REAL DATA from Yorkshire Water · City of Topeka · City of TorontoData sources: Yorkshire Water (Data Mill North, CC BY), City of Topeka, City of Toronto; open data also from City of Calgary and City of Kitchener. Outside simulators: LeakDB, BattLeDIM.
City water network: find leaks, rank pipes, show the money
CAPMAC watches the night flow of each metered zone, raises a leak alarm only when a checked rule fires, and orders pipes for renewal planning (the order is tested; a chance of breaking is not). The leak and renewal work below is tested on real utility data, then on other people's simulators, then on our own rigs.
The network on screen is a demo city of four zones. Its alarms come from the stored output of the zone leak check on our own simulated runs; the case studies hold the real-data evidence.
Real data first
Start here
Network map
Pipes are coloured by renewal rank (darkest red = top 10 % of the renewal order). Outlined zones carry an open alarm. Click a pipe, pump or reservoir.
Selected
Click an item on the map.
Asset tree
Selected
Choose a city, zone, pipe, pump or reservoir.
Live dashboard
Day 0 is the start of the stored run. Alarms appear on the day the check raised them.
Alarm list
Evidence card
Select an alarm.
Night-flow trends with normal bands
Audit log
When (browser time)
Role
Action
Alarm
Note
Case studies
What it does not claim
Checks
Check
Best evidence
What was shown
Evidence file
Every check listed is WORKING in our latest check record: it passed a held-out test on a fresh records. WORKING does not mean proven on real plant data; the evidence tag says what kind of data it passed on.
Money
Every price below isASSUMED. Replace it with your own. Water lost per day = leak size (L/s) × 86.4 kL a day × water price, counted from the day the check raised the alarm.
Water lost to open leaks
Alarm
Zone
Size L/s
kL a day
Rs a day
Rs if left the days above
Pipe renewal budget
The Topeka test (real data) found that the top 10 % of the ranked pipes held about 60 % of the breaks that came later. That is a record of one city's past, not a promise for yours.
Reports and exports
Pipe renewal ranking
Pipes ordered for renewal planning (material, age, length, earlier breaks); the order is tested, a chance of breaking is not. Remaining life for pipes is on the Remaining life tab as a range, not a date.
Order
Pipe
Zone
Material
Installed
Length m
Earlier breaks
Renewal rank
Decile
Connect your network
Read-only from your SCADA historian over OPC UA, or upload a CSV of district-meter flows and pressures. CAPMAC never writes to your plant.
CSV layout: timestamp (ISO 8601, with time zone), zone, inlet flow m3/h, critical-point pressure m. Shown times in this demo are browser time; the demo network runs in Asia/Kolkata (IST).
Learn this network
The normal bands on the trends are each zone's normal night flow, worked out from its own healthy weeks. No learning picture is hidden behind them.
Per-unit healthy levels and the physics constants are on the suite pageLearn this site.
How it works
1. Night flow
At night, customers use little. Each zone's inlet flow in the quiet hours is compared with what the other three zones imply for the same night.
2. Leak or use change
A steady extra draw that runs all night is a leak. A big customer's use falling is not, and is not priced.
3. Rank pipes
Pipes are ranked by material, age, length and earlier breaks. The ranking is tested city by city on later breaks.
What it does not do: promise a date. Remaining life is shown as a range on the Remaining life tab. Alarms report what the readings already show; the ranking orders pipes for renewal.
About and credits
Load your data
Add your readings and get a printable report. It takes three easy steps.
Looking for the file service...
1Add your readings file
Drop your file here, or choose a file A spreadsheet saved as CSV. One column for the time, and one column for each sensor.
This helps CAPMAC learn from your plant and sharpen remaining-life estimates. You can skip it.
Your dates are saved in this browser only. A named person can approve them later on the Remaining life tab.
3Get your report
Add your readings file first.
Advanced options
Most people do not need this. CAPMAC picks the columns, units and checks for you. Change them here only if something is wrong.
What is in the file
Add a file to see this.
Which column is what
Add a file to see this.
Which checks to run
Add a file to see this.
Report wording
Remaining life: limits and dates (optional)
A limit is the reading at which you would act. Leave it empty and CAPMAC uses a labelled assumed limit. Dates help it count from the last repair or cleaning.
Leakage management
DEMOGenerated demo city as of 19 Sep 2026 (day 80 of the stored run). The night-flow runs are our own simulated runs; the incident log, crew hours and audit inputs are generated samples. Price is assumed: Rs 20.75 a kL.
3 open leak incidentslosing 10.3 L/s together (890 kL a day, Rs 18,476 a day). Median awareness 6.2 days, location 30 h, repair run time 2.0 days over 14 closed demo jobs.
