Vision of Future Smart Control Center: Fusion of Large Language Models and Power System Intelligent Computing

The high stochasticity of renewable dominated power grids demands advanced dispatch strategies. However, the disparity between flexible operator intent and rigid legacy tools hinders resilient system operation. To address this challenge,this paper proposes a future smart control center architecture that fuses Large Language Models (LLMs) with Power System Intelligent Computing (PSIC). In this architecture, the LLM functions as the cognitive core for intent understanding, workflow orchestration, and causal reasoning. Concurrently, PSIC executes physical simulations and data-driven optimization tasks decomposed by the LLM. This integration creates a cognitive decision-making process where the adaptive decision-making capability enabled by LLM is rigorously bounded and verified by physical laws. Consequently, the architecture enables next generation human-machine collaborative dispatch, advancing grid operations from automation toward true intelligence and self-evolution.