Differerence in flow 1

[SonnieTP]

What could have cause this to differ in flow?

Our waste water centrifugal pump is discharging water at a normal rate, there is an installed flow meter on the upstream of the waste water disposal approximately 15 feet away from the pump discharge and the flow rate is 225 gpm.
On the downstream of the waste water line that goes into the injection well, 9.6 miles away, we also have another flow meter installed. The reading that I am getting on the downstream flow meter is 166 gpm.

I have both flow meter calibrated but the readings are far apart of each other. What could have cause this uncertainty in meter reading?

The downstream meter location is 9.6 miles away from the pump with an elevation of 1000 ft. The upstream location had a site elevation of 1330 ft. The pipe is a 6 inch carbon steel.

 

[katmar]

BigInch has highlighted a very important point in that the downstream meter is probably at a higher pressure than the meter at the pump. This raises the question of why put the pump where it is and have the whole 9.6 miles under high pressure. Would it be possible to drain the water by gravity to a reservoir tank at the injection well and place the pump there to give the required injection pressure? It also means that air would behave the way artisi suggested (i.e. be compressed along the pipe) and this could be a factor.

However I suspect the most likely cause is bends near the meters. According to the Rosemount manual "No K-factor correction is required if 35 straight pipe diameters upstream (35D) and 10 straight pipe diameters downstream (10D) are available." You have stated that the upstream meter is 15 feet from the pump discharge - which I suspect would be much worse that a pipe bend. Have you corrected the K-factors accordingly?

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[jmw]

Let's assume some things:
Let's assume these are the meters installed 9 years ago when the pipeline was installed.
Let's assume they are identical and the displays are identical.
Let's also assume that when first installed they agreed within acceptable limits.

So we need to ask why they now disagree.
Have there been any significant operational changes?
Any significant installation changes?
Has anyone been doing anything to the meters at all for any reason?

Is the upstream meter flow rate (225g/min) consistent with past history?
If it is suddenly indicating higher flows than previously why? If not then assume this meter is accurate. 225g/min is being pumped into the pipe (is there perhaps a small increase in flow? e.g. due to the lower pressure drop caused by some of the flow finding an easier path)

Is the outflow flowrate lower than usual?

The tests suggested above can and should all have been done prior to sending the meters away for testing as that takes time and may be unnecessary. At around 60g/min that's about 600,000gallons a week.... that's a lot of water to go missing.

Why have two meters? On long pipelines it makes sense as this is a good way to check for significant leaks.

I can't think there is a bit of Chernobyl syndrome evident.
Sure, you need to be sure the meters are correct but if the upstream meters reads more or less what it always does and no one has messed with the installations, and swapping the meters shows the error staying with the location, then maybe there is a leak?

These are Vortex meters. They have no moving parts and the accuracy should remain stable over time. You would not expect one or other meter to show any significant calibration shift, especially not one compared with the other. Failure is likely to be complete failure due to component malfunction or there would be some erratic behaviour.

JMW

http://www.ViscoAnalyser.com

 

[ione]

Could the measure be affected by possible cavitation phenomena?

 

[BigInch]

Would't think so, especially if the meter is after the pump somwhere in the discharge where there is high pressure. A cavitating meter would probably never show one flowrate long enough to read it.

Sit back for awhile until the OP posts the pressures and probs the dirt. If he doesn't have those, nobody's going to figure out what gives here.

It's not always instrument error.

 

[tr1ntx]

I agree with Katmar, since the OP stated that it has been like this from the beginning, and since it is ~15 ft from the pump to the upstream flowmeter, it makes me believe that it's an installation issue. Either the meter won't ever read properly at that location or possibly needs large correction factors. I've seen ghost-chasing too many times on flowmeter issues that come down to basics. I guess I had a good professor over Fluids Lab, that was one thing we studied: installation, bends, reducers, turbulence, etc. effect on meter's accuracy/ability.

 

[BigInch]

15'/6" = 30 diameters. If it's straight pipe, shouldn't be a problem, but a bend in the wrong place and you could be right.

 

[jmw]

OK, referring back to the information given, this has been the situation for 9 years..... ??????? that's not far short of 300,000,000 gallons of water unaccounted for...!!!

SonnieTP, I guess this is an operation that you've inherited....?

