| 418 | 13 | 620 |
| 下载次数 | 被引频次 | 阅读次数 |
为解决风电历史数据挖掘不充分导致的预测精度不高问题，提出一种基于特征工程、蝠鲼觅食优化算法(Manta Ray Foraging Optimization, MRFO)和极端随机树(Extremely Randomized Trees, ET)模型的短期风电功率预测方法。首先对时间特征提取小时属性特征，并通过对风速、风向和温度等原始气象特征进行特征创造，从而充分挖掘历史数据的隐含信息，同时通过PCA方法降低数据维度。其次，将降维后的数据输入ET模型，并利用MRFO优化ET模型的参数；最后，以新疆某风电场实测数据进行了算例仿真。结果表明：与5种典型机器学习模型相比，ET模型具有更高的风电预测准确度。与单一ET模型相比，特征工程-ET模型较大程度地提高了预测精度，验证了特征工程方法的有效性。在同等条件下，特征工程-MRFO-ET模型比使用特征工程-ET模型均方根误差和平均绝对误差分别降低了29.46%和36.54%,而拟合优度系数提高了3.97%。与此同时，特征工程-MRFO-ET模型也比特征工程-GA-ET模型和特征工程-PSO-ET模型拥有更高的预测精度。研究成果可为解决短期风电功率预测问题提供了一种新的思路。
Abstract：In order to solve the problem of low prediction accuracy caused by insufficient historical data mining of wind power, a feature engineering and MRFO(manta ray foraging optimization)-ET(extremely randomized trees) model-based method of short-term wind power prediction is proposed herein. Firstly, the time feature is extracted from the hourly attribute feature, and the original meteorological features such as wind speed, wind direction and temperature are created, so as to fully excavate the hidden information in the historical data and simultaneously reduce the data dimension by PCA. Secondly, the dimension-reduced data is input into the ET model, and then the parameters of ET model are optimized by MRFO. Finally, a calculation case simulation is carried out with the measured data from a wind farm in Xinjiang. The results show that the ET model has higher wind power prediction accuracy if compared with those from five typical machine learning models. Comparing with single ET model, the prediction accuracy is greatly improved by the Feature Engineering-ET model, and then the effectiveness of the feature engineering method is verified as well. Under the same conditions, the root mean square error and average absolute error of the Feature Engineering-MRFO-ET model are reduced by 29.46% and 36.54% respectively, while the fitting goodness coefficient is increased by 3.97%. At the same time, the Feature Engineering-MRFO-ET model also has higher prediction accuracy than those from the Feature Engineering-GA-ET model and the Feature Engineering-PSO-ET model. The study results can provide a new idea for solving the problem of short-term wind power prediction.
[1] 王勃，刘纯，冯双磊，等.基于集群划分的短期风电功率预测方法[J].高电压技术，2018,44(4):1254-1260.WANG Bo,LIU Chun,FENG Shuanglei,et al.Prediction method for short-term wind power based on wind farm clusters[J].High Voltage Engineering,2018,44(4):1254-1260.
[2] 孙勇，李宝聚，孙志博，等.融合RBF神经网络和集对分析的风电功率超短期预测[J].昆明理工大学学报(自然科学版),2020,45(5):49-58.SUN Yong,LI Baoju,SUN Zhibo,et al.Ultra short term wind power forecasting integrated RBF neural network and set pair analysis[J].Journal of Kunming University of Science and Technology (Natural Science),2020,45(5):49-58.
[3] 国家电网公司.风电功率预测功能规范：Q/GDW 588—2011[S].北京：中国电力出版社，2011.State Grid Corporation of China.Function specification of wind power forecasting:Q/GDW 588—2011[S].Beijing:China Electric Power Press,2011.
[4] 叶林，赵金龙，路朋，等.考虑气象特征与波动过程关联的短期风电功率组合预测[J].电力系统自动化，2021,45(4):54-62.YE Lin,ZHAO Jinlong,LU Peng,et al.Short term wind power combination forecasting considering correlation between meteorological characteristics and fluctuation process[J].Automation of Electric Power Systems,2021,45(4):54-62.
[5] 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，2017,41(12):3797-3802.ZHU Qiaomu,LI Hongyi,WANG Ziqi,et al.Short-term wind power forecasting based on LSTM[J].Power System Technology,2017,41(12):3797-3802.
[6] 周小麟，童晓阳.基于CEEMD-SBO-LSSVR的超短期风电功率组合预测[J].电网技术，2021,45(3):855-864.ZHOU Xiaolin,TONG Xiaoyang.Ultra-short-term wind power combined prediction based on CEEMD-SBO-LSSVR[J].Power System Technology,2021,45(3):855-864.
[7] 姜贵敏，陈志军，李笑竹，等.基于EEMD-ACS-LSSVM的短期风电功率预测[J].太阳能学报，2020,41(5):77-84.JIANG Guimin,CHEN Zhijun,LI Xiaozhu,et al.Short term wind power prediction based on EEMD-ACS-LSSVM[J].Acta Energiae Solaris Sinica,2020,41(5):77-84.
