Intelligent Wind Turbine Diagnostics Using SCADA Data, Machine Learning and Digital Twins

I am developing a project concept focused on intelligent diagnostics and condition monitoring of wind turbines using operational SCADA data, machine learning and anomaly detection methods.

The project aims to move beyond simple anomaly detection towards more interpretable and practically useful diagnostic tools. The planned research will investigate unsupervised and semi-supervised methods for identifying abnormal turbine behaviour, representation learning and latent-space analysis, and the relationship between detected anomalies, operating conditions, component degradation and potential failure modes.

A further objective is to combine data-driven methods with physical knowledge, engineering models and digital-twin concepts. Such an approach could support more reliable condition assessment, fault diagnosis, remaining useful life estimation and predictive maintenance.

Our team at Gdańsk University of Technology brings expertise in SCADA data analysis, anomaly detection, machine learning, multivariate time-series analysis and scientific computing. We are particularly interested in methods such as DeepSVDD, PyOD-based algorithms, representation learning and interpretable machine learning.

For the POLNORIS call, we are looking for partners from Norway, Iceland or Liechtenstein with complementary expertise in one or more of the following areas: wind turbine operation and maintenance, condition monitoring, fault diagnostics, drivetrain or structural monitoring, digital twins, physics-based modelling, predictive maintenance, machine learning, or access to operational and maintenance data from wind farms.

We are especially interested in collaboration with partners who can contribute domain expertise, physical or engineering models, industrial validation, or access to real operational datasets and maintenance records.

The final project concept will be developed jointly with the selected partners, with the goal of creating an interdisciplinary framework for reliable, interpretable and data-driven wind turbine health assessment and maintenance support.