high AoA in xfoil 1

[quibo]

Hi,

I want to use xfoil to get CL and CD data for all angles of attack, i.e. the full 360 degrees.

This is for a NACA 4412 profile, Re = 0.708*10^6.

Has anyone ever tried this?

I'm beginning to think it's impossible as I get bombarded by "non convergence" and "no stagnation point found" errors.

 

[rb1957]

does it work for low AoA ?

is Re = 0.7E6 "low", as in "XFOIL: An Analysis and Design System for Low Reynolds Number Airfoils, Conference
on Low Reynolds Number Airfoil Aerodynamics" ? (ie are there limits in the program ?)

 

[KENAT]

Um, does the software even work in the stalled portion of AOA spectrum?

From what I remember of my advanced aero course, I'd doubt it. The algorithms for the different regions are likely quite different.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

http://eng-tips.com/market.cfm?

 

[Dan320]

X-Foil documentation:

"XFOIL doesn't crash very easily, but it can happen.
A common cause is accidental specification of a silly flow condition, e.g.

ALFA 90

In viscous mode this will produce all sorts of numerical fireworks,
and possibly an arithmetic fault."

Don't be surprised if it won't work for alpha 90.

Dan

 

[YoungTurk]

Dare I ask why such an exercise would be necessary?

 

[quibo]

I read that in the documentation too

Suppose I'll just have to use the Viterna post-stall model, although I've never seen it used on a non-symetrical aerofoil. It can't be too dificult surely.

It's for a vertical axis wind turbine simulation.

 

[rb1957]

pardon my ignorance, but how does a vertical axis windmill work ? i'd've thought that the only way to get it to entrain air across the blades would be to mount the blade disc in a duct ??

 

[rb1957]

took my own advice and looked it up in wiki ... laughed to myself about the article explaining in detail how to resolve the velocity vectors to get the local airpseed and direction and then CL = ... somehow i think Xfoil isn't "man" enough a CFD program to handle this very complicated loading. i think looking at a 2D geometry is an over simplification, the arched blades would behave quite differently (IMHO).

 

[rb1957]

how do you account for trailing wakes ? i think this needs a supercomputer to analyze. don't you need to model the whole d'mn'd thing, including the velocity profile of the wind (the shear close to the ground) ... what if the wind hitting this thing is coming from two different directions ?

 

[quibo]

Even if the wake makes it invalid, I'd still like to have a way of roughly predicting CL and CD for a full 360 degree sweep incase its ever needed for something else (i've spent all week programming on this so I may as well finish!).

Just realised that Ive never seen the the viterna corrigan model used to predict CM. Could I predict CM somehow by making a pitching moment from lift and drag?

 

[Dan320]

Let the airfoils that are stalled be represented by flat plate? At alpha 90 or more one can argue they're not airfoils anymore, aerodynamically speaking.

 

[cpinz]

Xfoil was designed to examine in detail the transition laminar/turbulent in a flow, but is not suited for fully turbulent analisys as happens for a fully stalled airfoil.

These limits of the codes are well known,and in fact it has been used succesfully for laminar airfoil analysis but not for flow condition were only the solution of Navier/Stokes equation coupled with a turbulence model, can give physically acceptable results.

The argument of post stall behaviour is well treated in the publications of NREL, where this problem is studied in detail,since it's an important issue in wind turbine blades behaviour.

Have a look there, you will find also a code to obtain the cl/cd curve for a profile working on a rotating blade.

 

[quibo]

I think you're right Dan, I don't think the Viterna model is far off a flat plate assumption anyway.

However, if you were using a NACA 4412, which has a zero-lift angle of attack of about -4 degrees, would you shift the viterna model by 4 degrees to the left?

For some reason I can't get access to the original Viterna-Corrigan paper from NREL. Strange because most of their stuff is publically available in mt previous experience

 

[Dan320]

Don't know anything about the Viterna models, I'm way too old. However, using the zero-lift angle for as a starting point sounds very reasonable to me.