[8] 阎洁，许成志，刘永前，等.基于风速云模型相似日的短期风电功率预测方法[J].电力系统自动化，2018,42(6):53-59.YAN Jie,XU Chengzhi,LIU Yongqian,et al.Short term wind power forecasting method based on similar days of wind speed cloud model[J].Automation of Electric Power Systems,2018,42(6):53-59.
[9] ZHANG Y,CHEN B,PAN G,et al.A novel hybrid model based on VMD-WT and PCA-BP-RBF neural network for short-term wind speed forecasting[J].Energy Conversion and Management,2019,95(SEP):180-197.
[10] 王依宁，解大，王西田，等.基于PCA-LSTM模型的风电机网相互作用预测[J].中国电机工程学报，2019,39(14):4070-4081.WANG Yining,XIE Da,WANG Xitian,et al.Prediction of interaction between grid and wind farms based on PCA-LSTM model[J].Proceedings of the CSEE,2019,39(14):4070-4081.
[11] 胡志鹏，刘剑，张玻，等.基于风险关键特征量的输电线路运行环境风险评估[J].电力系统自动化，2017,41(18):160-166.HU Zhipeng,LIU Jian,ZHANG Bo,et al.Environmental risk assessment of transmission line operation based on the key risk characteristics[J].Automation of Electric Power Systems,2017,41(18):160-166.
[12] GE L,XIAN Y,YAN J,et al.A hybrid model for short-term PV output forecasting based on PCA-GWO-GRNN[J].Journal of Modern Power Systems and Clean Energy,2020,8(6):1268-1275.
[13] 王昕，黄柯，郑益慧，等.基于PNN/PCA/SS-SVR的光伏发电功率短期预测方法[J].电力系统自动化，2016,40(17):156-162.WANG Xin,HUANG Ke,ZHENG Yihui,et al.Short term prediction method of photovoltaic power genera-tion based on PNN/PCA/SS-SVR[J].Journal of Modern Power Systems and Clean Energy,2016,40(17):156-162.
[14] 束洪春，田鑫萃，安娜.采用实测数据和主成分分析的直流输电线路故障识别方法[J].电力系统自动化，2016,40(21):203-209.SHU Hongchun,TIAN Xincui,AN Na.DC transmission line fault identification method based on measured data and principal component analysis[J].Journal of Modern Power Systems and Clean Energy,2016,40(21):203-209.
[15] 曹旦旦，范书瑞，夏克文.共享单车短时需求量预测的机器学习方法比较[J].计算机仿真，2021,38(1):92-97.CAO Dandan,FAN Shurui,XIA Kewen.Comparison of machine learning methods for short -term demand forecasting of shared bicycles [J].Computer Simula-tion,2021,38(1):92-97.
[16] 龙艳芳.基于极限随机树集成的短时交通流预测模型研究[D].长沙：湖南大学，2017.LONG Yanfang.Research on Short-term Traffic Flow Prediction Model based on Ensembles of Extremely Randomized Trees[D].Changsha,China:Hunan University,2017.
[17] 王琎，宫铭举，白银，等.关于住宅建筑供热节能减排预测仿真研究[J].计算机仿真，2019,36(10):429-433.WANG Jin,GONG Mingju,BAI Yin,et al.Simulation research on energy saving and emission reduction heat load prediction of residential buildings [J].Computer Simula-tion,2019,36(10):429-433.
[18] 邱春辉.风力电站短期输出功率预测方法研究[D].武汉：湖北工业大学，2020.QIU Chunhui.Research on Short-term Output Power Prediction of Wind Power Station[D].Wuhan,China:Hubei University of Technology,2020.
[19] 习佳.实时风功率预测算法研究[D].包头：内蒙古科技大学，2020.XI Jia.Research on Real-time Wind Power Forecasting Algorithm[D].Baotou,China:Inner Mongolia Uni-versity of Science & Technology,2020.
[20] ZHAO W,ZHANG Z,WANG L.Manta ray foraging optimization:An effective bio-inspired optimizer for engineering applications[J].Engineering Applications of Artificial Intelligence,2020,87:103300.1-103300.25.
[21] 凤盛强.基于蝠鲼觅食优化算法的配电网故障区间定位[J].兰州文理学院学报(自然科学版),2021,35(1):19-23.FENG Shengqiang.Fault section location in distribution network based on manta ray foraging optimization algorithm[J].Journal of Lanzhou University of Arts and Science(Natural Sciences),2021,35(1):19-23.
[22] GEURTS P,ERNST D,WEHENKEL L.Extremely randomized trees[J].Machine Learning,2006,63(1):3-42.
[23] 赵泽谦，黄强，明波，等.基于多模型随机组合的水文集合预报方法研究[J].水力发电学报，2021,40(1):76-87.ZHAO Zeqian,HUANG Qiang,MING Bo,et al.Hydrological ensemble forecasting method based on stochastic combination of multiple models[J].Journal of Hydroelectric Engineering,2021,40(1):76-87.
基本信息:
DOI：10.13928/j.cnki.wrahe.2022.03.018
中图分类号:TM614;TP18
引用信息:
[1]康文豪,徐天奇,王阳光,等.基于特征工程和MRFO-ET的短期风电功率预测[J],2022,53(03):185-194.DOI:10.13928/j.cnki.wrahe.2022.03.018.
基金信息:
国家自然科学基金项目(61761049)