An asset health index for every transformer, breaker, cable, pump and main — each score traced to its test, its reading and its page
The same capital budget removes 2.5× the risk, every score traced to its evidence.
See one case, screen by screen ↓The condition evidence for one grid transformer comes from at least five places: the oil laboratory's certificates for dissolved gases, furans and moisture, two to four a year; test reports for bushing power factor, winding resistance and the tap changer; thermal surveys; load and hot-spot temperature in the historian; and the work order history in the EAM. Most of it arrives as PDF and is retyped months late, or not at all, so it is rarely put side by side.
So the capital plan replaces the oldest units first. On the sample fleet, an age-based plan spends $52.8M replacing the eight oldest grid transformers — four of which still score healthy — and misses Ferncliff T2, which is younger but has rising thermal gases and aged paper. The cost of guessing wrong is a failed transformer at peak with a lead time of about 120 weeks for a new one, and a plan the regulator can question because it cannot be traced to evidence.
Estimated split for one asset class's condition review and capital plan, in working days, by hand and with the solution.
Six specialist agents read the evidence, score, rank and forecast every asset; the asset manager signs the plan.
Monday 12 October 2026 at Lumina Utilities. The 2027–2031 capital plan for 64 grid transformers is due at the capital planning board on Thursday 29 October. Ferncliff T2 — 132/33 kV, 90 MVA, built by Calder Electric in 1979 — dropped three points overnight. Here is what happens next, screen by screen, in the working solution.
Imran owns the 64 grid transformers and their five-year plan. They sit as dots on a 5×5 matrix, health from H1 Good to H5 Very poor down the side and criticality from C1 Low to C5 Extreme across. Three sit in the red corner today — Dunmore T1 at 24, Ferncliff T2 at 31 and Saltmarsh T3 at 36 — carrying $3.9M a year of monetised risk.
In the red corner: “3 transformers with health below 40 and criticality 60+” · $3.9M a year, probability × consequence
The nightly run read 23 new documents — 9 lab certificates, 6 test reports, 3 thermal surveys and 5 work order closeouts — and tied each value to its asset. Four scores changed. Ferncliff T2 went from 34 to 31 on Brightwater Oil Laboratories certificate BOL-26-44817, and two changes of more than 10 points went to a reliability engineer instead of the plan.
“Health scorer re-scored 1,656 assets · 4 scores changed · Ferncliff T2 34 → 31”
The certificate for the sample taken on 14 September shows ethylene at 182 ppm against Lumina's alert limit of 50, up from 98 in March, with methane at 96 and acetylene at 1. The interpreter places it in zone T3 of Duval triangle 1 — a thermal fault above 700 °C — and works out the rise at about 14 ppm a month. The values it used are highlighted on the page.
Lab interpretation: “Duval triangle 1, zone T3 — thermal fault above 700 °C … Recommend resample in 4 weeks.”
Health starts at 100 under scoring method v3.2. Dissolved gases take 22 of a possible 30, paper insulation 14 for 2-FAL at 2,140 ppb (a degree of polymerisation of about 370), age 9 for 47 years against a 45-year design life, and load 8 for a 94% summer peak with a 118 °C hot-spot. The tap changer, bushings and moisture take the rest: 69 points, a score of 31, H4 Poor.
“7 deductions · every one opens to its page”
Ferncliff T2 serves 41,800 customers, Northgate General Hospital and Lumina's own Ferncliff water treatment works (120 Ml/d). Its sister unit T1 carries only 78% of peak alone, and its 38,000 litres of oil sit 60 m from Ferncliff Brook: criticality 88, Extreme. The forecaster takes it to 12 by 2031, with the probability of failure rising from 2.9% to 11.4% a year.
“Forecaster projected to 2031 · health 12 by 2031 if nothing is done · failure 2.9% → 11.4% a year”
Test report TR-2026-7716 on Corran T1 reads contact wear of 85% and arcing marks on the tap changer selector. The health scorer moves it from 61 to 50, but 11 points is more than the 10 the settings allow, so the change waits, with one button to ask Leah to confirm it and one to reject the reading.
“More than 10 points, so it waits for a reliability engineer before it reaches the plan.”
Imran drags the year slider to the end of the plan period. With no plan, every dot moves to where its forecast puts it, trails showing the path from 2026. Dunmore T2, Wexley T2, Kestrel Bay T1, Ashby Vale T1, Corran T1, Brackenridge T2, Northgate T3, Saltmarsh T1 and Holloway T1 join the three already there: 12 in all, and $10.7M a year of risk.
