Generator bearing temperature rise
Sharp temperature steps on the generator bearings — typically the drive-end or non-drive-end bearing — are one of the cleanest SCADA-detectable signatures in the entire turbine. The typical pattern is a 10–25°C jump above expected, against a calm power profile. The right action is usually a fast intervention (sensor / thermocouple / lubrication check) before the rise becomes thermal damage to the bearing itself.
Reviewed by Michael Tegtmeier, Founder & Managing Director · Last reviewed: May 10, 2026
What happens
Generator bearings carry the high-speed shaft and absorb axial and radial loads from the rotor and gearbox. They run at higher rotational speeds than the main bearing, so the thermal envelope is tighter. A clean temperature step — different from the slow drift on a main bearing — points at a discrete event: a thermocouple drift, a lubrication delivery problem, a cooling-system fault, or actual bearing damage. The combination of step size and step recovery (does it cool back down?) is what tells operations whether to intervene immediately or schedule a planned check.
Signs to look for
A step change of 10–25°C in bearing temperature, persisting across operating points (not a transient).[1,2]
Temperature differential against the model-predicted value blows past historical max.[2]
No corresponding wind / power explanation — the same load profile that worked an hour ago now produces a hotter bearing.[1]
Cooling-system temperatures may NOT shift, which is what distinguishes a sensor / lubrication fault from an actual cooling failure.[2]
Root causes
- Faulty thermocouple in the bearing housing or in adjacent components (slip-ring ventilation, etc.) — a sensor issue, not a bearing issue.[1,3]
- Lubrication system fault — failed grease pump, blocked grease line, lubricant aging.[3]
- Cooling-system fault — failed fan, clogged cooler, low coolant flow.[2]
- Actual bearing damage — fluting, electrical pitting from shaft currents, fretting, or fatigue spalling.[1,4]
How Turbit detects this
Step changes register with low latency in Turbit's per-turbine model: typical detection within hours of the actual event. The AI's root-cause classifier discriminates between sensor faults (no cooling-system shift), lubrication issues (slow recovery), and real damage (no recovery, accelerating trend). This shapes the operations team's response — verify-the-sensor vs. dispatch-a-crew.
From the Turbit fleet
Roughly half of generator-bearing temperature-rise alerts in Turbit's fleet resolve to non-bearing root causes (faulty thermocouples, lubrication faults, cooling-system issues). Catching them early matters anyway — left unattended, the same conditions can damage the bearing itself.
On a 50-turbine portfolio, generator-bearing temperature alerts are the most frequent generator-side detection. The quick-confirm workflow (verify with a service partner inside 1–2 weeks) consistently keeps these at zero downtime.
References
- Fault detection of a wind turbine generator bearing using interpretable machine learning — Frontiers in Energy Research (2023)
- Bearings faults and limits in wind turbine generators (2024)
- Wind Turbine Generator Reliability Analysis To Reduce Operations and Maintenance Costs — NREL (2023)
- A Review of Research on Wind Turbine Bearings' Failure Analysis and Fault Diagnosis — Liu et al. (2023)
See Turbit on your fleet
Backtest Turbit on the turbines where you already know something happened — or read how operators like VSB, Energiequelle, Enercity and Teut use Turbit on their fleets today.