GE Ultrasonic flow meter

[rcmac21]

I am using a GE ultrasonic flow meter that measures flow through 2 ultrasonic transducers. I have never used one like this before and it works good until the flow rate gets toward top end of scale. It is suppose to measure 650gal/min and I can not get above 425gal/min. It is showing signal strength and volecity but these values all jump around after approx 425gal/min. I am moving liquid at 200C of sodium metal and 5barg line pressure. I have instead of 10 times, I have 20 times pipe diameter before flow meter and 5 times after. This is 4" SS304L pipe and test loop holds about 20 gallons of liquid that simply circulates. Anyone who has any experience with this type of testing I would appreciate any thoughts as to trying to solve issue.

 

[MiketheEngineer]

Call GE??

 

[rcmac21]

I have contacted GE and they do not have anyone so far that has any experience with media of sodium metal. They have looked over the problem but they have a standard set of solutions to try and that did not yield successful results.

 

[btrueblood]

Any chance you are seeing a laminar-to-turbulent transition in pipe flow profile at around 425 gpm? What is the Reynold's number for the flow condition? Dunno, just a WAG...

Are the probe ends protruding into the flow, and is this the doppler type meter, that beams a signal along the flow axis?

 

[btrueblood]

Sorry, one last WAG. The type of hiccup you describe sounds like cavitation or entrained air, both of which should not (better not) be happening with liquid sodium. But is there some tramp gas space in your line, and could this trapped gas become entrained into the flow once you hit a certain velocity?

 

[rcmac21]

The reynolds number at 425gal/min is 32,280. It is funny you mention laminar flow because we have been discussing installing inside the pipe 8" long tubes all welded together becasue I need to install a gate valve that causes cavitation. The cavitation makes a sound that sounds like gravel and vibrates the piping. The valve is not currently install, I just need maximum flow first and I had enough cavitation that flow meter read -volecity and -flow. My flow meter is not a doppler, it is a flow cell with two transducer bouncing signal off opposite pipe wall and signal traveling in both direction back and forth. I do have nitrogen holding fluid in overflow tank in compression. I do this for expnsion because I need to run test to 600C, once I achieve max flow. I have to apply pressure to sodium in order to get signal strength to read at any flow. I assume this means I have some type of nitrogen in my line that I am compressing to get good signal strength.

 

[jmw]

Swirl maybe?
You have a succession of tight elbows. Possibly a flow straightener "eg crate" bundle or simple static mixer element, enough to break up the flow prior to the meter and far enough upstream so the flow can become more "organised" by the time it reaches the meter.

The other thing is, running at high temperatures, how good is the insulation? What are you using - calcium silicate? (In too many installations insulation is applied to protect operators and not to prevent significant heat loss.

The effects of poor insulation on high temperature installations with sensitive fluids can be profound and I have seen some bitumen installations that have suffered badly from an open window on a cold day close to a badly insulated pipe.

Temperature gradients may cause you to get some nice tunnelling effects with layers of fluid by the walls cooling, thickening and slowing and the flow increasingly down the centre of the pipeline.

This may mean you have quite a velocity differential across the pipe diameter and this may also lead to some unstable flow regimes at the meter at higher flow rates where the flow breaks up boundary layers briefly and which then re-establish themselves only to be broken up again.

Again, a static mixer may help this. Possibly Vortab or similar.

Not sure that GE should necessarily need liquid sodium experience to diagnose possible causes of flow measurement anomalies. They must know what causes such effects and then look back into the application to see if those causes could arise. But granted, experience of liquid sodium would speed up the diagnosis, so they may just be being cautious.




JMW

http://www.ViscoAnalyser.com

 

[rcmac21]

I have wrapped pipe with trace heaters and wrapped that with fiber wooven insulated fiberglass tape. The insulation that wraps this and pipe is ceramic fiber insulation rated to 800C. This ceramic insulation is very dense and is 1" thick as anything heavy would not wrap pipe tightly.
Yes my 90-elbow are fairly tight and I will look into the vortab concept.
My fluid max volecity is about 11ft/s and meter is rated for max 40ft/s. When I loss signal strength, I also losssound speed, volecity and flow. So once signal strength drops there are no valid valves.
My time issue with any changes to loop is this, it takes me 5 days to drain loop of sodium, cool loop down, clean sodium out, fix issue, heat loop up, pressure and vacuum test and then fill loop. Trial an error is a long process and I try to make sure I can make as many changes ay once as I can. This is very different from my past where we never made more than one change at a time.