Flow Meter Uncertainty

[triviaux1]

Hi,

A test performed on a flow meter indicate an uncertainty of 12% compared to the reference, knowing that the reference uncertainty is 5%.

Can we substract the uncertainty and assume that the real uncertainty of the flow meter is (12%-5%) 7 % ?

Thank you

 

[1503-44]

No. Using that "logic", if your reference was 12%, you would have a really great accuracy from your meter.

You can only assume that your meter accuracy could be up to 12/5 = 1.4 x worse than your reference accuracy.

You meter will still measure +/-12% from the real flow.


--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[triviaux1]

Thanks for your feedback,

Is there any standard to back calculate a flow meter "real" uncertainity taking into consideration the uncertainity of the flow meter used as a reference?

 

[LittleInch]

Well both of those numbers are truly terrible IMHO.

that's more like an educated guess than a meter.

Also normally you take a wide range of flows from 10% of max to 100% in 10% steps.

Your reference meter needs to be sent back and calibrated against a much much better meter and that uncertainty reduced to less than 1% for it to be of any use as a reference meter.

Or one or the other meters is operating at a small fraction of it's rated flow (<10%)

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[1503-44]

If your 12% meter reading was always within a certain percentage of the reference meter reading, say 2%, it would be statistically possible to infer that the 12% meter was 5+2 =7%, but it would be equally as possible to conclude the 5% meter has a worse accuracy than 5%, having an accuracy of 10-14%, only proving the old adage, "A man with two watches never knows the time."

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[LittleInch]

How was the reference uncertainty determined?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[triviaux1]

The reference measurement is a two phase test separator and liquid measurement was cross checked using a tank

 

[1503-44]

I thought you were talking about a properly calibrated and certified "standard reference meter" as in prover looped, or something. Tank volume measurements are never going to make a reference standard. Notoriously poor accuracy, temperature variations everywhere, pressure variations everywhere. How did you even do that? But it does explain why your reference is not better than 5%.


--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[LittleInch]

What type of meter are we taking about here?

What's a separator got to do with it?

Is this two phase flow?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[shvet]

The reference measurement is a two phase test separator and liquid measurement was cross checked using a tank

Note that part of separator liquid volume is occupied by popping up bubbles so gas-liquid level measurement and liquid density calculation are not simple issues.