Nuclearn Equipment AI Book a demo
PI Reads the sensor data and records you already keep

Equipment health,
explained.

Nuclearn Equipment AI reads your sensor trends, work orders, condition reports and operator rounds. It scores each piece of equipment with the findings and evidence behind the score, and checks your PM intervals against the industry PM basis. Your engineers see what needs attention first, and why.

Book a demo See what it does
  • Uses your PI, work order and CR data
  • Single-tenant, hosted by Nuclearn
  • Signs in through Nuclearn
Equipment AI's key findings for a pump in a simulated plant: it still starts and runs, but an open work order to replace its outboard bearing has been in progress 157 days and repeats an earlier replacement, both bearing temperature channels have stopped responding, its bearing and lube-oil PMs are overdue, and seal leaks recur.

Screenshots on this page show Nuclearn Equipment AI running on data from a simulated plant, not a real one.

  • Backed by NuclearnAI software built for nuclear plants since 2021
  • Over 70 facilitiesNuclearn products serve sites in North America and the U.K.
  • ISO/IEC 27001:2022Certified information security management
Why teams use it

From the site heat map to the pump that needs a look.

Your plant already records how its equipment is doing, across the historian, work orders, condition reports and rounds. Equipment AI reads them together for each location and writes down what it found.

System engineers

See what needs attention across the site

Health by site, unit and system on one heat map, the issues driving each score, and a click from the station down to the pump.

See site health → System engineers

Read why, with the evidence

For each location: a health score, the key findings with the sensor values and records behind them, recommended actions by priority, and how confident the analysis is.

See the analysis → Maintenance and PM owners

Check PM intervals against the industry basis

Each PM's defined interval and last completion, against the recommended interval for that equipment's criticality, duty and service.

See PM intervals → Monitoring and diagnostics

Triage alarms with likely causes

Each alarm firing sorted as nuisance, real and monitor, or real and act, with ranked causes, cited evidence, and actions marked online or out of service.

See alarms → Engineers troubleshooting

Ask about any set of sensors

Pick the sensors and locations you care about and ask in plain language. Answers come with plots, statistics and the work history behind them.

See questions → Equipment reliability

Fresh scores before each health review

Run the analysis daily, weekly or on your own schedule, and see which data sources have gone quiet.

See the details →
The problem

The data is already there. Reading it all is the job.

A system engineer's picture of a pump comes from the historian, the work order history, the condition reports, the operator rounds and the PM program, each kept in its own system. Pulling them together for a health report or a meeting is slow work, and it happens on a calendar, not when a trend turns.

Most of a plant's work is periodic, so knowing which PMs run more or less often than the equipment needs is part of the same picture.

Equipment AI reads it all for each location, scores what it sees, and shows its working.

80%+1
of the work done in a US nuclear plant is surveillance or preventive maintenance

INL, citing the Nuclear Energy Institute: more than 80% of the work historically performed in a US nuclear plant.

INL's figure describes the industry, not results from Nuclearn Equipment AI. Source at the foot of the page.

How it works

From your plant's data to a score you can read

Equipment AI works from the hierarchy and records you already keep.

Bring in your data

Load your PI Asset Framework hierarchy, then send sensor data, work orders, condition reports and operator rounds from your side.

Equipment AI reads it

For each location it follows your plant's analysis procedure: sensors first, then work and condition history and PM adherence.

Your engineers act

Scores roll up by system, unit and site. Engineers read the findings, add feedback for the next run, and decide what to do.

AI analysis

A health score you can read, not just a number.

Each location gets a score from 0 to 100, a verdict of healthy, watch or critical, and the reasons: key findings graded by severity, each with the sensor values, baselines, setpoints and records behind it. Recommended actions come ranked by priority, with the data the analysis reviewed and how confident it is.

  • Your plant edits the analysis procedure and the scoring rubric.
  • Engineers add feedback that the next run takes into account.
  • A written report with citations for each location.

An analysis is advice to your engineers, not an engineering disposition.

The executive summary of the AI analysis for the same simulated pump: health score 30, critical, confidence high. The biggest driver is a repeat bearing work order open 157 days; the pump's two bearing temperatures barely change between running and stopped, while the other train's change by 33 and 61 degrees; and the pump has not functionally failed.
The AI analysis for a pump in a simulated plant: the score, the confidence and the reasons.
Site health

From the station to the pump in a few clicks.

Health rolls up by site, unit and system, as a heat map or a table, coloured by health or by PM compliance. The top drivers list what is behind each score. Open a system to see its locations grouped as critical, watch or healthy, with their sensor trends beside them.

  • Show every sensor, or only the ones the AI has referenced. Filter by component type, or search for a location ID.
  • A health history for each location, with work orders, condition reports and rounds on the same timeline.
  • A notice when a data source has gone quiet or scores are out of date.
A sensor trend from a simulated plant: component cooling water heat exchanger outlet flow over a week, jumping from about 880 to 970 pounds per second on October 16 and settling near 930.
A sensor trend from a simulated plant's component cooling water system.
PM intervals

Your PM program, against the industry basis.

For each location, Equipment AI matches your PMs to the recommended tasks for its component type, criticality, duty and service. It compares the interval you've defined, and when the task was last done, with the recommended interval.

  • Compliant, or how much less often than recommended.
  • PM compliance by location, rolled up the hierarchy.
  • Export the comparison to CSV.

