Flow, metering and water loss

How accurate is a clamp-on flow meter?

• Accuracy is of reading, not of full scale. The two are very different at low flow.
• Entry level is around 1 percent. Higher specification instruments reach roughly 0.2 to 0.5 percent.
• Straight run decides the outcome. A common requirement is ten diameters upstream, five downstream.
• MCERTS asks for better than 8 percent of daily volume for permitted effluent flow, a far looser bar.

Of reading versus of full scale

The first thing to check on any quoted figure is what it is a percentage of, because the two conventions differ enormously in practice and both appear in datasheets.

Accuracy of reading means the error scales with the measured value. At 1 percent of reading, a flow of 100 units is measured to within 1 unit, and a flow of 10 units to within 0.1. The proportional accuracy stays constant across the range.

Accuracy of full scale means the error is a fixed proportion of the meter’s maximum. At 1 percent of full scale on a meter ranged to 100 units, the error is 1 unit everywhere, including at a flow of 10 units, where it is actually 10 percent of the reading.

For leak detection this distinction is decisive, because the flows that matter most are the small overnight ones at the bottom of the range. A specification of reading is what makes low flow data trustworthy. Clamp-on ultrasonic meters are normally specified of reading.

What actually degrades it on a real pipe

The laboratory figure assumes a fully developed, symmetrical, swirl free velocity profile. Everything that goes wrong in the field is a departure from that.

Insufficient straight pipe is the dominant cause. A bend, tee, valve or pump immediately upstream leaves the flow asymmetric and swirling, and the meter infers the wrong average velocity from the transit times it measures.

Pipe wall condition matters more than people expect, because the signal has to pass through the wall twice. Heavy internal scaling, corrosion, delamination or an unknown liner all attenuate and refract the ultrasound. So does the coupling between transducer and pipe: an air gap, loose clamping or degraded couplant will cost more accuracy than most other factors.

Entrained air and solids scatter the signal. Partially filled pipe defeats the measurement entirely, because the technique assumes a full bore. And the pipe dimensions entered at commissioning matter directly, since the meter calculates area from them: a wall thickness taken from a table rather than measured introduces an error that no amount of averaging removes.

How good does it actually need to be?

It is worth setting the required accuracy against the purpose rather than buying the tightest figure available.

For regulatory effluent flow, MCERTS requires the discharge to be measured with an uncertainty better than plus or minus 8 percent of the daily volume. Any competently installed clamp-on meter is comfortably inside that, which means the regulatory driver is rarely the reason to pay for a higher specification instrument.

For leak detection and water balance work, repeatability matters more than absolute accuracy. Finding a leak means detecting that overnight flow has changed, and a meter with a small consistent bias will do that perfectly well. A meter with excellent stated accuracy but an unstable installation will not.

For billing, sub-metering and cost recovery, absolute accuracy is what counts, because the number is being used directly. That is the case where the straight run, the wall measurement and the transducer mounting justify real attention.