Aircraft Sideslip Derivatives

[Ben H]

I am trying to calculate a rudder size for my aircraft and I need what this textbook calls "aircraft sideslip derivatives." I attached an image showing these two derivatives, dCn/dB and dCy/dB. 

From the VSPAero stability analysis I believe these two derivatives to be CMn wrt Beta and CMy wrt  Beta. Is this correct? I am getting strange results for my rudder size and I suspect I am using the wrong derivatives.

[Rob McDonald]

Pretty sure you are half right.

VSPAERO uses the following convention...

Body axis forces:

CFx          CFy          CFz

Body axis moments:

CMx          CMy          CMz

Wind axis forces:

CL           CD           CS

Wind axis moments:

CMl          CMm          CMn

In the context of sideslip and rudder sizing, 

CnBeta most likely means the derivative of the yawing moment due to sideslip.  So, in VSPAERO speak that is

CMn wrt Beta

CyBeta most likely means the derivative of the side force due to sideslip.  So, in VSPAERO speak that is

CS wrt Beta

You mentioned looking at CMy - which is the Body axis pitch moment -- which isn't what you want for static directional stability.

Rob

On Tue, Apr 6, 2021 at 6:10 PM Ben H <407yo...@gmail.com> wrote:

I am trying to calculate a rudder size for my aircraft and I need what this textbook calls "aircraft sideslip derivatives." I attached an image showing these two derivatives, dCn/dB and dCy/dB. 

From the VSPAero stability analysis I believe these two derivatives to be CMn wrt Beta and CMy wrt  Beta. Is this correct? I am getting strange results for my rudder size and I suspect I am using the wrong derivatives.

--
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/b16f69d4-6f01-4f0f-883f-410fa8fc22c9n%40googlegroups.com.

[Alan Teeder]

I am having problems with CYB as calculated by OpenVSP.

For reasonable values of incidence I expect CYB to be more or less
constant.

However OPenVSP  does not give me such a result. Attached is a screen
shot showing CYB over a range of incidences , as predicted by Datcom and
OpenVSP. The aircraft is the Vickers VC10. The screen shot is for the
panel method, but degen geometry shows similar results.

Can anyone shed any light on this?

Thanks.

Alan

[Rob McDonald]

Can you post your *.vsp3 file?

Rob

[Alan Teeder]

Here it is. I am using set 6 for analysis at the moment. This does not
include the engine nacelles.

Thanks for the quick reply .

Alan

[Rob McDonald]

I am not set up to run a sweep of stability runs to get at the meat of your problem.  However, I am able to take a quick look at your file and look for red flags.

The main thing that jumps out is that you have inadequate spanwise resolution in your model - particularly on the outboard main wing.

In my experience, the thin-surface mode needs substantially less chordwise clustering than you might think.  I often use it with chordwise clustering 'off'.

I would probably start this analysis with the fuselage and tail bullet fairing turned off (particularly for thin-surface mode).  Once I was happy with how things were working, I would turn them on and evaluate their effect.

I've attached an image of how I would start with the model -- once I had some success with this, I would take steps to add details and try out other features.

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 on the web visit https://groups.google.com/d/msgid/openvsp/1dba2e0c-422e-85eb-4e26-f6da85c3caa7%40gmail.com.

[Alan Teeder]

As suggested I have increased the spanwise wing resolution and am currently running a  test with just wing, tail , vertical-tail.

I will not have time to look at the results this evening, so will report back later.

My fin has a large area that is normally submerged in the fuselage, so there should be sufficient side area to generate CYB forces.

Alan

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/CAEppYpEXJonErQkoh7s07W8yCHUbN-cj4nCBfYSkFH8K2Rpybw%40mail.gmail.com.

[Alan Teeder]

The initial result is that CS-beta  shows the same trend as before, but with smaller values.

Alan

On 12/04/2021 18:56, Rob McDonald wrote:

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/CAEppYpEXJonErQkoh7s07W8yCHUbN-cj4nCBfYSkFH8K2Rpybw%40mail.gmail.com.

[Rob McDonald]

The next thing I would try would be to flatten out the large root twist in your wing.

Another good thing to do is to click through the different cases in the viewer and see if you can spot anything odd with either Cp hot spots or wake behavior.

