Meters, AMI and water-loss accounting

# Meters, AMI and water-loss accounting

> Your meters do two jobs at once: they earn revenue and they are some of the best leak detectors on your system. Meter types, what AMI adds over a monthly read, and how to build the water-loss numbers your exam and your board both want, are the ground floor.

Why the exam tests this

Meters and AMI sit inside System Components, the heaviest duty on the Water Distribution exam at 35 percent. Most operators know the meters running at their own utility and stop there.

The exam expects the whole meter family, from a residential PD meter to a production magnetic meter. It also expects you to know which one fits which job.

The bigger gap is accounting. Reading a meter is routine field work. Building a water audit, defining a District Metered Area, and telling apparent losses from real losses is a different skill.

That skill is where the exam gets harder fast between Class II and Class IV. Class I and II ask you to recognize the categories. Class III and IV ask you to work the numbers and defend a conclusion.

What you need to know

Meter types and what each is good at

Positive displacement (PD) meters use a nutating disc or piston and are the standard for residential and small commercial services, 5/8 inch to 2 inch. They measure water in discrete volumes, which gives them strong low-flow accuracy, the trait that catches a dripping toilet or a slow leak. Grit and sand wear the moving parts and cause under-registration over time.

Single-jet meters spin an impeller off one directed jet. They cost less than PD meters and hold decent accuracy at normal residential flow, but they miss small continuous flows a PD meter would catch, and wear or debris pushes them toward under-registration.

Compound meters pair a PD element for low flow with a turbine for high flow, with a valve that switches between the two. They fit sites with a wide flow range, hospitals, schools, and large irrigation accounts, where a single meter type would either miss the small leaks or choke on the peak demand.

Electromagnetic (mag) meters have no moving parts. They read velocity through a magnetic field, which makes them a strong fit for production and large-line metering. They need a full pipe and conductive water, and a bad ground or wiring fault causes more "bad meter" complaints than the sensor itself does.

Ultrasonic meters also have no moving parts and time sound pulses through the water to read flow. They give low headloss and strong data resolution for leak analytics, but air in the line and a poor straight-run installation both degrade the signal fast.

Propeller meters handle sustained high flows, mostly irrigation and other large dedicated uses, at low headloss. They read poorly at low or variable flow, so a seasonal irrigation account can under-register badly during its slow months.

Meter sizing has to match the customer's actual flow profile, not just the pipe size. An oversized meter misses small leaks and quietly under-bills for years; an undersized meter runs high headloss and draws pressure complaints.

AMR and AMI: what the extra data buys you

A manual read gives you one number a month. AMR, automatic meter reading, speeds up collection, usually a drive-by radio read, but it is still close to one number by the time you collect it. AMI, advanced metering infrastructure, is different: a fixed network reports on its own schedule, often every 15 to 60 minutes, with no truck required.

That interval data turns a meter from a billing device into an operations tool. Minimum night flow, the lowest steady flow in the early morning hours, is the cleanest leak signal available, because almost no legitimate use happens at 2 a.m. A rising minimum night flow on a meter, a DMA, or a whole system points to something running that should not be.

AMI data also flags high-flow spikes (irrigation, a stuck fixture, a possible break), continuous flow (a strong leak indicator), and reverse flow (which can mean a backflow event or an installation problem). None of these signals proves a leak on its own; each one tells you where to look first. Field confirmation still closes the loop before you dig.

District metered areas

A District Metered Area (DMA) is a section of the distribution system closed off by valves so every gallon entering it gets metered. Comparing that metered inflow to the billed use inside the boundary turns "the whole system is leaking" into "this one section changed." DMAs commonly serve 500 to 3,000 connections, sized to the system's hydraulics and geography.

Night flow analysis, usually run between roughly 2 and 5 a.m., separates real customer use from real loss inside a DMA. At that hour legitimate use is close to its lowest point, so most of what is still flowing is loss. A DMA is only as good as its boundaries; a drifted-open valve or an unaccounted intertie will throw off every number that follows.

At higher grades, expect a question that asks you to interpret a step test, where crews close valves one section at a time and watch the flow drop, to isolate which part of a DMA carries the loss.

The water audit and the loss categories

Build the water balance from the top down: System Input Volume is every gallon produced or purchased. Authorized Consumption is what customers used, billed or not. What is left over is Non-Revenue Water (NRW), and NRW splits into two very different problems.

Apparent losses never left the pipe. They are unauthorized use, meter under-registration, and billing or data-handling errors. They cost revenue, not water, and the fix is largely administrative and metering work, not a shovel.

