Non-revenue water management that sends leak crews where the loss is certain, not where a meter is in doubt
Leak crews sent where the loss is certain, not where a meter is in doubt.
See one case, screen by screen ↓Non-revenue water — real losses, apparent losses and unbilled authorised use — is spread across production meters, the billing extract, network records and field data. So the water balance is usually built once a year, in a workbook, for the whole system. By the time it is finished, the leak it points at has run for months. Billing reads lag the period by weeks, unbilled use from flushing and firefighting is a guess, and meter under-registration is assumed rather than measured.
Real losses are what is left once everything else is counted, so every other figure's error lands in them. A zone inlet meter that reads 10% high looks exactly like a big leak. Ranked by the headline figure, the zone with the doubtful meter comes first, and the survey budget goes on a loss that is not there while the real leak keeps running.
Estimated split for one balance of a supply area and its zones, in working days, by hand and with the solution.
Seven specialist agents build and check the water balance for every zone; the water loss manager decides where crews go.
Lumina Utilities supplies the Eastvale supply area: 18 district metered areas, 136,330 connections and 991 km of mains, with 61.99 M m³ put into supply in the 12 months to 30 Sep 2026. On Tuesday 13 October the agents rebuild the balance at 06:12. Here is what happens next, screen by screen, in the working solution.
Hannah owns the balance, the leak survey budget and the annual audit. Her home screen is the standard water balance for Eastvale: 61.99 M m³ put into supply, 45.51 M m³ billed, 0.79 M m³ unbilled authorised, 3.01 M m³ apparent losses, and real losses as what is left. Non-revenue water is 26.6% ± 1.2, worth $16.44M a year. Beside it, all 18 zones are ranked by the part of their real loss that is certain.
“Real losses are what is left once everything else is counted, so their band carries the doubt of every other figure.”
The meter data reconciler reads 4 production and 18 zone inlet meters — 2.1 million 15-minute flows — fills three missing nights for Weir Lane and applies each meter's tested error. The billing period aligner fits 1,084,212 reads to 1 Oct – 30 Sep and flags 1,910 estimated and 418 stuck meters. The water balance builder counts 412 hydrant permits, 96 mains-flushing jobs and the fire service log. The apparent loss estimator sizes meter under-registration from 2,412 bench tests. Then 19 balances are built, each with a 95% band.
Night-flow analyst: “16 zones agree within their band · Kestrel Fields (−41%) and Weir Lane (−27%) do not: inlet meters in doubt”
Ranked by likely value, Zone 07 Kestrel Fields comes first: real losses of 32.6%. But the band is ± 12.1 pts, from 20.5% to 44.7%, because its mechanical inlet meter was last tested on 19 Aug 2019. Night flow gives only 0.60 M m³ of real losses for the year against 1.02 M m³ in the balance — 41% apart. The action offered is an inlet meter test, $3,200 and one day, not a survey.
“The band is ±12.1 pts: most of this figure could be the inlet meter. A survey now could find nothing.”
Ranked by certain loss, Zone 12 Ashcombe Heights is first. Its 7,420 connections and 64 km of mains took in 3.41 M m³; real losses are 1.06 M m³, 31.0% ± 2.2, with a certain part of 28.8%. Its electromagnetic inlet meter was verified on 9 Sep 2026, so confidence is high. ILI 7.0, 390 litres per connection a day, $1.25M of lost water a year.
“Certain real loss 28.8% is above the 20% trigger, and the band is within ±3 pts.”
The zone map draws every district metered area by its certain real loss, with zones whose band is wider than ±6 pts hatched — Kestrel Fields and Weir Lane. Hannah can colour it by likely loss or by band width instead. Ashcombe Heights, on the River Ash, is the darkest zone, and its summary offers the leak survey.
Minimum night flow between 02:00 and 04:00 stepped from 128 to 186 m³/h on the night of 14 June, minutes after Pump 2 at the Ashcombe booster tripped and restarted, with a pressure surge from 58 to 81 m. No burst was reported. Bottom-up, the night flows give 1.02 M m³ for the year — within 4% of the balance's 1.06 M m³, inside its band. At today's rate the zone is losing 1.36 M m³ a year.
Intervention planner: “The step test of 1–2 Oct puts 86 of the 162 m³/h net night flow on Quarry Lane — 19 km of mains, mostly 1950s cast iron.”