1. DMA league table
Zones ranked by how far their night flow now sits above their own normal, over the last seven nights. Demo city, as of 19 Sep 2026. Money uses the assumed price of Rs 20.75 a kL.
#
Zone
Night-flow rise
m3 a day
Money a day
Confidence
Since
Trend, last 60 nights
1
Zone 4 Hill (tank fed)
6.06 L/s
524
Rs 10,867
High
25 Aug 2026
2
Zone 1 North
2.91 L/s
252
Rs 5,225
High
14 Sep 2026
3
Zone 3 Riverside
1.33 L/s
115
Rs 2,383
Medium
19 Aug 2026
4
Zone 2 Market
no sustained rise
0
Rs 0
none
no rise standing
2. Leak and burst alert cards
A01 · Zone 3 RiversideMediumLeak growing
Water leaving zone 3 unexplained at night. Rise began 19 Aug 2026. Stage now:Located.
Detected20 Aug 2026 06:12
Located23 Aug 2026 11:12
Dispatchednot yet
Repairednot yet
Verifiednot yet
Awareness time
18 h
start of the rise to detection
Location time
3.2 days
detection to located
Repair run time
27.0 days so far
located to repaired
Loss and money (assumed price Rs 20.75 a kL)
Loss now: 1.33 L/s = 115 kL a day = Rs 2,383 a day
Lost since it began: 3,144 kL = Rs 65,243
If left 30 more days: Rs 71,499
Size when first raised: 0.47 L/s
Where to start (zone 3)
A ranked starting list for the survey crew, not a map pin.
cast iron, laid 1960 to 1989 (2 pipes, 455 m)
asbestos cement, laid before 1960 (2 pipes, 458 m)
asbestos cement, laid 1960 to 1989 (2 pipes, 710 m)
Sections to check first:
Z3P03: cast iron, laid 1975, 345 m, 3 earlier breaks
Z3P12: asbestos cement, laid 1948, 339 m, 2 earlier breaks
Z3P13: cast iron, laid 1953, 220 m, 1 earlier breaks
A03 · Zone 4 Hill (tank fed)HighBurst
Water leaving zone 4 unexplained at night. Rise began 25 Aug 2026. Stage now:Repaired.
Detected28 Aug 2026 06:12
Located28 Aug 2026 16:42
Dispatched29 Aug 2026 08:12
Repaired05 Sep 2026 11:12
Verifiednot yet
Awareness time
2.8 days
start of the rise to detection
Location time
10 h
detection to located
Repair run time
7.8 days
located to repaired
Repair logged, but night flow is still 6.0 L/s above normal 14 nights later. Not verified; the incident stays open.
Loss and money (assumed price Rs 20.75 a kL)
Loss now: 6.06 L/s = 524 kL a day = Rs 10,867 a day
Lost since it began: 13,145 kL = Rs 2,72,755
If left 30 more days: Rs 3,26,012
Size when first raised: 5.92 L/s
Where to start (zone 4)
A ranked starting list for the survey crew, not a map pin.
cast iron, laid 1960 to 1989 (2 pipes, 362 m)
asbestos cement, laid 1990 and later (3 pipes, 1,086 m)
asbestos cement, laid before 1960 (1 pipes, 256 m)
Sections to check first:
Z4P13: asbestos cement, laid 1953, 256 m, 3 earlier breaks
Z4P11: cast iron, laid 1988, 173 m, 3 earlier breaks
Z4P08: cast iron, laid 1952, 373 m, 0 earlier breaks
A02 · Zone 1 NorthHighBurst
Water leaving zone 1 unexplained at night. Rise began 14 Sep 2026. Stage now:Located.
Detected16 Sep 2026 06:12
Located18 Sep 2026 15:12
Dispatchednot yet
Repairednot yet
Verifiednot yet
Awareness time
42 h
start of the rise to detection
Location time
2.4 days
detection to located
Repair run time
21 h so far
located to repaired
Loss and money (assumed price Rs 20.75 a kL)
Loss now: 2.91 L/s = 252 kL a day = Rs 5,225 a day
Lost since it began: 1,778 kL = Rs 36,892
If left 30 more days: Rs 1,56,763
Size when first raised: 2.85 L/s
Where to start (zone 1)
A ranked starting list for the survey crew, not a map pin.
asbestos cement, laid before 1960 (2 pipes, 218 m)
cast iron, laid 1960 to 1989 (2 pipes, 497 m)
cast iron, laid before 1960 (1 pipes, 343 m)
Sections to check first:
Z1P12: cast iron, laid 1948, 343 m, 1 earlier breaks
Z1P07: cast iron, laid 1975, 350 m, 1 earlier breaks
Z1P01: asbestos cement, laid 1997, 357 m, 3 earlier breaks
3. Water balance (IWA layout)
Annual figures for the generated demo city, worked out by the water-loss audit arithmetic check (volumes in megalitres). The demo network is 300 km of main and 26,000 connections at 35 m average pressure; production price Rs 20.75 a kL and retail price Rs 35.00 a kL are assumed.
