vspaero makes different results using GUI or command prompt

[David Eržen]

Hello,

I am trying to calculate a wing properties from NACA TN 1270 (technical report).

When using GUI, vspaero is giving me acceptable results (though I could't find anywhere how to change the air density) -> CL@-5AoA = -0.1

When I run the same case in command prompt (where I can change the density in setup fie) , the results come illogical -> CL@-5AoA = 0.025

I attached the case files and both results (.polar) fiels.

I really thank you for any help and instructions.

Regards,

David

p.s. when I tried to change the setup file and then use the GUI, the setup file is simply ignored

[Jasper]

Hi David,

When using the GUI, under the hood VSPAERO is launched pretty much the same as how you would do it yourself through command prompt. You can scroll all the way up through the VSPAERO output window, and will actually see the command that is send out to run VSPAERO - see attached image. As you can also see in this image, this command sends the Mach, alpha and beta lists at which the analysis should be performed and some settings you use for your aerodynamic analysis: start iteration for wake averaging, iteration to start calculating the wake and using the Jacobi preconditioner. Now, as you only use a single analysis point, the Mach, alpha and beta are also stored in the .vspaero file used when launching the analysis. Hence, when not passing these the solver will still solve at this flight condition. However, the AVG and SKIP wake iteration parameters are not stored in this .vspaero file and will have to be passed on through the command when you want VSPAERO to use them. Now I just noted that, even though the Jacobi preconditioner is stored in the .vspaero file, you still have to specify it manually, or otherwise it wont work. -- maybe Rob can tell if this is a bug?

When I run your case through command prompt whilst including the wake parameters and Jacobi preconditioner, I do get the same results as the GUI. However, when you do not specify these parameters, it will assume the default values for the AVG and SKIP wake iterations and the Jacobi preconditioner (=off); in this case you do indeed end up with the CL@-5AoA = 0.025. Actually it is only the Jakobi preconditioner that causes the difference I think. But at least this explains the differences you are getting between the GUI and Command prompt: they were actually running with different wake and preconditioner settings!

Regarding the density, and also flight speed velocity (they can be set in the .vspaero file, as well as in the GUI under the ROTOR tab): AFAIK they are not used in a normal VSPAERO analysis. They only matter when using actuator disks in your analysis, because they are used to determine the propeller performance, like thrust and power, based on the specified RPM, CT and CP. 

Hope that clarifies,

Jasper

[David Eržen]

Hi Jasper,

thank you for your prompt reply.

I did get some clarifications, especially about the "-jacobi" flag.

I do have another question about GUI: now I am experiencing a problem with Wake iteration. Although I set this parameter to 16, the vspaero keeps using the default value of 5. When I switch off batch calculation, it uses 16. Did I accidentally hit a button or set a parameter that switched off the iteration set?

I noticed a similar problem when using Command prompt: sometimes, all settings after AoA in NACA1270_DegenGeom.vspaero were ignored and default parameters values were used.

Any logical explanation? :)

Thank you.

Best regards,

David

[Jasper]

I just tried your file with and without batch calculation through the GUI, but in both cases it runs 16 wake iterations without a problem for me. So I cannot reproduce your problem. Also I am not aware of a second option or button that overrides the number of wakes specified.

Assuming you are using the latest version, I would guess that it has something to do with your VSPAERO file, as that is where the number of wake iterations is specified and you indicate you have more problems with it. 

A few quick things that I can think of:

Normally the GUI should write a new .vspaero file each time you start an analyses, so in case it goes wrong using GUI maybe the .vspaero file is write protected or antivirus is blocking something? In that case you can see if it actually writes a new .vspaero file and if it contains the correct parameters. Also make sure that the input DegenGeom file name under the advanced tab is still correct.

Otherwise maybe something went wrong when manually editing the .vspaero file and now it contains a mistake? Attached is mine for 3 angles of attack and 16 iterations, which for me also does run for 3 AoA at 16 iterations when running VSPAERO manually through command prompt. You can compare it to yours and see if anything is different. Also note the . and , specifications for decimals and next input respectively in the angle of attack list for example.

Jasper

[David Eržen]

Hello Jasper,

thanks again for the prompt and extended reply.

I have noticed that when I choose more than 18 AoAs, the number of wake iterations is reset to 5. Does the vspaero parser have a limit on AoA set? Does this make any sense? :)

Regards,

David

[Jasper]

Hi David,

I just tried an analysis for 20 AoA's and indeed the wake iterations is set back to 5, so it seems you found the limit :). Note that when you turn batch calculation off it will actually do the number of iterations specified, as it sends the cases one by one to VSPAERO. I think vspaero output files like .history and .lod will be overwritten each case, but you can still save all the results together in the .csv file (although the layout is a bit different/messed up compared to batch calculation on).

I am not sure what results your after; CL-alpha should be (nearly) linear, so you could do with only a few iterations. Cm and CD (and others) not so much linear ....

But I am not a developer, so if the AoA limit makes sense should be answered by one of them I guess ;)

Hope that helps,

Jasper

[Rob McDonald]

It looks like the AoA limit you're seeing is actually an artifact of a
maximum command line length. It is not an inherent limitation of
VSPAERO or OpenVSP.

However, your point is certainly a good one -- why would you need 20
AoA in a potential flow code?

Rob

> --
> You received this message because you are subscribed to the Google Groups
> "OpenVSP" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to openvsp+u...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[David Eržen]

Hi Rob and Jasper,

@Rob: The biggest reason fro 20 AoAs is that I am far from experienced openvsp/vspaero user and I am just exploring the program. :)

I did notice some slightly odd CL, CD, Cm results (like sudden offset at an aoa...) so I increased the aoa steps in that "region".

Anyway, I find openvsp to be excellent tool and looking forward to using it with my projects.

Thank you for your help.

Best regards,

David

[Jasper]

Hi David,

I had a similar problem: for a wing that I am analyzing I found that using the Jacobi preconditioner actually messes up the linear results (sudden increase in CL) in the lift curve after specific angles of attack (around 8-10 degr.) -- It was possible to mitigate this using different wake skip and AVG settings, but actually it completely disapeared by turning the Jakobi preconditioner off. Now, of course, results are very shape and case dependent, but if you get something similar you could try turning it off and see if it fixes your problem as well. 

Good luck with your projects!

[Rob McDonald]

The Jacobi preconditioner is the 'old' one. If you don't turn it on,
it uses a SSOR approach. The next release of VSPAERO will also come
with a matrix based preconditioner that will be 'on' by default, with
SSOR and Jacobi as fallbacks.

Rob