Custody metering of Oxygen >10:1 turndown

[DaveScott]

I have a rather expensive problem, and would like to ask for a sanity check.
There is a DN250 oxygen line feeding our plant (it's a Steel plant with Electric arc furnaces), and the existing metering system is Orifice based, with pressure and temperature compensation. The supplier has declared the measurement system to be accurate enough over a 10:1 turndown starting from the peak flow rate of 23,000Nm3/hr. Below the minimum, the flow shall be deemed to be at the minimum. That is, once the supply valve is open, we are charged at 2,300Nm3/hr, whether we use it or not. The trouble is, we often run at 500Nm3 to 1000Nm3/hr when the plant is on standby.
The oxygen pressure is around 20 bar.

I have suggested to the supplier that our company pay for the installation of a better metering system, so that we are more likely to pay for what we use (obviously he is not keen to do this, since now he is making a fine profit...), and to that end, I put forward several coriolis mass flow catalogues. Fortunately for us, the contract has one little line that says we can do this. Their answer was that 'head office' doesn't approve those type of meters for custody transfer, which I thought was 'curious'.

The questions that I would like to put to you are as follows:

What is so wrong with Coriolis measurement, that a major oxygen supplier would refrain from using it? I assume for the time being that the head office is being quoted correctly.

What other metering options/ piping systems are possible?

Of course, we have them checking what options there are, but all the same, I would like to get some parallel comments.

Thank you.

 

[JLSeagull]

NPS 10 is as big as available for any Coriois meter and much larger than most. If dealing with pure oxygen, the piping reductions create impingments that require monel etc. The meter tube might need to be monel. It sounds like a much bigger deal than multiple dp transmitters to accommodate wide turndown on the existing orifice plate.

 

[IRstuff]

Why not put in a secondary tank? You could fill it at a rate commensurate with your gas supplier's meter and take from it at the rate that you want.

TTFN

Eng-Tips Policies FAQ731-376

 

[dcasto]

Orifice meters are very accurate and 16 to 1 turndown is very acceptable with "smart" tranmitters. Make sure your supplier has the latest version of API chapter 14 and that they comply with it. I assume you witness calibrations and inspect the test equipment to see if it has been NIST traceable certified correct.

The next best option would be to install a ultra sonic meter, they have a 100 to 1 turndown. Again SAPI 14 will give all the specifications.

The next is a turbine meter, 50 to 1 turndown is resonable. I think API 8 covers these meters.

The Coriolis meter has not been widely accepted in vapor phase, the industry has performance standards, but you'll still need to calibrate it to one of the meters above.

There are ISO standards that correspond to the API. I would have my own meter installed as a check and back up if the suppliers meter fails. I hope the meters are all EFM too.

 

[hacksaw]

The coriolis and sonic meters would be pricey, forget turbine meters, unless you can get bearings rated for Oxygen service. A vortex shedder might be cost effective.

The orifice plate can be made to work, but you'll need two transmitters. Your flow range requires a 2000:1 d/p turndown.

an automatic valve for positive shut-off, could be used to establish "zero-flow" contact basis, but that won't be cheap.

after all that expense, if you are dealing with lox you may be paying for boil-off even is you eliminate your off-set reading...at a minimum you need two d/p transmeters to reasonable cover the flow range.

 

[DaveScott]

Thanks for the responses so far. I noted a couple of times that 2 Dp cells with one Orifice are mentioned. I must admit, I thought that the error at low flow comes from the very low DP across the orifice. However, I assume that the system works in the obvious way, and requires a totaliser that can swap between the two as required. This sounds rather promising on the face of it, since most of the hardware is in place, and it doesn't require much in the way of a shutdown. I shall have to read on this further.

For interest, our supply line is about 50kms long, and forms a ring main. There are several consumers, and several air separation plants. I don't think there is a boil off problem.

I also don't think there is much scope for a local accumulator vessel. Nice idea, though, but we are only talking about 20 bar. It would be HUGE.

