Get in Touch

A municipal hydro utility

Three hydro plants and a local grid, running on what they know about themselves.

A hydroelectric dam in a dry river valley

Utilities & Grid · Western Norway

A municipal hydro utility

Running Three Hydro Plants and a Local Grid on One Model

Screen capture of a hydro utility plant overview

The utility is a municipal energy company in western Norway: three hydro plants, a local distribution grid, power sales and broadband, run by a few dozen people. With NRY, the company brought its maintenance records, the plants' sensor data and the operators' logs into one asset model on Forge, and moved the plants from scheduled to predictive maintenance.

What changed across the plants:

Deviations predicted before they happen, the hours a plant does not loseYears of repair tickets read automatically to guide today's workMaintenance windows planned against forecast inflow and market pricesEvery resolution written back, so the model keeps learning from the plants

Building a Predictive Maintenance Engine for Every Asset

Calendar-based maintenance treats every asset as average. The utility's model treats every asset as itself: its history, its sensors, how it is run. Work happens when the machine needs it, and a planned outage lands when the market misses the megawatts least.

In a company of a few dozen people, the knowledge of a plant lives in a few heads. It is becoming the plants' shared memory. The model reads what an operator would have noticed, and says so.

Screen capture of a turbine health dashboard

Placing Every Outage Where It Costs the Least

Water rushing over a weir at night

Every turbine's vibration, temperature and output runs against its own learned baseline. Drift is flagged weeks before a failure threshold, with the evidence, the likely cause and the recommended window attached.

Forge weighs forecast inflow, demand and market prices to place each intervention where it costs the least generation. The planner proposes, the production team approves, and the plants keep earning.

All data presented herein is notional.