Each morning the night-flow analyst checks all 18 zones and lists what changed: the Ashcombe Heights step, Kestrel Fields' night flow too low for its balance, three missing nights and a pump-station meter reading 6–9% above its pump curve in Weir Lane, Copperhill creeping up 0.12 m³/h a night since May, and the Cardew Park burst on Fenwick Road, repaired and verified. Tomás has a step test booked in Copperhill for 20 Oct.
The night-flow analyst flags; it never closes a zone's alert.
Hannah sends the survey from the balance. The order goes to Calder Acoustic Surveys under framework contract LW-2024-07: 64 km of mains and 7,420 services, Quarry Lane first, acoustic correlation and ground microphone on the nights of Thu 15 Oct to Mon 19 Oct. It costs $19,840 — 64 km at $310 — within her $25,000 limit.
Target: “net night flow 162 → 70 m³/h · 0.77 M m³ and $479K a year”
Survey order LS-2610-12 moves to Booked. The plan shows every action with its water and value a year, cost, payback and owner: night pressure 71 → 55 m at PRV RE-2 in Riverside East for 0.31 M m³ and $192K a year, the Kestrel Fields meter test first, 1,240 customer meters over 20 years old being replaced in Sandown Edge, and the Fenwick Road burst repaired and verified.
The intervention planner proposes; it never dispatches a crew or spends money.
The audit report writer drafts the annual water audit from the same figures: the water balance in thousand m³ with its 95% bands, the performance indicators with their data grades — real losses of 255 litres per connection a day, ILI 4.9, 15 of 18 zones with high confidence — and a board note in plain words, every figure cited. Hannah sends it to Owen, and Owen signs.
Board note: “Two zones, Kestrel Fields and Weir Lane, show high losses we cannot yet trust: their inlet meters are in doubt. We will test those meters before sending crews.”
The same balance, seen by the five people who act on it — what their year looked like, and what it looks like now.
Seven specialist agents build and check the water balance for every zone; the water loss manager decides where crews go.
Reads production and zone inlet meter flows every 15 minutes, fills gaps of up to 3 nights from the same weekday profile and marks them, applies each meter's tested error and accuracy, and flags meters whose test is overdue or whose flow disagrees with pump hours.
Fits every read in the billing extract to the balance period, prorates consumption by days, and finds accounts estimated for 2 or more cycles, meters reading zero on occupied premises for 3 or more cycles, and supplied connections with no account.
Fits the under-registration of each meter age band from bench tests, applies it to each zone's billed metered volume, and adds read errors and an estimate of use without permission, each with a band.
Builds the standard water balance for the supply area and each zone — water put into supply, billed and unbilled authorised use, apparent losses and real losses as the remainder — gives each input its 95% band from its source and combines them.
Every day at 05:00 takes each zone's 02:00–04:00 minimum night flow, subtracts legitimate night use, finds steps held for 3 nights and slow creep, checks real losses bottom-up and looks for pump trips and pressure surges at the time of a step.
Proposes the action each zone's band allows — leak survey, inlet meter test first, pressure review or meter renewal — sizes the water won back from night flow, values it, and ranks actions by certain water won back per dollar.
Fills the annual water audit table, the performance indicators and their grades, and a one-page board note in short sentences, from the balance.
Sends orders to $25K; the director signs the audit. The agents propose; a named person decides.
Ask why a zone ranks where it does, what a fix is worth, or tell it what to change. Every answer points to its meters, tests and records.
Why is Ashcombe Heights first and not Kestrel Fields?
Because Ashcombe Heights' loss is certain. Its real losses are 31.0% ± 2.2: the inlet meter was checked on 9 Sep and reads within 0.8%, and night flow agrees within 4%. Kestrel Fields reads 32.6%, but ± 12.1. Its mechanical inlet meter was last tested in Aug 2019, and its night flow implies only 0.60 M m³ against 1.02 M m³ in the balance. Much of that loss may be the meter.
What would fixing Ashcombe Heights be worth?
About $479K a year. Net night flow has been 162 m³/h since the pump trip on 14 Jun. Back to 70 m³/h keeps 0.77 M m³ a year in the network — water for about 4,300 homes — valued at $0.62/m³. Survey $19,840 and repairs about $46,000: payback in about 1.6 months. Quarry Lane carries 86 m³/h of the night flow, so the survey starts there.
Which zones need a meter test before any survey?
