When Will Each Wind Turbine's Gearbox Fail?
An operator manages 1,100 wind turbines across 40 farms. Replacing a gearbox costs €250,000 and requires a giant crane that is booked 3 months in advance. If the failure arrives by surprise, the turbine is down for 4 months. Most turbines have not failed yet. Maintenance management wants to know when each one is likely to fail in order to plan cranes and spare parts.
Where the money comes from
What to do now
- 1Unify the failure history and SCADA data.
- 2Survival model per turbine.
- 3Plan crane campaigns by farm using the 6-month risk.
For your committee
One page with the decision, the impact, the risks, the steps and the KPI to track.
The print dialog opens: choose "Save as PDF".
Watch out: more volume is not always more profit
Replacing gearboxes too early wastes the useful life of a €250,000 part. Intervene when the risk x cost of the outage exceeds the value of the remaining useful life.
KPI for the committee
Fleet availability (%) and maintenance cost per MWh.
Detail for your data team
Proposed technique: Survival Analysis
Why: The question is 'when?', and most turbines have not failed yet (censored data). Survival analysis (Cox model or accelerated failure time models) uses that partial information correctly and returns a risk curve over time for each turbine.
Cost of getting it wrong: Intervening too early wastes the useful life of a €250k part; too late means 4 months of idle turbine and unsold energy.
Data needed: 10-year history: installation date and gearbox failure date (if it occurred).; Sensors (SCADA) every 10 minutes: oil temperature, vibrations, power, wind.; Metal particles in the oil (quarterly analysis), manufacturer and model.