When diagnosing something specific like this, you might want to skip the 'stability' run and instead run a matrix of alpha / beta in the region that you're interested in.  It makes it easier to click through in viewer and see trends -- you can also plot the cases in the Results Manager GUI.

Rob

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/33036778-cb96-099b-940b-f2bfe90c3a40%40gmail.com.

[Alan Teeder]

I have just checked my .vsp3 file and the wing has zero twist at all
stations. The whole wing is set at 3deg incidence.

I have defined a different root section as this matches what I have seen
in photos of the fuselage without the wings. The rest of the wing has a
supercritical section with most of the camber near the trailing edge.

Is the incidence and change of section giving a false impression of twist?

I will look more closely for pressure peaks. With the additional wing
chord sections there are more wakes which look nice and smooth. I will
look more closely.

Alan

[Rob McDonald]

In the file you sent me....

The entire wing is rotated 4 deg about the Y axis

The root section has 4 deg of incidence - that rotates the root section, but not the rest of the wing depending on the twist rel/abs setting

The twist of both wing sections is 0 deg.

In the end, this means that your root section is at 8 deg, while the outboard sections are both at 4 deg.

When combined with the dihedral and planform break change in taper, it created a visually striking shape that would cause a non-planar wake.  In my experience, this sort of thing can sometimes cause somewhat counter-intuitive results.

When watching the wakes, I would watch to see if (for example) a wing wake crosses over the vertical or horizontal tail at a particular angle of alpha/beta where you are seeing changes in behavior in the forces / moments.

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 on the web visit https://groups.google.com/d/msgid/openvsp/fd331638-865b-c4b3-2dbf-0e245b422253%40gmail.com.

[Alan Teeder]

I  eventually found the root incidence setting. A second pair of eyes, especially if they know what to look for, was needed to find that problem. I have set the root incidence back to 0 deg. Thanks.

I have been doing some more checking an with VSPAero and find that CYB is affected by wing incidence. Rotating the wing about the Y axis clearly demonstrates this.  My various old text books ignore the incidence effect or say that it is very difficult to estimate, but that it can be larger than the tailfin contribution. Given that this derivative does not have a dominant effect on stability  I can close this subject now.

With the high tail configuration the tail is always well clear of the wake. The engine nacelles however are much more closely coupled to the wing. Perhaps I will ask questions about this later.

Thanks again

Alan

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/CAEppYpFWAHYbPZVZQksVw9-dZXZZRL6E-Suhfg9iqLAc6JwVaw%40mail.gmail.com.

[Rob McDonald]

Glad this worked out for you.

In general, I find that VLM-like methods are very good at estimating stability derivatives - and they will often calculate contributions that handbook methods (Datcom, Smetana, etc) would ignore.

In almost all of these cases, I find that it is useful to back the VLM model off to a more simple case -- hopefully match the handbook method -- and then start adding the added features in one at a time.  Usually that allows you to isolate that some geometric feature is causing the observed change in derivative.

Once you've isolated that causality, you can then search for something in the literature about that contribution.  You can generally figure out whether it is important or not etc.

VSPAERO's thin surface method is not strictly a VLM method.

I find that VLM is more tightly defined in practice vs. panel codes and CFD (each representing a broad class of methods).

With VLM, people expect (for example) the wake lines to go straight back in +x or possibly be aligned with alpha (certainly not beta).  VSPAERO's wakes are either relaxed with the flow (in steady mode) or are convected in a time accurate manner for unsteady calculations.

When you start to really drill into the differences in sideslip derivatives between AVL, VORLAX, and VSPAERO, these differences can make a substantial difference.

When someone says 'Panel Method', everyone knows there are many implementation decisions that need to be understood when comparing results.  When someone says 'VLM', they seem to expect perfect agreement across tools.

VSPAERO's thick surface (Panel) and thin surface (VLM) solver are mostly the exact same code.  In fact, we are moving towards a unified view where models can mix thick and thin surfaces - or be all thick - or all thin.  From this standpoint, there really won't be a difference between Panel and VLM solver -- but instead a difference in how you set up your model.  I'm excited about this because the thin surface representation does a great job with lifting surfaces - while it has always been a bit unsatisfying for bodies and fuselages.  This mixed mode approach should give the best of all worlds.

Rob

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/5383abc1-5b3f-fa62-0e39-979b90ef8853%40gmail.com.