Orifice plate flow measurement.

[Calistodwt]

Take note of item 4 given by burnerman. Because of the square root relationship between flow, differential pressure, and transmitter accuracy the turndown on a single transmitter is 4 to 1 on flowrate as a maximum.

 

[geniusintraining]

GIT,
You're just a little confused...

That's a under statement...

Thanks mellis, now I got it.

 

[burnerman]

Thanks for the heads up Burnerman, the trouble is that my boss sells these ideas and poor sods like have to make them work!
I would be grateful for some part numbers for ratio controllers that you have used successfully before this goes too far!!

Most burners work well with a pressure to pressure based ratio control, e.g. Kromschroder GI or GIK but if there are significant fluctuations in chamber pressure a DP to DP ratio system may be required, e.g. Landis & Staefa SKP55 or Kromschroder Moduline valves have ratio control options.

There's nothing wrong with electronic ratio control per say, I just find that sometimes it's overkill and you've got to be aware of the pitfalls - gas is a dangerous thing.

 

[pw3]

Hello-

I work in a plant where we meter both NG and COG flows.

First you need engineering sheets listing what your max. DP equates to in Flow. This is vital and should have been provided to you by the O-Plate vendor, or your Project/Plant Engineer. Someone someplace sat down and calculated the DP for the flow range they wanted and bought the plates.

Next, how accurate of a reading do you want? The NG supply pressure greatly effects the flow reading (as well as temp) and should be compensated for-- Considering the cost of NG nowadays I would use a "multi-variable" Flowmeter. Foxboro makes a model now (I've used them all at this point) that can be calibrated rather easily compared to other models. Your DP to flow calculations are entered into this device and the 4/20ma output is compensated flow. Square-rooting is also done there.

Another choice would be a stand-alone flow computer that do this and output a compensated flow signal. KEP makes several (we use the Supertrols) and Prosoft makes a flow computer card for the CLogix. They do however require separate process meters for DP, Pressure, and temp.

If accuracy is not an issue, the easiest solution is to calibrate your flowmeter for straight (linear) DP, then square-root and scale the value to flow in the PLC. I typically use this formula--

Flow coming in (counts): 0/4095; LIN
Range needed: 0/5000 CFM; SQR

(SQR(Flow*0.02442)/10)*5000

The 0.02442 figure is 100/4095, this reranges the counts to 0/100. The SQR function then yields a 0/10 square-rooted range, and dividing by 10 yields a 0/1 range that can be multiplied with the max range to give you the square-rooted range.

You may also want to add totalizers to your flow as most companies are fairly shrewd about NG consumption and it would be a handy tool for them to see how much NG each boiler uses in a month's time.


Good Luck!

 

[danw]

The NG supply pressure greatly effects the flow reading (as well as temp) and should be compensated for-- Considering the cost of NG nowadays I would use a "multi-variable" Flowmeter.

The value of multivariable flow meter applies even more to combustion air, because natural gas fuel usually comes from underground piping that keeps the gas within a fairly narrow temperature range, but the density of the combustion air temperature varies not only seasonally, but day to night and with the weather.

An orifice plate produces accurate flow values only at its "design" conditions, those 'design' temperatures and pressures specified by someone, someplace when the orifice plate was spec'd. Slight deviations from those design temp & pressure can result in fairly large errors as the graphs below illustrate:

I became a believer when I saw the the large differences in readings when running a DP in series with a multivariable (it was one long meter run on 4" pipe). Those happened to be Honeywell transmitters, but the principal is the same.

Dan