Leveraging generative AI to facilitate Knowledge Management across Substation Lifecycle

The electric power industry is confronting a critical situation due to the mass retirement of experienced engineers, which threatens the loss of valuable tacit knowledge accumulated over years. Traditional knowledge transfer methods, such as on-the-job training (OJT) reliant on limited human mentorship or manual-based approaches, are unsustainable because they fail to bridge the critical gap between codified knowledge and practical application. This study fundamentally addresses this structural issue by leveraging the transformative capabilities of generative AI. The objective is to integrate generative AI with the SECI model, a cornerstone theory in knowledge management, to develop an efficient and systematic knowledge transfer mechanism and demonstrate its effectiveness. The methodology strategically embeds AI into the SECI cycle: AI facilitates Socialisation by capturing and transcribing OJT dialogues; it executes the most critical phase, Externalisation, by employing specialized prompt engineering strategies to analyse the transcripts, systematically reconstructing expert thought processes and transforming practical insights into structured explicit knowledge, including underlying decision-making rationales; this knowledge is then used in Combination by integrating it with internal documents and finally, novice engineers achieve Internalisation by interacting with the developed "AI Virtual Mentor". The performance of this tool was validated through a case study on substation busbar design. The quantitative results demonstrated dramatic effectiveness.