alan: Looks like very sound work! Setting up the files correctly seems to be a bit of an art. Some principles are known, but for me there are still some mysteries.

DATCOM also assumes rigid structure, doesn't it?

Meanwhile I was experimenting a bit with Carlson pressure correction. Not better than cl_max. At least it gives a start.

@Eagle: Told you so...(sorry!). You might want to remember our CD0 discussion as well.

@dany: Excellent! Two quick thoughts:
-Elevator and fuselage have lift components as well. This could add up to 1.8x.
-Report 233 seems to be untwisted wing.

Deeper Looks when at home
User avatar
Ysop
 
Posts: 1327
Joined: Thu Oct 25, 2018 10:06 pm
Version: 2020.3.18
OS: ubuntu 22.04

Re: JSBSim Piper J-3 Cub

Postby Alant » Mon Sep 11, 2023 11:45 am

Thanks. We seem to be reading from the same hymn book. Thanks for the additional references.
At the moment I am concentrating most on flight in the normal flight envelope. Datcom, wind tunnel and VSPAero all are in good agreement in the +/- 5 deg incidence range. After that they are still reasonable up to over 10 degrees. I do not see that we should put too much emphasis on precise stall behaviour - normal flight is much more important. As long as the stall behaviour is more or less correct we should be OK. Look at it like driving a car - all cars can skid, but skidding is not part of normal driving training or the driving exam. We are normally advised to avoid skidding. and the same applies to stalls and spins in aircraft. Flying training (mine did) includes "recovery from unusual attitudes" and how to recognize the onset of a stall, but most of it is about safe flying.
Alant
 
Posts: 1209
Joined: Wed Jun 23, 2010 6:58 am
Location: Portugal
Callsign: Tarnish99
Version: from Git
OS: Windows 10

Re: JSBSim Piper J-3 Cub

Postby Alant » Mon Sep 11, 2023 11:48 am

Ysop - yes it is a bit of an art.
All of my work is with rigid structure aircraft. Calculating aeroelastic effects is another subject, and even less published data is available.
Alant
 
Posts: 1209
Joined: Wed Jun 23, 2010 6:58 am
Location: Portugal
Callsign: Tarnish99
Version: from Git
OS: Windows 10

Re: JSBSim Piper J-3 Cub

Postby dany93 » Mon Sep 11, 2023 1:23 pm

Alant wrote:At the moment I am concentrating most on flight in the normal flight envelope. (...). I do not see that we should put too much emphasis on precise stall behaviour - normal flight is much more important. As long as the stall behaviour is more or less correct we should be OK

I agree. The post-stall behaviour is not the main issue. All the more that the spin departure is not implemented.
However, the 3-point landing is an issue. It is important for short landing (e.g. in the bush). And slats will be useless otherwise. This point is one of the most interesting features of this aircraft.

In order to be able to do a decent 3-point touchdown, we must have the possibility of holding the aircraft during a short time at 11 to 13 deg (at least) pitch up angle grazing the ground with a weak descent rate. Reminder: it means 13 to 15 deg wing AoA.
dany93
 
Posts: 861
Joined: Mon Sep 07, 2009 4:43 pm
Location: France (Paris region)
Version: 2020.4.0
OS: Linux Mint 20.3

Re: JSBSim Piper J-3 Cub

Postby Alant » Mon Sep 11, 2023 2:59 pm

The perfect 3 point landing is the holy grail for tail dragger pilots. Just hold the aircraft so that all three wheels are the same height above the ground, close throttle and hold that pitch attitude until speed reduces and there is not enough lift to support the aircraft any more. Sounds easy when put like that ;)
We to be able to pitch the aircraft to that attitude at low airspeed. Another thing to check in my experiments.
Alant
 
Posts: 1209
Joined: Wed Jun 23, 2010 6:58 am
Location: Portugal
Callsign: Tarnish99
Version: from Git
OS: Windows 10

Re: JSBSim Piper J-3 Cub

Postby dany93 » Mon Sep 11, 2023 5:54 pm

Ysop wrote:-Elevator and fuselage have lift components as well. This could add up to 1.8x.
-Report 233 seems to be untwisted wing.

- Elevator: I see.. Also the horizontal stab. I've already started this. Their combined contributions should be weak concerning the effect on the stall airspeed.
Fuselage: Good idea. Can you give a "CL vs alpha due to fuselage" contribution which could be added to the FDM lift axis? If possible for positive and negative alpha values. A table or a linear slope (for +/- 10 or 20 deg?).

