Scholarlink

This study proposes a novel approach for short-term solar power prediction using CEEMDAN, Sample Entropy (SE), and Long Short-Term Memory (LSTM) model. CEEMDAN decomposes solar energy time series into intrinsic mode functions (IMF) with adaptive noise, enhancing adaptability. SE quantifies complexity, combining IMFs with similar entropy values. LSTM forecasts each combined IMF, and results are aggregated for overall prediction. The CEEMDAN-SE-LSTM model is evaluated using RMSE, MAE, and R2, demonstrating superior performance over LSTM-only models. Testing on Indian solar power plant data showcases efficacy of the model, with RMSE values of 16.8787 and 17.7847 in univariate and multivariate tests respectively, surpassing LSTM-only RMSE of 26.3272 and 29.1333. This highlights the effectiveness of the proposed approach in forecasting short-term solar energy production and its potential for seamless integration into the main electricity grid.

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

Short Term Solar Energy Prediction Based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and Sample Entropy

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Muhammad Danish Asyraf Syamsir Alam,
  • Norashikin Yahya,
  • Suhaila Badarol Hisham,
  • Saiful Amin Muhammad Mokri

摘要

This study proposes a novel approach for short-term solar power prediction using CEEMDAN, Sample Entropy (SE), and Long Short-Term Memory (LSTM) model. CEEMDAN decomposes solar energy time series into intrinsic mode functions (IMF) with adaptive noise, enhancing adaptability. SE quantifies complexity, combining IMFs with similar entropy values. LSTM forecasts each combined IMF, and results are aggregated for overall prediction. The CEEMDAN-SE-LSTM model is evaluated using RMSE, MAE, and R2, demonstrating superior performance over LSTM-only models. Testing on Indian solar power plant data showcases efficacy of the model, with RMSE values of 16.8787 and 17.7847 in univariate and multivariate tests respectively, surpassing LSTM-only RMSE of 26.3272 and 29.1333. This highlights the effectiveness of the proposed approach in forecasting short-term solar energy production and its potential for seamless integration into the main electricity grid.