Research on Large Language Model-Based Management and Operation of Distributed Integrated Energy Systems

With the growing significance of Distributed Energy Systems (DES) in the energy sector, efficiently integrating crossdisciplinary knowledge has become a critical research challenge. This paper proposes a large language model-based management and operation framework for the integrated distributed energy systems, aimed at enhancing research efficiency and application effectiveness through vertical domain-specific Large Language Models (LLMs). First, we constructed a domain-specific corpus and knowledge graph for DES; then, we used a multi-agent framework to generate the instruction dataset. Subsequently, we fine-tuned the pre-trained model using the Low-Rank Adaptation parameter-efficient technique, significantly reducing computational and Graphics Processing Unit memory demands. Experimental results demonstrate about a 6% improvement in domain-specific accuracy compared with the base model after finetuning, validating the efficacy of our approach. This work provides a novel solution for intelligent DES management while offering valuable insights for LLMs development in other vertical domains.