Steam pressure sensing line noise 3

[DickWGINT]

We have been asked by a client to suggest changes to the existing pressure sensing line configuration to reduce the pressure oscillations seen in the transmitter output. The transmitters in question are mounted at almost the same elevation as the taps on the steam line and are located many feet away horizontally. Moving them is not currently considered to be an option. The transmitters use a wet sensing leg with the tubing routed up to a point approximately 15 feet above the transmitter using a series of vertical and sloped horizontal runs and then back down to the instrument itself. There are no condensate pots installed on these lines.
The determination that there are excessive pressure oscillations generated in these sensing lines (flashing of condensate is one suggested explaination) is based on comparison with another transmitter monitoring a similar point which was relocated years ago such that the line tees off of a 2 foot vertical capped section of pipe about 1 foot above the top of the steam line and then slopes almost continuously down from this point to the transmitter.
In all cases the sense lines are 1/2 in stainless steel tubing and are uninsulated after they leave the steam line.

Our client is looking for some assurance that we understand what is generating the additional noise and, based on this understandin, suggested modifications to reduce it short of relocating the transmitters.

 

[Skogsgurra]

I would look for non-steam issues. Like EMI.

Gunnar Englund

http://www.gke.org

 

[m777182]

We had excessive HF oscillations on a pressure transmitter monitoring air flow in a dryer. We put an pneumatic RC filter into the sensing lines with great success.Later it was find out that the oscillations came from mechanical vibrations of the entire dryer.
m777182

 

[Scotsinst]

skogsgurra could be right. I would however eliminate the 15ft rise of the impulse line. Noise could be caused by condensate running back down the impulse line and burping its way back to the pipe.
Steam flow measurement hookup is similar to liquid and the transmitter should be about 2 ft from the taps, lower than the taps with no pockets.
Impulse line is insulated for personnel protection and heat traced if minimum ambient is below freezing.
If they won't move the transmitter then route the impulse line with a gradual slope (as low as possible) to a horizontal condensate pot above the pipe and drop it down to the taps.

 

[JimCasey]

I agree with the suggestion of change-of-state.

At least at the vapor/liquid interface the liquid is at saturation. If a slug of condensate starts back down the pipe and pulls a small vacuum it can cause the condensate in the line above it to flash. Also when a slug of condensate starts back down toward the steam line, it can be reheated and flash, propelling a part of the slugh back out toward the instrument and causing a pressure spike.

I hate it when customers want me to fix stuff but they won't let me change anything. Or they want to repeat the same mistakes they made originally with a different label on the components, as if their system can read.

Come off the steam line preferably in the upper quadrant with an enlarged-diameter line (1/2 NPS) up to a condensate pot. Slope this line-do not let it be vertical or horizontal. Protect operators with a heat shield-not insulation. Look at the heat shield on a truck muffler for inspiration. Then the water-filled impulse line from the condensate pot slopes downward to the instrument. MUST slope at all points-particularly no high points are allowed. Prefill the line from the condensate pot to the instrument, and you are more assured of a bubble-free fill if you fill it from the instrument to the condensate pot.

 

[danw2]

First, I'm not exactly clear as to whether this is line pressure (gauge pressure) or dual impulse lines for flow measurement (DP). The term "lines" is used, so I'll assume it's DP with dual impulse lines so that's why there's a rough simulation of an orifice plate shown in the steam.

Although you described the impulse tubing as a wet leg, and it does have steam in it, the reason there is no condensate pot is that there is no way to hold liquid in a condensate pot. The steam that condenses in the rising leg A (see below) is going to flow by gravity back into the steam line connection. I suspect that the noise the customer is seeing is bursts of condensed water reheated to boiling, as described by SCotsinst & Jim Casey. I'm actually surprised that there is no complaint of continuing, increasing offset error caused by condensation in the B leg.

The reason the other situation works is that the entire C leg is a filled with condensate.

Usually steam flow DP transmitters are installed below the elevation of the steam line so that the impulse lines can remain filled with condensate. The alternate installation achieves that by making the C run an impulse line with the transmitter below the tap point, so the condensate is trapped.

My suggestion is to create a condensate trap, fairly close to the steam line. Something like either a pigtail siphon (pictured below) or a section of pipe replumbed with a vertical drop to accomplish the same thing (bold pipes below). The vertical drop leg trap will hold condensate at elevation D and prevent steam accessing the entire length of the A riser impulse line. Doing so, preventing steam from accessing the entire run, will limit the condensate that flows back into the line.

A filler tee can be installed so that water can be poured down the A line to fill the trap. Any excess will just flow back into the steam line.

The closer the trap is to the steam line, the less impulse line remains for condensate to form and flow back into the pipe.

