Pressure Transmitter for Turbine Back Pressure

120 Deg C approaches the limit of temperature rating for the pressure sensor bodies. The spec sheet on my desk shows one model with a 'rated' limit of 110 Deg C with the operative limit at 125 Deg C.

I read that as continuous duty at 110 Deg C; with momentary excursions to 125 Deg C, with some accuracy error above the rated limit. I'd isolate the transmitter with condensate, regardless of the spec'd temp. It's heat that kills electronics.

I don't know whether your 'backpressure' is a gauge pressure measurement or a DP measurement referencing some other tapped point. If gauge, I'd probably use a DP transmitter and leave the low side open to atmosphere unless you get a gauge pressure transmitter that doesn't need 50:1 or 100:1 turndown (becxause it has a 35 bar native span) in order to get down to the 0.5bar range you need. The typical smart DP transmitter will do your 0.5bar range with only a 2:1 or 2.5:1 turndown from its native full scale.

In practice, steam service DP flow transmitters are almost always mounted below the tap(s) to allow condensate to isolate the transmitter from the steam temperatures.

But gauge pressure transmitters are sometimes mounted on a pig tail siphon above the pipe tap.

The 'above or below' argument is opinion, like what your preferred car brand is. If you use a siphon though, you have to make sure it's got water in it.

If mounted below, then the impulse tube liquid level has be maintained at a constant height; if it varies, it's a direct measurment error. Either an upper elevation Tee or condensate pot should be at or higher in elevation than the tap, so that any condensate between the Tee and the pipe drains back into the port/tap. You only want condensate in the vertical leg down to the transmitter.

Some installations use condensate pots, others don't. They're easy to fill with water. If there's no condensate pot then a Tee at the high point (with one plugged port facing up) is needed for filling the impulse tube with water.

Some installations put a bend in the impulse line running down to the transmitter to allow the bend to absorb the expansion and contraction with comes with heat and cooling. A bend does not affect 'calibration'; only elevation and the head pressure of the water column in the impulse tube affects the measurement/calibration.

When measuring gas,always locate transmitter above pipe,to prevent condensation from entering transmitter.

When measuring steam,always locate transmitter below measurement point to prevent hot steam from contacting sensor element.

When locating below is not possible,install a pig tail fitting and bleed the air from the pigtail to prevent erroneous readings.

(Paraphrased from Combustion Engineering manual.)

I opened the thread to read HiTekRednek's post and realized I'd missed a critical element - this is vacuum range: -1bar.

There's a problem using condensate legs to isolate the transmitter at vacuum pressures because the lowered boiling point of condensate at vacuum pressures could allow the condensate water to flash off, as the table below shows.

The situation depends on the condensate temperature and the applied vacuum.

The condensate level has to remain constant to ensure an accurate pressure measurement. A condensate level in the impulse tubing/condensate pot that changes is reflected immediately as a pressure measurement error.

The 0.5 bar steam temperature falls within the operating limit of the example spec I posted above. If the pressure transmitter you pick has a similar rating, I'd suggest installing it on a length (depends on accessibility, ambient conditions, radiant heat affecting the location) of slightly inclined impulse tubing slanting up from the tap to the transmitter so that condensate drains back into the tap by gravity.

Attempting to maintain a condensate level otherwise, could prove to be high maintenance.

that is precisely how we do it around here, and I presume some young Ghengus figured this out somewhere else a long time ago, because he don't work here no more.