Some years back, a DP type flow meter was installed in a Gas line and the linear output of DP Transmitter was fed to a "Linear" integrator. Since then we've been considering those values of Gas flow. The flaw was recently discovered that either the Integrator should have been a Sq. type Integrator or the transmitter itself should have been sq. root type. Now that the mistake has occured is it possible to work upon the "linear" data of yester years and get correct readings?

GST

 

I believe if you restate the data as a fraction of full span, take the square root, then reconvert to engineering units you should be fine.

Actually, what you have is the DP just expressed in unusual (for DP) units.

Example:

Full span is 100"wc DP = 20 lps.

Your reading is 15.3 lps, which is 76.5% or 0.765 (or 76.5 "wc). Square rooting this is .8746 which converted back to flow is 17.5 lps. In all cases, your flow was understated.

In the future, I would recommend leaving the meter as linear, and perform the square rooting in the controller/integrator. This makes instrument calibration easier for the technicians.

Also, if all you have are integrated readings (such as 1.2MG over a day interval, just take the average, i.e., 13.89Gal/Sec and do the conversion as above, then multiply by the interval again to get the new integrated reading. For the example above, 1,321,920 litres over 24 hours is 15.3 lps old reading, which corrects to 17.5 lps over 24 hours, or 1,511,382 litres.

 

The square root of flow is proportional to DP.
So the DP squared is proportional to flow.

Since the square root was not being extracted, the readings from past years are proportional to DP not Flow.

To get the correct flow, just take the fraction of fullscale and square it.

Eg. If fullscale=100 ft3/min and the (prop to DP) reading you got was 50 ft3/min, the fraction of FS = 1/2

(1/2) x (1/2) = 1/4
so the real flow is 100 x 1/4 = 25 ft3/min

Half the pressure drop is a quarter of the flowrate.

GC

 

Correction to my previous post...

The square root of DP is proportional to Flow.
So the Flow squared is proportional to DP.

Since the square root was not being extracted, the readings from past years are proportional to DP not Flow.

To get the correct flow, just take the fraction of fullscale and take the square root it.

Eg. If fullscale=100 ft3/min and the (prop to DP) reading you got was 25 ft3/min, the fraction of FS = 1/4

SQRT (1/4) = 1/2
so the real flow is 100 x 1/2 = 50 ft3/min

Half the flowrate causes a quarter of the pressure drop.

GC

 

I am curious. Your question seems to indicate that you wish to ascertain the correct(corrected) values based on some historical values (uncorrected) that are recorded in some sort on historical data solution.

Am I correct?

CTTech

 

Yes that's what I precisely wish to do. My integrator (Linear) and DP Transmitter (Linear Output) combination has measured values for a certain period of time. Now I wish to ascertain the correct values based on these "incorrect" values. Can it be done? (The integrator values are in terms of Daily Gas Consumption. A Six digit or five digit figure. I want to arrive at a corrected figure from these figures)

thanks for the responses this post received.

 

Please note: The integrator (if connected corectly) gives you the volume of consumed gas for the past interval you say. the integrator works on this principle that it takes a sample of the inputed flow rate (from your dp transmitter) and multiplies it by the interval of sampling time(very short interval of time).

the result is the volume of gas consumed for that short interval. Then the integrator goes for the next sample and calculates the volume in the same way and adds this volume to the previous volume and so on... Here you have made a mistake and instead of flow rate, you have given the square of flow rate to the integrator. so you have got the wrong number in the output of the integrator. There is no way to get the correct number from this wrong number(you should forget it). But I can suggest a way to calculate the approximate volume of the gas consumed for the past period and it is as follows:

1-you divide the past times in some intervals(seasons or months).
2-try to guess an average flw rate for that months or season.
3-multiply this average flow rate by the time of that interval.
4-Add up the numbers you get for all intervals and so you get a rough number showing the gas consumed for the past period.

Good luck.