Tank Calibration

Ah South Africa! Two years ago my wife and I spent 25 wonderful days in the Zulu region, Jo-B, Cape Town, all the way to Good Hope, and special memories of Robbin Island. Love the country and can't wait to get back. Now for your answer - the prover is a calibrated tank, usually smaller enough to be portable for carting it around to your job sites. If your wine tanks are 5000L, then a prover tank of 500L would work well. The master meter is engaged and liquid sent through into the prover tank until full at line 500L, then you compare the meter to ensure the meter reads 500L throughput. If not, calibration of meter is required. Ambient temperature is factored into the prover tank's full measurement line, and specs for what F or C temp it was marked should be available from the prover tank's supplier. Here in the states, many wine companies rent a prover tank to check their master flow and fill meters, maybe you can do that there and if not available, now you have another business opp you could start!

Apologies, I intended on offering a comment on the calibration of the sizes of tanks you actually mentioned, the 40K-L to 250K-L tanks. If your firm will be hired to calibrate many of these tanks, and possibly re-inspect others' meters many times, then investing in a calibtrated prover tank instead of renting one might be prefered. That is a "rent/lease - or - buy" financial decision you would have to analyze.

I spent 10 in the U.S. Air Force working in Fuels, then got a civilian job on a pipeline.

In the AF, we used master meters and hoped for the best with our meters. Our tanks had been calibrated mathematically, I assume by measuring the circumference of the tank every so many increments up the side then figuring the area and volume. This is difficult and has a margin of error. We also used API tables to compensate for thermal expansion.

You might have better luck using water and a turbine meter with a prover loop. You can look up prover loops online, they usually are set up with turbine meters. I looked in to this online awhile ago and saw a trailer mounted setup, so you should be able to find a company that will come with a complete unit on a trailer.

The way the prover loop works is by comparing how much fluid passes a known volume of pipe to how much went through the meter. They use a calibrated section of pipe in a loop and either a plug or a ball that slides (not rolls) down the pipe. The ball does not allow the fluid to pass by and sets off sensors at the start and end of the run. The known quantity between the sensors is compared to the readout from the meter.

Unless you always filled your tank to a certain mark you would still need strapping charts for your tank. If you can find a company that has a portable prover unit, perhaps they would draft strapping charts for you. If not, you will need to accurately mark the tank at increments as it is filled. I would look into a automatic tank gauging system that you could move from tank to tank because you should only have to do this once for each tank. They make some nifty laser tank gauging devices.

Here is a link to a pdf I found with info on prover systems and it has an image of a truck mounted one. Good luck!

Drew

Matt - glad to hear from someone in USA that knows sunny SA! About time to visit again!

As you probably are aware - the economy in SA is tiny compared to you guys. I have not yet come across provers for rental.

Drew - thanks for the pdf link. As mentioned above there probably are provers around, but not any that is available for rental. The system we have used to date (for meter calibration) was to have the flow meters (in our case - magnetic flow meters) calibrated at a flow laboratory that is accredited for flow calibration. This calibration of the master flow meter is obviously done at a known temperature. Once in the field the temperature of the calibration liquid may be at a different temperature. Without a prover to adjust for local temperatures this presents a bit of a challenge. One alternative would be to calculate the density variation between the master flow meter calibration and the tank calibration temperatures (assuming that the water for both is the same)

From both posts it is obvious that two types of provers exist - tank system and pipe system. The tank seem to be a easier concept to implement, but I am still curious as to how the compensation for temperature is made - is this tank calibrated at various temperatures with corresponding differences in volume and thus marked, or is some other system used?

In the wine industry here, 99% of tanks are fitted with sight glasses and we perform calibration of the tank by marking this sight glass with the corresponding volume at that level. This is done at intervals that allows convenient reading of the volume. This unfortunately makes calculations wrt temperature and density somewhat difficult as this sight glass probable moves (expands, contracts) with the tank.

How often is magnetic flow meters used in USA for this type of work? Any idea on how you do calibration of your wine tanks in California?

Jan

Jan,

Washington state rivals California in wine production, not quantity of course, but WA has about 300 wineries and our latitude and soil is equal to the Bordeaux region of France. Here in WA the compensation for temperature is made by one of several methods - typically a chart is supplied from the vendor for the meter and/or tank - but the winery then uses (rents) a prover tank and checks the data. I just phoned a gentleman I know at Columbia Crest (I saw the brand in SA) and he said they check every single tank they purchase. He also said they have developed a proprietary business data log of volume at level at temp per each tank.

He said they also use sight glasses and calibrate the tank by marking the glass with the volume at that level. He agreed that there is some difference and as an engineer he has to let it go! His words not mine!!

I asked him about magnetic meters and he has not used them.

Hope this helps some! Yes I should schedule a trip back to SA soon. I particularly enjoyed the "Stellenbosch" region but pardon my spelling if I got that wrong. I am going to New Zealand this winter - for their summer, I like to go south for the winter to enjoy someone else's summer!

Matt

To my knowledge there are 2 types of proving...Gravimetric Proving, which measures what your meter reads compared to the mass of the amount run through it (density of the fluid vs the mass = the volume). The other method is volumetric, where an exact volume is run through your meter and this compared to the meter reading.

