Steam DP Flow Measurement - Long Impulse Lines 3

[Niwot]

Looking for anyone with experience in DP steam flow measurement with impulse lines (wet legs) 30-40 ft. in length. Having issues with the steam flow tracking the condensate flow. We're looking at a possible imbalance in the liquid legs that isolate the DP transmitter from live steam. When the legs are filled and the transmitter is zeroed at start-up, the steam measurement is reliable. After a certain period of time, the steam measurement tails off (appears to read low) and doesn't track the condensate output in a closed loop system. The DP transmitter is mounted below the primary DP flow sensor and the tubing (1/2" SS) slopes continously downward. I believe that the vent valves of the DP transmitter are oriented correctly to avoid trapped air in the diaphragm. In addition, we have found no evidence of a leak path across the equalizer valve in the manifold.

Additional Notes: Pipe Size: 30". Pipe run: Horizontal. Steam Pressure: 100 PSI. Steam Temp: Slightly above Saturation. Reading low DP's during turbine test phase(understanding that there will be an error induced at the liquid/vapor interface in wet legs due to a dynamic equalibrium). Condensate (reference) Meter: Mag

We have three other sensors in service with no problems, the only difference being that the liquid legs are only 4-5 feet in length. For what it's worth, we're measuring geothermal steam (probably not related). Has anyone experienced issues with running liquid legs 30-40 feet or more using a DP flow meter on steam? We are not concerned with absolute accuracy at start up. The problem is that the steam meter tracks the condensate meter, then falls off dramatically. Any help or advice would be greatly appreciated.

 

[hacksaw]

sketch up your installation it doesn't sound like an installation commonly required for saturated steam

how much condensate are you losing done stream of the meter & how much superheat are you using

 

[danw2]

If "steam meter (snip) falls off dramatically" means that the steam flow values decrease from expected values (compared to condensate), then it sounds like the steam's high side impulse fill is flashing off more than the low side, resulting in a low DP/flow reading.

A note in "Best Practice Guide: Impulse Lines for Differential-Pressure Flowmeters" by UK's NEL group, Page 13, section 2.9:

Steam recommendations:
Spirax Sarco do not recommend the use of condensate pots except for superheated steam applications . . .

http://www.idc-online.com/technical...entation/Differential pressure meters_nel.pdf

Do you have condensate pots installed?

 

[Niwot]

Hacksaw, the steam line is 30-40 feet above grade with no scaffolding - DP transmitter is at ground level. We try to make sure that legs are filled and bled (no entrained air). I'll provide a sketch or digital photo of installation.

Dan, thank you for the valuable link. We do not recommend condensate pots either. Visiting the jobsite in a couple of weeks to see if they were installed any way.

What you're saying about the high side flashing and driving down the DP makes sense.

I have some charts showing steam vs condensate measurements. I'll try to post tose as well.

Thanks guys.

 

[hacksaw]

use larger diameter impulse lines, say 1/2" ss tubing, small bore impulse lines can be a real issue

you may also need to consider pre-filling filling the impulse lines from grade, you'll need block and bleeds at the meter connection and at the transmitter for that and a hand pump for the fill fluid

with saturated steam cannot imaging that flashing a a problem unless you are experiencing large swings in pressure

by the way, are you using drain holes on your plate...if not the condensate build up up stream alters the meter coefficient, resulting in less d/p for a given flow...

 

[EthyleneInstrument]

it can be condansation problem with low side of transmitter.
If u sketch up your installation with writing of elevations
we can help you.
Thankyou

 

[Niwot]

