Sanity check of NACA 0012 gives wrong coefficient of lift
Moderator: oliveroxtoby
Forum rules
and Helpful information for the FEM forum
and Helpful information for the FEM forum
Re: Sanity check of NACA 0012 gives wrong coefficient of lift
I confess, a couple of things are still unclear to me.
In the linked http://www.ae.metu.edu.tr/tuncer/ae443/ ... All-Re.pdf
They give the velocity as 68ft/sec which is 21 m/s
You have adjusted this to match the reynolds number to 33.4m/s ( due to the linked example using a pressurised wind tunnel ? )
How did you get from Re 3.12*10^6 to V=33.4 ?
If I want to solve a 3d Shape, do I calculate the reference area values based on the surface area of the shape ?
For example if I create wing with a 1.4m chord, and a 5m depth, is the surface calculation then 140cm x 500cm = 70000cm2 ?
I solved some values using your equations for pressure - I solved for 55m/s and 52m/s and 21m/s
Example solved by Thomas for 33.4 m/s
q = (1.2 kg/m^3) / 2 * (33.4 m/s)^2
= (1.2 / 2) * (33.4^2)
= 0.6 * 1115
= 669 N/m^2
Solve for 55 M/s
q = (1.2 kg/m^3) / 2 * (55 m/s)^2
= (1.2 / 2) * (55^2)
= 0.6 * 3025
= 1815 N/m^2
Solve for 21 M/S
q = (1.2 kg/m^3) / 2 * (21 m/s)^2
= (1.2 / 2) * (21^2)
= 0.6 * 441
= 264.6 N/m^2
Solve for 52 M/S
q = (1.2 kg/m^3) / 2 * (21 m/s)^2
= (1.2 / 2) * (21^2)
= 0.6 * 2704
= 1622 N/m^2
I did an example solve for a 40mm mesh
Solve for 40mm mesh
140cm x 4cm = 560cm2
I note the reference area rounds up the values.
In the linked http://www.ae.metu.edu.tr/tuncer/ae443/ ... All-Re.pdf
They give the velocity as 68ft/sec which is 21 m/s
You have adjusted this to match the reynolds number to 33.4m/s ( due to the linked example using a pressurised wind tunnel ? )
How did you get from Re 3.12*10^6 to V=33.4 ?
If I want to solve a 3d Shape, do I calculate the reference area values based on the surface area of the shape ?
For example if I create wing with a 1.4m chord, and a 5m depth, is the surface calculation then 140cm x 500cm = 70000cm2 ?
I solved some values using your equations for pressure - I solved for 55m/s and 52m/s and 21m/s
Example solved by Thomas for 33.4 m/s
q = (1.2 kg/m^3) / 2 * (33.4 m/s)^2
= (1.2 / 2) * (33.4^2)
= 0.6 * 1115
= 669 N/m^2
Solve for 55 M/s
q = (1.2 kg/m^3) / 2 * (55 m/s)^2
= (1.2 / 2) * (55^2)
= 0.6 * 3025
= 1815 N/m^2
Solve for 21 M/S
q = (1.2 kg/m^3) / 2 * (21 m/s)^2
= (1.2 / 2) * (21^2)
= 0.6 * 441
= 264.6 N/m^2
Solve for 52 M/S
q = (1.2 kg/m^3) / 2 * (21 m/s)^2
= (1.2 / 2) * (21^2)
= 0.6 * 2704
= 1622 N/m^2
I did an example solve for a 40mm mesh
Solve for 40mm mesh
140cm x 4cm = 560cm2
I note the reference area rounds up the values.
Re: Sanity check of NACA 0012 gives wrong coefficient of lift
-
- Veteran
- Posts: 3156
- Joined: Sat May 20, 2017 12:06 pm
- Location: Germany
Re: Sanity check of NACA 0012 gives wrong coefficient of lift
I used the chord-length as reference:
Thats the way I would do it.
Re: Sanity check of NACA 0012 gives wrong coefficient of lift
Thank you Thomas, that is a great explanation.
I tried to get the NACA0012 aero foil simulation working to see if it would give me different results to the 2412 test. I used some some the techniques from your 2D 2412 file like the mesh volume, it was very useful to work from that file, thank you.
Values of 10 degree angle of attack
I recalculated pressures for 52m/s as 1622 PA
Reference area of 5m2 ( 1m chord and 5m length )
It solves out to Cl of just over 0.7 ( Should be around 1 )
We are definitely getting somewhere... the results are in the right area, just off by around 30% each time which is promising.
I wondered if 30% was a reasonable loss from it not being a " perfect " aerofoil of infinite aspect ratio, I searched and I found this publication :
https://zenodo.org/record/4068002#.Y7899JBws6M
They test NACA 0012 with very short aspect ratios at the same 52 m/s inlet velocity.
This graph shows three wings with the Aspect Ratio ( AR ) 0.5 of the chord, 1x of the chord and 2x of the cord. My test wing has an aspect ratio 5 times that of the chord
Interestingly, the efficiency drop is surprisingly low, less than 10%
.