To my knowledge there are 2 methods...bench proving, which typically uses water and then calcs made to compensate for the difference in S.G. The other is field proving which is done with the same medium being measured. All combinations of which have been mentioned by previous posts.

I worked for contract company once and we built our own gravimetric prover. This consisted of a tank (about 1200 litres) which sat on a scale. The scale needed to be certified yearly and proof of period verifications were required. The cost of the build was minimal compared to the money made contracting our services. The same would hold true if your own demands are great enough.

I like the idea of weighing the tank. You could get an intermediate tank and place it on a scale between two of your tanks. If you wait until the weather settles and fill both tanks half way, wait till the temp is the same between both tanks. then transfer one tank to the other until the first tank is empty. Then take the full tank and fill the intermediate tank on the scale to its limit and start annotating the weight and compute the volume as it fills, marking the side in increments. As it fills you get it calibrated by the weight of the fluid (just make sure you know the density of your fluid). Then as long as the tempature doesnt change much you should be able to calibrate the second tank as it is filled from the intermediate tank. Repeat the process until the full tank is emptied. You can mark the volumes on the emptying tank as you fill the intermediate tank, this way you can knock out two tanks at the same time.

This is the best way I can think of unless you want to order and set up your own prover / meter system. You could do that and make it portable and rent out your services to other companies that need it.

It took me two or three sessions to write this so if it is confusing, I will clarify.

Drew

What the guys do in WA is what we do for the wineries locally. The tank manufacturers do not supply any calibration data or charts - once again we provide the service. In South Africa it is required by law to have the wine tanks calibrated, although no standard or specification is provided. In the past the tanks were calibrated by entering a small volume of liquid into the tank - just enough to reach the cylindrical section of the tank (assuming a vertical cylindrical tank - majority). The rest of the tank was calibrated by mathematical method (using measured tank geometry). This type of calibration did not allow the top of the tank to be calibrated due to the non-linear shape of the top.

The advent of flow meters allowed the calibration of tanks from empty to full capacity. No attention was however given to temperature effects and it worked well for the size of tanks in a winery and still does.

The above do not hold true for larger tanks that are more common nowadays and especially for tanks used for spirits storage compared to wine storage. We thus have a situation where we have to develop some cost effective way of implementing a method (hopefully from a established industry such as petroleum) to calibrate tanks and to compensate for temperature. This is were the tank on a scale sounds to me like a good idea. I think enough data is available that specifies the density of water at various temperatures (the ISO standards specifies a algorithm for water density) that makes this option viable.

The use of a sight glass with volume indications that is converted to a temperature related volume (as is used by the winery in WA, USA) is obviously what we are aiming for. Question remains - should the sight glass indicated volume or should a dip tape be used to determine the level and convert that via a spread sheet calculation to volume.

What we are discussing the petroleum industry has refined to a fine art. We need to modify (if necessary) this to suit us.

Alex - I would have loved to visit your company, but it is a bit far for now given that you are in Argentina.

All the comments are starting to define a method - thanks for the help!

Every month while working on the Pipeline in Georgia (USA) we had to do tank calibration. This consisted of dipping the tank with a calibrated tape and annotating the temperature. This had little to do with the calibration of the tank, but rather the computer system that read the totals from the meters (which were temp compensated) and the liquid level meters on the side of the tanks.

We would take the "stick" readings and temperature back to the office and use the API conversion tables to convert the level to 60°F (15°C). We used one table to convert the specific density, and from that we used the other table to determine the correction factor.

Some of our tanks were over 1 million gallons so the correction factor although small accounted for a large quantity of fuel.

Drew

Hi there,

I think you would be wasting your time with magnetic flow meters. Why - simply put, the ambient temperature can and will change, the density will not always be consistent and will vary, even by a tiny amount. This will affect your flow readings.

Rather opt for a Coriolis type of flow meter. With this meter you do not worry about temperature and density compensation, and you will get true mass flow. This is especially great where you have custody of transfer. (Some meters can even give you volumetric flow if required)

In petro-chem we either use Coriolis or positive displacement meters as they are very accurate and can be ISO accredited for custody of transfer. The problem with positive displacement is mechanical wear, whereas Coriolis has no maintenance issues. For certification - send them off to Eskom, they will issue you with a certificate stating the inaccuracy of the meter.

My Coriolis of preference - Khrone.

Regards,

Craig

I have used coriolis meters in the past when working on a project for a client. The calibration industry in the wine industry is however rather small and it is difficult to justify using a coriolis. I agree with you, Craig on the exceptional features these instruments have.

Using a mass flow meter such as a coriolis does not yet solve the issue of tank dipping the way the wine industry does - all in volume terms, not mass. All wine and related products are sold on volume basis and all stock take and customs and excise duties are determined on the volume of alcohol.

I maintain that the processes and procedures established in the petroleum industry is very relevant to the wine industry. The problem I still have is the compensation for temperature during the tank calibration. This, as pointed out, can be solved by using a coriolis or as previously mentioned, a scale & tank setup.

