Scholarlink

In the context of today’s energy resource shortage, the Virtual Power Plant (VPP) represents an innovative operational model that significantly enhances the flexibility and efficiency of the power system. By integrating optimal dispatch, demand-side management, demand response, and accurate load forecasting technologies, VPP improves both the flexibility and economic performance of power systems. To boost load forecasting accuracy, the VPP leverages feature parameters derived from data screening, combined with a Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) model optimized with Attention mechanism through Crown Porcupine Optimization (CPO). This approach allows for precise power load forecasting, providing reliable support for scheduling decisions. Experimental results demonstrate that the proposed method outperforms both the standard CNN-LSTM-Attention hybrid model and the CNN-LSTM-Attention model optimized by Particle Swarm Optimization (PSO) in terms of prediction accuracy.

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

Electricity Load Forecasting of Virtual Power Plant Based on CPO-CNN-LSTM-Attention Neural Network Model

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Yan Huang,
  • Yongzheng Cong,
  • Haibo Du,
  • Min Zhu,
  • Mengjue Xu,
  • Shaohua Wang

摘要

In the context of today’s energy resource shortage, the Virtual Power Plant (VPP) represents an innovative operational model that significantly enhances the flexibility and efficiency of the power system. By integrating optimal dispatch, demand-side management, demand response, and accurate load forecasting technologies, VPP improves both the flexibility and economic performance of power systems. To boost load forecasting accuracy, the VPP leverages feature parameters derived from data screening, combined with a Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) model optimized with Attention mechanism through Crown Porcupine Optimization (CPO). This approach allows for precise power load forecasting, providing reliable support for scheduling decisions. Experimental results demonstrate that the proposed method outperforms both the standard CNN-LSTM-Attention hybrid model and the CNN-LSTM-Attention model optimized by Particle Swarm Optimization (PSO) in terms of prediction accuracy.