Scholarlink

Under the dual driving force of global fossil energy shortage and “Carbon Peaking and Carbon Neutrality Goals”, accurate prediction of solar power generation share is crucial to optimize China’s energy structure. Aiming at the modeling challenges of small-sample data, this study innovatively integrates the gray fractional order FGM(1,1) model with particle swarm optimization (PSO) algorithm to construct a dynamic rolling window prediction framework: based on the annual data from 2017–2024, the model’s accuracy and adaptability to short-term fluctuations are verified with a training window of four years. According to energy development forecasts, China’s solar energy proportion in total power generation is expected to demonstrate continuous growth, with projections indicating it will reach 5.95% by the target year of 2025. Accordingly, policy recommendations are proposed to provide decision support for the government to coordinate energy security and low-carbon transformation.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Rolling Window Integrating FGM(1,1) and PSO for Forecasting Solar Power Generation Share in China

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yuhan Li,
  • Zhiheng Huang,
  • Wei Dai,
  • Hongyan Tao,
  • Fengjuan Wang

摘要

Under the dual driving force of global fossil energy shortage and “Carbon Peaking and Carbon Neutrality Goals”, accurate prediction of solar power generation share is crucial to optimize China’s energy structure. Aiming at the modeling challenges of small-sample data, this study innovatively integrates the gray fractional order FGM(1,1) model with particle swarm optimization (PSO) algorithm to construct a dynamic rolling window prediction framework: based on the annual data from 2017–2024, the model’s accuracy and adaptability to short-term fluctuations are verified with a training window of four years. According to energy development forecasts, China’s solar energy proportion in total power generation is expected to demonstrate continuous growth, with projections indicating it will reach 5.95% by the target year of 2025. Accordingly, policy recommendations are proposed to provide decision support for the government to coordinate energy security and low-carbon transformation.