Two. Kestrel Fields (± 12.1 pts — inlet meter last tested in 2019) and Weir Lane (± 8.1 — the pump-station meter has read 6–9% high since April). Testing both costs $5,000 and narrows each band to about ±2.5 pts.
13 screens from the working solution, on its sample data. Pick one to see it large.
The survey rule, the prices, the confidence bands, the night-flow assumptions and who approves are settings, chosen by the utility.
| Setting | Default | Choose from |
|---|---|---|
| Propose a leak survey when the certain real loss is at least | 20% of input, band within ± 3 pts | Any trigger and band width — when the band is wider than ± 6 pts, an inlet meter test is proposed first |
| Valuation | $0.62/m³ and $2.85/m³ | Real losses at variable production cost · apparent losses at the average retail tariff, both in $/m³ |
| Water per home | 180 m³ a year | Any volume, used to state water saved in homes |
| Confidence bands | 95% · High ± 3 · Medium ± 6 | Confidence level, and the band width in pts for high and medium confidence · ± 30% on estimated unbilled volumes |
| Night flow | 02:00–04:00 · 2.6 L/conn/h | Night window, legitimate night use per connection per hour, and a night-day factor of 22–24 hours per zone from pressure logging |
| Step change alert | 15 m³/h | Any step in m³/h, held for 3 nights |
| Survey and meter-test orders | Hannah Okoye, up to $25,000 | Above that, Owen Mbeki, Director, water network, approves |
| Annual water audit and rebuild | Signed by Owen Mbeki · rebuilt on the 13th at 06:00 | Monthly, and on demand |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (Liemberger & Wyatt, Quantifying the global non-revenue water problem, Water Supply, 2019 — about 126 bn m³ and $39B of non-revenue water a year worldwide). Method context: the IWA water balance and ILI (Lambert et al.); AWWA M36 valuation — real losses at production cost, apparent losses at retail. Homes at 180 m³ a year. Lumina Utilities, its zones, contractors and records are fictional, and people named on this page are characters in the working solution.
Software that works out how much of the water a utility puts into supply is not paid for, and why — real losses from leaks and bursts, apparent losses from meters, reads and theft, and unbilled authorised use. Non-Revenue Water Management builds the standard IWA water balance for every zone each month, with a 95% band on every figure, and ranks zones by the part of their loss that is certain.
Seven agents read the production and zone inlet meter flows, fit billing reads to the period, count unbilled use from hydrant permits, flushing and the fire service log, and size apparent losses from meter bench tests. The water balance builder then builds a balance for the supply area and every zone, with real losses as the remainder and a band on every figure.
Two ways. Every figure's band shows how much of a zone's real loss could be meter doubt, and the night-flow analyst checks real losses bottom-up from minimum night flow. In the sample case, Kestrel Fields looks worst at 32.6%, but its band is ± 12.1 and night flow is 41% apart, so an inlet meter test comes first; Ashcombe Heights' 31.0% ± 2.2 agrees with night flow within 4%, so it gets the leak survey.
Yes. The balance and the annual audit show non-revenue water as a percentage of input and by value, real losses in litres per connection a day and the infrastructure leakage index for the supply area and each zone. Real losses are valued at the variable production cost and apparent losses at the retail tariff, both of which you set.
The apparent loss estimator sizes meter under-registration by meter age from bench tests, and the billing period aligner finds stuck and estimated meters and supplied connections with no account. In the sample, 418 meters read zero on occupied premises for 3 or more cycles. Stuck meters go to billing for exchange in one list; suspected unauthorised connections go to revenue protection, and no customer is accused.
Yes. The agents have read-only access to meter, SCADA and billing data, and propose only. The water loss manager sends every survey and meter-test order up to $25,000, the director approves above that and signs the annual water audit, and every order and signature is recorded.
Yes. The survey trigger and band width, the valuation prices, the confidence bands, the night window, legitimate night use and night-day factors, the step change alert, the approval limits, who signs the audit and when the balance is rebuilt are all settings.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your water loss methods, zone list, meter registers and a sample of meter, billing and night-flow data — and it goes live once every quality gate has passed. We will walk you through it on your own zones first.
Exceedances caught the hour they happen, and low chlorine seen two days ahead.
1-min demoLeak Detection AnalyticsSend crews to real leaks, not false alarms.
1-min demoWater Quality Complaint InvestigationBrown-water calls traced to their cause in minutes, the flush and every reply drafted.
1-min demoWe'll build Non-Revenue Water Management's balance on a sample of your own meter, billing and night-flow data.