An operator of a 24-turbine onshore farm schedules service by interval and crew availability, and books each stop as the stopped turbine's own lost output. This case follows ten weeks of planning maintenance, production settings and one siting question in a single model of the farm.
When to stop which turbine, and whether to steer wakes
Duration
Ten weeks, then a full season
Runs in
The operator's private cloud
> The starting point
Three Calendars,One Farm
Maintenance is planned by the service team against intervals and crew rosters. Production settings belong to the control room, and the question of reblading row C sits with asset management, in a spreadsheet nobody else opens.
Each team is right on its own terms. But a stop on a windy day costs far more than the same stop in a calm, a wake steered off one turbine lands on the one behind it, and none of that is visible from a single desk.
How the decision is made today
Who
Three teams, three tools
On what
Service intervals, crew availability, nameplate output
When
Weekly, by calendar
Lost output
Booked per turbine, wakes ignored
# What could not move
What The FarmWould Not Allow
Settled in the first week, before any model was built. Each one decided part of the architecture.
The controller stays in charge
Turbine control is certified, and it belongs to the manufacturer and the control room.
So the system recommends; it never writes to a turbine controller.
No stop just to test a plan
A turbine taken offline to verify a recommendation costs what the recommendation was meant to save.
So the system is validated by replaying past seasons, not by stopping machines.
The farm's data stays with the operator
SCADA records, forecasts and work orders are operational data the operator does not share.
So the system runs in the operator's private cloud, isolated to this deployment.
Crews are the hard limit
There is one crew, travel takes time, and access depends on wind and ground.
So the system plans around crew, access and travel before it plans around output.
% The architecture
One Model Of The Farm,Three Decisions In It
The farm is represented once. A plan is weighed against the wind it will meet, the loads it will cause and the access it needs before it reaches anyone's schedule.
T-07Main-bearing inspectionLowest lost production this week; crew already on site for T-09
T-03No wake-steering offset in the west sectorDownstream T-04 is close to its fatigue allowance
T-11Leading-edge repair moved to the next access windowDay 2 access is marginal; erosion loss at current wind is small
Row CReblading versus repowering, to evaluateHighest wake exposure and fatigue history in the farm
Stop for maintenance
Production setting
Siting review
Why T-07 on day 3
The forecast drops below cut-in speed for most of day 3. Stopping T-07 then costs a fraction of any other day, and the crew travelling for T-09 can take both jobs on one visit.
In the product, every line opens onto the forecast, records and load estimates it rests on. The plan is a recommendation; the operator decides.
~ What gets measured
Measured OverA Full Season
Wind does not repeat week to week, so the outcome is measured over a full seasonal cycle. Results are published with the operator's approval.
Measure
Baseline
How it is measured
Result
Production lost to planned stops
The same season last year, at farm level
Wakes included, normalised for the wind actually seen
Published on approval
Availability
The previous seasonal cycle
Time-based, per turbine and for the farm
Published on approval
Crew visits per job
Last season's work orders
Visits and travel per completed job
Published on approval
Fatigue on protected turbines
Load estimates before the engagement
Damage-equivalent load estimates on the turbines a setting protects
Published on approval
What stays with the operator
The farm model, documented, in the operator's own cloud
Every plan with its reasoning, and every decision taken on it
One record of wakes and loads the three teams now share
Close to yours if
Maintenance, production and siting sit with different teams
Stops are scheduled by calendar and crew availability
Lost production is booked per turbine
You are weighing reblading or repowering
Your operation, in the same format
The technical briefing is where your specific constraints get discussed, including which of these four your operation resembles. Results are published with each customer's approval, never before.
