Review of AI-driven augmented planning methods for distribution systems

Driven by the “dual-carbon” transition, new-type distribution systems are evolving with emergent characteristics, including transmission-distribution-microgrid coordination, primary-secondary system integration, and the interactive connectivity of diverse flexible resources. However, traditional planning methodologies relying on “physical modeling plus manual coordination” face fundamental limitations in complex modeling, high-dimensional uncertainty analysis, and multi-modal data fusion. Specifically, mechanistic models fail to fully capture the dynamic interactions within the “generation-grid-load-storage” nexus; heuristic experience offers limited decision support for high-dimensional spatiotemporal correlations; deterministic approaches are ill-suited to the stochastic nature of renewable energy. This paper presents a systematic review of the deep integration of artificial intelligence (AI) and distribution system planning. First, it deconstructs the defining features of the new system, such as integrated transmission-distribution-microgrid operation, primary-secondary coordination, and distributed resource interaction. Second, it analyzes how AI methodologies break through traditional paradigms—specifically using deep neural networks to surrogate complex physical models, reinforcement learning to optimize dynamic game-theoretic decisions, and knowledge graphs to fuse and interpret multi-source heterogeneous data. Subsequently, a “human-in-the-loop” hybrid augmented intelligence framework is proposed. Characterized by “embedded physical constraints, data-driven feature processing, and iterative intelligent decision-making”, this framework aims to synergize AI’s data-driven capabilities with the physical interpretability of mechanistic models—transcending mere substitution—to achieve a systemic enhancement in planning efficiency. Finally, research prospects for AI-augmented applications across the entire planning lifecycle are outlined. By systematically aligning AI capabilities with the demands of distribution system planning, this study establishes a theoretical foundation for adaptive planning in high-uncertainty environments and provides methodological support for overcoming cross-hierarchical coordination challenges.