Open Access

Cansu Kahveci, Kajan Kandiah, Max-Ferdinand Stroh, Wolfgang Boos

• The advancing climate change and the resulting need to restructure the energy supply industry mark the beginning of an era of significant upheavals, profoundly impacting energy distribution. The ongoing integration of renewable energy sources introduces new strains on power grids, as they are now confronted with bidirectional energy flows as well as significant fluctuations in energy generation and demand. These developments pose unprecedented challenges to the physical integrity and operational efficiency of distribution networks. Therefore, electrical distribution grid operators (DGOs) must balance these new challenges while ensuring supply security and sustainability, as well as considering economic efficiency. To meet these new challenges, the development of innovative maintenance strategies is essential. In this context, the digitalisation of power grids emerges as a key technology, offering solutions through the introduction of holistic assistance and prognosis systems (APS). This paper focuses on the collection and analysis of requirements leading to the development of an assistance system for maintenance workers in distribution networks based on Large Language Models (LLM). The LLM aims to support workforce management by providing maintenance workers with targeted information and real-time recommendations. The methodology involves conducting expert interviews to gather essential requirements, followed by a thorough analysis of the collected data. Based on this, the foundation for developing the LLM-based assistance system is established. The results demonstrate how LLMs can enhance the efficiency and effectiveness of maintenance management by optimising workflows and supporting personnel in decision-making processes.