Stall modelling using 2DClmax

[Séraphin Remy]

Hi everyone, 

Is it possible to have more information about the stall model with the 2D Clmax please ? 

I made some testing and results are good overall : 

- ASK 21 computed stall speed flaps retracted : 74km/h vs 74km/h in flight manual

- DR400 computed stall speed flaps retracted : 94km/h vs 94km/h in flight manual

2D Clmax are computed with XFLR5.

However, I noticed some behaviours that I don't really understand : 

 - computed local Cl is almost always overshooting the 2D Clmax by ~10% : why is that ? is the 2DClmax the value of the maximum or the maximum value in linear range ? 

 - how is the Cl evolution calculated beyolnd 2DClmax ?

 - what is the influence of Cl0 supposed to be ? 

Also, I made some trials with flaps deflected (on the DR400 model) : 

 - computed CLtot is totally wrong but mainly due to a very negative local Cl at junction between wing, so I took the .lod file and recomputed the total CL manually (2*sum(Cl*dArea)/S_wing... good ?). But even like this, taking the raw data, the CL computed is higher than the CLtot from VSP. Do you know where it could come from ?

Thank you in advance, best regards,

Séraphin REMY

[Séraphin Remy]

Anyone has any idea ? My results seem too good to be true, it's hard to fully trust them without more infos... 

[Brandon Litherland]

There are already a few posts where people have asked about the 2D Cl_max behavior in VSPAERO.  First search for those and have a look at the responses.  If those don't answer your questions or if you have a different question, you can follow up here and we can see about providing more information.

[Mike Vivaldi]

In my case I have an aircraft with multiple airfoils across the span. The suggestion from rob was to: 1)find the airfoil which matched the span location where there was the highest maximum value on the load distribution. 2) Then using XFLR identify the CLMAX and the CL=0 alpha angle. Finally 3) input the CLMAX2D and the CL=0 alpha value. If you only have one airfoil skip to step 2.

[Brandon Litherland]

Point of clarification, the Clo2D is the airfoil Cl at alpha=0, not the alpha at Cl=0.  It's the Y-intercept of the Cl v alpha curve.  This should be quickly apparent if you use a stall model that thinks the Cl at alpha zero is something like -5.

[Rob McDonald]

This is not correct.

Clo2D is the lift coefficient that corresponds to the lowest CD value for the airfoil.

It will usually be approximately the ideal lift coefficient / design lift coefficient for the airfoil.

Rob

[Rob McDonald]

Most significantly -- it influences the drag term.  It has nothing to do with the stall calculation.

Rob

[Brandon Litherland]

Okay, can we get a description update on that one then because the Parm description at the moment says "Zero alpha Cl for airfoil", hence my apparent confusion.

[Rob McDonald]

Oops.  I guess that is just there to throw you off the course...

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.
To view this discussion visit https://groups.google.com/d/msgid/openvsp/bf9aa5f3-8e5f-4b6b-8bdc-7fef051633a7n%40googlegroups.com.

[Séraphin Remy]

Thank you for your answers. I arrive a bit late, I think you already answered the question by saying that Cl02D should actually be Cl(Cdmin)2D, is that it ? 

I tried modelling an almost infinite wing (AR=500) to compare the local Cl0 at center for a NACA 5210 with the results from XFoil/XFLR5, as the Cl02D (ie lift coef of airfoil at zero aoa) is taken into account anyway with the airfoil curvature, correct ? Results are very similar : 0.5 for XFLR5, 0.47 for VSP. 

It seems then that the parameter is affecting parasite drag, but not induced drag, correct ? 

Also, it seems that even when the parameter is set to zero, clicking "stall model on" changes the drag model. I think I'm getting a bit confused here.

Here are some results for drag coefficients : 

[Séraphin Remy]

Correction : I understand that the sharp increase in CD0 at ~9°aoa is due to stall (wing begins to stall at around this aoa according to VSP). But I thought Cl(minD)2D would rather affect the induced drag, as per S. Gudmundsson "GA aviation aircraft design" adjusted drag model for wings : CD=CD0+[CL-CL(minD)]^2/(Pi*AR*e). Maybe this particular analytical drag model is not so good ?

[Rob McDonald]

Are you sure nothing else is changing between your cases?  You might extend your alpha range (in the negative direction) to capture the minimum of the behavior in that direction.

Also, the labels in your charts don't make a lot of sense, some things are confused.  Finally, you should produce these sorts of charts with the same y-axis limits if you want to compare results from chart to chart.

The Clo2D term should not change the induced drag term.

The Clmax and stall model (on/off) should not change the parasite drag term -- it should only influence the induced drag term.  It should only influence the induced drag term when you've approached clmax on some part of a wing.  In that situation, the local circulation will be reduced and the lift distribution will change -- which will change the induced drag.

Rob

[Séraphin Remy]

Hi Rob, 

Yes, parameters are unchanged. However, I should have mentioned that Clo2D is set to zero in the "stall model off" curves. I thought it wasn't taken into account when stall model is off.

I tried extanding the alpha range to negative values, and yes it appears that Clo2D shifts the drag minimum from CL=0 to CL=0.8 in this case. However, this shift appears on the CDwtot and CD0 curves, as seen on the above charts, not on the CDi or CDiwtot curves. 

Labels are as found from VSPAero output (apart from the upper right chart, where "tot" is missing in the legend), that's why I'm confused. No data extraction mistake is apparent.

Maybe my interpretation is wrong and CDi is not for induced drag ? And CDo is not for parasite drag ? 

I will try to compute induced and parasite drag with VSP, XFLR5 and analytics for a simple wing and compare results. I must be missing something here... 

PS : I put all charts to compare what I understood to be two different approaches, surface integral and wake integral. Values are a bit different but the overall behaviour seems to be the same.

[Rob McDonald]

Clo2D has nothing to do with the stall model on/off and the Clmax input.

The Clo2D should have an influence on the CDo and CDtot.  That is expected.

VSPAERO has no output labeled CDiwtot.

CDi and CDiw are two methods of estimating induced drag.

CDo is the parasite drag

CDtot = CDo + CDi

CDwtot = CDo + CDiw

Rob

[Séraphin Remy]

Thank you Rob for the calrifications.

There is one last thing I don't quite understand. It seems that the use of nonzero positive value for Clo2D shifts the minimum CDo of the wing to some positive CL, but the behaviour looks symmetrical about the CL=0 axis, so there are two minimals for CDo. Is that expected ? 

[Rob McDonald]

Not really, no.  I'll pass it on to the developer of VSPAERO.

For now, I would just set Clo2D to zero.

Rob