| 542 | 25 | 31 |
| 下载次数 | 被引频次 | 阅读次数 |
为了提高风电场输出功率预报的精度,提出一种将随机森林(random forest, RF)与门控循环单元(gated recurrent unit, GRU)神经网络相结合的超短期风电预测模型,并以云南李子箐风电场10台风电机组作为研究对象,在风电场配属的1座测风塔4个高度(10、30、50、70 m)上进行风速测量。将测风塔所测的4层风速数据和数值天气预报(numerical weather prediction, NWP)系统的风速输出数据进行归一化处理;然后,计算4个不同高度的实测风速数据和风电场总输出功率的皮尔逊相关系数,确定出与风电场输出功率相关的最显著实测风速高度;接下来,构建数值预报模型输出的70 m风速与风电场测风塔70 m高度处风速之间的RF订正模型,训练数值预报风速产品,建立二者之间的映射关系;最后,用该映射关系订正后的数值预报风速输入GRU神经网络,预测风电输出功率。试验结果表明:所提方法的预测精度较传统误差反向传播(back propagation, BP)神经网络方法有了很大的改进与提高,有利于进一步提高风电并网功率规模。
Abstract：In order to improve the accuracy of wind farm output power forecasting, this paper proposes an ultra-short-term wind power forecasting model which combines the random forecast(RF) and the gated recurrent unit(GRU) neural network. In this paper, 10 wind turbines of a wind farm in Yunnan province are taken as the research object. The wind speed is measured on four heights(10 meters, 30 meters, 50 meters and 70 meters) by a wind measuring tower attached to the wind farm. We normalize both the wind speed data which measured from the four altitude of the wind measuring tower and the historical wind speed data from the numerical weather prediction(NWP) system output; Then we calculated the Pearson correlation coefficient between the total wind farm output and the measured wind speed data from the four altitudes, to determine the most significant measured wind speed altitude related to the wind farm output power. Then we construct a RF correction model based on the measured wind speed at the altitude of 70 meters and the wind speed output from numerical weather prediction at the altitude of 70 meters. Furthermore, we establish a mapping relationship between the measured data and the NWP results by training the model. Finally, the corrected wind speed has been inputted into GRU neural network to forecast the wind farm output power. The experimental results show that the forecasting accuracy of the paper's method is much better than that of the traditional back propagation(BP) neural network and is benefit for enlarging the scale of wind power in power grid system.
[1] 李博文,张靖.风电场风速及风功率预测研究综述[J].贵州电力技术,2017,20(5):9-13.LI B W,ZHANG J.The summary of wind speed and wind power forecast technique research [J].Guizhou Electric Power Technology,2017,20(5):9-13.(in Chinese)
[2] 姚辉.风电场风电功率预测模型及应用研究[D].南京:东南大学,2015:6-7.YAO H.Study on the model and the application of wind power prediction [D].Nanjing:Southeast University,2015:6-7.(in Chinese)
[3] 彭小圣,熊磊,文劲宇,等.风电集群短期及超短期功率预测精度改进方法综述[J].中国电机工程学报,2016(23):14-25.PENG X S,XIONG L,WEN J Y,et al.A summary of the state of the art for short-term and ultra-short-term wind power prediction of regions [J].Proceedings of the CSEE,2016(23):14-25.(in Chinese)
[4] MILLER N W,SANCHEZ-GASCA J J,PRICE W W,et al.Dynamic modeling of GE 1.5 and 3.6 MW wind turbine-generators for stability simulations [C]// Power Engineering Society General Meeting.Toronto,Canada:IEEE,2003.DOI:10.1109/PES.2003.1267470.
[5] BOSSANYI E A.Short-term wind prediction using Kalman filters [J].Wind Engineering,1985,9(1):1-8.
[6] MOHAMED A M,SHAFIQUR R,TALAL O H.A neural networks approach for wind speed prediction [J].Renewable Energy,1998,13(3):345-354.
[7] 方市彬,厉虹,王丽婕.基于NWP单点聚类分析与BP神经网络的短期风电功率预测[J].电气应用,2017(15):67-70.FANG S B,LI H,WANG L J.Short term wind power prediction based on NWP single point clustering analysis and BP neural network [J].Electrotechnical Application,2017(15):67-70.(in Chinese)
[8] 韩爽,孟航,刘永前,等.增量处理双隐层BP神经网络风电功率预测模型[J].太阳能学报,2015,36(9):2238-2244.HAN S,MENG H,LIU Y Q,et al.Study on optimization of bp neural network wind power prediction model with two hidden layers [J].Acta Energiae Solaris Sinica,2015,36(9):2238-2244.(in Chinese)
[9] 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(3):456-462.XUE Y,WANG L,WANG S,et al.An ultra-short-term wind power forecasting model combined with CNN and GRU networks [J].Renewable Energy Resources,2019(3):456-462.(in Chinese)
[10] 朱乔木,李弘毅,王子琪,等.基于长短期记忆网络的风电场发电功率超短期预测[J].电网技术,2017(12):68-73.ZHU Q M,LI H Y,WANG Z Q,et al.Short-term wind power forecasting based on LSTM [J].Power System Technology,2017(12):68-73.(in Chinese)
[11] 康重庆,夏清,刘梅.电力系统负荷预测[M].北京:中国电力出版社,2007:208-213.KANG C Q,XIA Q,LIU M.Power system load forecasting [M].Beijing:China Electric Power Press,2007:208-213.(in Chinese)
[12] 许杏花,潘庭龙.基于KPCA-RF的风电场功率预测方法研究[J].可再生能源,2018(9):1323-1327.XU X H,PAN T L.Wind power prediction based on KPCA-RF [J].Renewable Energy Resources,2018(9):1323-1327.(in Chinese)
[13] HOCHREITER S,SCHMIDHUBER,JüRGEN.Long short-term memory [J].Neural Computation,1997,9(8):1735-1780.
[14] CHO K H,MERRIENBOER B V,BAHDANAU D,et al.On the properties of neural machine translation:encoder-decoder approaches [EB/OL].[2018-06-05].https://arxiv.org/abs/1409.1259.
[15] CHUNG J Y,GULCEHRE C,CHO K H,et al.Empirical evaluation of gated recurrent neural networks on sequence modeling [EB/OL].[2018-06-05].https://arxiv.org/abs/1412.3555.
基本信息:
中图分类号:TM614
引用信息:
[1]赏益,高志球,韩威.基于RF和GRU组合算法的超短期风电功率预测[J].中国科技论文,2020,15(09):987-992.
2020-01-22
2020
2020-09-27
2020
1
2020-09-15
2020-09-15