Internet of Things (IoT) and Industry 4.0 are recent concepts that allow for the incorporation of a massive number of sensors into industrial machinery and processes, and thus, enable the machine owners to constantly supervise and analyse the machines.

The Wind Energy industry is not an exception, and new wind turbines have many sensors installed that provide information on most of their components.

This information is typically obtained and processed by the SCADA loggers installed in the turbines but can also be transferred (streamed) in real-time to other locations, gaining deep insight into the operational status of the machines.

Real time connections to wind farms and wind turbines are challenging, as the high volume of data has to be transferred reliably and brought into a uniform format for further use in the analyses.

This colossal volume of data will, however, allow the development of data-driven Digital Twin models to mimic, analyse and forecast the wind turbine operation. The models will be based on Deep Learning (DL) algorithms and will be fed and updated in real-time by incoming data streams, allowing accurate knowledge of the health status of the turbine.

The fellow will be trained on establishing streaming pipelines, the use of streamed data, and DL techniques to generate the digital twin models, which will allow to dynamically monitor the health status of the turbines and their components.

As the industrial result, the digital twin will be developed to be part of the virtual wind farm hub and validated against benchmarking tools.

Two academic secondments. UNIZAR for training and collaboration on data science and completing mandatory PhD courses (Prof. Julio J. Melero, M13-18). TU-DELFT for joint work on data-driven Digital Twins (Prof. Simon J. Watson, M28-30).