Wind energy capacity is growing rapidly and in addition, there are several wind turbines
approaching their end of life. Once wind turbines reach this stage, there are typically
three options available: decommissioning, re-powering and lifetime extension. This
research investigates lifetime extension, specifically of drivetrains, which has not been
extensively studied, as opposed to the supporting structures. Therefore, the aim of this
work was to develop a methodology for determining lifetime extension of wind turbine
drivetrains. Initially, existing research and industrial guidelines were reviewed to enhance understanding for application in the wind industry. Based upon these findings
and a systematic approach, a methodology for wind turbine drivetrain life extension
was developed. This proposed method required identifying the most vulnerable components, to which the methodology could be applied. It was realized that vulnerability
maps of mechanical drivetrain components already exist, so this work used a datadriven approach to produce a vulnerability map for a power converter. Next, the proposed methodology was tested using data from an onshore wind farm, focusing on the
rear generator bearings, which were identified as problematic. SCADA data, particularly temperature readings, were utilised due to their wide availability. Seven years of SCADA data were analysed using machine learning models to predict remaining useful life (RUL) and failure metrics. Two approaches were implemented: deterministic
and probabilistic, using Monte Carlo simulations to improve accuracy. The probabilistic approach employed two methods: discrete confidence interval-based and continuous probability distribution-based, to characterize model errors. Results indicated that while SCADA data can inform predictive maintenance by identifying failure thresholds, its reliability for long-term RUL prediction is limited. This research concludes that combining SCADA data with more detailed condition monitoring methods, such as vibration analysis, is essential for robust drivetrain lifetime extension assessment. Future work will focus on validating these findings through integration and analysis.
| Date of Award | 19 May 2025 |
|---|
| Original language | English |
|---|
| Awarding Institution | - University Of Strathclyde
|
|---|
| Sponsors | EPSRC (Engineering and Physical Sciences Research Council) & University of Strathclyde |
|---|
| Supervisor | Abbas Kazemi Amiri (Supervisor) & James Carroll (Supervisor) |
|---|