Laying the foundations for context-aware and AI-ready fault diagnosis with the Operations Ontology

In wind farm operations, the full value of operational data is not realised due to obstacles in understanding and integration. With vast amounts of data in operational silos, reference ontologies provide a common framework for describing and connecting heterogeneous data, establishing a shared meaning that is machine-readable, human-understandable and a foundation for applying AI. As part of IEA Wind Task 43 and within the WeDoWind ecosystem, we have formed a public working group of experts across multiple disciplines to develop a foundational ontology of operations. Starting with foundational entities in the field of operations and maintenance, such as maintenance process and alarm system, we develop a section of the Operations Ontology (OpOn), focusing on a demonstration use case involving diagnostics and troubleshooting of rotor over-speed protection alarms. Here we illustrate how data annotation and integration become more intuitive and efficient with the use of the ontology, and we describe how this builds a solid foundation for the use of modern AI.