NACA-6-digit series airfoil

306 views
Skip to first unread message

Patrick Hammer

unread,
Aug 25, 2022, 12:08:12 AM8/25/22
to ope...@googlegroups.com
Good morning,

I was wondering if you could add the capability of superposing two NACA 6-digit series camber-lines. It's discussed in NACA report 824 as a way of effecting desired characteristics using the 6-digit series and I was wondering if that is something that can be implemented in future versions.

Thanks,

Patrick

Brandon Litherland

unread,
Aug 25, 2022, 1:52:55 PM8/25/22
to OpenVSP
https://ntrs.nasa.gov/api/citations/19930090976/downloads/19930090976.pdf Page 6, yes?

I'm not sure if the effort of implementing a mathematically accurate airfoil for these cases would be worth the amount of potential use cases.  In the meantime, you could implement the equations yourself to create the airfoil points and make an AF or DAT file to read into VSP as an AF FIle.  I'm unsure if the UIUC database or Airfoils Tools, etc. will have these combined camber line airfoils.

Rob McDonald

unread,
Aug 26, 2022, 12:52:49 AM8/26/22
to OpenVSP
It certainly isn't impossible...

The 6-series airfoils are not as analytical as the 4-digit and 5-digit and 16-series.

The 6-series thickness distributions are not described by a simple equation like the others.  Instead, they were developed from an iterative process based on conformal transformations.  OpenVSP does not duplicate that process.  Instead, we do what is commonly done in that we work from a tabulated set of points for some of the thickness forms.  In particular, we actually use the old Fortran program from NASA TM-X 3069 to do our 6-series airfoils.  That Fortran has been converted to C and we all that C routine from OpenVSP.

It actually looks like the routine we use does support summing multiple camber lines -- so some plumbing would be required, but all the 'hard math' is in there.

The f2c converted code has been stripped down a bit -- it no longer reads from a namelist file, etc.  So, it may take some work to make it work again for multiple camber lines.  Worst case, you could go back to the Fortran and re-convert it with f2c and then look at what was ripped out (and how) to guide what needs to be put back.

I suspect the biggest challenge would actually be figuring out the UI.  How do you express this idea to the user?  Do you limit them to two camber lines, or allow an arbitrary number?  What is the meaningful way to set this up?  The Propeller component allows you to parameterize a blade in terms of design lift coefficient -- if you are constructing a foil from multiple camber lines, how would you want it to scale by cli?

I will certainly add this to my stack of airfoil pet problems.  I enjoy implementing ever more esoteric airfoil parameterizations in OpenVSP.  For example, I've somewhat recently found that there are analytically defined forms for reflexed 4-digit and 5-digit foils.

I have another idea for my own airfoil parameterization that I'd like to do one of these days.  I would most likely use CST for the thickness form, but then then the user specify an arbitrary load distribution -- that I will use to construct a camber line.

Is this something you have a specific need for?

Rob

Patrick Hammer

unread,
Aug 26, 2022, 10:54:49 AM8/26/22
to ope...@googlegroups.com
Rob,

I didn't know there are also 4-digit reflex. I'm aware of 5-digit reflex, but the 4-digit I didn't know about. One of the reasons for looking at the 6-digit is because the max thickness can be pushed further aft, which is something I'd like to do for some things I'm working on.

I'm currently trying to use two camber-lines to create a reflex airfoil from NACA 6-digit so that it is based on simple formulas comprised of two variables: design CL and the transition location between uniform and tapering down dCP. I'm certainly aware that I can read in geometry files, but the challenge for me was applying a 6-digit thickness distribution. Since it's built into OpenVSP, I thought perhaps it could be "easily" updated to include a second camber-line since the thickness distribution is independent of the camber-line. I hadn't accounted for the fact that it might be using that old naca456 program. But since you said the framework is there within the old fortran code, it sounds like it'd be do-able.

Thanks for the insight,

Patrick

Patrick

--
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/18514716-3736-41bb-afd6-4dbfb126de90n%40googlegroups.com.

Rob McDonald

unread,
Aug 26, 2022, 11:43:42 AM8/26/22
to ope...@googlegroups.com
I think the code we're calling is much older than the naca456 code (:

As I understand it:  Ladson first wrote a code for the 6-series foils, released with TM-X 3069 in 1974.  We are using that code.  He then wrote a code for 4,5,16 foils, released with TM-X 3284 in 1985.  The combined NACA456 code came out as TM 4741 in 1996.

I'm sure all the 6-series code is essentially identical, so it really doesn't matter for those airfoils.  However, I wanted to make it very clear that we aren't using any of that code for our 4,5,16-series foils.  All of our other NACA airfoils are all generated with our own native implementations from the ground up.  In fact, there are a few somewhat innovative things that don't always get handled....

Rather than use the published polynomial coefficients for 4-digit foils, we go back to the formative equations and solve for the coefficients ourselves.  The published coefficients are to four or five sig figs, re-calculating them gets us the coefficients to machine precision.

The published 4-digit coefficients result in a 0.2% thick TE.  Practitioners commonly sharpen this airfoil by either skewing points (to maintain chord), extrapolating (extending chord) and sometimes then scaling down in the chordwise direction to restore the chord.  In OpenVSP, by default, we re-solve the formative equations to generate coefficients (to machine precision) that generate a sharp trailing edge with unit chord and the desired thickness directly.  While this is the default, we leave the option to generate the original NACA 0.2% thick TE.

Of course all of OpenVSP's TE modification tools can be applied to any airfoil.

The published equations for the modified thickness forms rely on some coefficients that are only published for a small number of points (making a continuous parameterization impossible).  Riegels proposed a curve fit to the tabulated data that has become the standard way of doing this.  Of course OpenVSP has to put a bit of a spin on it...

Also, while the 4-digit foil camber is parameterized in terms of geometric camber, I derived the expression for the design lift coefficient of those camber lines.  I then inverted that equation to enable OpenVSP to instead treat design lift coefficient as an input parameter for the 4-digit foils

All of our 4,5,16-series airfoils are built up from the equations for the thickness forms and the camber lines in our own implementation.

Otherwise, our airfoil code is all pretty standard.

I don't have my hands on the 4-digit reflex paper right now.  I know I saved it away somewhere.  I believe it was motivated by rotorcraft applications....

Rob




Patrick Hammer

unread,
Aug 29, 2022, 10:28:34 AM8/29/22
to ope...@googlegroups.com
Rob,

Thanks for the extra info. All great and interesting stuff.

Patrick

Reply all
Reply to author
Forward
0 new messages