We have been calibrating and changing meter factors in our coriolis flowmeter for HFO and LPG in our refinery. However, after some time of custody transfer, we experience more than 1% variance in the accuracy of our meters. Then we go back to calibration again and changing meter factors but the same situation occurs. What are other possible concerns that we should look at in order to properly address this recurring situation? What are tests that we should conduct to somehow determine the root cause of this?

 

I'm no expert on Coriolis meters, and haven't been around many new installations. A number of older applications that I have seen used flexible connections to the meter inlet and outlet, to reduce mechanical strain on the meter, and to reduce the transmission of piping/structure vibration to the meter. I have also seen installations with rubber feet serving as vibration isolation mounts.

Good luck.

 

Do we assume the manufacturer has been unable to assist?

We have two options:

1) Is it a meter problem such that irrespective of the fluid, the calibration drifts. In that case, what is the history.... has it always been this way or it is a recent phenomena,; has it been progressive or did it just happen one day?

2) Is it an application problem such that if the meter is calibrated for LPG and then you change to HFO you discover a drift? LPG and HFO are quite dissimilar fluids and it is possible that some factor relating to the fluid or the process equipment related to that fluid alone is causing the problem.

Early coriolis meters had a number of sensitivities that needed to be considered. These included

a) vulnerability to noise where low frequency operating meters could be sensitive to pipe borne noise and to interference from other mass meters; meters that operate at a higher frequency are less suceptible since the noise at that frequency tends to attenuate more the further away from the source.

b) pipe stresses where any stresses caused by misalignbment could affect the reading for which reason the pipe runs before and after should be carefully aligned and supported to prevent distortion.

c) pressure effects. bent tube designs are sensitive to high pressure pulsations, the tubes tend to stiffen and to try and straighten out.

d) zero drift over a period of time the zero tends to drift and requires recalibration
e) flow balance: the flow spltter at the inlet and outlet is a critical factor as it is intended to split the flow equally between the two flow paths.

To a large extent the vulnerability of the meter to these effects depends on age since these are features that have been the focus of much development work over the years.
However, it is worth looking at the conditions that exist when metering HFO and when metering LPG to see if there are any common problems or where they differ.

One other factor to consider is entrained air or bubbles. I would usually suspect the HFO as if this has been aerated in any way then the bubbles will tend to remain in the fluid for much longer. Bubbles could give rise to significant variations. If you are only looking at the performance over a period of time you may only notice a net effect but if you log data at 10second intervals and log denisty as well as mass flow then if you have entrained air you will see this evidenced by an unstable signal. (visit www.viscoanalyser.com/page35a.html you will see some plots of density (and viscosity) from a vibrating element viscometer that illustrate the effects of air both as a dispersion and as large pockets.
However, LPG might exhibit this too if due to the high pressure drop across the meter some gas breakout occurs. Even fine bubbles in small quantities can be a problem as they attach to the tube walls and insulate the flowing liquid from the tubes. This was a problem with vibrating tube density meters where bubbles in the media also caused a density error due to velocity of sound effects (though later versions are far less vulnerable, indeed, the Entrained gas versions can work with pto 100% entrained gases).

If you suspect air then it cold be bein drawn into the system through leaking gaskets on the pump inlet or otherwise, it could be due to the pressure drops across the meter itself causing gas breakout. It could affect both LPG and HFO.
Recent work by BP on a 12" mass meter (Foxboro?)on HFO showed that errors of around 3% could arise due to gas bubbles.

 

I observed our LPG coriolis flow meter (CMF 400) during custody transfer and heard intermittent rumbling noise (minor) in the flow tube. is this normal for LPG? Also, will the stream temperature and pressure effect the accuracy of the meter??

 

Coriolis force flow meter is usually under sized. At higher temperatures and high pressure drop, the LPG can evaporate (cavitation can also appear). You hear as if small pieces of stone or sand hits the tube inside the meter. It causes that driver control is not stable, and of course the measurement is not correct. Check the LPG vapour pressure point and the process pressure and temperature. In some cases the pumps are not sized correctly

 

I think the mechanical vibration causes some volatile fluids to flash off at lower temperatures than normal. I have seen this with Sulpher Dioxide. Also you can get an uneven coating of sludge on the inside of tubes which gives a zero drift.
Roy

 

Hello,
I believe you are flashing the LPG off your LPG thru the meter. You may have to increase the back pressure. It will not be the meter that causes this or the tube vibration as someone else mentioned. It will due to the inherent properties of your LPG. Vp of commercial propane @ 38 Deg. C. is approx. 192 psi or so....it will vary depending on its composition.

Same thing would apply to your HFO. Do you happen to have PT's or at least gauges before and after the meter? You should not have to repeatedly calibrate your Micro Motion meter. Another way to check this is to look at the drive gain of your tubes (you will need a hand-held or ProLink II) while you are hearing the "rock" pinging noise. I would also not be changing your meter factors, UNLESS, you are absolutely positive about the changes. What are you comparing it to?

Hope this helps....cheers!!

 

Hi
Sounds like you already identified gas bubbles as cause of the problem. It certainly sounds like. (I'm assuming this thread is still live/problem exists as no closure by originator)

Perhaps the meter density reading can verify if bubbles exist at the tube measuring section?

Best solution is to irradicate the cause of the gas entrainment (bad gaskets/flashing/...)

To address this at the sensor I can only suggest orientation and location. As a liquid application, the bend in the Coriolis tubes should be below horizontal (assuming we are not in a vertical line here). Orientation nearer horizontal, but still below (rotate flanged sensor to one hole below horizontal) will allow the trapped bubbles to pass through.

Location - of course, make sure the sensor is not in the highest section of pipework, where gas pockets are likely to collect.