Hi there,

We are trying to do the calibration on a de-aerator tower sometimes under max vacuum and sometimes at normal atmospheric pressure. Max vacuum is when the vacuum pumps are running and atmospheric pressure is when they are stopped.

The level transmitter is a normal 3051 Rosemount DP transmitter with instrument tubing connected to the tapping points. The transmitter is installed 300mm below the bottom tapping point and the top tapping point is 400mm above the bottom tapping point. This is a wet leg configuration and the measurement works fine so now problem there.

Here’s the problem.

At atmospheric pressure easy enough to check the zero and do the span calibration.

Now the vacuum pumps start up and max vacuum is applied to both HP and LP legs so the transmitter should still read zero right? It doesn’t. (zero in this case being the negative wet leg pressure on the LP side)

I have eventually found the problem but don’t know how to compensate for it. The problem is, due to some flow straightening restriction plate inside the vessel, exactly between my tapping points, the vacuum on the LP leg is slightly different from the HP leg. Due to the small span, this slight difference have quite a dramatic effect on the transmitter’s zero and therefore span measurements.

Ok easy enough, with constant max vacuum on both legs, use that negative value as zero again, even if the two vacuums are different the meter will still work. Fair enough I have done that, and the meter works, but now when the vacuum stops the meter show a incorrect atmospheric zero reading again due to the negative zero value used with the two different vacuum pressures on the HP and LP legs. So we have now taken the error of the process zero and moved it to the atmospheric zero.

If we don’t fix this production will continuously call us and tell us the meter is wrong since it now still show a level even after the pumps are stopped. If we correct the atmospheric zero the meter will give a inaccurate level measurement during vacuum again, now what?
It is not possible to move the meter to another area so that is not a option.

Your thoughts and ideas on how to compensate for this error will be appreciated. Can you think of a way to maybe split this error?

 

You don't say where the level calculation is being done. If in a DCS or PLC, then you could automatically switch between the max vacuum and atmospheric offsets based on whether or not the vacuum pumps are running, assuming the vacuum pumps status is also monitored in the same DCS or PLC.

 

Hi there,

I thought my post was clear enough in that it is a normal DP transmitter calibrated for -700 to +300mmH2O and 4 to 20mA going to the DCS. So there is currently just the one signal to the DCS.

Your reply however did give me another idea and that is to install a second transmitter and then it might be possible to switch between the two in the DCS depending if the pumps are running or not. The drawback of this idea is that approvals will have to be done in head office as well as new cables and drawings for the new loop. I will leave this as a last resort. Anyway thanks for the response.

 

Assuming that you are measuring the bottoms level, at least for the dearator towers in my experience, the low side connection should be connected below the plate you've described.
good luck,

 

Sir, i think you have misunderstood otised, You can use the software function of the DCS, use the software to make a compensation

 

Hi again,

Ok I get it now about the offsets and switching between them with the stop and start of the pumps. Thanks, I think this will work since the differential vacuum is there and nothing can change it.

The measurement is only 300mm half way up the tower and not at the bottom, sorry should have mentioned this but regardless where it is, the calibration is not the problem it was this differential in HP and LP leg vacuum that kicked my butt all over the place.

Thanks to all, your replies have all helped and I think we will be able to get it working accurately now during run and stop of the pumps.