I take the points about the coriolis meter being rather new and unaccepted. My internet travels were perhaps a little misleading. Most of the sites claimed rather good accuracy (since they measure mass directly I suppose, and not Pressure, DP, and temperature).

Again, thanks.

 

[IRstuff]

Seems to me though, even a 10 m^3 tank could possibly buffer the usage over a long enough time to keep the flowmeter, and you, happy.

TTFN

Eng-Tips Policies FAQ731-376

 

[Ashereng]

DaveScott,

Some of the turndowns mentioned in posts above seem a bit wide. I am not disputing they are possible, just a bit wide.

This is just my rule of thumb:

For most orifice meters, including custody transfer, we use 5:1.

For ultrasonic, which has much less accuracy and repeatability, we use 20:1.

For turbine meter, and I have heard sales pitches giving 100:1, we use 15:1.

For coriolis, we usually go with published turndown. For some reason, they seem to be close. We have seen 50:1 will good accuracy and repeatability.

If a client has specific experience with a specific model, then we will of course go with their experience. But, if we are designing from scratch, we go a bit conservative.

For oxygen service, the reaction of oxygen with the material will be a problem, as a poster already mentioned. I guess that is why you are currently using an orifice and DP.

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[dcasto]

With DP's you can use stacked transmitters, 0 to 10, 0 to 50, and 0 to 200. Squareroot 200 = 14 squareroot 1 = 1 turndown is 14.1 to 1 Spread it out more with a 0 to 250 and 16 to 1 is ok. We used this to measure extermly valible gases, one station metered $4,000,000/day.

 

[quark]

I agree with Ashereng about the turndowns. I came across a paper that suggests methods of improving orifice meter turndown (by using smart transmitter), presented by one of Rosemount's employees. The guaranteed turndown is 9:1 and then error increases asymptotically and steeply after 9:1 (1% is minimum possible accuracy for gases).

Multiple orifices is an option in your case. I have seen thermal mass flow meters being used in oxygen lines for fermentors. Accuracy and turndown, both are good.

 

[JLSeagull]

First define the accuracy requirements. Then evaluate the measurement accuracy considering the precision statement of the transmitter, the orifice plate, installation etc.

Once can discuss the transmitter precision improvements. The Foxboro E11 dp transmitters claimed +/- 0.5% accuracy at full scale. The accuracy quoted as flow percentage would be the same at full scale. However at 50% of the flow the dp is 25% full scale. The transmitter error becomes more significant for small flows. We regarded an orifice flow installation suitable for 3:1 turndown in the 1970's. The transmitter precision is far superior today. One could claim a significant turndown improvement on that basis.

The orifice plate is calculated based upon certain fixed values of density, viscosity, pressure, temperature, etc. The process conditions rarely match the values used for the calculation.

You must define the turndown in terms of accuracy requirements. Define the accuracy requirements first then evaluate the potenial errors. Then sum all potential errors. With the accuracy defined first, I doubt that a single transmitter could credibly meet most peoples requirements beyond a 7:1 ratio. Separate transmitters selected for the maximum rate and minimum rate would improve turnaround. In the dual transmitter case with a large range difference, perhaps 14:1 is credible.

 

[dcasto]

Think about the error and the cost to find it. At the low rates a 2% error wil be in someones favor and it may only amount to a to hundred dollars a per day, in the buyers favor.

The calculations for flow coefficient are less than .1% over a HUGE range with the latest equations and mechanical set ups (flow straiteners for example). The ol orifice meters are only 2% accurate is out (with your wide ties too). The latest API has examples for performing error analysis. I've measured supercritical fluids with 50 to 1 turndowns and even more when the consumer's plant was coming up over a 2 day period. All the talk here about orifices ASSUMES no plate changes. You want to change plates (a second tube), then 200 or 300 to 1 is possible.