Recommendations come from the PM basis database you provide. Changing a PM interval stays with your PM program.

PM intervals for a pump in a simulated plant: lube oil sampling and bearing inspection run less often than recommended, by seven weeks and 1.4 years; the overhaul has no defined schedule; lube oil change, vibration analysis and seal inspection are compliant.
PM intervals for a pump in a simulated plant, against the recommended interval for its type and duty.
Alarms and questions

A first look at each alarm. An answer about any sensor.

Equipment AI gives each alarm firing a first-pass disposition for your monitoring team, and answers engineers' questions about the sensors and locations they pick.

Equipment AI chat
Why has service water pump B's motor current been climbing since Tuesday?
Query_Time_Series_DataQuery_Rounds_Readings
Motor current rose from about 92 to 108 A over three days while discharge pressure fell from 63 to 54 psig, which points to the pump, not the motor. Rounds noted higher bearing noise on Wednesday. The quarterly lubrication PM is overdue. Next step: a vibration reading on the pump bearings before the next shift.

Example conversation with synthetic details. Answers come with plots and the records behind them; your engineers decide what to do.

Alarm triage

For each alarm firing, Equipment AI proposes a disposition with its reasoning, for your monitoring team to review.

NuisanceNoise or a setpoint that needs a look
Real — MonitorWith a proposed setpoint and its justification
Real — ActionA real condition to act on
Likely causesRanked, with cited evidence and ruled-out causes
Online or out of serviceActions sorted by what the plant can do now
Export PDFMarked AI-generated and advisory

Alarms are grouped by your PI Asset Framework tree, and each one has an investigation plan your engineers can edit. Setpoint changes go through your own engineering approval.

Built for your data

Your hierarchy, your records, your procedure

PI Asset Framework hierarchy

Import your PI AF export to build the site, unit, system and equipment tree, with sensors attached where they belong.

Your records

Work orders and work order logs, condition reports, operator rounds and narrative logs, alarms, sensor data and manual readings, each as its own source with when it last updated.

Scheduled analysis

Daily, weekly or on your own schedule, so scores are fresh for each review.

Custom systems

Group any sensors and locations for a question, and share the view by link.

Your analysis procedure

The steps and scoring rubric each analysis follows are your plant's to edit.

What it is not

Equipment AI doesn't predict failure dates, and it isn't your maintenance program. It reads your data, explains what it sees and recommends what to look at; it doesn't change PM intervals or setpoints. Your engineers and your programs decide.

Deployment and security

Your plant data stays in your Nuclearn deployment.

Sensor data, records and analyses live in your plant's datasets on the Nuclearn platform, in Nuclearn's own datacenter, and access goes through the platform's sign-in and roles.

Single-tenant, hosted by Nuclearn

Each plant gets its own instance, hosted by Nuclearn in our own datacenter on our own hardware. Azure Government or your own servers on request.

Built for 10 CFR Part 810

How Equipment AI is hosted, operated and supported is set up for 10 CFR Part 810 export controls.

Data sent from your side

A small broker on your network sends the data Equipment AI asks for. The platform doesn't reach into your plant network.

Sign-in and roles

Single sign-on through your identity provider, or the platform's own accounts. Platform roles decide who can view Equipment AI, change its configuration and manage scheduled analyses.

Advisory by design

Analyses and alarm dispositions are recommendations. Equipment AI doesn't change setpoints, PM intervals or work orders.

A certified company

Nuclearn is ISO/IEC 27001:2022 certified. Equipment AI is switched on per deployment.

FAQ

Questions plants ask first

What data does it need?

Your PI Asset Framework hierarchy and the sensor data under it, plus work orders, condition reports and operator rounds. Narrative logs, alarms and manual readings add to the picture. Equipment AI works with the sources you have and tells you when one goes quiet.

Does it connect to PI directly?

No. A small broker on your side sends the sensor data Equipment AI asks for, in PI Web API format, so the platform never reaches into your plant network.

How is the health score worked out?

An AI analysis follows your plant's analysis procedure and scoring rubric: sensors first, then work orders, condition reports, rounds and PM adherence. It records the score, a verdict, its key findings with evidence, recommended actions and how confident it is. Your plant can edit the procedure, and engineers' feedback is used on the next run.

Where do the PM recommendations come from?

From the PM basis database you provide. Equipment AI matches your PMs to its recommended tasks for each piece of equipment and compares the intervals.

Does it predict when equipment will fail?

No. It shows the current health of each location and how it has trended, with the reasons. It doesn't forecast failure dates.

Where does it run, and who can sign in?

Each plant runs in its own single-tenant instance, hosted by Nuclearn in our own datacenter on our own hardware, with hosting, operations and support set up for 10 CFR Part 810 export controls. Azure Government or your own servers on request. People sign in through the platform, with single sign-on if you use it, and platform roles decide what each person can see and change.

Book a demo

See a station's health, explained.

We'll walk through a synthetic station: the heat map, a critical pump's analysis and its PM intervals, and how alarms get a first-pass disposition. Then we'll answer your IT team's questions about data, deployment and sign-in.

  • A live walkthrough on synthetic equipment data
  • Your questions on data, deployment and sign-in
  • Which of your systems to start with

Nuclearn Equipment AI is backed by Nuclearn, whose products serve over 70 facilities in North America and the U.K.

Book a demo

Tell us a little about your site and we'll set up a time.

Prefer email? contact@nuclearn.ai