“12 · +9 since 2026 · transformers with health below 40 and criticality 60+”
Switched to the risk-ranked plan, replaced units jump back to the top, refurbished ones rise a band and monitored ones carry a ring. One transformer is left in the red corner in 2031 — Kestrel Bay T1, on an online gas monitor from 2027 and due for replacement in 2032 — and the class's monetised risk falls to $3.6M a year. Ferncliff T2 reads 96 in 2031.
“1 · −2 since 2026 · transformers with health below 40 and criticality 60+”
The capital plan puts three options side by side. Replacing the eight oldest removes $6.7M of risk over 2027–2031 and leaves 8 in the red corner. The risk-ranked plan — 7 replacements, 6 refurbishments and 10 gas monitors — removes $16.9M, leaves 1 and defers 4 healthy units worth $26.4M. Ferncliff T2's order goes in by February 2027 to meet the 120-week lead time.
“Total $52.8M of the $52.8M envelope · risk removed 2027–2031 $16.9M · 2.5× replacing by age”
The sign-off shows what he is signing: $52.8M, 23 interventions and 1 transformer in the red corner in 2031, built on 64 scores with every deduction cited. The two score changes still waiting for Leah are left out of the evidence pack. The first order to place is Saltmarsh T3 in November 2026, and his note to the board says Ferncliff T2 must be ordered by Feb 2027 to meet the 2029 summer.
The plan builder recommends; the asset manager signs and the capital planning board approves the spend.
The same transformer, seen by the five people who score it, plan it, defend it and test it — what their month looked like, and what it looks like now.
Six specialist agents read the evidence, score, rank and forecast every asset; the asset manager signs the plan.
Reads lab certificates, test and inspection reports, thermal surveys and work order closeouts as they arrive, extracts every measured value with its unit, date and page, and ties it to the asset by serial number, site and position.
Interprets dissolved gases and furans with IEEE C57.104 and IEC 60599: each gas against the alert and action limits, its monthly rate of change, Duval triangle 1 and Rogers ratios, and 2-FAL converted to a degree of polymerisation.
Starts at 100 and deducts up to each condition factor's weight under the approved scoring method, listing every deduction with its source page. Where a factor has no evidence it scores from age and labels the score estimated.
Sets each asset's consequence of failure, from 1 to 100 and in dollars, from customers served, critical sites, network redundancy, environment and safety, using the network model and the critical customer list.
Projects health and probability of failure year by year to the end of the plan period from age, load growth and condition trends, drawing on the failure history from 1996 to 2025 — the grid transformer failure curve is calibrated on 31 failures.
Chooses the replacements, refurbishments and gas monitors that remove the most monetised risk inside the budget envelope, checks supplier lead times and outage windows, places each order date, compares the plan with doing nothing and replacing by age, and drafts the justification.
Signs; engineer confirms big changes; board approves. The agents propose; a named person decides.
Ask about any asset, score or forecast in plain words, or say what to change in the plan. Every answer points to its evidence.
Why is Ferncliff T2 scored 31?
Seven deductions take it from 100 to 31. The largest are its gases: ethylene is 182 ppm, rising 14 ppm a month, and the Duval triangle puts it in zone T3, a thermal fault above 700 °C, so 22 points. Its paper is aged: 2-FAL 2,140 ppb, about DP 370, 14 points. It carried 94% of nameplate on 22 Jul with a hot-spot of 118 °C. With criticality 88 — 41,800 customers, the hospital and the water works — that is $423K a year of risk.
Which grid transformers reach the red corner by 2031 if we do nothing?
Nine join the three already there, so 12 by 2031: Dunmore T2 (41 → 15), Wexley T2 (42 → 20), Kestrel Bay T1 (44 → 20), Ashby Vale T1 (47 → 26), Corran T1 (50 → 27), Brackenridge T2 (50 → 32), Northgate T3 (48 → 32), Saltmarsh T1 (56 → 38) and Holloway T1 (52 → 38). Monetised risk for the class rises from $3.9M to $10.7M a year.
How does the risk-ranked plan compare with replacing by age?
Both spend $52.8M. Replacing the eight oldest leaves 8 in the red corner in 2031 and removes $6.7M of risk over 2027–2031; the risk-ranked plan leaves 1 and removes $16.9M, 2.5× as much. It defers four old units that still score 64–71 ($26.4M) and replaces Ferncliff T2, which age misses.