 

[igorcornejo]

An orifice plate who's required turn down is 10:1 means that the transmitter has to have a turn down ratio of at least 100:1 (Since flow is proportional to dP^1/2). Although some transmitter manufacturers are going to boast that they have transmitters with far more rangeability than this, are you going to take their word for it?

Your situation is something that should be studied from an economic point of view. You need to ask yourselves the following questions:

1) What is your acceptable loss? Assume the flow rates that this flow meter is registering compared to what flow rates you're being charged at. How much difference is between the two? This difference is what you can assume is your metering systems accuaracy. Is this amount acceptable? If this is > 2% then you should seriously consider means and methods of proving this metering system or ask for justifications on the discrepancies.

2) Whatever technology you use to measure flow, in replacement of what you presently have, you will need to perform an accuracy calculation to determine the economics of switching. Will the accuracy you achieve eventually pay for the meter as well as save you money? How long will it take for you to start saving money. Just a word of caution - the economics should be based on total installed cost (TIC) or plant life cycle cost rather than just the price of the meter because there will probably changes to your piping, etc. Also, the meter may require maintence, etc.

3) Is your supplier charging you based on a standard flow (a temperature corrected flow) or are they charging you based on actual flow? This makes a significant difference. Put it this way, when you buy gas for your vehicle you are buying it based on a standard temperature (15 Deg C) and maybe even pressure. Think about what happens when there is a change in temperature each and every time you buy gas and what it does to the actual amount of gas you buy . . .

4) Is your custody transfer meter a calibrated meter run or simply a mass flow measurement? If it's a calibrated meter run, what standard was used to certify it? In your industry API isn't really relevant, but you should check out AGA Report 3. This report specifies what needs to be done in order to certify a calibrated meter run, according to the American Gas Association. Calibrated meter runs have a high accuracy of measurement but only as long as all the rules have been adhered to. If not, then there is no real way to quantify the measurement error without a proving system.

5) Calibrated meter runs are, in essence, an exact replica of a laboratory setting where they've proved an accuracy to within an acceptable percentage of error. Therefore all variables that could possibly introduce errors are controlled. If any of the lab conditions aren't met (including the ambient air conditions and the temperature of the measured fluid) then there is room for error. Remember that you are only as accurate as your least accurate measurement. Does the temperature of the O2 vary? If so, by how much does it swing? What is the accuracy of the temperature and pressure measurement? All of these introduce more errors if the O2 temperature varies significantly. There are correction factors built into the AGA 3 formula to correct for this, but these are just correction factors, not accurate measurements.

6) What is your turn down ratios and are they within the accepted turn down ratios as stated by AGA report 3? Turn-down ratios are a factor in custody transfer and you should see what certification was used for your custody transfer and if your turn-down ratios are within the limits specified.

7) Traditionaly turn down ratios of 3:1 are what's been specified as the maximum turn down for dP devices. With new smart transmitters it might be acceptable to say that 5:1 is achievable within a high degree of accuracy but I'd steer clear of anyone who says you can achieve 10:1 turn downs with one dP transmitter because of the reason I stated at the begining. Even if you use two transmitters to satisfy a 10:1 turn down, you're looking at doing some fancy logic in your control system so that it knows when to switch from one transmitter to the other.


Ok, now for coriolis mass flow meters:

Advantages:

1) Gives you true mass flow without temperature and pressure compensation.

2) Is impervious to changes in density (which could occur if the temperature of the O2 varies significantly) as well as viscosity.

3) Has a fairly high turn down ratios.

4) No straight run or upstream/downstream requirements.

Disadvantages:

1) Max meter size is 6", ANSI CL 600. You might have to put two 6" meters in parrallel (or two 5", something like that), as well as change your piping system to accomodate.
2) Because you're in O2 service, special cleaning and materials selection is required. This could cost you a pretty penny.
3) These are 4-wire devices so you'd probably have to bring out 120 Vac to the device, as well as your signal wiring.
4) Coriolis meters have a fairly large "foot print", as a ball park estimate in size - think of an equal size globe valve with a pneumatic actuator. If you have tight piping contraints then this would become a factor.


