airflow in a tube

[dreamer]

This is a generic engineering question that I've been pondering.

Is there a way to figure the airflow velocity in a tube given air passing over one end of the tube? The parameters are as follows:

Tube diameter: 6"
Tube length: 12'
Air speed past end of tube: 15 knots

[daiquiri]

You mean a 15 kts flow perpendicular to the tube opening? And you mean to figure out a mathematical relationship?
Then a very generic answer is no - it depends very much on the geometry of the hole, inclination, edge bevels, air turbulence, distance from surrounding obstacles, pressure difference between the ends of the tube and probably something else.
Unless someone did the empirical measurements of the particular configuration you're considering, in which case you can interpolate results and try to figure out a mathematical relationship between the perpendicular velocity outside of the tube and the axial velocity inside the tube, given the static pressure differential.
The only other viable way, as far as I know, is CFD.

[BMcF]

The length of that tube would allow you to make a reasonably close meaurement using a pitot probe and wall static pressure together to determine the dynamic pressure in the tube..whihcis equal to 1/2 x rho x Velocity-squared. However, unless the velocity is reasonably hihg, getting an accurate measurement of the dynamic pressure, even via the use of an inclined water manometer may include large errors.

We routinely use the pitot-perssure technique in long exit tubes to calibrate the airflow of fans to be used in hovercraft and surfce-effect-ship scale models.

[dreamer]

Thanks guys.

Clearly I need to answer more questions. I do have CFD capabilities.

How about a more general question, would 15kts perpendicular to the tube opening generate any airflow at all? If so, would it be minimal? (.1x) or would it be anywhere close to the actual wind speed?

Thanks for the info BMcF. The velocities will not be higher than average, outdoor wind. So a range of 0 to 30+kts ?

I can model up the tube I'm referring to. My goal is to have as much velocity in that tube as possible. There will be venturi in the tube as well.

[daiquiri]

Originally Posted by dreamer

How about a more general question, would 15kts perpendicular to the tube opening generate any airflow at all? If so, would it be minimal? (.1x) or would it be anywhere close to the actual wind speed?

The velocity inside of tube will be generated essentially by the pressure differential between the two ends. So it is not enough to know the conditions at just one end of the tube. At the side of the tube where fluid is flowing perpendicular to the opening (let me call it side "A"), a certain static pressure will be established - call it p(A). It's value will depend on geometrical characteristics of the tube end (as I stated in the previous post), on the surrounding ambient, turbulence scale etc.
At the other side of the tube (side "B") you will have other fluid conditions (velocity, ambient pressure etc.) which will create the static pressure of different value, p(B).
The difference dp = p(A)-p(B) is the driving force for the fluid flow inside the tube. If there is a pressure differential, there will be a fluid flow. You need a high velocity inside the tube? Then you need a big dp.

But dp alone is not sufficient to find the velocity of this flow. It will depend on friction against tube's walls, which in turn depends on velocity.
The variables involved in the calculation of the friction are:
- diameter of the tube
- surface roughness
- length of the tube
- presence and number of joints, knees, restrictions, etc.
- even the temperature of the tube walls, if very different from fluid temperature, can play a role.
Once a flow inside the tube is established, it will modify the conditions at both ends of the tube, and the corresponding pressure differential. And hence it will modify the flow velocity itself (and hence the friction, again).
So in order to find the velocity you (or your PC) will have to perform a sufficient number of iterations, until the solution converges to some stable value.

You might try to do it manually, as it was done in the times when only NASA had computers capable of performing some rudimental CFD. You can do it by using Moody's diagram (www.google.com), provided that you are able to estimate the loss coefficient at both ends of the tube, and the loss coefficient for every restriction, joint, knee etc. influencing the flow.
Last note: if your fluid is air, take care of compressibility effects, which will arise wherever airflow Mach number exceeds (even locally) the conventional value of M=0.3 .

If you decide to analyze the problem with CFD, take care of modelling accurately the geometry of the assembly, the wall conditions (mainly temperature and roughness), boundary conditions, and to choose the correct turbulence model (which is really the most painful part of CFD). You can find a good and quite comprehensive summary of various common turbulence models here: http://web.njit.edu/topics/Prog_Lang...ml/node328.htm