You have to wonder why the problem was not investigated before.

Given that no one has ever bothered to find the cause of the error, why have the meters been calibrated every year??
Apart from anything else, I'd guess no one has checked to see by how much calibration shifts each year either - probably, even if they were no differences to account for, it would have been found that the calibration shift was pretty small and far less frequent checks would be acceptable (no moving parts sensors with no erosion and no deposition should exhibit pretty stable calibration).

If we accept there are no leaks, because the difference has remained constant (is this the case?), then either one or both installations are bad or there is some simple span and bias scaling issue.

As has been suggested, it could be the flow conditions downstream of the pump. A couple of bends before the meter and that could generate a lot of swirl. Turbulence will dissipate quite quickly but not swirl.

If there is reason to suspect swirl then you need to install flow straighteners upstream of the meter. These needn't be to big of a problem to install as you can have an straightener assembly mounted to a thin plate to locate between flanges. If there is a straight pipe location further downstream you could install the meters, move the upstream meter further away from the pump.

But I'd first be sure that the error is with the location and not the meters and it would be nice to know which meter is wrong. I'd suggest hiring a clamp on ultrasonic meter and measure the flow rates at various points along the pipe.

JMW

http://www.ViscoAnalyser.com

 

[zdas04]

Lot of conjecture, not much data.

This sort of problem is caused by one of three things (or a combination of the three):

1. One or both meters is reading wrong.
2. There is actually a different mass flow rate at the two points (i.e., a leak or an undocumented source).
3. You are mistaking a variable stream for a steady stream.

The third one is really common--the pump is surging between 150 and 300 gpm and the software is displaying snapshots. You can generally catch this by using daily cumulative volume instead of snapshot readings.

The second one is really unlikely--someone would have noticed 300 million gallons of water in their basement.

Everyone above is focusing on the first possible source. I would too. A vortex meter infers a flow rate from the frequency of vibration of an element in the downstream side of a bluff body in the flow. The vibrations are proportional to mass flow rate. It becomes a volume flow rate by inputting a fluid density into the computer. Errors in this sort of input are really common. I'd print out the meter station parameters and compare them line for line.

David

 

[tr1ntx]

"Lot of conjecture, not much data."

Yeah, we need those pressures BigInch asked for. It occurred to me that because the downstream flowmeter is 330 ft below the upstream meter, and all we know is that it "goes into the injection well", the pipe might not be running full at the downstream meter. We've all assumed that because it is an "injection well" that the waste water is being injected by the pipeline pressure. That may not be the case. Perhaps the pipeline discharges to an injection pump reservoir, so that it is vented to atmosphere and thus allows the elevation drop to accelerate the pipeline flow velocity. 225 GPM equates to 2.5 fps. I don't have any rules of thumb or quick calcs handy, but 330 ft head on a 6" pipe will flow a heck of a lot more than 225 GPM.

 

[zdas04]

The line needs 925 gpm to run full. I'd guess that since it is moving (based on upstream measurement) 2.496 ft/sec there is considerable opportunity for flow variability with 5.6 hours required for a chunk of water to travel the 9.6 miles.

It really only makes sense to look at daily (or weekly) cumulative flow.

David

 

[hydtools]

Why have the downstream meter? All the input goes to the injection well. Since it has never read the same as the upstream meter, it serves no purpose. Use it as a spare for the upstream measuring station.

Ted

 

[BigInch]

Well, it could tell you how much is leaking from the pipeline. Such leak detection systems are quite common in themselves. Additionally, in most jurisdictions you must measure whatever you are disposing into any kind of a disposal well. This is wastewater, maybe with some toxic components, and the gross amount injected per year or something could have to be closely controlled, etc. etc. etc. Even just salt water injection wells are closely controlled. It seems to be in California, so I would imagine that its very much controlled. Which brings up the question, how did they get away with the mismatched flows for 9 years. Anyway, there's a lot of reasons why they might need a out meter, but none of them matter to us here.

Nobody mentioned if there is a high point somewhere between inlet and outlet. Without knowing the exact elevation profile (not just at the inlet and outlet) and the pressure profile of this pipeline, the outlet meter might even be in slack, or "cascade" flow, (2 phase flow in a downhill segment) and measuring the passage of both wastewater liquid and its vapor phase too. Even more likely if the pipeline outlet is a turndown into the well.

