Intelligent Assessment Method for Regional Load and Distributed Energy Capacity Considering Spatiotemporal Coupling
Year: 2026
Type: conference-paper
Abstract: To address the planning challenges posed by high penetration of distributed energy in modern power systems and achieve coordinated and accurate evaluation of regional load and distributed photovoltaic (PV) capacity, this study proposes an intelligent assessment method that considers spatiotemporal coupling. The method first constructs a multi-scale spatiotemporal graph neural network to automatically extract complex spatiotemporal interaction features between load and PV generation. It then designs a conditional generative adversarial network (GAN) framework incorporating power balance and line capacity constraints, generating multi-scenario sequences that both conform to data distributions and satisfy operational safety requirements. A meta-learning strategy is adopted to enable the model to quickly adapt using only a small amount of data from new regions. Experiments based on actual provincial grid data show that the proposed model achieves a minimum root mean square error (RMSE) of 2.87 for load prediction, reducing errors by 19.5% compared with the second-best baseline model. The recommended configuration scheme can shorten the capital payback period in typical regions from 6.3 years to 5.1 years and increase renewable energy utilization by 12.7 percentage points. These results validate the effectiveness of the proposed method in improving assessment accuracy, ensuring feasibility, and enhancing economic performance, providing new insights for optimized planning of distributed energy.
Authors: Ning Luo, Mao Miao, Yongkui Guo, Chen Chen, Zhen Li
Institution: Guizhou Power Grid Co., Ltd. (China)
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7204465614