System input volume1,900.0 ML
Authorised consumption1,320.3 ML
of which unbilled15.3 ML
Water losses579.7 ML
Apparent losses (meters, data, theft)32.9 ML
Real losses (leakage)546.8 ML
open incidents now, per year if left325.0 ML
Revenue water1,305.0 ML
Non-revenue water595.0 ML (31.3% of input)
1.63Infrastructure leakage index (ILI)
335 ML a yearUnavoidable annual real losses (UARL)
Rs 1,13,47,037Cost of real losses a year at the assumed production price
4. Loss volume and money, open incidents
Assumed price: Rs 20.75 a kL (production cost; replace it with yours on the Money tab). Lost since the rise began = the nightly excess carried over 24 hours.
Incident
Zone
L/s now
kL a day
Money a day
Lost so far, kL
Lost so far, money
A01
Zone 3
1.33
115
Rs 2,383
3,144
Rs 65,243
A03
Zone 4
6.06
524
Rs 10,867
13,145
Rs 2,72,755
A02
Zone 1
2.91
252
Rs 5,225
1,778
Rs 36,892
Total
890
Rs 18,476
18,067
Rs 3,74,889
5. Repair prioritisation
Dispatch order for incidents with no crew yet: water saved per day for each crew hour spent. Crew hours come from closed jobs of the same kind in the demo log.
Order
Incident
Zone
Stage
kL a day
Crew hours
kL a day per crew hour
Why
1
A02
Zone 1
Located
252
15
17.2
Zone 1 loses about 252 kL a day (Rs 5,225); a repair is taken as about 15 crew hours (from closed jobs of the same kind in the log), so each crew hour saves about 17.2 kL a day. Send the crew here first.
2
A01
Zone 3
Located
115
8
15.3
Zone 3 loses about 115 kL a day (Rs 2,383); a repair is taken as about 8 crew hours (from closed jobs of the same kind in the log), so each crew hour saves about 15.3 kL a day. Next after the incident above.
in progress
A03
Zone 4
Repaired
524
15
34.9
Already with a crew (repaired); not re-ordered. It loses 524 kL a day, so keep the repair moving.
6. Location hints are inside each alert card above.
7. Burst, valve closure or demand surge
A sudden change in a zone's night inflow is named as one of these, with the reason in words. Pressure at the zone's critical point is a generated demo reading.
Event
Zone
Named as
Why
A01
Zone 3
Leak growing
Inflow climbed over many days with pressure flat or lower: a slow leak, not a sudden burst.
A03
Zone 4
Burst
Inflow jumped within a day or two, pressure at the far end fell or stayed flat, and it has not come back.
A02
Zone 1
Burst
Inflow jumped within a day or two, pressure at the far end fell or stayed flat, and it has not come back.
A04
Zone 2
Demand fall
Inflow fell with no pressure rise: customers are using less. This is not a leak.
The three side by side (generated sample)
Sample
Named as
Why
Pipe burst
Burst
Inflow jumped within a day or two, pressure at the far end fell or stayed flat, and it has not come back.
Valve closed by mistake
Valve closure
Inflow fell while pressure rose: the water is being held back, not lost.
Demand surge (event or filling)
Demand surge
Inflow rose but has already come back within a day or two, so it looks like use, not a break.
Real data from Yorkshire Water
Yorkshire Water DMA night flows and mains repairs, Data Mill North, CC BY
Question tested.Does a logged mains repair show as a step down in the zone's night flow (a retrospective check on past repairs)?
Repairs confirmed 0.30 against placebo 0.03 on one test set, and 0.30 against 0.05 on a second.
A third set of 247 zones (202 repairs): 0.262 against 0.034.
Leak start dated before a confirmed repair in 4 of 4 cases.
Alerts are tuned for a low false-alarm rate; trial on your own history recommended.
Result as stored in the case library. This card is real data and is not part of the demo city numbers above.
This view reports what the night flow is doing now and what the log says was done; remaining-life ranges are on the Remaining life tab.
Remaining life: city water
Faults this app caught, by equipment
Rows with a Replay button come from a recorded run with a named planted fault. Rows without one are checks that caught every planted fault in their scored tests.