Circuits already serve load. Outages and switching show up after the event has started. Substation and feeder assets fail before the crew is staged. The utility still has to restore.
01
High-risk transformers found in the test history, not after the failure
Predict health index from oil, DGA and paper-condition data the utility already collects, with class rebalancing so Poor and Very Poor units are not buried under overall accuracy.
02
Inspections aimed at the accounts consumption data already points to
Rank likely theft and non-technical loss from metering history and confirmed inspection outcomes, so field effort follows pattern rather than chance.
03
NIS2 obligations mapped before the reporting clock starts
Structure applicable cybersecurity requirements, documentation gaps and incident-reporting workflows from regulatory text and existing compliance records — for human review, not as a legal determination.
In the control room and in the field
Utility operators lose time and money in three places that rarely share one screen: a switching sequence that differs by crew, a feeder or transformer failure found after customers are already dark, and an approved procedure that sits in a shared drive while the event is live. Most of that is already in EMS, OMS, SCADA and CMMS — after the fact.
The networkPublished national analyses put the customer cost of major U.S. outages in the tens of billions of dollars a year, with counts and durations rising. That cost starts as a grid event — congestion, storm damage, equipment failure — before it becomes a call-center queue.
OutagesIsolation, LOTO and restoration sheets often sit apart from the live incident. When the approved pack is hard to find, shifts and contractors invent the next step — and the outage clock keeps running.
AssetsTransformers, breakers, reclosers and lines decide whether capacity is real for the customer. When condition is read reactively, an unplanned stop cascades into prolonged interruption and overtime.
Restoring by habit is the easy correction. It is also how the night is spent.
INSUS works from SCADA and OMS tags, outage tickets, switching packs and equipment logs a utility already keeps. If the gap is an unread transformer alarm, a feeder that never settled, or a clearance sheet that never joined the event, that is what we will say. Recommendations stay advisory. Existing EMS, SCADA and protection systems stay in charge.
Industry figures
Four external figures, each with its source. Nothing on this page is an estimate we made up. Where we cannot cite a number, we describe the situation in words instead.
Customer cost
$121B
estimated cost of major U.S. power outages to electricity customers in 2024 in Oak Ridge National Laboratory analysis — with a seven-year average above $67 billion and major outage counts rising about 40% from 2018 to 2024.
Source: Oak Ridge National Laboratory / U.S. DOE, analysis of major outage customer costs
Major outages
+40%
increase in the number of major U.S. outages between 2018 and 2024 in the same ORNL analysis (from 4,666 to 6,533), with average major-outage duration lengthening from 9.6 to 11.8 hours.
Source: Oak Ridge National Laboratory / U.S. DOE, analysis of major outage customer costs
C&I impact
$6,031
average cost of power outages to commercial and industrial customers in 2024 reported in that ORNL analysis — a reminder that each prolonged feeder or substation event has a customer bill beyond the utility’s overtime.
Source: Oak Ridge National Laboratory / U.S. DOE, analysis of major outage customer costs
EFORd spread
0.15–5.85%
EFORd range for combined-cycle resources across FERC Common Metrics respondents (2023) — showing how much availability can differ even within one technology class. Utility planning still depends on whether capacity shows up when demand does.
Source: FERC, 2023 Common Metrics Report
Common problems
None of these show up as a single alarm. They show up in the transformer that looked fine on average accuracy, the inspection that found nothing, and the report assembled after the clock has started.
Healthy assets dominate historical test datasets. A model can look strong overall while still missing Poor and Very Poor units — the ones that drive outages and capital risk.
See: Transformer Health Index PredictionRandom or complaint-triggered visits spread capacity across accounts with no data-driven ranking. Consumption anomalies theft produces are rarely modelled continuously across the customer base.
See: Electricity Theft DetectionNIS2 and electricity Network Code obligations sit apart from incident-response packs. When an event hits, teams rebuild classification, evidence and reporting steps under the early-warning deadline.
See: NIS2 Cybersecurity ComplianceUse cases
Each is a specific starting point — this transformer fleet segment, this customer territory — and a next conversation if it is worth your time.
Predict transformer health index from oil quality, DGA and paper-condition data your utility already collects, with attention to rare high-risk classes. Advisory only.
See the full use case
Detect likely electricity theft and non-technical losses from consumption data your utility already bills against, and prioritise field inspections for human review. Advisory only.
See the full use case
Map NIS2 cybersecurity obligations and electricity Network Code requirements, surface documentation gaps and prepare incident-reporting workflows for human review. Not a legal determination.
See the full use caseFAQ
A practical look at what INSUS needs, where it fits and what your team can expect.
The work that fits is the data a T&D utility already produces: SCADA and OMS tags, outage tickets, switching packs and equipment condition signals. Predictive Maintenance can start from signals you already log. Closed-loop protection and automatic switching are not promises we will make on day one.
Energy & Utilities is the parent overview across power, grid, renewables, water and waste. Power focuses on generating units — heat rate and plant availability. Utilities focuses on transmission and distribution — outages, switching, feeders and substations.
Usually not for SCADA, OMS and maintenance. Most utilities already have feeder and substation tags plus trouble history. If a critical sensor has been drifting or an alarm code is unused, that comes before any model.
No. The person who knows this feeder, this substation and this protection scheme stays responsible. The aim is to stop the next event repeating the same restoration miss, and to flag a developing equipment issue before it prolongs the outage.
No. It advises the control-room or field team. Existing EMS, SCADA and protection systems stay in charge of the loop. Protection systems are not touched.
No. Reliability depends on weather, vegetation, equipment condition and operating practice. We will not invent a SAIDI or CAIDI target for your system.
They are published U.S. national estimates — not measurements of your circuits. Your OMS, SCADA and work orders are the numbers that matter.
Yes. One feeder, one substation, or one outage program is usually the practical start. Measure that, then decide whether to widen.
Contact
