Creating and Analyzing Ducted Rotors

[Lansing Wei]

Hi,

I'm trying to use VSPAERO to analyze aircraft configurations which make use of ducted fans/ducted props. Ultimately, my goal is to be able to capture the flow interactions between the fuselage, the duct, and the wing. 

I have tried using both the Duct and using a Fuselage whose design XSec Order is set to "LOOP" and can't seem to get good results with either the vortex lattice or panel methods. I created a model by sticking a duct around a fuselage and putting a Disk element inside the duct. Then I put a wing just behind the rotor. The idea is that with the wing that close, I should see extremely localized and very clear interactions between the flow coming out of the duct and the wing itself. However, if I use a Duct with the Panel Method, I show interactions on the wing tip, which doesn't make much sense. If I have an open rotor with no duct, the same result occurs. If I switch to using a fuselage and the panel method, I get better results, but the degree of influence is still in excess of what I would expect. Using the VLM with the Duct crashes VSP, and the fuselage fails to compute anything.

The flow conditions I'm attempting to run this at are:

Alpha = 3 deg

Beta = 0 deg

Mach = 0.2

The Rotor states are: 

Vinf = 223.29 fps

Rho = 0.002377 slugs/ft^3

Dia = 6

HubDia = 0

RPM = 2000

CP = 0.6

CT = 0.4

My reading around this Google Group suggests that I should be able to do this. Is there a suggested way of approaching this problem? I have attached some images illustrating my concerns and my VSP Models.

[Rob McDonald]

My biggest question would be about whether the actuator disk model is
appropriate for modeling with a duct. We use what we call a 'Conway'
approach to the actuator disk model. Conway came up with solutions
for an isolated actuator disk -- that means we know the velocity
everywhere caused by the disk.

To couple that with the flow, we superimpose that on the freestream
and solve the boundary conditions / flow that exists in that
interaction. This means we have a somewhat one-way coupling of the
flow -- the actuator disk influences the aircraft -- but the aircraft
does not influence the actuator disk. So, if you have a case with
strong slipstream changes, the Conway approach won't capture it. If
the streamtube is not very affected, then this approach is probably
OK. A duct is a bit of a tough one -- the whole point of the duct is
to influence the streamtube of the disk...

Have you tried using a Body of Revolution component type? It was
recently added to make modeling ducts better...

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.

[Lansing Wei]

Rob,

Thanks for your quick response! Apologies for the relatively slow reply, I had not tested using a body of revolution yet, and was out of the office over the weekend. I've just run a few cases using a wing floating behind a rotor with and without a Body of Revolution. I've attached the VSP file and screencaps of the cases I ran, as well as a wing only just for reference. Unfortunately, by my eye, it looks as if the case with a body of revolution ducting the rotor is effectively identical to the case without a body of revolution ducting the rotor. It's possible that a more detailed examination of the Cp distribution could turn up differently, but I'm not terribly hopeful. What is your opinion?

Lansing

[Rob McDonald]

Hard to say.

That far downstream, the difference would really be small.

However, like I said up-front -- I don't expect it to really work. So
I wouldn't be surprised if you don't see the desired difference.

Rob

[Lansing Wei]

Rob,

I spent a few days checking my work to make sure I wasn't doing something silly with CP and CT, but I see what you mean regarding ducts. I did find that once I started putting in reasonable CP and CT values (0.05 and 0.02, respectively) that some changes occurred. Do you know of anyone else who has modeled, or has attempted to model, ducted fans in VSPAERO? If so, do you know their contact information?

Thanks again for all the help you've provided, I'd been banging my head against this wall for a while.

Best,

Lansing Wei

[Rob McDonald]

I am working with Dave Kinney on an approach to modeling rotors by
actually spinning the blades (not an actuator disk). Right now, it
only works for isolated propulsors located on the x-axis. I've done
some work there with ducts. It is promising, but not a turnkey
solution -- and it won't integrate onto a full aircraft at this time.

I also have written a standalone actuator disk / duct / body code
based on a vortex panel code -- it is a very simple and elegant code
that is well documented in a textbook. I can provide more information
if you would like. However, it too won't handle full aircraft as the
approach is axisymmetric in nature.

Rob

[Lansing Wei]

If you're able, I'm definitely interested about the standalone actuator disk/duct/body code. What is the name of the method you used?

Also, have you seen the paper "Solution of the flow over a non-uniform heavily loaded ducted actuator disk", by Bontempo and Manna? It was published in 2016 and presents an extension to ducted rotors of the nonlinear Conway model. I'm curious if you guys are going down that route at all. 

Lansing Wei

[Rob McDonald]

The standalone code was a 2D axisymmetric panel code based on:

R.I. Lewis, 'Vortex Element Methods for Fluid Dynamic Analysis of
Engineering Systems', 1991.

Lewis pretty much lays it all out for implementation. I never had
good luck with the inverse design techniques he describes, but all the
analysis stuff worked great.

Bontempo's work is pretty nice -- I started by trying to get their
code -- it ended up easier to write my own based on Lewis' text.

I don't foresee VSPAERO going down the path of an integrated duct/disk
superposition -- though it could be done. The rotor influence term
part of VSPAERO is pretty well isolated, someone else has looked at
implementing something different.

Rob

[Lansing Wei]

Rob,

Ah, very cool, thanks for the information! I'll take a look. Thank you also for the info about where VSPAERO is going for the foreseeable future. 

Lansing Wei