Problem with Level Measurement of a Vessel Under Vacuum Using DP Transmitters

screamingsilence:

(1) We use dP transmitters on hot wells on all our power generating units.

(2) the main requirement on the reference leg is that it remains full to the level of the return pipe on the hot-well (beneath the condenser tubes). This can be by maintaining the reference leg in a cold state (well below the hotwell temperature). The sensor itself will keep fluid from flowing out of the reference leg (unless it is a faulty sensor with a bad diaphragm or something like that). The flooded (reading) side is just what it sounds like: flooded constantly by liquid condensate from the hot-well that exerted a head pressure dependent upon the level of condensate in the hot-well.

To answer the leak detection question: Yes there are several ways to undergo leak detection:

(1) Make sure the turbine back-pressure is consistent with the observed turbine exhaust temperature - since you are taking vacuum pressure readings on the hot-well/condenser. If not perfectly consistent, then you need to figure out why.

(2) Make sure oxygen is not ingressing the system (and other non-condensible gases) through excessive requirements for make-up water (loss of condensate return, steam leaks, excessive boiler blow-down rate, etc.). This is evidenced by a high residual concentration of oxygen (above an arbitrary number of say 30 ppb oxygen in water solution) in hot-well condensate.

(3) If the source of non-condensibles is not make-up water rate, then you are most likely dealing with a direct air-ingress on the gland seals, at the turbine exhaust, through allowing air leakage into a sub-atmospheric pressure steam extraction (if one exists on your system), but these are seldom employed. Other sources of air ingress can be through leaking condensate pump seals, and even through high pressure boiler feed-pump seals (as a result of a Bernoulli effect at points where high fluid velocity exists). Leak detector systems can be brought on site by selected vendors to diagnose your systems ( I am not offering to provide advertisement for them here).

One old option that was used in power plants back 30-40 years ago (before we got all sophisticated and starting using Mass Spectrometers) was simple shaving cream applied to the area where the leak was "suspected" to exist. If that area is leaking, it will suck in the shaving cream. Now either helium leak detection or sulfur hexafluoride gas is used with a fancy detector, by spraying the test gas near the suspected leak, and waiting for the detector operator to declare a "hit".

This is one area that cannot and should not be ignored with respect to cycle chemistry, in that leaks or air will cost downtime, and already cost in lost power generated. Sufficient air binding can even exist in the larger condensers of big plants, that whole sections of condenser tube become "air-bound", and the exhaust steam from turbine cannot reach cold metal, hence it is as though the condenser were heavily fouled. One can also use effective U coefficient of the condenser as a long-term diagnostic to the quality of vacuum in a system in those cases.

I don't think you will have to apply a cooling coil over the reference tube of the dp transmitter, but if all else fails, that could be an option.