| 顾慧杰,周华锋,彭超逸,赵化时,胡亚平,聂涌泉,陈根军.基于CEEMDAN-BERT-LSTM的现货市场电力价格预测[J].电力需求侧管理,2025,27(5):112-117 |
| 基于CEEMDAN-BERT-LSTM的现货市场电力价格预测 |
| Prediction of spot market electricity prices based on CEEMDAN-BERT-LSTM |
| 投稿时间：2025-06-01 修订日期：2025-06-28 |
| DOI：10. 3969 / j. issn. 1009-1831. 2025. 05. 017 |
| 中文关键词: 现货市场电价预测 完全集合经验模态分解 双向编码器 长短期记忆网络 |
| 英文关键词: spot market electricity price forecasting CEEMDAN BERT LSTM |
| 基金项目:中国南方电网有限责任公司科技项目（000005KK52220024）；国家自然科学基金项目（U23B20129） |
|
| 摘要点击次数: 217 |
| 全文下载次数: 53 |
| 中文摘要: |
| 准确预测现货市场电价对电力市场参与者的利益保障具有重要作用。原料价格和气候因素均会影响现货市场的电价波动，再者，当前有大量风光能源参与到现货市场的交易当中，导致现货市场的电价预测更具挑战性。为此，提出了一种结合CEEMDAN、BERT和LSTM的现货市场电力价格预测模型。首先，采用CEEMDAN算法分解原始的电价数据；随后，采用BERT算法对原料价格、气候条件和可再生能源3类外生特征文本数据进行处理，提高模型的预测精度；其次，将电价分解子序列与外生特征处理结果相结合，利用LSTM对模型进行预测，并将预测结果相叠加获得最终的电价。最后，通过仿真验证了所提方法的有效性，结果显示，基于CEEMDAN-BERT-LSTM的预测模型显著提升了电价预测的精度。 |
| 英文摘要: |
| The accurate prediction of spot market electricity prices plays a crucial role in protecting the interests of participants in the electricity market. Both raw material prices and climate factors can affect the fluctuation of electricity prices in the spot market. In addition, a large amount of wind and solar energy is currently involved in spot market transactions, making electricity price forecasting in the spot market more challenging. Therefore, a spot market electricity price prediction model that integrates CEEMDAN, BERT, and LSTM is proposed. Firstly, the CEEMDAN algorithm is used to decompose the original electricity price data；Subsequently, the BERT algorithm is used to process the text data of three exogenous features：raw material prices, climate conditions, and renewable energy, in order to improve the prediction accuracy of the model；Next, the electricity price decomposition subsequence is combined with the results of exogenous feature processing, and LSTM is used to predict the model. The predicted results are then overlaid to obtain the final electricity price. Finally, the effectiveness of the proposed method was verified through simulation, and the results show that the CEEMDAN-BERT-LSTM prediction model improved the accuracy of electricity price prediction significantly. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |