In the context of optimized Operation & Maintenance of wind energy infrastructure, it is important to develop robust decision support tools. A major challenge lies in the multiple uncertainties in the aggregated data, the available knowledge with respect to the wind turbine structures, and sub-components, as well as the constantly varying operational and environmental loads. We propose a decision tree learning algorithm, running on telemetry data, to detect faults, damage patterns, and abnormal operation. The telemetry includes condition monitoring (CM) information and data from specialized structural health monitoring systems (SHM). This is a Big Data problem with continuous sampling at high rates from thousands of wind turbines in the field. We train an ensemble Bagged decision tree classifier on a large condition monitoring dataset from an offshore wind farm of 48 wind turbines, and use it to automatically link identified faults to their possible root causes (Abdallah et al. 2018). We further propose an architecture to implement decision tree learning in the context of cloud computing, namely involving a cloud based Apache Hadoop software for large data storage and handling, a cloud based Apache Spark for efficiently running machine-learning algorithms, and the object-oriented based decision tree concept.
• Abdallah, I. Dertimanis, V. Mylonas, H. Tatsis, K., Chatzi, E., Dervilis N., Worden, K. Eoghan, M. (2018), “Fault Diagnosis of wind turbine structures using decision tree learning algorithms with big data”, to appear in the European Safety and Reliability Conference ESREL, 17-21 June 2018, NTNU, Norway
Contact: Prof. Dr. Eleni Chatzi