- Untwisted wing for Report 233 : yes, it's almost sure.
Twisted wing effect might be approximated by extending the CL vs alpha curve at a constant value beyond or on each side of the stall angle. By how much? How much is the twist angle? (1 to 2 deg?).
Same issue, currently: the aircraft must be able reach and hold the "beyond the stall angle" pitch up.
dany93
 
Posts: 861
Joined: Mon Sep 07, 2009 4:43 pm
Location: France (Paris region)
Version: 2020.4.0
OS: Linux Mint 20.3

Re: JSBSim Piper J-3 Cub

Postby S&J » Mon Sep 11, 2023 6:09 pm

I'd look at the way ground effect is modelled.

I'm not convinced it's correct
"Stay away from negative people.They have a problem for every solution." - Albert Einstein
S&J
 
Posts: 769
Joined: Wed Aug 26, 2020 7:31 pm

Re: JSBSim Piper J-3 Cub

Postby Ysop » Mon Sep 11, 2023 8:06 pm

Fuselage: OpenVSP just makes a NACA0015 out of it. DATCOM will certainly give better approximation.
Wing twist and stall: Considering it is not a distinct point of stall, but a stretched zone due to wing twist, it should be half the twist angle to each side of the stall point. VERY theoretical, but at least an approach. Exact value of twist is in Alan's model. 2° or so?

It might be a bit of trial and error for the 3-pointer smoothness.

Just working on something which is not interfering with the other work packages. Hopefully it ends up in a contribution.

edit: wording and last sentence.
User avatar
Ysop
 
Posts: 1327
Joined: Thu Oct 25, 2018 10:06 pm
Version: 2020.3.18
OS: ubuntu 22.04

Re: JSBSim Piper J-3 Cub

Postby S&J » Mon Sep 11, 2023 8:47 pm

Create a wing with multiple panels, run it through the range of AoA in a virtual wind tunnel recording the cl and cd ?

Of course it's still very rough as the AoA along the wing varies with roll rate and U velocity and circulation.
"Stay away from negative people.They have a problem for every solution." - Albert Einstein
S&J
 
Posts: 769
Joined: Wed Aug 26, 2020 7:31 pm

Re: JSBSim Piper J-3 Cub

Postby dany93 » Thu Sep 14, 2023 10:25 am

Hi,

Hopefully for an improvement and after the discussions and suggestions in the forum, I tried and propose these changes:
- Change the CoG location (a bit aft).
- Change the current wing lift curve (close to Xfoil prediction, 2D) for a measured one from NACA Technical Reports. Slightly extended and smoothed close to the stall angle to take into account for the wing twist.
- Add lift due to horizontal stab incidence
- Add lift contribution from the fuselage (Ysop suggestion).

I've created a branch on GitHub https://github.com/wlbragg/J3Cub/tree/Issue-8

For more explanations and details, see https://github.com/wlbragg/J3Cub/issues/8

Dany
dany93
 
Posts: 861
Joined: Mon Sep 07, 2009 4:43 pm
Location: France (Paris region)
Version: 2020.4.0
OS: Linux Mint 20.3

Re: JSBSim Piper J-3 Cub

Postby Alant » Thu Sep 14, 2023 10:57 am

I already have most of these points addressed in my current Datcom/OpenVSP work. This is being backed up by a few flight cases run in OpenFoam and any published aero data that comes to light. I do not think that is a good idea to arbitrarily change these coefficients just to make the aircraft feel better. I leave that approach to Yasim aficionados.
Tail lift is already included in the Datcom/OpenVSP results. However the geometric model used in both of these methods is very much simplified due to constraints in the modelling facilities that each method offers. I am revisiting all such problem areas in my previous work.
One major problem that I have is the inconsistency of wing root incidence in the various Cub original drawings. In many of them it appears that the lower wing surface, which is almost flat, is set at 0 degrees incidence. The incidence datum is (according to the aircraft rigging instructions) the top of the cockpit door). However other drawings show the incidence as 2.5 degrees, which is actually specified on one of them.
I have weight and balance forms for the PA11 and PA12 cub and have recently found one for the J3. [url]file:///D:/fg/AircraftDev/PiperJ3Cub/Drawings/J-3%20weight%20&%20balance%20calculation%20_%20specs%20for%20J3C-65%20engine.htm[/url]
Alant
 
Posts: 1209
Joined: Wed Jun 23, 2010 6:58 am
Location: Portugal
Callsign: Tarnish99
Version: from Git
OS: Windows 10

PreviousNext

Return to Aircraft

Who is online

Users browsing this forum: No registered users and 1 guest