Dan

 

[DickWGINT]

I want to thank everybody for responding, particularly Scotsinst, Jim Casey and Danw2. Additional information that I didn't provide in my initial posting is that the instrumentation in question consists of gage pressure transmitters and pressure switches monitoring 1000 psi superheated steam in a nuclear power plant. The transmitter signal loops are 4-20 ma current loops, primarily Rosemount 1151s. The instent is to maintain the sense lines to the high points as wet legs by back filling with demin water after maintenance. The responses confirmed our opinion that the excessive pressure oscillations seen in the transmitter outputs by the client are due to condensate flashing in the sense lines. One of the 4 installations was changed years ago to a configuration similar to the working installation diagram provided by Danw2 and this configuration has about 1/3 of the oscillation amplitude seen in the other transmitter outputs. However, as stated in the initial posting, the client is reluctant to make similar changes to the other installations unless there is an absolute guarrantee of significant improvement because of the large cost of running new cable and tubing.

 

[danw2]

I don't know if this is far-fetched or not, but there's an NRC document on the web that goes over problems found in a steam pressure impulse line where there is steam and water (two phase) and the "instrument line itself (was) basically interfering with our ability to accurately measure what was happening in the main steam system." (page 46)

The measurement in question had an pressure impulse line 100' in length. (page 46)

They had a condensate pot (transcribed as condensate part (sic)) with liquid leg to transmitter (page 63).

These are the notes I made a year ago when I stumbled onto it. Something about acoustic frequency noise creating fluctuations (page 45). There's lots about instrumenting with strain gauges and accelerometers to get frequency signatures.

Is the case here? Dunno. Too esoteric for my run-of-the-mill problems. Dan

http://www.nrc.gov/reading-rm/doc-c...arch="steam "condensate pot" practice theory"

 

[danw2]

Thanks for the details, Dick. I thought it was 14# low pressure steam flow.

I fail to see how the A leg (in the 'At present' sketch) can be a wet leg. What am I missing?

The B leg is dead-ended at the transmitter, so it will hold liquid up the 15' elevation point.

But the bottom of the A leg is open to the pipe. So the A leg has to be a drain line, even if prefill liquid comes out in glugs (like when one tries to empty of full 5 gallon gas can: glug of gas, glug of air, glug of gas, glug of air: glug, glug, glug, glug . . . . .)

How can A be a wet leg? (without re-plumbing the whole thing)

Dan

 

[Skogsgurra]

When there are problems like this, the poor sparky often is blamed for everything. I am glad that you (for once} found something else to blame.

Gunnar Englund

http://www.gke.org

 

[DickWGINT]

Dan

I probably mis-led you. The A leg isn't wet, as you pointed out, but rather a steam/water mixture depending on how far away from the main steam line tap you are. It is interesting that you located a NRC proceedings document that mentioned that their sense lines were over 100 feet long and that they couldn't correlate the pressure oscillations to what was happening in the main steam line. The lines that we are analyzing are typically 150 feet from the transmitter or pressure switch to the main steam line tap. As a result there are some fairly long horizontal runs with minimum slope (13 ft horizontal with 6 inch change in elevation in one case) but I think most of them are on the water side of the high point. The transmitters which were relocated such that they are below the main steam line tap, and which have significantly reduced oscillations in their output, are also much closer, approximately 50 feet, with the line continuously sloping down from the "t" located about 2 feet above the steam line tap.
I appreciate your continued input on this. It never hurts to get another opinion.

 

[danw2]

1) I withdraw my suggestion for the trap - not a good idea.

2) Re-read the post from m777182 above - vibration induced high frequencies. On a 150' run, with all that 1/2" tubing hanging or fastened to who knows what, is that a possibility?

3) Is the switch activating on the noise pressure pulses? Is it an electromechanical switch (diaphragm actuated microswitch) or one of those newer electronic sensor/solid state switches? If electromechanical, is the mechanical sensor filtering/damping the noise?

4) Is damping an option? Either pneumatic or electronic?
Sometimes filtering noise, as opposed to getting rid of the noise source, is a solution.

a) There are commercial pulsation dampers, like the one at

http://www.midwestinstrument.com/limiting/limiting_main.asp

Simple - a tapered needle valve and a check ball.

b) The 1151S can run from 0.2 to 16 second time constants on its output. What is the setting now? Does it have to be that low?
Rosemount 1151S manual (page 6):
Damping
Numbers given are for silicone fill fluid at room temperature. The
minimum time constant is 0.2 seconds (0.4 seconds for Range 3).
Inert-filled sensor values would be slightly higher.

Output Code S (S = 4–20 mA with Digital Signal based on HART Protocol)
Time constant is adjustable in 0.1 second increments
from minimum to 16.0 seconds.

Dan