Rank flowmeters in order of accuracy.


Thread Starter


While on a job interview for an instrumentation technician job I was asked to rank the following flowmeter types in order of accuracy:

Mag meter
Mass flow
Nuceiating disk
Venturi tube
Parchal flume
Oriface plate

I was a little dumbfounded given no further information about application or specifications of the meters in question. Lets just assume the process is drinking water treatment. Can anyone give me a definitive rank on these meter types?
Well, you're being asked to do an apples-and-oranges thing here, but I can probably help.

The most inherently accurate and repeatable of the devices you list is the Coriolis Mass Flow meter. Thermal mass flow devices are not quite as accurate. This is because the meter actually measures mass flow directly. All the other meters in your list are either volumetric or height (delta-P) differential meters that do not take density or temperature variations into account.

The most accurate non-invasive device is the mag meter. On completely full pipe, without significant air entrainment (=/- 0.01% air or gas) you can expect around 0.5% of indicated flow rate to 1.0% of indicated flow rate depending on the model, size, installation conditions and range of the meter. Magmeters are often used for slurries and solids bearing flows like primary sewage and RAS. It is the most used flowmeter of all closed pipe flowmeters in large diameter water and wastewater applications.

The nutating disk meter is a positive displacement meter, and in clean fluids it is accurate enough to be used as a fiscal meter. Most of the cold water meters in front of your houses are nutating disk or rotary piston or single jet PD meters (and I know that single jets aren't really PD, but for this discussion they act like them). They don't like rocks and bottles or string.

Parshall Flumes (and Palmer Bowlus and other style devices) are used for open channel measurements and the most accuracy you can get from a properly installed Parshall flume is around 2%. Most are not installed properly and can easily be 10-15% off of actual flow rate.

Orifice plates produce a differential in pressure across the plate which is measured by a differential pressure transmitter, and converted to flow using a square root function. Properly installed and new, they have a maximum possible installed accuracy of around 2-5% of rate. As they age, or if they aren't properly installed, or if the flow fluctuates greatly from min to max, the accuracy of the orifice plate and differential pressure flow transmitter fluctuates too.

David Spitzer and I have written extensively about this in our "Consumer Guide" series (visit www.spitzerandboyes.com for information), and I have made some instructional videos about several of these flow meter types that you can find on YouTube or www.controlglobal.com.



Walt Boyes
Life Fellow, the International Society of Automation
Editor in Chief
Control and ControlGlobal.com
555 W. Pierce Rd Suite 301
Itasca, IL 60143

[email protected]
Not just apples and oranges but do they want mass or volume?
In general it doesn't affect Walt's reply but it does raise the caution that not all mass meters are coriolis and not all coriolis are equal, no more than any other technologies.

If it were volume that was of interest, then coriolis mass meters would convert the measured mass to volume using the integral density measurement and here there is a wide difference between mass flow meters.

Nutating disc meters are used widely in the US for domestic water metering.

In the Uk and much of Europe the oscillating (rotary) piston meter is mostly used.

It should be said that some PD meters are as capable of fiscal accuracy as coriolis meters and one of the most accurate is the Avery Hardol multi-piston meter.

The nutating disc meter is not so accurate (but has better dirt handling than most and has a better first time start survival rate).

So the list technologies is interesting in that while they say mass meters they do not say coriolis but they say nutating disc and not positive displacement.

A clever and interesting question that to me, suggests a little more thought is required.

Like all such questions, you have to wonder if they want you to simply give a ranking or explain your thinking. Explaining your thinking reveals a great deal more about your knowledge of flow measurement.
That was the reason I mentioned thermal mass flow meters as less accurate than coriolis. While there are differences between coriolis meters, the generality are so similar in specs that they can really be counted as a typical set. I agree that the way density is measured causes differences in specs.

Quite a few PD meters are accurate enough for custody transfer. Every cold water meter is by definition a custody transfer meter, and there are numbers of industrial grade PD meters such as those from Avery Hardol. Siemens, McNaughton, Brooks, and many others make custody transfer PD meters of a variety of kinds.

Your suppositions about the types of questions are spot on, I think. This was perhaps a "trick" question.


Walt Boyes
Life Fellow, the International Society of Automation
Editor in Chief
Control and ControlGlobal.com
555 W. Pierce Rd Suite 301
Itasca, IL 60143

[email protected]
I recently raised a the self same issue in an article about 6-8 months ago, comparing meter technologies and applications and I specifically included reference to water meters as an example of a low accuracy meter used for commercial transactions.

Different commercial transactions have different objectives.

Custody transfer is about high accuracy and taxation and error balancing from one transaction to another.
These are often applications with the same supplier and the same client (and the taxman taking a share).
In that case, meter errors could be expected to average out, even though there is a great deal of care to ensure the best accuracy or minimum errors whichever side they err on in a particular transaction.

Forecourt metering is different.
Not everyone purchases their fuel via the same forecourt pump nor even in the same forecourt. So less assurance of errors balancing out.
The accuracy requirement is to deliver to the consumer at least as much as is indicated, just as with any consumer purchase.
Of course, the supplier would like to keep the "give-away" to a minimum. They have an inbuilt bias in favour of the consumer.

Water meters are another story.
In the UK, and I think it is a 1942 act, the requirement is (or was) for metering to be "equable".
Meaning fair rather than necessarily accurate. At least, in terms of a single transaction.
The meter is dedicated to a single property and over the meter life, the expectation is that neither party should benefit unduly from measurement errors. The meter does not necessarily accurately measure the flow but no one consumer should be more nor less fairly treated than another nor should there be any inherent advantage to either the consumers or the suppliers.

Not only doesn't this mean accuracy per transaction but it means accuracy over the life of the meter and it also accounts for flow that takes place at below the accurate measuring range.
The trickle flow isn't measured, but it is "paid" for based on what is measured.
In the UK this is important (mostly in the south of England) because, dating back to the Napoleonic wars, the invasion fears meant that houses had a cold water gravity tank designed as a reservoir for if supplies failed.
Fitted with a simple ball valve it meant that every use of water resulted in a significant proportion flowing at very low flow rates, especially as the ball valve seals would wear.

Hence the use of rotary piston meters rather than inferential single or multi-jet or the nutating disc because of their better low flow performance and as the meter wears it does not significantly worsen in its accuracy except a low flows.

But water meters are not intended to be replaced every other year nor to be routinely tested and adjusted (there is no adjustment on most).

So a valuable property of rotary piston meters is that as they wear they tend to bed in and then stabilise but such wear as occurs over the life of the meter mostly influences the low flow performance and it means that they may over record by some amount in their early life and under-record later but with a balance, over the life of the meter, somewhere in the middle.

So there are many sometimes overlooked parameters in meter selection and understanding the way performance is reported has to also include the way their performance relates to the application and expectations which may vary including within legislation.

This "bedding in" is particularly noticed in Truck metering where rotary piston meters were favoured by many over the competing PD technologies which as they wore, continued to drift away from the initial calibration.