Cascade flow can be steady state over the long term, but transient over the short term, running full, then with vapor slugs. We just don't have any idea what the hell it might really be doing. I'm just saying that if you do the hydraulic model first, you can tell if you should look for a leak somewhere, change the operating pressures or flow, or fix the instruments.

 

[hydtools]

Can a vertical pipe run be installed so the downstream meter is in an always full pipe?
Interchanging meters produced the same result, downstream reading less than upstream.

Ted

 

[BigInch]

If a sufficient flowrate is there, it's possible, although it is obviously much easier to guarantee full x-sect flow at any flowrate and at any pipe slope when a backpressure valve on the outlet is part of the design.

At any slope or any backpressure, whether provided by any combination of discharge pressure and pressure drop from a restricted outlet, atmospheric pressure included, is less than the exit pressure of the fluid, theoretically it will discharge across the full cross section. Gravity helping on the end of the pipe will only accelerate out so much. If a pump is feeding it at that rate or more, full cross-sectional flow occurs. Think of a garden hose without a nozzle on it. Turn it downwards and turn on the water. I bet its pretty easy to get full x-sect flow, because the x-sect is so small in relation to the pressure available at the last high point. Cut back the flow enough and you'll eventually see water just coming down in a ring from the connection wall and the water making a coaxial flow pattern with air in the center as the air tries to enter and run back into the hose.

 

[Artisi]

BigInch, you bet me to it, my comment was to be, fit a discharge valve to ensure the pipe runs full.
Anyway, seems SonnieTP may have stopped listening.

 

[BigInch]

Wouldn't be the first time the OP did a runner.
Does it really matter? The only real important thing is that WE work it out one way or another, right?

 

[Artisi]

100% correct, half the time we are probably better off without OP's

 

[BigInch]

Naw. 'cause we wouldn't have anywhere to start. Need some seed topic. Then we can grow it into something really interesting.

 

[hydtools]

BigInch, if flow is low enough to not fill the hose(pipe) won't gravity cause the flow to partially fill the pipe from the bottom up while air goes to the top rather than upstream through the center of the flow?

Adding the valve to restrict flow, as you say, should be an easy add-on.

Ted

 

[BigInch]

Gravity would cause the fluid to accelerate out the pipe allowing space for air to reach the high point. The same as saying flowrate is so low that backflow of air is not prohibited by the water's velocity, or flowrate across the full x-section (same thing really).

Say that flowrate was slow and that there was an orifice through an end cap that restricted water outflow enough and still allowed air inflow, so that the bottom of the pipe could fill as you suggest. Any air entering would tend to backflow upstream in relation to the slow liquid velocity. If the air reached a high point in the pipeline and could not travel any further upstream (now downhill), even against the minimal flowrate, it would tend to collect there, at first along the top wall of the overbent pipe. Further collection would at first just increase the velocity of fluid below the bubble, perhaps drawing some air downstream with it, but if still not fast enough to do that, the bubble would grow larger. But while restricting liquid flow and increasing the velocity, pressure drop might also be increased to the point where upstream pressure built up and then just blew out the air bubble and all downstream liquid to restart filling with liquid again. Or .. the bubble could continue to grow and fill up the downhill leg until it reached the next uphill segment and start the bubble up-and back filling process there too. It's possible that a number of air-locking and pressure build cycles would occur and initiate a 2-phase flow pattern that would remain for the duration of that particular flowrate. Of course what exactly would happen would be very dependent on the precise elevation of all points of the pipe along its entire length, and if upstream pressure could build, or remained more or less constant. If it remained constant, perhaps a vapor lock situation might eventually develop over multiple overbends where the sum of water column pressures in uphill segments, without the help of a counteracting dense fluid in the downhill segments, finally balanced the pipeline's inlet pressure.

That's why they usually need a lot of air release valves in water pipelines, because water holds a relatively high amount of air that tends to easily come out of solution with reduced pressure and either increase some local velocities and pressure drops, or tend to cause vapor lock ups. Fortunately liquid hydrocarbons do not absorb so much air and generally operate at higher pressure than water pipelines, so that problem can usually be limited, as long as you don't mechanically introduce air into the line by mistake or by not purging after repairs.