Power Grid Model Generation Based on the Tool-Augmented Large Language Model

Power grid models are essential for power system research and AI-based applications. However, due to security reasons, publicly available power grid models are very limited, making the generation of models with specific topologies and power flow properties crucial. As these models are typically stored in organized text, their generation can be framed as structured text generation, which can be effectively handled by Large Language Models (LLMs). Inspired by their remarkable performance, this letter proposes a novel method for generating power grid models based on the tool-augmented LLM, providing users with an intuitive way to generate power grid models according to their needs. Numerical results show that the proposed method can effectively generate power grid models that meet specified requirements expressed in natural language.