Thread: DE Measurements

To do a matrix flow, would you suggest, based on your last post, to set the instrument on timed average and move it from one side of the duct to the other, slowly, and capture the average in both a vertical and horizontal position? Might try that tomorrow.

I have already done a "stepped" single point vertical and horizontal set of measurements.

Originally Posted by Lappa

To do a matrix flow, would you suggest, based on your last post, to set the instrument on timed average and move it from one side of the duct to the other, slowly, and capture the average in both a vertical and horizontal position? Might try that tomorrow.

I would do that a few times to see what sort of differences you get.
In practice I found measuring along just one diameter was usually all that was needed.

You could average all the timed averages at the different positions for speed, but it's unlikely to give a reliable answer and although you may not be interested in this it also won't enable a reliable uncertainty to be determined

Of course you do what you like but here's how I would do it with that instrument in a pipe with high velocity.

1) Performed repeated multiple timed averages at systematic distances across a 150 mm pipe.
eg 5mm from wall, 15mm from wall, 25, 35, 45 etc up to the middle i.e. 75mm and then across the other side
Go back and forth at least twice (more if you can stay it), so 15 locations and 4 timed averages for each location.
The time based average feature of the probe should take care of the variation due to turbulence.
The repeated back and forth should take into account how well you can locate and orient the probe into same position every time.
It may pay to make a location/orientation jig to get this reproducible

2) Calculate the mean and standard deviation for each point. Then use the std dev to decide if at most any one of the individual timed averages is an outlier. Reject the outlier and recalculate

3) Average the corresponding X mm from the RHS side of the duct with that of the X mm from the LHS of the duct.

4) A simple way to find the flow is to use the 5mm average to find the total flow in the 0 to 10 mm cross section of the ppie, the 15 mm average to find the flow in the 10 to 20 mm section , etc and sum all these this will give a reliable measure of flow (but it won't provide meaningful data for the uncertainty) This is normally what I do.

To find the uncertainty correctly I would do that by fitting a polynominal function to the data - average air speed as a function of distance from the centre to the middle of the pipe.
The integral of that function will give the most reliable air flow value and the uncertainty can be determined from the difference of fit between the real data and the polynomial - won't go into that here.
The uncertainty is critical for deciding if there is a difference between change ducting/port setups etc - otherwise changes may be being made that are worse rather than better or no different.

If you decide to give even just the data collection a try and you can send me a set of data I wouldn't mind having a play with it.
This sort of thing was the bread and butter part of my former day job..

Thanks. re Std Dev, would you use the sample or population formulae for this one?

Cheers

Originally Posted by Lappa

Thanks. re Std Dev, would you use the sample or population formulae for this one?

Cheers

Sample

Interesting new Testo air flow measuring gear for AC hoods
testo 420 Volume Air Flow Hood

Not sure if it would be much good for anything related to DE but it shows what sort of effort and cost is needed for accurate air flow measurements.

Originally Posted by BobL

Interesting new Testo air flow measuring gear for AC hoods
testo 420 Volume Air Flow Hood

Not sure if it would be much good for anything related to DE but it shows what sort of effort and cost is needed for accurate air flow measurements.

I have a TSI flow hood Bob. Same thing as that Testo. TSI make for others which are Brand named Dwyer, Alnor etc. There is a chinese made one I found when searching for one to purchase but which does not have L/s in the measuring which the AC industry measures from supply air registers, and another one called Shortridge.