Benoît G. asked 3 months ago
Hello,
I am extracting the results of a 3D Analysis on a whole plane via right click -> Export data -> to file, then importing into an excel. I understood the concept of body forces/ far field forces, but I can’t match the forces I get from the .csv to the ones displayed in the software for a particular operation point.
I am comparing the results at 70m/s, 0° angle of yaw & +5° angle of attack. From the pictures I took ( which I don’t manage to send on the forum cause local=no link, but I will send them by mail ) I get:
1) The forces from the display ( total forces, body axes) almost match the farfield forces, which is already weird for me as I was expecting the body axis to be different from the farfield axes ( which I assumed were wind-relative ). But the Fx doesn’t match for unknown reasons.
2) The body forces don’t match between the two methods. Even when projecting using Fx_2=Fx_1*cos(5°)+Fz_1*sin(5°) I am far off.
So I was wondering if I was missing something, or is it a bug of sort?
Benoît G. answered 3 months ago
With aoa=0°, yaw=0°, vel= 70 m/s
Total forces, body axes ( bottom left corner display):
fx=10.5 [N] , fy=-0.436 [N] , fz=74.5 [N] , mx=-55.6 [Nm] , my=20.4 [Nm] , mz=0.314 [Nm]
From extracted .csv:
FF_fx=2.8359 [N] , FF_fy=-0.43572 [N] , FF_fz=74.457 [N] , mx=-55.563 [Nm] , my=20.353 [Nm] , mz=0.31364 [Nm]
As you can see, it seems there is a 8[N] for the x-axis force,which I fail to understand the origin of, while the others seem to align pretty well.
Best,
Benoît
techwinderJust to make sure: we are talking about thee export of operating point data and not polar data, right?
The operating point’s .csv does not contain far-field force information but only panel data, so I guess it should read FPressure_fx instead of FF_fx in the comment above.
There are two potential causes for the mismatch.
1. The pressure forces on the wing’s tip are not included in the .csv file but they are by default in the flow5 display. This should not matter because the tip forces are 0 in the x-direction and they cancel on both sides in the y-direction barring numerical errors. Their contribution is non-zero in the z-direction if there is dihedral. They can be excluded explicitly from the results by deactivating the checkbox “Include the contribution of wing tip pressure forces” in the definition of the analysis. Since the wing tips are locations of high pressure gradients and velocity which would be dampened by viscosity in reality, it is likely best to exclude them.
2. In the flow5 display of properties, the part forces are the far field forces, converted to body axes, and not the pressure forces. This is a bit confusing for sure, so I’ll modify the display to something like “Far_field forces on parts, body axes:”
Hope this helps.
Benoît G.
replied 3 months ago Hello,
I am talking about the .csv I get when I follow a right click on the 3D analysis ( T8-VLM2 ) -> Export data -> to file. Which gives me a .csv in which I have several lines corresponding to the changing alpha and the forces etc. associated. In those columns I have both ” Fx_FF_(N) Fy_FF_(N) Fz_FF_(N) Fx_sum_(N) Fy_sum_(N) Fz_sum_(N) ” (copy/paste directly from the excel).
I just tried to remove the wing tip contribution, it doesn’t change a thing. I still have Fy_FF, Fz_FF and the moments perfectly aligned between the .csv and the display, but the Fx_FF still have a 8 [N] difference ( Fx(display): -18.3 [N] / Fx_FF(.csv): -26.125 [N] )
Best,
Benoît
Ps: As a comparison ( Fy(display): -0.734 [N] / Fy_FF(.csv): -0.73374 [N], Fz(display): -396 [N] / Fz_FF(.csv): -395.15 [N] )
Benoît G.
replied 3 months ago Another question regarding the .csv export, when I read the .csv of a specific point ( alpha=x°, beta=x°, vel=x m/s ) I don’t understand what the F.z is. When I plot it against the y(m) it looks very weird.
I’m lost.
My testing shows that the forces exported in the csv are the same as those displayed in the operating point properties barring viscous effects.
F.z is the vertical component of the force, but I don’t understand how yuou can possibly plot it a
I’m lost.
My testing shows that the forces exported in the csv are the same as those displayed in the operating point properties barring viscous effects.
F.z is the vertical component of the force, but I don’t understand how you can possibly plot it against y using only the exported polar data.
André
Benoît G.
replied 3 months ago Hmm that is weird indeed… I will send a quick screen recording later today to illustrate all the points previously mentionned, maybe there something I am doing wrong.
Best,
Benoît
Hi Benoît,
I suspect that the answer to both questions is related to the viscous drag.
The values which are displayed in the bottom left box include the viscous drag.
Also they are the far-field values for the forces, but the on-body forces for the moments which may indeed be a bit inconsistent.
Note that in potential theory, the body does not create either drag or lift due to d’Alembert’s paradox. Therefore any residual inviscid force acting on the body is due to numerical errors, barring viscous drag.
Hope this helps,
André
Hi André,
Ok thank you, but then if I understand correctly, the forces displayed in the bottom left corner including the viscous drag should match the far field forces right? As a matter of fact it does, quite well for Fx/Fz and even the moments Mx/My/Mz but I still don’t understand where the -8 [N] discrepancy comes from for Fx. I tried to run the simulation without the viscous calculation on the fuselage ( from the analysis option ) but I keep getting the same difference.
Kind regards,
Benoît
The far fields forces are in wind axes, whereas the exported forces are in body axes.
Please try with aoa=0° so that axes are the same.
André
With aoa=0°, yaw=0°, vel= 70 m/s
Total forces, body axes ( bottom left corner display):
fx=10.5 [N] , fy=-0.436 [N] , fz=74.5 [N] , mx=-55.6 [Nm] , my=20.4 [Nm] , mz=0.314 [Nm]
From extracted .csv:
FF_fx=2.8359 [N] , FF_fy=-0.43572 [N] , FF_fz=74.457 [N] , mx=-55.563 [Nm] , my=20.353 [Nm] , mz=0.31364 [Nm]
As you can see, it seems there is a 8[N] for the x-axis force,which I fail to understand the origin of, while the others seem to align pretty well.
Best,
Benoît