dynamic error- pressure sensing furnance draft snubber

[davefitz]

Kindly advise if you can provide a reference that describes how to determine the dynamic sensing error of a pressure sensor with long narrow sensing tube.

IN our application we have a large coal fired furnace. The furnace pressure ( "draft")is sensed by a pressure transmitter, but this transmitter is separated from the furnace by a 30' long x 1" dia sensing tube. Large changes in furnace pressure ( from 14.7 psia to 13.7 psia ) , as during a main fuel trip at full load, are predicted to casue a 1 psig drop in pressure within 2 sec but the sensor is not detecting more than a 0.5 psig drop in pressure.

I hope I don't need to make a 1 dim finite diff model to determine this error- this type of error must have been addressed by now.

 

[itsmoked]

It's a fluid in a tube.. The whole tube should experience the same pressure as I think you expect.

A couple of points to consider:
1) If the tube is not angled correctly in the stack you could get a ram effect or some other problem that masks the change you expect.

2) A tiny leak at the end of that tube would make a big difference.

3) Your sensor could be bad. Have you put an actual manometer on the end and checked what's really being seen at the sensor?

4) While the pressure should be the same in the entire tube it won't be the same instantly. I suspect it will follow and exponential decay. Perhaps your sensor is NOT sensitive to the slow end of that exponential.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[davefitz]

itsmoked:

The sensor at the end of the tube is expected to read significantly lower magnitude transient pressure than the actual furnace pressure during fast spikes or transients- the long run of sensor tubing will act as a "snubber" or "gauge saver" . Normal furnace draft is controlled at -0.25 in wc, and a transient to -30.0 in wc is not within the range for the presure switch or sensing element. So, wehn the furance pressure actually spikes to -30. in wc during a full load MFT event, the pressure recorded by plant data archive system does not acurately reflect the true pressure spike.

I am searching for a standard error correction formula to allow me to confimr that the order of magnitude or the error is 100% of gauge reading for the described event.

 

[itsmoked]

davefitz; I see now what you are after.

I will say your plan seems shaky to me as the result you are reaching for is going to be very dependent on the time that that event lasts. A -30 event that lasts 1 second may indeed be integrated into a -27 event whereas the same -30 that lasts 2 seconds may translate to a -29.5 event. Seems that you can't know the actual value without somehow knowing the actual time. I am guessing you don't maybe care as perhaps anything over -26(?) is fine to be called a full load MFT. What about a 50% load MFT?

Might I suggest you cross post to forum378 as I believe you will be more likely to gain the chart/table/rule you seek.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[davefitz]

Well, I estimated that the order of magnitude of the error is negligible in this case.

A draft excursion to -27" wc implies that 7% of the air in the sensing tube must flow into the furnace in the 3 second period of the event. The frictional loss of that mass flow is truly miniscule and no correction is required.

Thanks anyway.

 

[hacksaw]

there was a lot of theory and experiment on this in the early '50s by Page Buckley.

you need to check the damping adj on the transmitter, the low draft range types have a lot of damping just to stabilize the reading. to monitor the 30" swing, you might have to go with another transmitter in parallel.