| 引用本文: | 肖白,高文瑞,李道明,綦雪松,阚中锋.基于3σ-CEEMDAN-LSTM的空间负荷预测方法[J].电力自动化设备,2023,43(3): |
| XIAO Bai,GAO Wenrui,LI Daoming,QI Xuesong,KAN Zhongfeng.Spatial load forecasting method based on 3σ-CEEMDAN-LSTM[J].Electric Power Automation Equipment,2023,43(3): |
|
| 摘要: |
| 为有效利用从配电网采集的海量数据以及改善空间负荷预测效果，提出一种基于3σ准则、自适应噪声完备集合经验模态分解(CEEMDAN)和长短期记忆神经网络(LSTM)的空间负荷预测方法。基于3σ准则对每个Ⅰ类元胞的实测负荷数据进行奇异值检测和处理；运用CEEMDAN技术将处理后的Ⅰ类元胞负荷数据分解为若干个频率和幅值均不同的本征模态函数(IMF)；分别对每个IMF分量构建LSTM模型进行预测；将所有IMF分量预测结果进行线性叠加，得到目标年基于Ⅰ类元胞的空间负荷预测结果，在此基础上使用空间电力负荷网格化技术求得基于Ⅱ类元胞的空间负荷预测结果。算例分析结果验证了所提方法的正确性和有效性。 |
| 关键词: 空间负荷预测 元胞 地理信息系统 自适应噪声完备集合经验模态分解 长短期记忆神经网络 |
| DOI：10.16081/j.epae.202208016 |
| 分类号:TM73 |
| 基金项目:国家自然科学基金资助项目(51177009)；国家重点研发计划项目(2017YFB0902205) |
|
| Spatial load forecasting method based on 3σ-CEEMDAN-LSTM |
|
XIAO Bai1, GAO Wenrui1, LI Daoming1, QI Xuesong2, KAN Zhongfeng2
|
|
1.School of Electrical Engineering, Northeast Electric Power University, Jilin 132012, China;2.Jilin Power Supply Company of State Grid Jilin Electric Power Company Co.,Ltd.,Jilin 132001, China
|
| Abstract: |
| In order to effectively use the massive data collected from distribution network and improve the forecasting effect of spatial load, a spatial load forecasting method based on 3σ rule, complementary ensemble empirical mode decomposition with adaptive noise(CEEMDAN) and long-short term memory neural network(LSTM) is proposed. Based on 3σ rule, the singular value detection and disposition for measured load data of each Class Ⅰ cell are carried out. CEEMDAN technology is applied to decompose the processed load data of Class Ⅰ cells into several intrinsic mode functions(IMFs) with different frequencies and amplitudes. LSTM model is built for each IMF component for forecasting. The spatial load forecasting based on Class Ⅰ cells at the target year is obtained by linearly superimposed of the forecasting results of all the IMF components, on this basis, the spatial power load grid technology is used to obtain the spatial load forecasting results based on Class Ⅱ cells. The correctness and effectiveness of the proposed method are verified by case analysis results. |
| Key words: spatial load forecasting cell geographic information system CEEMDAN LSTM |