Real losses are water that actually escaped: main breaks, service-line leaks up to the meter, and tank leaks or overflows. They cost water, treatment chemicals, and pumping energy, and they need field work to fix.

There is no federal cap on distribution loss. Most states that regulate it set an unaccounted-for-water threshold around 10 to 15 percent of production, though the exact definition and number vary by state.

At higher grades, know that the older unaccounted-for-water percentage can mislead. A jump in production, a new subdivision coming online, for example, can shrink the percentage even while the actual lost gallons stay flat. The Infrastructure Leakage Index (ILI) fixes that by comparing Current Annual Real Losses (CARL) to Unavoidable Annual Real Losses (UARL), the theoretical floor if every best practice were applied.

ILI equals CARL divided by UARL. An ILI near 1.0 means losses are close to the practical minimum. Utilities that actively manage loss typically land in the 1.5 to 2.5 range.

Worked examples

For the general flow, area, and unit-conversion formulas behind these calculations, see the Water Distribution math module. These examples use the loss and audit relationships specific to this topic.

Legitimate night use versus real loss: a DMA serves 1,200 connections. The utility estimates legitimate night use, mostly toilet flushing with no irrigation running, at 1.5 gallons per connection per hour.

Legitimate use = 1.5 x 1,200 / 60 = 30 gpm. AMI shows the DMA's actual minimum night flow at 90 gpm. Real loss is roughly 90 minus 30, or 60 gpm, the number worth chasing down.

Water balance and real loss: a utility's production meter reads 500,000 gallons for the day. Billed metered consumption is 460,000 gallons, and the utility already knows 15,000 gallons went to authorized, unbilled hydrant flushing that day.

Water Losses = System Input - Authorized Consumption = 500,000 - (460,000 + 15,000) = 25,000 gallons. If 5,000 of those gallons trace to a known meter-reading error, an apparent loss, Real Losses = 25,000 - 5,000 = 20,000 gallons for the day.

Infrastructure Leakage Index: a utility calculates its Current Annual Real Losses at 75 million gallons a year and its Unavoidable Annual Real Losses at 50 million gallons a year. ILI = CARL / UARL = 75 / 50 = 1.5, inside the 1.5 to 2.5 range typical of a utility that is actively managing losses, not yet at the practical floor of 1.0.

Check your own math against the free calculator at /tools/practice before you trust an answer on a timed test.

Common traps

• Treating one high flow reading as proof of a leak. It is a triage flag, not proof; confirm with interval data or a field check first.
• Mixing up apparent and real losses on category questions. Apparent losses are a billing and metering problem; real losses are water physically gone from the pipe.
• Quoting the old unaccounted-for-water percentage as the current standard. The exam wants ILI language and the CARL-to-UARL relationship at higher grades.
• Trusting a DMA without checking its boundaries. A boundary valve that is not fully closed throws off every number built on that DMA.
• Flipping the ILI formula. ILI is CARL divided by UARL, current losses over the unavoidable floor, not the reverse.
• Assuming a bigger meter is the safer choice. Oversizing kills low-flow accuracy and turns into a slow, silent revenue loss nobody notices until the audit runs.

Practice

Run the Water Distribution practice test at /tools/practice and filter to System Components questions on meters, AMI, and water loss. Drill the DMA night-flow math and the water-balance categories until you can sort a scenario into apparent or real loss without pausing. If a question asks for ILI, write CARL over UARL before you touch a number.

Quick reference

• PD meters: best low-flow accuracy, the residential standard, worn by grit and sand over time.
• Single-jet meters: lower cost, fine at normal flow, miss small continuous flows.
• Compound meters: PD plus turbine for a wide flow range, more maintenance points.
• Mag meters: no moving parts, need a full pipe and conductive water, grounding is the top failure point.
• Ultrasonic meters: no moving parts, air is the enemy, need good straight-run installation.
• Propeller meters: strong at sustained high flow, weak at low or variable flow.
• Manual read equals one number a month. AMR speeds up collection but is still close to one number. AMI reports automatically, often every 15 to 60 minutes.
• Minimum night flow, roughly 2 to 5 a.m., is the cleanest leak signal AMI provides.
• A District Metered Area is a metered, valve-bounded section, commonly 500 to 3,000 connections.
• Water balance: System Input Volume minus Authorized Consumption equals Non-Revenue Water.
• Apparent losses: water billed wrong, no water physically lost.
• Real losses: water physically escaped from mains, services, or tanks.
• ILI = CARL divided by UARL. Near 1.0 is the practical floor; 1.5 to 2.5 is typical for an actively managed system.
• Most states that cap unaccounted-for water set the threshold around 10 to 15 percent of production.