Home » News » Researcher in Residence: Introducing Chat-Offshore Wind (OSW)

Researcher in Residence: Introducing Chat-Offshore Wind (OSW)

As the cornerstone of the UK government’s ambitious 2050 Net Zero target, offshore wind  stands as a pivotal sector poised to revolutionise the energy landscape. Amidst this transformation, stakeholders, including the Offshore Renewable Energy Catapult (OREC), have underscored the indispensable role of digitalisation, particularly the integration of Artificial Intelligence (AI), in facilitating swift and informed decision-making amidst time constraints and uncertainties. Together with the University of Hull’s Nina Dethlefs, this project looks to gain a systematic understanding of the key requirements and human factors that need to be considered to establish natural language as an effective medium of AI-expert communication.

In response to this clarion call, this proposal embarks on a pioneering exploration into the transformative potential of natural language to unravel the intricate connection between environmental conditions and faults in OSW assets. By harnessing the power of natural language, Nina seeks to illuminate the underlying relations, enhancing the explainability and confidence in critical decisions.

While the advent of natural language-enabled information access, exemplified by recent advancements like ChatGPT, has garnered widespread acclaim, its utility in safety-critical and commercially-sensitive settings remains limited. Recognising these constraints, this project charts a course to (1) systematically delineate the key requisites and human factors essential for establishing natural language as a robust medium for AI-expert communication; (2) craft a prototype AI-system dubbed ‘ChatOSW’ from a trove of domain-specific documents—ranging from technical manuals to maintenance reports—incorporating explicit knowledge representation to circumvent the pitfalls of general-purpose models, thereby laying the groundwork for a highly specialised AI tailored for information-access and decision-support in OSW; and (3) forge a comprehensive techno-economic framework to evaluate the efficacy and cost-effectiveness of the ChatOSW tool, with a keen focus on its applicability to future OSW and renewable energy ventures across the UK.

Nina, speaking about the project, said: “This project is an exciting opportunity to bring cutting-edge AI technology, such as recent advances in conversational AI and large language models, to the offshore wind sector. There is a vast amount of hidden knowledge in text, for example, on past situations, operations, maintenance, repairs, orders and parts, etc. that can be immensely helpful to decision making and enhance productivity. Unfortunately, this knowledge is mostly buried in documents and therefore not easily  accessible. This project aims to offer solution in the form of an intelligent framework that allows intuitive and rapid information access via natural language, while at the same time integrating formal knowledge representation to make outputs reliable, transparent and accountable.”