Best way to measure velocity approx 5m/s +- 0.1m/s

[es335]

Hi.

What is the best way to measure velocity approx 5m/s +- 0.1m/s in air approx 50C ?

A pitot tube would have to measure a difference of 0.005Pa to detect a 0.1 m/s change, and that is properbly impossible.

I think that a hotwire tool is the best solution, but I am not sure.

Any comments on the problem is appreciated.

 

[balu1]

You are right that a hot wire anemometer would be better than a pitot in this case. But +/-0.1m/s accuracy could be a little bit too much I believe. It is exactly at the best assuracy limit of +/-2% (see

http://www.topac.com/anemometer.html)

 

[hacksaw]

you can get about as much accuracy as you can afford.

the really accurate meters(the very best) suitable for custody transfer of single phase fluids start at 100k or so.

most meas. arn't so demanding and do not require traceability, etc. can work also for you.

avg. pitot (annubar and other variants), multipoint pitot, traversing pitot, hot wire will under good conditions give you better than 1%. The catch is the intangibles involved.

 

[ccfowler]

Assuming that this is for flow in some sort of pipe or duct, what are its dimensions and how much straight run is exists both before and after the measurement section?

If the conduit is relatively large, how is the flow distributed across the measurement section? How stable is the flow rate and distribution across the measurement section?

What is the purpose of measuring the velocity? Are you looking to measure the flow in a relatively large duct but means of a velocity traverse pattern?

 

[es335]

Thanks everybody for your answers.

A observation:
I just tried to hook a pressure transmitter up to a small system, and I noticed that when I held my hand around (just around, not bending, shaking or anything else) and heated the small plastic tube, the absolute pressure rose from 50mmWG to 54mmWG.
That is a pretty visual demonstration of one of the error sources a pitot system can have. A small temperature change in only one of the pressure tubes (static and stagnation) result in a massive error – massive compared to a transmitter accuracy of 1% of 20mmWG.
I guess that I prefer hotwire now Unless, of course, someone can see an error in my observation.

Well, back to the questions:
Assuming that this is for flow in some sort of pipe or duct, what are its dimensions and how much straight run is exists both before and after the measurement section?
The dimensions are not given yet, but the setup should be used for different sizes of ducts.
The straight run has a length of about 50-100cm, depending on the device. The flow exits into a plenum.

If the conduit is relatively large, how is the flow distributed across the measurement section? How stable is the flow rate and distribution across the measurement section?
How the flow is distributed across the section I do not know. The duct may be as small as 5cm and as large as 30cm. Ideally the distribution is uniform, but due to wall gradients in the small duct and upstream anisentropy in general the flow may the skew. But hopefully it is not.
The flowrate should be stable.

What is the purpose of measuring the velocity? Are you looking to measure the flow in a relatively large duct but means of a velocity traverse pattern?
I am looking for differences in flowrate across the duct. And I will traverse the probe across the duct.


Btw, if anyone got links to manufactures of hot-wire equipment (other than the one already given), please share.

 

[ccfowler]

Your duct sizes will be small enough that the mere presence of any velocity sensor may disrupt the flow to some noticeable extent. You may want to keep this in mind particularly for the smaller duct sizes.

You may want to consider the use of a very small diameter Pitot-Prandtl style tube for your traverses. They have a potentially useful characteristic for your proposed application where the local flow direction may not be parallel to the axis of the duct at the measuring section. The Pitot-Prandtl style tube inherently indicates the local flow velocity component parallel to its axis where the actual flow velocity is actually misaligned to a significant amount. Check the calibration charcteristics (including flow alignment effects) for the specific probe that you wish to use.