XFoil Re and NCrit help

[Zipp]

Hi

This is my firts post at this forum. but I have read quite many of the treads about XFoil and foils.

Now I have a question about how to calculate the Re number in XFoil. If I use the normal formula Re=v/V for water at 20deg, I end up with Re=2.5e6 for a boat speed of 5 knots. Many treads are talking about Re in the range of 0.2e6 to 0.6e6 for a "normall" keel monohull. This value is 4-10 times smaller what I expect it to be. Any comments on this?

The other question is the value of NCrit. The default value is 9, but most recomendations says 3 for water tank like conditions. What about normal dirty lake water?

I have seen that the typical behaviour (CL/CD behaviour) of a foil at NCrit of 3 and 9 makes the Re value to be changed a factor of 5-10 e.g NCrit=3 and Re=75.000 has the same typical foil behaviour as NCrit=9 and Re=750.00 (with an offset on Cd).

Are there any other parameters that should be adapted to reflect water fluid instead of air in XFoil?

/Fredrik

 

[daiquiri]

There's no "normal" Reynolds number for keels, it depends on speed and reference length (the chord length, in this case).
The formula you've written is incorrect, the reference length is missing. The correct formula is Re=(V*c)/visc, where visc=1e-6 m^2/s is the water kinematic viscosity, "c" is the chord in meters and "V" is the speed in m/s.

So, for a 1 m chord keel at 1 kts (0.514 m/s) the Reynolds number is 514000 (5.14e5). for other chord lengths and other speeds, just multiply the above "base" Reynolds number by the actual chord and speed (in knots).

As for your question about Ncrit, the problem is not so much the water cleanliness as is the keel surface cleanliness and roughness. If the hull in question will stay in water for longer periods, the marine growth will soon make it become a rough surface, so you can imho assume Ncrit=1.
The difference between the towing-tank water conditions and lake conditions is mostly in the turbulence level, which is higher in the latter case. So if Ncrit = 3 is recommended for towing testing, then a smaller number should be necessary for real-life keels. I would go with 2 or again 1, but perhaps you should hear Tom Speer or Mark Drela on that issue.

Cheers

 

[Zipp]

Thanks for your reply.
In the manual of XFoil it states that they are NOT using the normal formula Re=(V*c)/visc, instead they use Re=V/visc so 5 knots => Re(XFoil)=2.500.000 independent of the chord length?

From the XFoil manual:
"the XFOIL Reynolds number RE is defined with the freestream velocity
and viscosity, and an implied unit chord:

CL = L / q | V = freestream speed
CD = D / q | v = freestream kinematic viscosity
CM = M / q | r = freestream density
RE = V / v | q = 0.5 r V^2

The conventional definitions are

Cl = L / q c
Cd = D / q c
Cm = M / q c^2
Rc = V c / v

so that the conventional and XFOIL definitions differ only by
the chord factor c or c^2.!"

/Fredrik

 

[daiquiri]

The Reynolds numbers of the real keel and the keel simulated with XFoil (or any other software) have to be the same or as close as possible, in order to obtain the necessary flow similarity.
So if you have a, say, 3 meters chord keel moving in water at 10 m/s, it's Re will be 3e7. This will be the Re you need to obtain in XFoil too - for example by scailing the speed (the first thing that comes to my mind) of a unit-chord keel to 30 m/s. Or by changing the fluid properties, if you want to go trickier way.

I use XFLR5 for that job. It is the GUI version of XFoil and it scales the model automatically in order to get the correct Re.

Cheers

 

[Zipp]

OK :idea:,

Thanks a lot for this information. Then the low Re numbers (e.g. Re=2e5) seen in other threds are because of fins and keels with short chord lengths.

One other question: I have read some Eppler E836 vs. Wortmann FX-L-V-152 discussions, and they mostly prefer the Wortmann profile. In my XFoil comparisment for a wide range of Re's, I think that the E836 is the one to prefer (angle of attack 0-7 deg). Any ideas what I'm doing wrong? Could it only be different personal design criterias?
I have made some measurements on my boat, and the "normal" steady state maximum leeway is appr. 4 deg. so I think that a range from 0 to 4 should be my first design goal (find a nice bucket for this area), then check for risk of stall at higher angles?

/Fredrik

 

[markdrela]

Thanks for your reply.
In the manual of XFoil it states that they are NOT using the normal formula Re=(V*c)/visc, instead they use Re=V/visc so 5 knots => Re(XFoil)=2.500.000 independent of the chord length?

You are misreading the manual.
XFOIL makes no attempt to determine the reference chord from your coordinates. It simply assumes that YOU have normalized the coordinates with your chosen reference chord c. And this "c" doesn't have to be the physical chord. It can be anything:
* Root chord of the keel (rather the local chord)
* Length of the boat
* 1 meter
* ...
So the x,y coordinates don't necessarily run from (0,0) to (1,0).

Whatever "c" you have chosen, the Reynolds number that you must specify is V*c/v. If you have chosen c = 1 meter, so that the x,y are in meters, then Re = V*1m/v = V/v

The output force coefficients use the same implied "c":
CL = L / 0.5 rho V^2 c , and same for CD

And again, if c = 1 meter, then
CL = L / 0.5 rho V^2

There are good reasons why you absolutely do not want XFOIL to deduce the chord. For example, say you modify your airfoil by deflecting the flap 30 degrees, the LE-TE distance will decrease. Do you really want XFOIL to be clever and use this reduced distance as the new "chord"? Surely not -- it will really complicate intepreting the deflected-flap polars with the zero-flap polars, because they would need to be rescaled by the chord ratio before being compared.

 

[Zipp]

:idea::idea:

Double thanks. Now I understand (I hope) the underlying meening of the normalized XFOIL Re.

/Fredrik