Propeller-Blown Wing Simulation with Zero Free-Stream Velocity

300 views
Skip to first unread message

Supakorn Suttiruang

unread,
Sep 28, 2021, 2:38:20 AM9/28/21
to OpenVSP
Hi all,

I'm working on building an aerodynamic database of a VTOL tailsitter aircraft.
The aircraft relies heavily on the propeller slipstream over its control surfaces for attitude control during ascent and descent. It appears that VSPAERO can handle propeller-blown wing and there is some validation data in this recently published paper. However, I wonder if VSPAERO can perform a similar simulation where the free-stream velocity is zero (or very close to zero) as it would be during hover.

Thank you all in advance!

Supakorn

Brandon Litherland

unread,
Sep 28, 2021, 7:03:11 AM9/28/21
to OpenVSP
There have been a few questions about this aspect of VSPAERO recently and I encourage you to search the conversations here for the topic.  The way to make this work is to set the Vinf to something very low and set your Vref according to the blade tip speed.  Depending on if you are wanting rotating blades or actuator disks, the settings will vary somewhat but otherwise the analysis will run and provide results.  If you want the efficacy of the control surfaces, you'll want to do a stability run.

Supakorn Suttiruang

unread,
Sep 28, 2021, 10:58:14 PM9/28/21
to OpenVSP
Hello Brandon,

Thank you for your swift response and apologize for duplicating the question.

I did some numerical experiment and followed your procedure. Below is the result.
The model is a 4-m span and 0.3-m chord NACA 0012 wing with a 12-in disk propeller at 0.25D from the leading edge. The geometry is shown below.
geom.png
(FIG1: Case geometry)

I ran VSPAERO without prop and the results are as expected:
clean_cl.png
(FIG2: Span-wise lift distribution of the clean configuration [prop off])

clean_dcp.png
(FIG3: dCp and trailing wakes of the clean configuration [prop off])

And then I turned on the propeller at 5000 RPM with Vinf = 0.01 m/s and Vref = 79.8 m/s (propeller tip speed at 5000 RPM). The lift distribution looks as expected but the trailing wakes look a bit funny:
prop_on_cl.png
(FIG4: Span-wise lift distribution of the prop-on configuration [5000 RPM])

prop_on_dcp.png
(FIG5: dCp and trailing wakes of the prop-on configuration [5000 RPM])

Here are my follow-up questions:
1) What does VSPAERO do to the Vinf input? I noticed that it will only be enabled if propeller solver is activated. From this thread, Rob McDonald mentioned that Vinf is only used to convert the propeller coefficients into dimensional quantities, so I thought the Vinf input is relevant only to the propeller and does not affect the wing. This is why at first I didn't think setting Vinf to low values would simulate zero-freestream condition over the non-blown sections of the wing. My experimental results clearly says otherwise.
2) Has there been any validation study on how VSPAERO performs with wing-blown propeller at zero freestream velocity?

Rob McDonald

unread,
Sep 29, 2021, 12:27:56 AM9/29/21
to ope...@googlegroups.com
Your final picture looks somewhat more odd than it might otherwise because you have it rotated the wrong way.  I.e. the trailing edge is near the viewer and the leading edge is away from the viewer.  You can turn on visualization of the actuator disk to help make sure you have context 

The thread you cite about the use of Vinf is a little outdated -- it comes before we had a Vinf and Vref -- before VSPAERO had support for static (or hover) cases.

Vinf is the freestream velocity.  Usually Vref is equal to Vinf.  However, in cases where Vinf is zero, you can set Vref to something else for normalizing coefficients.

Here is a recent paper that looked at using VSPAERO on blown wings.

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/b5147506-b891-4722-8d3f-12c2ce20bb7en%40googlegroups.com.

Supakorn Suttiruang

unread,
Sep 29, 2021, 12:42:57 AM9/29/21
to ope...@googlegroups.com
Thank you Rob. I indeed rotated the viewer incorrectly!

If I understand correctly, the AIAA 2021-2510 paper that you linked validates a forward flight (Vinf = 27.84 m/s) case and compares the VSPAERO results with those obtained from OVERFLOW RANS code. I was hoping there's some validation work with static cases against experimental data but could not find any. I'm working on validation work of my own against NACA TN-3307 in which data for a somewhat close to static case was collected.

You received this message because you are subscribed to a topic in the Google Groups "OpenVSP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/openvsp/7ZzNVBM7ITs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openvsp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/CAEppYpFz%2Be_ZV_qx9%2Bu3_mkhsf-Ly1qMGo2_zicSMrEBq4KQHw%40mail.gmail.com.

Rob McDonald

unread,
Sep 29, 2021, 1:17:14 AM9/29/21
to ope...@googlegroups.com
I am not aware of a published case where VSPAERO was compared for a static blown wing.

However, "It should work".

Rob

Supakorn Suttiruang

unread,
Sep 29, 2021, 3:11:34 AM9/29/21
to OpenVSP
Very well. I'm beginning validation against TN-3307 now. A little follow-up question: In the propeller solver mode (actuator disk in this case), the solver uses Vref (which is set to the linear tip velocity of the propeller) to non-dimensionalize lift, right? Since the TN-3307 uses slipstream dynamic pressure (calculated from propeller induced velocity), I should be using the propeller induced velocity as Vref instead of the propeller tip speed. Do I understand this correctly?

Brandon Litherland

unread,
Sep 30, 2021, 6:58:36 AM9/30/21
to OpenVSP
If they list the induced velocity in the paper, then you can use that or you can "correct" the results proportional to Vref.

The TN-3307 paper references the idea of turning flow rather than "blown wing" per se.  It relies on the concept that a large, highly deflected flap/vane system can deflect the prop slipstream downward to produce lift rather than increasing the dynamic pressure over the wing for lift augmentation.
I'll give you a heads-up that this is really tricky to model in VSPAERO.  Trying to model this in VLM with a large section of the wing as a control subsurface will run but give you poor results.  Creating the geometry in a way that will run in Panel mode will be more realistic but continue to have problems and will have a really hard time converging.  Deflected slipstream is one of those problems that can technically run under the assumptions of potential flow solvers but resides outside the realm of "normal" application where you should expect reasonable results.  The best advice that I can give you is to try several different methods and geometries and see if you can get it working.  However, even if you get it working, the results aren't going to be accurate.  For this case, half of the slipstream is deflected downward and half is blowing over the top of the wing/flap system which then impacts the vanes and is hopefully turned downward as well.  The flow over the top gets really messy and has separation at multiple locations which VSPAERO will not capture.  

Best of luck.

Supakorn Suttiruang

unread,
Sep 30, 2021, 7:08:12 AM9/30/21
to ope...@googlegroups.com
Thank you, Brandon.

Reply all
Reply to author
Forward
0 new messages