This work is from my master thesis: Condition Monitoring with Machine Learning: A Data-Driven Framework for Quantifying Wind Turbine Energy Loss.
Paper available at: https://arxiv.org/abs/2506.13012
Wind energy significantly contributes to the global shift towards renewable energy, yet operational challenges, such as Leading-Edge Erosion on wind turbine blades, notably reduce energy output. This study introduces an advanced, scalable machine learning framework for condition monitoring of wind turbines, specifically targeting improved detection of anomalies using Supervisory Control and Data Acquisition data. The framework effectively isolates normal turbine behavior through rigorous preprocessing, incorporating domain-specific rules and anomaly detection filters, including Gaussian Mixture Models and a predictive power score. The data cleaning and feature selection process enables identification of deviations indicative of performance degradation, facilitating estimates of annual energy production losses. The data preprocessing methods resulted in significant data reduction, retaining on average 31% of the original SCADA data per wind farm. Notably, 24 out of 35 turbines exhibited clear performance declines. At the same time, seven improved, and four showed no significant changes when employing the power curve feature set, which consisted of wind speed and ambient temperature. Models such as Ran- dom Forest, XGBoost, and KNN consistently captured subtle but persistent declines in turbine performance. The developed framework provides a novel approach to existing condition monitoring methodologies by isolating normal operational data and estimating annual energy loss, which can be a key part in reducing maintenance expenditures and mitigating economic impacts from turbine downtime.
The end goal of this research is to quantify the potential energy loss resulting from performance degradation over time. To asses such a state, it is necessary to establish a healthy model state, representative of the turbines expected output. However, identifying such a state is a non-trivial task, even for a single turbine, if event logs are not provided as part of the data, and even sometimes with provided logs. An extensive framework to identify said periods of healthy model states and clean the data of anomalies is, therefore, a necessary step for further research.
An example of anomalous behavior can be observed in the following figure of the power curve:
The extensive framework implemented to accommodate the data cleaning and identifying healthy states can be observed:
The Predictive Power Score (PPS) is implemented to work in a temporal setting and utilized as a direct measure of the predictive quality of the data on the power before and after NB-filters. The PPS gives a direct insight into the effectiveness of the filtering applied to the data and allows for the selection of healthy model states. Such an example can be observed on:
where the PPS is not only improved across the operating history of the turbine, but further history can potentially be utilized in the following quantification of energy loss assessment.
Finally, when the desired turbines and periods have been identified, it is possible to measure the deviation from the expected production:
| Δ | Model | MAPE r | 2020 | 2021 | 2022 | 2023 |
|---|---|---|---|---|---|---|
| All | RF | 4.9 | 7.2 | 7.1 | 7.8 | 7.7 |
| XGBoost | 4.8 | 5.9 | 6.1 | 6.2 | 4.9 | |
| KNN | 5.1 | 2.8 | 2.8 | 3.0 | 1.4 | |
| MLP | 7.0 | 2.9 | 2.9 | 3.0 | 1.9 | |
| PC | RF | 6.6 | -0.3 | -0.7 | -0.5 | -1.4 |
| XGBoost | 6.7 | -0.1 | -0.7 | -0.4 | -1.5 | |
| KNN | 7.5 | -0.1 | -0.8 | -0.3 | -1.4 | |
| MLP | 8.4 | 0.2 | -0.9 | 0.1 | -0.5 |
Overall, the entire framework is a success and makes the selection and quantification of energy production loss trivial.