| 96 | 0 | 46 |
| 下载次数 | 被引频次 | 阅读次数 |
针对目前光伏功率短期预测方法精度低、运算慢等特点，提出一种基于特征重组与改进Informer-ARIMA的光伏功率短期组合预测方法。先利用完全自适应噪声集合经验模态分解(CEEMDAN)将光伏序列分解为若干个分量，经过特征重组得到特征分量和趋势分量，再分别输入改进Informer和差分自回归移动平均模型(ARIMA)中，叠加2个模型的输出值以得到预测结果。此外，提出双尺度概率稀疏化因果空洞注意力机制，在不显著降低训练速度的情况下增强数据特征挖掘能力，并且缓解异常数据造成特征偏差的问题，进一步提高了模型的预测精度。实验结果表明：与6种主流的预测模型相比，本文所提模型在3～5 d预测任务中预测效果较优，速度较快。
Abstract：Aiming at the characteristics of low accuracy and slow operation of current short-term prediction methods of photovoltaic power, this paper proposed a combined prediction method based on feature reorganization and improved Informer-ARIMA. Firstly, the complementary ensemble empirical mode decomposition with adaptive noise(CEEMDAN) was used to decompose the photovoltaic sequence into several components, and the feature components and trend components were obtained by feature reorganization. Then, they were input into the improved Informer and ARIMA models respectively, and the output values of the two models were superimposed to obtain the prediction results. In addition, a dual-scale probabilistic sparse causal hole attention mechanism was proposed to enhance the ability of data feature mining without significantly reducing the training speed, and to alleviate the problem of feature deviation caused by abnormal data, which further improved the prediction accuracy of the model. The experimental results show that compared with the six mainstream prediction models, the model has better prediction effect and faster speed in the 3～5 d prediction task.
[1] 国务院办公厅.国务院办公厅转发国家发展改革委国家能源局关于促进新时代新能源高质量发展实施方案的通知[J].中华人民共和国国务院公报，2022(17):24-27.
[2] 李芬，孙凌，王亚维，等.基于CEEMDAN-GSA-LSTM和SVR的光伏功率短期区间预测[J].上海交通大学学报，2024,58(6):806-818.
[3] 沈俊豪.基于多种模型融合的光伏发电功率预测[D].无锡：江南大学，2021.
[4] WANG K J,QI X X,LIU H D.A comparison of day-ahead photovoltaic power forecasting models based on deep learning neural network[J].Applied Energy,2019,251:113315.
[5] KALPAKIS K,GADA D,PUTTAGUNTA V.Distance measures for effective clustering of ARIMA time-series[C]//Proceedings 2001 IEEE International Conference on Data Mining.Piscataway:IEEE,2001:273-280.
[6] 周满国，黄艳国，段锦锋.基于GRU-RF模型的太阳辐照度短时预测[J].太阳能学报，2022,43(7):166-173.
[7] VASWANI A,SHAZEER N,PARMAR N,et al.Attention is all you need[C]//Proceedings of the 31st Annual Conference on Neural Information Processing Systems (NIPS).Cambridge:MIT,2017:6000-6010.
[8] ZHOU H Y,ZHANG S,PENG J,et al.Informer:beyond efficient transformer for long sequence time-series forecasting[C]// Proceedings of the AAAI Conference on Artificial Intelligence.Menlo Park:AAAI,2021,35(12):11106-11115.
[9] 徐柔，章杰，赖松林，等.结合改进的DCGAN和Attention-LSTM的光伏功率预测[J].福州大学学报(自然科学版),2023,51(4):498-504.
[10] JIANG Y Q,DAI Y,SI R,et al.Short-term state electricity load forecasting based on transfer-informer[C]//2022 IEEE 2nd International Conference on Digital Twins and Parallel Intelligence (DTPI).Piscataway:IEEE,2022:1-6.
[11] ZHOU T,MA Z Q,WEN Q S,et al.Fedformer:frequency enhanced decomposed transformer for long-term series forecasting[C]//International Conference on Machine Learning.Cambridge:PMLR,2022:27268-27286.
[12] CHEN M H,PENG H W,FU J L,et al.Autoformer:searching transformers for visual recognition[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.Piscataway:IEEE,2021:12270-12280.
[13] 廖文宇，郭鹏，赵冰，等.基于小波分解与ARMA-Bayes LSTM组合模型的风向短期预测[J].电力科学与工程，2023,39(5):17-23.
[14] 周满国.基于深度学习的光伏发电功率预测研究[D].赣州：江西理工大学，2022.
[15] 滕陈源，丁逸超，张有兵，等.基于VMD-Informer-BiLSTM模型的超短期光伏功率预测[J].高电压技术，2023,49(7):2961-2971.
[16] HUMPHREY W,DALKE A,SCHULTEN K.VMD:visual molecular dynamics[J].Journal of Molecular Graphics,1996,14(1):33-38.
[17] CAO J,LI Z,LI J.Financial time series forecasting model based on CEEMDAN and LSTM[J].Physica A:Statistical Mechanics and Its Applications,2019,519:127-139.
[18] 李练兵，高国强，吴伟强，等.考虑特征重组与改进Transformer的风电功率短期日前预测方法[J].电网技术，2024,48(4):1466-1480.
[19] 王振浩，王翀，成龙，等.基于集合经验模态分解和深度学习的光伏功率组合预测[J].高电压技术，2022,48(10):4133-4142.
[20] HUANG N T,WU Y Y,CAI G W,et al.Short-term wind speed forecast with low loss of information based on feature generation of OSVD[J].IEEE Access,2019,7:81027-81046.
[21] JIAO J Y,TANG Y M,LIN K Y,et al.Dilateformer:multi-scale dilated transformer for visual recognition[J].IEEE Transactions on Multimedia,2023,25:8906-8919.
[22] ZENG A L,CHEN M X,ZHANG L,et al.Are transformers effective for time series forecasting?[C]// Proceedings of the AAAI Conference on Artificial Intelligence.Menlo Park:AAAI,2023,37(9):11121-11128.
[23] 刘洋，王聪，张宏立，等.一种风矢量分解和RobustSTL-TimesNet-BiGRU的复杂地形风向预测[J].太阳能学报，2025,46(3):576-588.
基本信息:
DOI：10.13624/j.cnki.issn.1001-7445.2025.1195
中图分类号:TM615
引用信息:
[1]黄艳国,刘春华,张慧敏,等.基于特征重组与改进Informer-ARIMA的光伏功率短期组合预测方法[J].广西大学学报(自然科学版),2025,50(06):1195-1208.DOI:10.13624/j.cnki.issn.1001-7445.2025.1195.
基金信息:
国家自然科学基金项目(72061016)
2025-12-25
2025-12-25