I tried to get the NACA0012 aero foil simulation working to see if it would give me different results to the 2412 test. I used some some the techniques from your 2D 2412 file like the mesh volume, it was very useful to work from that file, thank you.
Values of 10 degree angle of attack
I recalculated pressures for 52m/s as 1622 PA
Reference area of 5m2 ( 1m chord and 5m length )
It solves out to Cl of just over 0.7 ( Should be around 1 )
We are definitely getting somewhere... the results are in the right area, just off by around 30% each time which is promising.
I wondered if 30% was a reasonable loss from it not being a " perfect " aerofoil of infinite aspect ratio, I searched and I found this publication :
https://zenodo.org/record/4068002#.Y7899JBws6M
They test NACA 0012 with very short aspect ratios at the same 52 m/s inlet velocity.
This graph shows three wings with the Aspect Ratio ( AR ) 0.5 of the chord, 1x of the chord and 2x of the cord. My test wing has an aspect ratio 5 times that of the chord
Interestingly, the efficiency drop is surprisingly low, less than 10%
.
Re: Sanity check of NACA 0012 gives wrong coefficient of lift
I re-did the NACA 0012 as a 2D planar simulation with the same 52 ms/s calculations.
It got closer, around 0.9 Cof L , so the maths looks plausible
I wonder if it is the reference area value that we do not have full understanding of......
It got closer, around 0.9 Cof L , so the maths looks plausible
I wonder if it is the reference area value that we do not have full understanding of......
Re: Sanity check of NACA 0012 gives wrong coefficient of lift
I tried the full scale 3D sim of the NACA 0012 at a much higher mesh resolution and with more boundaries........very slow, took 15 hours.
It *is* better, CoL increased to 7.5
I estimate that a target of around 8.5 would be ideal, if the 2D gets a result of CoL 0.9, that would account for the efficiency loss of the aspect ratio.
It *is* better, CoL increased to 7.5
I estimate that a target of around 8.5 would be ideal, if the 2D gets a result of CoL 0.9, that would account for the efficiency loss of the aspect ratio.
-
- Veteran
- Posts: 3156
- Joined: Sat May 20, 2017 12:06 pm
- Location: Germany
Re: Sanity check of NACA 0012 gives wrong coefficient of lift
Below you can see the definition of the variables in the NACA-report.
They differentiate between area "S" an area of wing "Sw". The definition of drag-/lift-coefficient
uses area "S". Unfortunately there is no further explanation in the following text.
My best guess is that area S=chord-length * wing-length
and area Sw is the area of the total wing-hull.
However, you can define your reference area as you like. Important point is, that everybody uses the
same definition for comparison.
At the moment I am running your files...
-
- Veteran
- Posts: 3156
- Joined: Sat May 20, 2017 12:06 pm
- Location: Germany
Re: Sanity check of NACA 0012 gives wrong coefficient of lift
Here is my result on intel i7 running all 12 threads.
Cut-off residual was set to 0,0001. Convergence after 740 iterations.
In my opinion the result is nice, compared to NACA.
I edit the control-dict after case writing and used the correct 1622 N/m^2 as pRef-value.
Maybe this gives the difference to your result.
Next step: running your 3D-case (on sunday, mate...)
Have a nice weekend!
Greetings Thomas
BTW: there where updates on cfdof-wb, cfmesh and Hisa last week.
Re: Sanity check of NACA 0012 gives wrong coefficient of lift
Thanks Thomas, I am really keen to get a working wind tunnel figured out, there are just so many practical uses for one, I am running a test now on the NACA 0012 as a 2D planar simulation - I changed the physics model to Transient and the results are near perfect in the report :
Running the 3D model now with a fairly efficient mesh and an improved volume mesh in the wake.
As an aside, I tried the Spalmart physics model and the results was a CoL of 4.2 so something is off there, probably not Freecads fault though.
Running the 3D model now with a fairly efficient mesh and an improved volume mesh in the wake.
As an aside, I tried the Spalmart physics model and the results was a CoL of 4.2 so something is off there, probably not Freecads fault though.
Re: Sanity check of NACA 0012 gives wrong coefficient of lift
Regarding Wing area, I wondered if it was wetted area rather than projected area, but this reference states it is projected.
https://wright.grc.nasa.gov/airplane/geom.html
States "
The wing area is the projected area of the planform and is bounded by the leading and trailing edges and the wing tips. Note: The wing area is NOT the total surface area of the wing. The total surface area includes both upper and lower surfaces. The wing area is a projected area and is almost half of the total surface area.
"
https://wright.grc.nasa.gov/airplane/geom.html
States "
The wing area is the projected area of the planform and is bounded by the leading and trailing edges and the wing tips. Note: The wing area is NOT the total surface area of the wing. The total surface area includes both upper and lower surfaces. The wing area is a projected area and is almost half of the total surface area.
"