I've attached a scan of the charts for the "calculated" vs. the measured steam. I was led to believe that a mag meter measuring the condensate was the reference for the calculated measurement. I just found out that there may be a "fudge factor" at the user's end to arrive at this number (I don't know what that means yet). I was originally told that the three steam measurement points were lower than the total condensate measurement by 10-13%. I apologize for the resolution, but the pens on the charts are light (red and blue). I'm not concerned with the charts HP and LP. They track the calculated flow and I learned that these numbers may come up and track closer based on the actual density. The SP chart is the "problem child". It tracks the calculated value then falls off. The transmitter is re-zeroed and it tracks again then falls off again. This is the installation where there is 40 ft. of 1/2" tubing leading down to the DP transmitter (the other 2 meters have 4-5 ft of tubing). The photo of the sensor in the pipe is that of a hot tap averaging pitot tube with HP and LP instrument lines. They couldn't get me a photo of the actual meter (40' above grade), but it looks identical (same DP transmitter installation as well). Note that the pipe is insulated but there is no insulation around the sensor or impulse lines. Steam is saturated (pressure = 117.6 PSIG Temp = 348-350 F. Not sure how clear the photos will appear, but I can provide a sketch if this will help.

 

[Niwot]

Hacksaw, we will make sure that we pre-fill the impulse lines from grade as you suggest. This is what we recommend, but often it is not done this way in the field. Thanks for the reminder. To answer your other question, it is not an orifice plate but it works on basically the same principal and equation except it averagaes the velocity using two separate chambers (HP and LP).

Thanks.

 

[hacksaw]

Niwot, you have an annubar or averaging pitot measurment from the picture,

in steam we always put them in the lower quadrant of the pipe i.e. +/- 40Deg from bottom, you are close enough, the big issue with these meters is low d/p and increased sensitivity to level off-sets

check you vendor drawing of the device if it is available to see the installation guide lines.

good luck with your project

 

[roydm]

I have seen something similar with long impulse lines, I put it down to slightly different density in the fluid, remember with square root extraction the DP cell is very sensitive at the low end.
The solution was relocating the transmitter to just below the DP element
Geothermal steam - perhaps something like sulphur settles out in one line more than the other.
You don't mention condensate pots, these are very important for a flow application, just using a tee wont gaurantee the level in both lines is the same

 

[Niwot]

roydm, Thank you. And thanks to all who have contributed to this thread. I'll let you know what we discover.

 

[lacajun]

I've run long impulse lines in steam with no issues.

Pamela K. Quillin, P.E.
Quillin Engineering, LLC

 

[Niwot]

Thanks for your responses. I've returned from the sight with some photos. Although NEL recommends the shortest impulse lines possible for gas and steam flow, I am in concurrence with Pamela and we have had success with with long impulse lines (where it can't be avoided). Under the new Eng-tips format, I'm trying to figure out how to post photos. When I figure it out, I'll post some.

I can tell you that the major problem we found was that the impulse lines ran horizontal from the DP instrument head with no slope for roughly 30 feet (with 3 horizontal bends) BEFORE sloping downward roughly 45 feet to the DP transmitter. For liquids and steam flows with liquid legs, the impulse lines should slope continuously downward (at least 4 in. per foot) free from peaks, dips or sharp bends where entrained air can collect).

 

[lacajun]

I ran my impulse lines straight down to 4' above grade from the orifice plate way up in the pipe rack. I'd have to think about the horizontal run's impact on the measurement because head is head and horizontal doesn't contribute to head, which you know better than me.

Pamela K. Quillin, P.E.
Quillin Engineering, LLC

 

[roydm]

I don't know why you would run long distances with the impulse lines, it's asking for trouble.
Remember that the signal is square root, it only takes a fraction of an inch to throw the zero out the window.
Transmitters are for transmitting signals so use it.
I also feel condensate pots are essential for a steam flow although "Experts" seem to feel to the contrary.

Roy

 

[hacksaw]

the length of impulse lines is often driven by factors beyond your control, but you are tasked to obtain a good measurement none the less, it's the nature of the business

 

[lacajun]

I used condensate pots once in an application where others were pushing for very short tubing runs, which meant the transmitter would be very close to the pipe. To ensure the transmitter would not be exposed to high temperature and condensate would form, I used condensate pots. Shortly thereafter, that company began to realize they needed to stop the practice of mounting transmitters in pipe racks for safety reasons. I was trained early in my career to never mount a transmitter in a pipe rack for safety reasons primarily. This plant had other installation issues that caused premature transmitter failure but helping them understand the correlation was non-trivial.

Pamela K. Quillin, P.E.
Quillin Engineering, LLC