A flow meter error cost is easy to work out and almost nobody does it. Take the volume that goes through the meter in a year, multiply it by what each unit is worth, then multiply by the error. That’s the money. On most sites it’s paid quietly, every month, by whoever is on the wrong side of the reading.
David Spitzer put it plainly in Flow Measurement (2001): where product changes hands, the measurement should be as accurate as possible, given the economic value of what passes through the meter. We’d go one step further. Every flow meter that feeds a bill, a batch record or a dosing pump is a custody transfer meter of sorts, even if no contract says so.
Working out your flow meter error cost
Annual cost of error = annual throughput × value per unit × error.
That’s it. The hard part is being honest about the inputs, so here are three examples with the assumptions shown. Swap in your own figures.
A filling line giving product away. A batching meter delivers 2,000,000 litres of a finished liquid a year, worth £1.50 a litre at cost. If it under-reads by 0.5%, every batch is slightly over. That’s 10,000 litres a year handed out free, or £15,000.
Bulk fuel coming in. A site takes 400,000 litres of gas oil a year, at an assumed 75p a litre. A delivery meter over-reading by 0.5% bills you for 2,000 litres that never reached the tank. £1,500 a year, every year, until someone checks.
Effluent going out. A food plant discharges 120,000 m³ of trade effluent a year, at an assumed combined charge of £1.80 per m³. An effluent meter reading 3% high adds 3,600 m³ to the bill. That’s £6,480 for water you never sent.
None of those meters would look broken. Each one reads a sensible, steady number.
“1%” might not mean 1% of what you think
Check the datasheet before you trust the headline figure. Some meters state accuracy as a percentage of reading. Others state it as a percentage of full scale, and that changes everything at low flow.
Say a meter ranged 0 to 100 m³/h is good to ±1% of full scale. That’s ±1 m³/h anywhere in the range. Run it at 25 m³/h and the same ±1 m³/h is now ±4% of what’s actually flowing.
Plenty of meters spend most of their working life well below their top end, which is why a calibration should test several points across the range you really use, not just one. We explain how that works in our guide to the flow meter calibration procedure.
Where the error usually comes from
Not always from the meter. The Environment Agency’s MCERTS standard for self-monitoring of flow notes that a partly closed gate valve directly upstream of a sensor can produce a flow measurement error of up to 18%. Pumps too close to the sensor fall in a similar band, 2% to 18%. A perfect meter in a poor spot still gives you a poor number.
Then there’s wear. Turbine bearings loosen, electrodes coat up, orifice plate edges round off. The drift is slow, so the readings never jump and nothing on the screen looks odd.
And sometimes it’s simply a wrong number in a setting. A measurement error report published through the Joint Office of Gas Transporters describes an orifice meter stream at a gas offtake where one pipe diameter reading was misread from the original data books in 2003. Nobody spotted it until an audit in 2020. The error turned out to be tiny, well under the 0.1% threshold below which no reconciliation is required, so nothing had to be settled. The lesson is the timescale. A fixed error can sit in a system for seventeen years if nobody compares the meter against something they trust.
What the rules allow
Some limits are set for you. Under the Measuring Instruments Regulations 2016, meters for liquids other than water carry an accuracy class. Class 0.3, used on pipelines, allows the meter itself an error of 0.2%. Class 0.5, which covers fuel dispensers and road tanker meters for low-viscosity liquids, allows the meter 0.3%. So the 0.5% in our gas oil example sits right at the legal limit for a road tanker system. These limits are tight because money changes hands on the reading.
Effluent sits at the other end. MCERTS asks for the total daily volume to be measured to a target uncertainty of ±8% or better. That’s a regulatory floor, not a commercial target. On the effluent example above, a meter comfortably inside ±8% could still be costing several thousand pounds a year.
So “compliant” and “accurate enough to protect your margin” aren’t the same question. Ask both.
Setting a calibration interval on cost, not habit
Annual calibration is a habit, not a rule. A better starting point is the exposure figure you worked out above.
If a meter carries £15,000 a year of exposure per half percent, checking it every twelve months is probably too slow. If it carries £50, a longer interval is fine. The international guidance on this, ILAC-G24 / OIML D10, recommends adjusting intervals from the instrument’s own history: if a meter keeps coming back within tolerance, the interval can stretch; if it keeps drifting out, it should shrink.
That history lives on the certificate. The as found reading tells you how far the meter had drifted while it was in service, which is the number that was quietly costing you money. The as left reading tells you where it is now. If your certificates don’t show both, or don’t show the uncertainty of the test, they aren’t telling you enough to set an interval. Our guide on how to read a calibration certificate covers what to look for, and traceability in calibration explains why the reference matters.
Calibrating without stopping the line
Pulling a meter out isn’t always practical, especially on a process that runs around the clock. We calibrate air, water and other media flow meters at our Southport premises, and our engineers can calibrate many flow meters on your site to keep downtime down. You can read more about how that works on our on-site calibration page. If we find a fault during calibration, we can repair meters from all manufacturers, and everything is quoted beforehand. If a meter is beyond economic repair, we also supply flow meters and switches.
Frequently asked questions
How accurate does a flow meter need to be?
Accurate enough that the error costs less than keeping it right. Work out the exposure (throughput × value × error) first. Legal minimums apply on top for some uses, such as the Measuring Instruments Regulations accuracy classes for liquid meters, or MCERTS for effluent.
Is a flow meter equally accurate across its whole range?
Rarely. Meters specified as a percentage of full scale lose accuracy, in percentage terms, at low flow. That’s why a proper calibration tests several flow rates across the range you actually use.
How often should a flow meter be calibrated?
There’s no single answer. Start from the manufacturer’s advice and any regulatory requirement, then adjust using the meter’s as found results over time. High-value meters that drift need shorter intervals.
Can a flow meter be calibrated without removing it?
Often, yes. Many meters can be calibrated on site. Some need a rig in a laboratory to reach the accuracy required. We’ll tell you which applies when we quote.
What does “as found” mean on a calibration certificate?
It’s the meter’s reading before any adjustment, so it shows how the meter was performing while it was in use. Compare it with the as left reading to see how much it had drifted.
Does a new flow meter need calibrating?
It arrives with the manufacturer’s calibration, but that was done on a test rig, not in your pipework. Installation effects such as nearby valves and pumps can shift the reading, so verifying it once it’s installed is worth doing on any meter that matters to your costs.
Put a number on it
Pick the three meters on your site that carry the most value per year. Do the sum for each one. If the answer is bigger than the cost of a calibration, that’s your priority list.
When you’re ready, our team can quote for flow meter calibration at our premises or on your site. Get in touch and tell us what you’re measuring.