Scholarlink

<p>The growing integration of intermittent renewable energy sources (RES), especially wind energy, presents substantial hurdles for the reliable and economical execution of microgrids. This paper presents a robust hybrid system that integrates wavelet transform-oriented signal decomposition with ML to improve wind power forecasting and optimize energy dispatch, addressing the uncertainty of wind generation and its effects on energy scheduling. The proposed model uses an innovative “cardiac wavelet” function for improved feature extraction and denoising, in combination with ANNs developed by chaotic shark smell optimization, to provide precise short-term wind predictions. This forecasting engine has been integrated into a grid-connected microgrid system that includes 18 wind turbines, 20 rooftop solar systems, 9 electric vehicles, and 26 prosumers, indicating a realistic dynamic energy exchange model. Adaptive demand response solutions that use time-of-use and RTP are combined to improve user participation and grid adaptability. Simulation outcomes showed significant improvements, with wind forecasting attaining an RMSE of 0.033 and an R² of 0.97. Grid dependency decreased from 64% to 49%, peak demand costs decreased by 30.5% and overall energy savings reached £427. The integrated system performed consistently and was scalable under practical implications, offering a cost-efficient and notable solution for modern energy management in smart microgrids.</p>

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

Wind Power Estimation Under Uncertainty Using Wavelet Transform for Optimal Dispatch in Microgrids with Renewable Energy Integration and Demand Response

  • 期刊论文

  • 发表日期：

    2026年2月7日

查看全文
  • Yan Liu,
  • Jinfeng Liu,
  • Yushang Qi,
  • Jie Deng,
  • Panpan Wu,
  • Kun Zhang,
  • Xiong Chen

摘要

The growing integration of intermittent renewable energy sources (RES), especially wind energy, presents substantial hurdles for the reliable and economical execution of microgrids. This paper presents a robust hybrid system that integrates wavelet transform-oriented signal decomposition with ML to improve wind power forecasting and optimize energy dispatch, addressing the uncertainty of wind generation and its effects on energy scheduling. The proposed model uses an innovative “cardiac wavelet” function for improved feature extraction and denoising, in combination with ANNs developed by chaotic shark smell optimization, to provide precise short-term wind predictions. This forecasting engine has been integrated into a grid-connected microgrid system that includes 18 wind turbines, 20 rooftop solar systems, 9 electric vehicles, and 26 prosumers, indicating a realistic dynamic energy exchange model. Adaptive demand response solutions that use time-of-use and RTP are combined to improve user participation and grid adaptability. Simulation outcomes showed significant improvements, with wind forecasting attaining an RMSE of 0.033 and an R² of 0.97. Grid dependency decreased from 64% to 49%, peak demand costs decreased by 30.5% and overall energy savings reached £427. The integrated system performed consistently and was scalable under practical implications, offering a cost-efficient and notable solution for modern energy management in smart microgrids.