Just of the top of my head, I'd say if the economics suggest it, put in some coriolis meters. Although they'd be fairly pricey - they were designed for your type of service. If your O2 supplier doesn't approve of them then you should ask why and what basis do they have? Talk to the AGA guys and ask them if they have a report on using coriolis meters for custody transfer. I'm almost positive they do or at least they could provide you their points of view on it. In the oil and gas industry colriolis meters are gaining in poplarity over calibrated meter runs when it comes to custody transfer because this meters is robust.

The only limitation this meter has, and why some people give it a bad name, is because of it's supposed ability to measure two-phase or multi-phase flow. Based on the principle of how this meter works, what it's actually measuring is a change in density. It has no idea what is flowing through it, only that it's density has changed. Based on this notion coriolis vendors keep trying to sell this meter as a multi-phase flow measurement system but don't buy that arguement. Unless you know your % compositions up front, all this meter will do is tell you that there has been a change in density. If all you care about is what the flow rate is, then this meter works wonders. But, if you want to know the void fractions of each composition - then you're in trouble.

Hope this helps.

 

[Ashereng]

igor,

Coriolis meters come in sizes larger than 6".

http://www.endress.com/eh/sc/america/ca/en/home.nsf/systemcontentview/index.html

The Promass F is available up to 10".

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[dcasto]

JLSeagull, modern flow meters have every variable put in with equations of state that calculates Cp, rho, speed of sound, everything. If you want your orifice meter right on, use a solitron densitometer. I'll bet the O2 industry has a EOS that only requires Temp and press to get density. The equation for mass flow is then Q = Fb* 6.02 * sqrt ( Hw * rho), finished no other correction factors with rho either from a densitomer or EOS.

igor, API is the standard reference for all gaseos and liquid measurements in the US. AGA 3 is a sub group to API.

 

[hacksaw]

believe the problem is not accuracy, the contract specifies the meter, what is causing grief is the totalizer runup at "zero flow".

this is normal inspite of the d/p turndown requirement, but the totalizer setting also has a cutoff, that defines zero flow.

also check your d/p transmitter. some of the older units extracted the d/p in the transmitter. while the d/p range was within spec the square root hardware was not. we set all transmitters for d/p only ant perform the extraction the the dcs.

 

[JLSeagull]

Dennis,

I rarely recommend an orifice plate where other technologies have a better fit. Review CGA 4.4 for oxygen related issues that limit the use of many flow measurement technologies. If an orifice plate exists, multiple transmitters can help with the turndown.

Hacksaw makes a good point. I don't know that the square root extraction is better in the DCS than the transmitter but it is worth looking into.

 

[DaveScott]

Thanks everyone, I *am* reading your responses, and checking through some of the ideas. I hope to be able to reply to most of you. At the moment, I quite like the idea of multiple DP 'smart' cells (since I see that 'smart' dp cells claim to give an improvement from 3:1 to 10:1 rangeability, I assume that the oxygen supplier is already using one - needs checking). I also need to check about multiple orifice plates.

But in my internet travels, so far, I have not seen much reference for using multiple dp cells or multiple orifices, and subsequent totaliser. Again, I have to check what is there (it's a small receiving gas station, and I doubt it has anything like a DCS system). Again, another check is required. When I find out, I will post details.

Again, thanks for the responses. Igorcornejo - thanks for the essay!

 

[JLSeagull]

Search for a "flow computer". I found that Emerson still has a flow computer division. These will have the AGA calculations, etc. I did not see one that can select from multiple pressure differential transmitters for turndown. Call the sales person and ask.

 

[JLSeagull]

http://www.documentation.frco.com/groups/public/documents/brochures/d350857x012.pdf

Page six reflects a "Typical Gas Orifice Application" that reflects multiple dp transmitters on one orifice run.