Bring Kestrel Bay T1’s replacement forward to 2031
Added. Brought forward from 2032 to 2031 it costs $6.6M, and the order must be placed by Feb 2029 for the 120-week lead time. The red corner in 2031 drops to 0, but the plan is now $6.6M over the $52.8M envelope. Switch a line off, or take the difference to the board.
13 screens from the working solution, on its sample data. Pick one to see it large.
The red corner, the forecast, who confirms a score and how often the fleet is re-scored are settings, chosen by the asset management team.
| Setting | Default | Choose from |
|---|---|---|
| Red corner — health below | 40 | Steps of 5 on the 1–100 health scale |
| Red corner — criticality at or above | 60 (Very high and Extreme) | 40 (High and above) · 60 (Very high and Extreme) · 80 (Extreme only) |
| Forecast horizon | 2031 (end of the plan period) | 2029 · 2031 (end of the plan period) · 2036 (ten years) |
| Load growth in the forecast | 1.8% a year | 1.2% a year · 1.8% a year · 2.5% a year (electrification case) |
| Score change that needs an engineer | More than 10 points | More than 5 points · More than 10 points · Any change of band |
| Reliability engineer | Dr. Leah Morrow | Dr. Leah Morrow · Samir Haddad — confirms score changes for transformers and breakers |
| Scores without evidence | On | On · Off — when on, an age-only score is shown as estimated, never as measured |
| Nightly re-score | Every night at 05:30 | Every night at 02:00 · Every night at 05:30 · Weekly, Sunday 05:30 |
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 fleet · “target” = the design goal, measured in the live solution · “estimated” = our arithmetic, inputs shown ($26.4M = 4 units × $6.6M installed, the sample fleet's quoted unit cost; the working-days split is our estimate) · “typical” = published figures (Wood Mackenzie: power transformer lead times about 120 weeks in 2024, up from about 50 in 2021; prices up 60–80% since 2020). Standards named as context: IEEE C57.104 (2019), IEC 60599, the UK common network asset indices method (CNAIM v2.1). Lumina Utilities and the people, sites, suppliers, laboratory and records named on this page are fictional — characters and sample data in the working solution.
An asset health index is a score of each asset's condition — here from 1 to 100 — built from its test results, inspections, age and load. Put next to criticality, the consequence of failure, it shows which assets carry the most risk. Asset Health Index scores every asset this way and shows the evidence behind every point.
Six agents read each lab certificate, test report, thermal survey and work order closeout as it arrives, interpret the oil results, score the asset under your approved method, set its criticality and forecast it to the end of the plan period. The fleet sits on a health × criticality matrix, and the plan builder ranks interventions by risk removed per dollar. People confirm, sign and approve.
Health starts at 100 and each condition factor deducts up to its weight: dissolved gases 30, paper insulation 20, age 15, load and temperature 10, tap changer 10, bushings 8, moisture in oil 5, visual and thermal 5. Gases are judged with limits based on IEEE C57.104 and IEC 60599, including the Duval triangle; furans give the degree of polymerisation. The weights and limits are yours to set and approve.
No. The sample utility scores 1,656 assets in six classes with one method: grid and primary transformers, 132 kV circuit breakers, 33 kV underground cable, water pumping stations and trunk water mains. Each class has its own condition factors and its own forecast to 2031. Across the six, the red corner holds 25 assets today and 86 by 2031 if nothing is done.
It shows the case. In the sample fleet, four of the eight oldest grid transformers still score 64–71, so the risk-ranked plan defers them ($26.4M) and spends the money where the risk is. A deferral is only real when your planning team and the board accept it — every score behind it opens to its evidence.
Yes. The plan builder checks supplier lead times and outage windows and places each order date. With a quoted 120 weeks for a 90 MVA unit, Ferncliff T2's replacement must be ordered by Feb 2027 to be in service for the 2029 summer.
Yes. A reliability engineer confirms any score change of more than 10 points, the asset manager signs the plan and the capital planning board approves the spend. Agents never commit spend, change a weight or a gas limit, or lower the criticality of a site on the critical customer list.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your asset register, scoring method, oil results, test reports and failure history — and it goes live once every quality gate has passed. We will walk you through it on your own fleet first.
Defects caught while they grow: every structure compared with last year's flight.
1-min demoMaintenance Knowledge SearchDecades of fixes, searchable before the expert retires.
1-min demoTransformer DGA InterpretationEvery oil result trended the day it lands, with advice set by how fast the gas is rising.
1-min demoWe'll run Asset Health Index on a sample of your own asset register, oil results and test reports.