人工智能在新型配电系统规划中的应用综述OA

[Objective]Driven by the carbon peaking and carbon neutrality goals and energy transition,the large-scale grid connection of multiple entities such as high-share distributed generation,flexible loads,and new energy storage has promoted new distribution systems to exhibit characteristics including deep generation-network-load-storage coupling and dynamic evolution of morphological structures.Traditional experience-dependent planning models are no longer capable of addressing complex challenges such as the surge in system uncertainty and the coupling of multi-objective constraints.There is an urgent need to establish a new planning paradigm adapted to the development needs of the new distribution system by relying on artificial intelligence(AI)technologies.[Methods]This paper systematically sorts out the core processes and technical pain points of new distribution system planning,analyzes the key challenges in links such as planning knowledge graph construction,generation-load forecasting and scenario generation,power balance,capacity planning and network optimization,and generation-network-load-storage coordinated planning,and comprehensively elaborates on relevant research progress and technical application paradigms.This paper summarizes and analyzes the existing problems of AI-based new distribution system planning technologies,such as low efficiency in multi-source heterogeneous data fusion,insufficient cross-scenario generalization capability of models,lack of interpretability in planning decisions,difficulty in dynamic adaptation of multi-objective optimization weights,and inadequate consideration of the engineering feasibility of schemes.Corresponding key technical evolution directions are highlighted,including enhancing model adaptability through the integration of graph learning and physical constraints,strengthening cross-scenario and small-sample application capabilities via transfer learning and data augmentation,deepening data mining based on multi-modal fusion technology,improving decision interpretability by combining causal reasoning,and building a human-machine hybrid intelligence system to enhance planning practicality.[Results]Compared with traditional planning methods,AI technologies,relying on their powerful data processing and intelligent decision-making capabilities,provide efficient and accurate solutions for new distribution system planning,and have significant advantages in multi-scenario adaptation and multi-objective coordination.However,they still face a series of bottlenecks in technical implementation and engineering application.[Conclusions]In the future,it is necessary to continuously promote the deep integration of AI and new distribution system planning,tackle core technical challenges,and provide technical support and decision-making references for the high-quality development of distribution networks.