SR22 Model

[Chris Regan]

I'm doing some prep work for some students, and reacquainting myself with OpenVSP. I'm also finding all kinds of new functions since I last created models! 

I pulled the Cirrus SR22 model off the hangar site. It was the old v2 format, so I converted it. It didn't take to the conversion well. I mucked about with the model, removing a lot of cross-section definition in favor of blending/skinning. I used the dimensions listed in the SR22 Maintenance Manual for all the major dimensions (dated 2007). And guessed at the airfoils (wing is modeled as a NLF0414, Cirrus used that on the VK-30, but I don't know what the SR20 and SR22 actually use). The dihedral is correct at 5.5 deg. LE and TE sweep angles and tapers are correct on the wing and H-tail (V-tail was under-dimesioned but it's probably really close). I guessed at the wing twist and fiddled with the wing tips until it looked nice. NACA 4-series for the H-tail and V-tail.

I couldn't get flaps to transition through the wing cuff section, so I compromised and made the flap span shorter and increase the % chord to maintain the proper flap area. The flaps aren't simply hinged on the real aircraft, so it's all a compromise. Ailerons are placed and sized properly. The SR22 has control horns on the Rudder and Elevators, for modeling in VSP I just truncated the surfaces (so they are smaller than the real aircraft, but might have about the same effectiveness, they'd just have higher forces on the stick). The rudder hingeline is really wonky, it has a weird wrinkle in it, and the little control surface direction indication is at a strange angle. I'd like to figure out how to fix that, or at least know the effect it has on aero estimates. The fuselage to V-tail blending resulted in the rudder being tricker than I anticipated.

Fuselage cross-sections are all best-guess. I used dimensions where available, but otherwise just tried to get it looking about right. Landing gear wheels are place correctly, struts and pants were just based on pictures, to look about right.

Diameter, placement, and the number of blades are probably the only things correct about the propeller. I only use actuator disk, so it's good enough for that. All the propeller stuff is new, and I haven't had a chance to dig into it.

My biggest issue is that I don't trust the VLM results. I just did a spot check at a couple angles of attack. I'm getting zero vortices off the wing tips and H-tail at a wide range of angles of attack (the picture below is at alpha=-6!). So, that doesn't seem right. The spanloads look somewhat reasonable; elliptical-ish, so downwash is being computed but the trailing wake just looks like a sheet. I've only mucked about with a few things so far, I'm hoping that someone else recognizes the symptom and can guide me to a resolution.

Once I get these couple issues resolved, I plan to post it on the hangar.

- Chris Regan

[Rob McDonald]

Sorry the v2-v3 conversion didn't work more smoothly -- in some ways it is a miracle it works at all...

I wouldn't be too disturbed that the wake looks like not much is happening.  I've found that some configurations don't show a lot of rollup.

In general, I do think your main wing sections need a bit more spanwise resolution.  I would try to avoid the abrupt change in spanwise resolution between the outboard wing and the winglet by refining the outboard wing and decreasing the outboard clustering parameter.

I played with a case at 20degrees AOA.  I found that you have a nearly triangular lift distribution with these airfoils, chord, and twist.  I suspect you need to unload the center of the wing a good bit.

Sometimes I go with a heavy sideslip case to convince myself that the wakes are working properly.  Vertical tails have a lot lower aspect ratio than most wings...

You can also run VSPAERO in 'Fixed Wake' mode to convince yourself that the wake relaxation is actually doing something.  That said, for a well behaved model, I often find that the fixed wake and relaxed wake solutions are very similar.

In general, the VLM does not need LE/TE clustering.  The Thick-surface mode does, but thin-surface does not.  You can reduce or eliminate the LE/TE clustering if you're sticking with VLM for the most part.

If you want to model flaps with a gap and/or a fowler action, you'll have to explicitly model the flap as a separate component.  There are some tricks, but it certainly complicates the model a lot.

The rudder hinge line problem is because it is located on the blended center segment of the rudder.  You might try breaking the rudder into three sections - one for each wing panel that it traverses.  The rudder is laid on top of the surface in U,V parameter space -- note the grid lines are constant U,V.  So, when you want a straight control surface boundary in 3D space, OpenVSP has to jump through hoops to solve for the curved U,V line that does the best job.  In this case, it isn't doing a good job, but if it only had to work on one part at a time, I think it would do better.  Raw control surfaces are put into 'groups' that are then deflected by VSPAERO.  This can be used for control mixing, flaperons, and all sorts of good stuff.  You will also be able to group the three sub-rudders into a single rudder control that should work for you.

If you change the vertical tail to a transparent color, you will see that the hinge line is drawn.  The hinge line is always a straight line from control surface root to tip. The angle direction indicator should be drawn perpendicular to the hinge line.

In VLM, the fuselage will not contribute to forces -- but it will contribute to moments.  However, it can change the forces on the other surfaces.  Your fuselage doesn't look like it is causing any problems.

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/4ad30498-5d27-47da-86dc-fd24412f379bn%40googlegroups.com.

[Chris Regan]

Rob, the transparency trick on the V-tail is great, seeing the hingeline made it all make sense. Wire frame makes it really tough to distinguish the hingeline from all the other lines. The aft fuselage and V-tail were re-done with the fuselage extending back farther and the V-tail was made to be more consistent along its span. Now the rudder is in a happier place. I attached the revised. 

I'm still having issues with Aero though. I made a straight low AR wing (AR = 3, NACA 0010) and ran it at aoa = 20 deg. 

When I run with v3.25 I'm not seeing any wing tip vorticity. I get the same characteristic of the SR22 model, no tip vortices in the wake at all.

I pulled up an older version of VSP (v3.12 because I created a bunch of models back with that version) and ran VLM. There I can clearly see the tip vortex.

The surface pressures and lift distribution look about the same between the two versions. Which is strange since their wakes are quite different. Could it just be a plotting issue? I'm sure a lot has changed between 3.12 and 3.25.

[Rob McDonald]

Thanks for doing that test and pressing the issue.  I'll run it by Dave.

Rob

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/761d8a37-6f89-481a-ae62-6ec7513248b1n%40googlegroups.com.

[Rob McDonald]

Chris,

Here is the reply I got back from the VSPAERO developer:

> The new code has more damping for the wakes - specifically
> it has the vortex core model implemented that was absolutely
> necessary for the unsteady rotor cases. You can extend the
> wake distance out further (default was likely set to -1)
> and see what happens to the roll up further down stream.
> For me it looks reasonable on a model I just ran. The
> tip vortices are not going to affect the loading that much,
> if they did then all the previous VLM codes out there would
> be all getting really bad answers... since most don't adapt
> the wake at all. Further, at AR = 3 it's going to be
> elliptic pretty much no matter what you do.

Rob

[Chris Regan]

Thank you Rob! I really appreciate the support.

I posted to Hangar as: http://hangar.openvsp.org/vspfiles/559

I debated posting as a revision to the prior SR-22 model, but decided against. That version was a great place to start, but I ended up changing almost everything. I placed a note in: http://hangar.openvsp.org/vspfiles/147. My apologies to Mark Moore. If there is a desire to post as a revision, I'm 100% on board.

- Chris Regan

[Rob McDonald]

It would be great to post it as a revision -- it provides a quick cross-reference between like models.

I have no doubt Mark would be thrilled that someone had found that model useful as a starting point and had brought it out of the v2 dark ages into the v3 light!

Rob

[Mark Moore]

Awesome job - love it!   Now how about a Cirrus Vision Jet 😊

Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/d0139deb-0eef-43a7-bbe4-7cf5abdaae37n%40googlegroups.com.