water balancing accuracy/tolerances 1

[Waramanga]

Hi Guys,

I am looking at the CIBSE code W for flow tolerances. if my coil needs a flow of 0.1574XXXX l/s and I have a tolerance of +- 10%, what decimal point goes in the coil schedules? I am thinking at if the flow needs to be within 10% then I need the third decimal place, ie 0.157 l/s, because +10% would be 0.172 and -10% would be 0.142 and without the third decimal place you could be anywhere between 0.140 and 0.149. Am I correct in concluding that if a accuracy of 10% is required, three significant numbers are needed? Does this make sense in that Kv numbers on balancing valves only go to three significant figures?
Would someone be nice enough to confirm or explode my understanding?

Thank you

 

[KiwiMace]

Don't specify unmeasurable quantities.

 

[DRWeig]

Kiwi is right, a third significant digit on a flow measurement is a random number unless you're in a laboratory. Often, the second significant digit is questionable as well.

You can figure out from the pipe size and configuration, fluid velocity, measuring instrument accuracy, and measuring instrument precision what sort of precision in flow rate can be realistically expected at those low flow rates.

Or, just call it 0.16 - 0.2 l/s so they err on the upper side for safety.



Good on ya,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[Waramanga]

Thats what I was getting at but i was hoping for a more specific answer. As far as i can tell, its possible to measure down to 0.01 Kpa on an alnor manometer, so if you put that kind of resolution through the formula for flow through a STAD valve with Kv of 1.47 (10mm valve), you you can measure a flow variation of 0.0000X l/s at 100kpa dP which would be around 0.006 Kw on a hot water dT of 15k. So its measurable but not significant.
At a 1Kpa dp accross the smallest valve, the smallest measureable change in flow is 0.00X l/s which is .6kw or so, which is clearly significant on a small terminal unit.

At this point I am leaning towards four decimal places for the smallest flows of 0,0Xl/s, that way on the smallest measureable flows on the smallest stad valves the flow can be checked to be within 10%.

any other toughts on this?

 

[KiwiMace]

For general purpose HVAC use, I would go 2 sf at this flow range, and most other ranges <100 of any units. Above about 100 I would be inclined to be rounding to nearest 5. That's a 2-3% error right there, but well within what they will be operating to.

While you may be able to "measure" 5dp, you will have no accuracy or repeatability even at 3dp. Secondly, even if you could measure it, you can't control to it. Fluid flow itelf is not that stable. Watch the meter for a while. Try closing the valve and opening it back up to the same position.

 

[DRWeig]

I was a little more conservative the Kiwi is discussing since coil control (or balancing) valves are typically very close to piping elbows, service valves, and the exit from the coils themselves. As he described, that puts a real strain on accuracy and repeatability. Note that I'm a controls and design guy, not a balancing guy, though.

I usually try to err on the upper side of the spec, in order to insure that I don't leave a coil with under-capacity, which leads to call-backs. But not too much, which may lead to not enough available flow to satisfy all coils. Hopefully, there is enough excess pump capacity to push all coils above spec a few percent.

Good luck on your project!

Good on ya,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[KiwiMace]

Agree with your setting at the high end. The control valve might close down a bit from full open, but they don't open much more.

 

[Waramanga]

thanks guys, kiwi's point on repeatability and actual field variability is interesting. I will make contact with a commissioning specialist I know (30 years in the industry and now an independent reviewer of projects). He will know for sure. Not withstanding what is repeatable etc, I am thinking that if you specify the flows at a resolution that can be theoreticaly obtained with the valves and instruments, there isnt any harm done right?
My conclusion at this point is three significant figures so you have the resolution to see the +-10% effectively.
I will keep everyone posted.

thanks