| 基于小波包分解和长短期记忆网络的短期电价预测 |
| |
| 引用本文： | 刘 达,,雷自强,,孙 堃.基于小波包分解和长短期记忆网络的短期电价预测[J].陕西电力,2020,0(4):77-83. |
| |
| 作者姓名： | 刘 达 雷自强 孙 堃 |
| |
| 作者单位： | （1.华北电力大学新能源电力与低碳发展研究中心，北京 102206； 2.华北电力大学智慧能源研究所，北京 102206； 3.华北电力大学 经济与管理学院，北京 102206） |
| |
| 摘 要： | 在电力市场环境下，精准的短期电价预测可以保障电网优化调度和安全稳定运行，但实时电价具有非平稳性和非线性的特点，加大了预测难度。针对这一问题，提出了一种基于小波包分解（WPD）和长短期记忆（LSTM）网络的短期实时电价预测方法。将实时电价序列分解，得到最高频细节部分和低频趋势部分，剔除波动性高、无效信息多的高频细节部分，再采用LSTM网络对有效信息最多、更能体现电价序列的趋势部分进行实时电价预测。使用所提方法对美国PJM市场某地区实时电价数据进行预测实验，结果表明所提方法相比随机森林、BP神经网络、支持向量机电价预测方和传统的LSTM网络电价预测方法具有更高预测精度。
|
| 关 键 词： | 小波包分解 LSTM网络 短期电价预测 电力市场 |
Short-term Electricity Price Forecasting Based on Wavelet Packet Decomposition & Long-Term and Short-Term Memory Networks |
| |
| LIU Da,' target="_blank" rel="external">,LEI Ziqiang,' target="_blank" rel="external">,SUN Kun.Short-term Electricity Price Forecasting Based on Wavelet Packet Decomposition & Long-Term and Short-Term Memory Networks[J].Shanxi Electric Power,2020,0(4):77-83. |
| |
| Authors: | LIU Da " target="_blank">' target="_blank" rel="external"> LEI Ziqiang " target="_blank">' target="_blank" rel="external"> SUN Kun |
| |
| Affiliation: | （1. Key Laboratory of New Energy and Low-Carbon Development， North China Electric Power University，Beijing 102206， China 2. Economics and Management School， North China Electric Power University，Beijing 102206，China；3. Institute of Smart Energy，Beijing 10 |
| |
| Abstract: | Accurate short-term electricity price forecasting can ensure optimal grid dispatching & safe and stable operation in the power market environment，but real-time electricity prices are characterized by non-stationary and non-linear characteristics， which increases the difficulty of forecasting. A short-term real-time electricity price forecasting method is proposed based on wavelet packet decomposition （WPD） and long-term short-term memory （LSTM） networks. The real-time electricity price sequence is decomposed to obtain the highest-frequency detail part and the low-frequency trend part， and the high-frequency detail part with high volatility and invalid information is removed. Then the LSTM network is used to forecast the real-time electricity price for the trend part which has the most effective information and can better reflect the electricity price sequence. The proposed method is used to predict the real-time electricity price data of a certain area in the US PJM market. The experimental results show that the proposed method has higher prediction accuracy compared with the random forest price forecasting model，BP neural network price forecasting method，support vector electromechanical price forecasting method and traditional LSTM network. |
| |
| Keywords: | wavelet packet decomposition LSTM network short-term electricity price forecast electricity market |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《陕西电力》浏览原始摘要信息 |
|
点击此处可从《陕西电力》下载全文 |
