Re: [OpenVSP] Static Margin From .stab file

[Rob McDonald]

You are welcome to post your file so we can see if there are any issues with it.

You might try removing the fuselage -- I assume it was not part of your hand calculation.

VSPAERO calculates the lift and pitching moment coefficients at two angles of attack.  From that, it calculates dcm/dcl which is the definition of static margin.

Make sure you have the CG in the right place for VSPAERO.

Rob

On Thu, May 20, 2021 at 8:20 PM Jess Terr <jesst...@gmail.com> wrote:

Hi,

I was wondering how the Static Margin gets computed in VSPAero?  What equations are used?  I was doing a hand computation given an OpenVSP aircraft geometry as a sanity check but the VSPAero .stab Static Margin differs by a significant amount. I hand computed a static margin of 15%  and VSPAero computed a SM of 40%.... I am not sure why VSPAero is overestimating the SM. 

I am using the empennage, wing, and fuselage for the VSPAero stability analysis.  

Thanks,

Jess

--
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 on the web visit https://groups.google.com/d/msgid/openvsp/6321fbf2-00c8-4c34-8afa-0558b056daf2n%40googlegroups.com.

[Rob McDonald]

Perhaps you can upload a similar model that isn't sensitive.

Used properly, a VLM model like VSPAERO will do a very good job at estimating the static margin.  Something is wrong with what you are doing.

For the CG, what matters is you use the same CG in your VSPAERO calculation as you do for your hand calculation.  We really don't care what it is, just that it is the same.

The mass properties tool can accurately calculate a CG, but to do so with a complex model takes a lot of work and many many inputs.  Frankly, it often isn't worth the trouble.  Early in the design process, the designer has a great deal of control over where the CG is.  They can generally move components around such that the CG can be placed as they desire at a later stage in the design.  So, to start with, I would just put the CG where you want it to be.

If you are designing a missile or a fighter or other tightly packed aircraft, placing the CG becomes important earlier.  If you are designing a big transport aircraft, you can keep freedom in placing the CG a while later.  Either way, where you place the internal systems is very much determined with weight and balance in mind.

Rob

On Fri, May 21, 2021 at 8:10 AM Felipe Valdez <vald...@gmail.com> wrote:

Hi Rob,

Thanks for the response.  Unfortunately, I can’t upload the vsp model since it’s part of my thesis work that isn’t completed yet - sorry.  

I actually did include the fuselage as part of my hand calculation. And I feel like this calculation is giving a more accurate SM value...

For the CG in VSPAERO, I am using the entire aircraft CG (which includes the landing gear and  nacelles) computed by the MassProp toolbox (200 slices).  Is that the right approach?

Thanks,

Jess

On May 20, 2021, at 10:24 PM, Rob McDonald <rob.a.m...@gmail.com> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/CAEppYpGkUVdCceKhVbwQHvsRVGciPrXWZNc2tZYmwPjNnBKxOw%40mail.gmail.com.

--
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 on the web visit https://groups.google.com/d/msgid/openvsp/3135629A-D92B-40EA-9E31-4DBCFE008059%40gmail.com.

[Rob McDonald]

I de-clustered the points around the airfoil - VLM tends to not need strong LE/TE clustering.  I also made some other adjustments to the resolution and spanwise clustering here and there.

I reset the wake settings to 5 iterations and 64 wake points.  If you want it to run faster, for this case you can simply use the 'Fixed Wake' option.  It is best to not reduce the number of wake iterations too low as the solver relaxes the convergence tolerance on the first wake iterations on the assumption that there will be more of them to eventually come to a fully converged solution.  When you explicitly choose 'Fixed Wake', it tightens the tolerance on the first solve to ensure convergence.

No big red flags in the model - no degenerate components or whatnot.  The results look reasonable in the results manager and in the viewer.

My result changed some from yours - but not as much as you were looking for.

# Result                 Value      Units            
SM                      0.4506580 no_unit            
X_np                   39.4257269 Lunit    

I see no reason not to trust the VSPAERO results.  I would take another look at your hand calculations and try to see if you can pinpoint where there are differences.

Rob

On Fri, May 21, 2021 at 10:58 AM Jess Terr <jesst...@gmail.com> wrote:

Hi Rob,

I was able to simplify the aircraft by just including the lifting surfaces in the vsp model. I also kept the desired CG values in the VSPAero settings. I have attached the file.  

I hand computed a SM of 16% (without the fuselage contributions). In VSPAero, I am now getting a 51% SM with the lifting surfaces only, big difference...

Thanks.

Jess

On Friday, May 21, 2021 at 9:12:40 AM UTC-7 Jess Terr wrote:

I tried to model the aircraft with all its components to have a desired CG so that's why I am using that CG for VSPAero as well as my hand computation.

I am not sure what I am doing wrong for VSPAero to give such a large SM estimate.  Is there a way I can send you part of the vsp model to you personal email?

Thanks.

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/fae19723-b0b6-41f4-b718-d846a490b49en%40googlegroups.com.

[Rob McDonald]

There is some significant effect to the z-displacement of everything in your model.  I suspect your hand calcs ignore the z components of everything.

If I zero out the z location of the wing, tail, and CG, I get the following result...  Still not enough to match, but one explainable source of difference.

# Result                 Value      Units            
SM                      0.3495906 no_unit            
X_np                   38.5244040 Lunit        

Rob

 

[Rob McDonald]

I am not saying that you should zero out the wing and tail locations -- I merely did that as an expeditious way of comparing to your hand calcs (which I assumed ignored z-components).  The z-effect is real, but for many aircraft it is small and can be ignored.  That may not be the case for your aircraft.

You can use whatever CG location you want as your moment center location -- just be sure to use the same location in VSPAERO as in your hand calcs.  You really shouldn't be focused on the static margin right now.  Instead, it is probably better to think about calculating the location of the aerodynamic center.

Once you reach confidence in your calculation of the aerodynamic center, you can then back-calculate a desired CG location to give the static margin that you want.  Then you arrange the components of the aircraft to get the CG as close to the desired location as possible.  Finally, you add ballast to fix whatever went wrong.

If you want to get a little more complex, you can look at moving your wing along the fuselage as you size the tail and elevator.  When you do so, you need to look at your landing gear position as well -- consider tip-back angle, pitch stability, elevator authority, load on the nose wheel, etc.  Depending on your vehicle, you'll need to look at forward and aft cg locations too.

When calculating cl_alpha in your hand calculations, are you using any form of aspect ratio correction -- or are you just using 2*pi per radian or 0.1 per degree?

Rob

On Fri, May 21, 2021 at 12:35 PM Jess Terr <jesst...@gmail.com> wrote:

Thanks for the help and suggestions Rob.

Making those changes and including the fuselage, I get an SM of 35% (5% less than previous result).  

As you noticed, I am using the desired cg location in VSPAero as my datum... should I be using a different datum?  Also, the wing's Z_displacement value is not zero because I choose my fuselage to be at (0,0,0) and the wing is displaced upwards in referenced to the fuse. I dont think I can do much in zeroing out the wing's Z_location, unless I shifted all other aircraft components down... but as you said, that zeroed Z_location of the wing had a significant effect on SM... 

I'll keep digging and see if I made a mistake in my hand computations. I did ignore the z-components of everything...

Thanks.

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/fe8084a4-299b-40e4-a1c5-82d897cf1df7n%40googlegroups.com.

[Rob McDonald]

My next step would be to use an aspect ratio correction for the lift curve slope in your hand calculations -- both for the wing and the horizontal tail.

a is the corrected 3D lift curve slope.

a0 is the 2D lift curve slope - 2*pi

AR is the overall wing aspect ratio

a = a0 / (1 + a0/(pi*AR))

Your wing is high AR, but your tail is pretty low.  This correction will reduce the tail effectiveness - thereby moving the AC forward, increasing the SM.

Rob

On Fri, May 21, 2021 at 1:28 PM Jess Terr <jesst...@gmail.com> wrote:

Thanks Rob. That makes sense. I did play with moving the wing along the fuse X_Loc and was able to reach a desirable SM in VSPAero, but the hand computation SM changed.  I will follow your approach and see if that works best, thanks.

for CL_a I am just using 2*pi per radian as an approximation.

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/896b1792-ce8e-41e9-ab17-800e2f74f7den%40googlegroups.com.

[C P]

Hi, 

if this can help, in addition to Rob's advices. 

 Looking at the tail geometry this seems to be too large for the SM you are expecting (16%), supposing the hand calculation is your reference. It means that you should change the tail volume, possibly reducing the tail area (if the arm to the wing is a constraint). This contributes to decrease that part of the longitudinal static stability equation where tail volume acts, leaving more room to play with the position of center of gravity respect to aerodyn. center.

Morevover, the airfoil used for the wing ( if I've not missed something in the discussion) has a moment coefficient very low ( there is some reflex in the geometry it seems), then there is no reason to over-balance the wing with such large tail. Statically it works, if you need it for some reason,  but dynamically it could  result an airplane too stiff.

Corrado