YANG Rui,WEN Wu,XU Hong.Short-term Wind Power Prediction based on PCC-CNN-GRU[J].Journal of Chengdu University of Information Technology,2022,37(02):165-170.[doi:10.16836/j.cnki.jcuit.2022.02.009]
基于PCC-CNN-GRU的短期风电功率预测
- Title:
- Short-term Wind Power Prediction based on PCC-CNN-GRU
- 文章编号:
- 2096-1618(2022)02-0165-06
- Keywords:
- wind speed prediction; PCC; CNN; combined forecasting model
- 分类号:
- TP301.6
- 文献标志码:
- A
- 摘要:
- 可靠的风功率预测对于电力部门制定电力调度计划、维护电网的安全运行具有重要意义。这项任务极富挑战性,因为影响风功率预测准确率的因素较多,如地理因素、环境因素、人为因素等。将环境因素考虑在内,提出一种基于深度学习的组合预测模型PCC-CNN-GRU皮尔逊相关系数法(Pearson correlation coefficient)-卷积神经网络(convolutional neural networks)-门控循环单元(gate recurrent unit)。该模型首先使用皮尔逊相关系数法分析输入数据中不同因素与风功率之间的相关关系,剔除与功率无关的因素,重构新的输入数据并进行归一化处理,并使用一维卷积神经网络对数据的深层特征进行提取,最后将提取的特征送入GRU神经网络进行预测。实验使用新疆某地风场实地采集数据仿真,结果表明,该方法的预测误差最小,预测能力最强。
- Abstract:
- Reliable wind power prediction is of great significance for the power sector to make power dispatching plans and maintain the safe operation of the power grid. This task is very challenging, because the accuracy of wind power prediction is affected by many factors, such as geographical factors, environmental factors, human factors and so on. In this paper, a combination forecasting model PCC-CNN-GRU(Pearson correlation coefficient-convolutional neural networks-gate recurrent unit)based on deep learning is proposed, taking environmental factors into account. Firstly, the model uses the Pearson Correlation Coefficient method to analyze the correlation between different factors and wind power in the input data, eliminate the factors not related to the wind power, and reconstruct the new input data and carry out normalization processing.Then the deep features of the data are extracted by using one-dimensional convolutional neural network, and finally the extracted features are sent to GRU neural network for prediction. The experimental results show that the method has the least prediction error and the strongest prediction ability.
参考文献/References:
[1] 谭显东,刘俊,徐志成,等.“双碳”目标下“十四五”电力供需形势[J/OL].中国电力:http://kns.cnki.net/kcms/detail/11.3265.TM.20210408.1657.006.html,2021.
[2] 郭敏晓,杨宏伟.围绕“碳中和”愿景能源与环境领域将呈现六方面的变化趋势“十三五”能源与环境形势及“十四五”展望[J/OL].中国能源:http://kns.cnki.net/kcms/detail/11.2587.TK.20210329.2130.008.html,2021-04-14.
[3] 马吉科,石星煜,王黎明,等.融合多源信息的短期风电功率预测方法研究[J].计算机仿真,2020,37(7):137-143.
[4] Landberg L. Prediktor: An on-line prediction system[A]. Wind Power for the 21st Century, EUWEC Special Topic Conference[C], Kassel,2000.
[5] Lange M U, Focken D. Heinemann previento—regional wind power prediction with risk control[C]. Proceedings of the World Wind Energy Conference, Berlin,2002.
[6] 王瑶,杨培宏,李忠.基于RBF算法短期风功率预测[J].绿色科技,2020(6):256-257.
[7] 陈岩,周晨,侯群.SVM算法在风力发电机功率预测中的应用研究[J].电子世界,2021(4):100-102.
[8] 刘军涛,欧阳孜.基于ARMA模型的短期风功率预测研究[J].西北水电,2017(6):100-104.
[9] 黄冬梅,唐振,胡安铎,等.基于VMD-GRU的电力短期负荷预测方法[J].物联网技术,2021,11(4):67-70.
[10] 谭彦聪,王海云,王江江.基于IFA-LSSVM的短期风功率预测[J].水力发电,2021,47(4):112-116.
[11] 田波,朴在林,郭丹,等.基于ARMA-GARCH模型的超短期风功率预测研究[J].电测与仪表,2016,53(17):12-17.
[12] 周盛山,汤占军,王金轩,等.EEMD和CNN-XGBoost在风电功率短期预测的应用研究[J].电子测量技术,2020,43(22):55-61.
[13] 张爱枫,段新宇,何枭峰.基于CNN和LightGBM的新型风电功率预测模型[J].电测与仪表:2021.
[14] X R Zu,R X Song.Short-term Wind Power Prediction Method Based on Wavelet Packet Decomposition and Improved GRU[J].Journal of Physics:Conference Series,2018,1087(2).
[15] Meng X,Wang R, Zhang X,et al.Hybrid Neural Network Based on GRU with Uncertain Factors for Forecasting Ultra-short-term Wind Power[C].2020 2nd International Conference on Industrial Artificial Intelligence(IAI).2020.
[16] 亓晓燕,刘恒杰,侯秋华,等.融合LSTM和SVM的钢铁企业电力负荷短期预测[J].山东大学学报(工学版),2021(4):91-98.
[17] 薛阳,王琳,王舒,等.一种结合CNN和GRU网络的超短期风电预测模型[J].可再生能源,2019,37(3):456-462.
[18] 胡汉.风电场风电功率概率预测研究[D].南京:东南大学,2016.
备注/Memo
收稿日期:2021-06-30