From all the posts and comments it looks like the wine industry (or more specific - the potable spirits industry) in South Africa may have to adjust the way dipping is done, by adopting petroleum standards. As seems to be the case in USA.

Hi there,

I am a little curious though. Are your tanks closed to atmosphere i.e. pressurised or are they open to atmosphere?

Is your process medium single phase or multi phase i.e. is the alcohol, water and whatever completely mixed and blended or do you have a phase of something sitting on a phase of something else?

If you could provide some information on the tank and the medium, I might have an eloquent solution for you that we also use in Petrochem, but can be used for custody of transfer.

Regards,

Craig

The tanks are all open to atmosphere. The process medium is single phase mixed, ie water and alcohol and some other components that constitute wine/brandy/whiskey. Temperatures are mainly ambient, but tanks are occasionally cooled to 10degC. Highest temperatures are in the region of 40degC (for areas such are Upington and Worcester).

In order to preserve the wine, CO2 is often used to blanket the contents of the tank, but not pressurised.

Looking forward to your comments.

Regards,

Jan

I don't have access to the API conversion tables anymore, but if you know the specific gravity of each tank, they might correspond to the API tables. Perhaps someone else has the books and can check to see if they correspond. It will take someone more knowledgeable than me to tell you for certain if the tables work off of the specific gravity of any liquid or just fuels. My guess would be they will work. Some of the lighter gasolines should be similar to your alcohols.

If the tables work, you can calibrate your tanks by any method: either buy one of the mass meters (and rent out its service to justify purchase) or by weighing an intermediate tank. Then you can stick your tanks with a tape and bob (fuel term for metal measuring tape and weight). Once you have a good stick, use the temperature and conversion tables to obtain the volume correction factor, apply it to your quantity and then you have a volume converted to 15°C.

Drew

Hi there,

I am going to make a few assumptions here (the mother of all f-ups).

As you have a temperature variation - you will have a change in density. This will affect your reading. There is a way to compensate for this - more later.

I don't think your tanks are linear - they would be more like barrels on their side? Regardless - you will need an up to date strapping table for the tank.

As your tanks are open to atmosphere, you could use the following type of level measurement system. I have used it for accounting grade level measurement within a Petrochem environment, but I think it will work well for you.

First of all you would need two pressure transmitters, preferably DP cells and one RTD, preferably with a thermowell.

The RTD would be installed at the lowest point of the vessel - to read the temperature of the liquid. PT1 would be installed at the lowest part of the vessel to get the pressure. PT2 can be installed about 1/3 way up the vessel at a fixed known height.

Now for the easy part. Pressure = Height * Density * Gravity.

To get the density (as it will change and we need it to get the level)

(P1 - P2) / Known Height = Density.

So now we have a fluctuating density that we can measure and use to get an accurate level i.e. we have density compensation. Don't forget to add the gravity constant to your calculation.

Now to get level.

P1/Calculated Density = height.

Height entered in the strapping table of the tank - gives you accurate volumetric level. Again don't forget to add the gravity constant to your calculation.

You could also use radar, but there are considerations to take into account for radar if you need high accuracy like roof deflection and thermal tank expansion.

Regards,

Craig

The use of pressure transducers to determine density and subsequent volume will work well, but the shear number of tanks and the relative small sizes may make this option not very viable. We have calibrated just over 1200 tanks in the past two years -of which I would think about 10% are for spirits. Given the availability of what is known as "alcohol tables" there is no desire or need for installation of density measurement (from the clients), unless a SCADA or equivalent system is used by the client (and there are none yet). In using "alcohol tables" a sample of the liquid is taken, the temperature noted and the subsequent density is then determined. This all works well, but temperature is not included in the actual dipping, i.e. expansion of materials due to temperature.

Two different aspects were mentioned in this thread - calibration and dipping - and both involves temperature compensation. To date we where only involved in the calibration of tanks, but the moment a client questions the calibration, calibration and dipping become part of the problem. We now have to ensure that we calibrate correctly and accurately and that the client dips correctly and accurately

From all the posts in this thread the following is to emerging that I need to address:

Calibration of a tank

• An onsite calibration of an already calibrated flow meter to compensate for temperature difference between original calibration and site calibration (unless a coriolis meter is use)

Dipping of tank (client responsibility?)

• Calculations to convert the dipping to a reference temperature

Seems easy enough, but applying this in practice causes some confusion in my mind. I am still trying to formulate the ideas that emerged from the posts. Thanks for all help!

1) WHAT THEY MEAN BY A PROVER YOU CAN USE A VOLUMETRIC MEASURE CAN AS TO USE TO CERTIFY YOUR METER SO IT DOES NOT SHIFT DURING VOLUMN FILL SO ALL YOU DO IS DIPS PER MM SAY EVERY 10000LTS DO A DID AND SO ON THAT WILL TELL YOU IF THE METER IS SHIFTING AND ALSO TELL YOU IF THERE ARE BULDGES ON YOUR TANK

CHEYNE

I have not used calibration tanks too much http://www.youngcalibration.co.uk/glossary.html may help.