基于机器学习的数值天气预报风速订正研究
|
摘要
对风速进行准确预测是精细化天气预报服务(如风能发电、冬季奥运会赛场条件保障等)的重要环节。本文基于三种机器学习算法(LASSO回归、随机森林和深度学习),对数值天气预报模式ECMWF预测的华北地区近地面10 m风速进行订正。首先利用LASSO回归算法提取对10 m风速有重要影响的气象要素特征集,将其作为三种机器学习算法的输入,建立相应模型对ECMWF预测的风速进行订正。用提取后的气象要素特征集建模有助于减少计算量和存储开销,并减小模型的复杂性,从而提高模型的泛化能力。将订正结果与传统订正方法模式输出统计(model output statistics,MOS)得到的订正结果进行对比。结果表明,三种机器学习算法的订正效果均好于MOS方法,显示了机器学习方法在改善局地精准气象预报方面的潜力。
Accurate prediction of wind speed is crucial for local weather forecasting services(e. g., dealing with wind power industry and the Olympic Winter Game). Based on three machine learning algorithms(LASSO regression, random forest and deep learning), this paper demonstrates three models for adjusting the 10 m wind speed in North China predicted by the numerical weather forecast model of ECMWF. Firstly, the LASSO regression algorithm is applied to identify the features which significantly affect the nearsurface wind speed, among all the available meteorological elements. The extracted feature set is used as input for each machine learning algorithm to establish a model to adjust the ECMWF-predicted wind speed.Feature extraction helps to reduce the amount of computation,storage overhead and the complexity of the model, hence to facilitate the generalization of the model. The results of the three machine learning algorithms are compared with that of the traditional MOS method. All the three machine learning methods show a better performance in adjusting the wind speed than that of MOS, indicating great potential of the machine learning methods in improving local weather forecast.
Accurate prediction of wind speed is crucial for local weather forecasting services(e. g., dealing with wind power industry and the Olympic Winter Game). Based on three machine learning algorithms(LASSO regression, random forest and deep learning), this paper demonstrates three models for adjusting the 10 m wind speed in North China predicted by the numerical weather forecast model of ECMWF. Firstly, the LASSO regression algorithm is applied to identify the features which significantly affect the nearsurface wind speed, among all the available meteorological elements. The extracted feature set is used as input for each machine learning algorithm to establish a model to adjust the ECMWF-predicted wind speed.Feature extraction helps to reduce the amount of computation,storage overhead and the complexity of the model, hence to facilitate the generalization of the model. The results of the three machine learning algorithms are compared with that of the traditional MOS method. All the three machine learning methods show a better performance in adjusting the wind speed than that of MOS, indicating great potential of the machine learning methods in improving local weather forecast.
引文
邓华,张颖超,顾荣,等,2018.基于PCA-RBF的风电场短期风速订正方法研究[J].气象科技,46(1):10-15. Deng H,Zhang Y C,Gu R,et al,2018. Research on Short-term wind speed correction method of wind farm based on PCA-RBF[J]. Meteor Sci Tech,46(1):10-15(in Chinese).
韩耀风,覃文锋,陈炜,等,2017. adaptive LASSO logistic回归模型应用于老年人养老意愿影响因素研究的探讨[J].中国卫生统计,34(1):18-22. Han Y F,Qin W F,Chen W,et al,2017. Study on the application of adaptive LASSO logistic regression model to the influencing factors of the elderly's willingness to support the elderly[J]. Chin J Health Statis,34(1):18-22(in Chinese)
胡海川,黄彬,魏晓琳,2017.我国近海洋面10 m风速集合预报客观订正方法[J].气象,43(7):856-862. Hu H C.Huang B,Wei X L, 2017. An objective correction method for the ensemble prediction of 10 m wind speed near the ocean surface in China[J]. Meteor Mon,43(7):856-862(in Chinese).
孔令彬,赵艳茹,王聚杰,等,2014.基于支持向量机风速订正方法的研究[J].西南大学学报(自然科学版),36(5):194-200.Kong L B,Zhao Y R,Wang J J,et al,2014. Research on wind speed correction method based on support vector machine[J]. Southwest Univ(Nat Sci),36(5):194-200(in Chinese).
李丽辉,朱建生,强丽霞,等,2017.基于随机森林回归算法的高速铁路短期客流预测研究[J].铁道运输与经济,39(9):12-16. Li L H,Zhu J S,Qiang L X,et al,2017. Research on short-term passenger flow forecasting of high-speed railway based on random forest regression algorithm[J].Rail Way Transport and Economy,39(9):12-16(in Chinese).
戚双斌,王维庆,张新燕,2009.基于支持向量机的风速与风功率预测方法研究[J].华东电力,37(9):1600-1603. Qi S B,Wang W Q,Zhang X Y,et al,2009. Research on wind speed and wind power prediction based on support vector machine[J]. East China Electric Power,37(9):1600-1603(in Chinese).
孙军波,钱燕珍,陈佩燕,等,2010.登陆台风站点大风预报的人工神经网络方法[J].气象,36(9):81-86.Sun J B,Qian Y Z,Chen P Y,et al,2010. Artificial neural network method for landing typhoon site gale forecast[J]. Meteor Mon,36(9):81-86(in Chinese).
陶祖钰,范俊红,李开元,等,2016.谈谈气象要素(压、温、湿、风)的物理意义和预报应用价值[J].气象科技进展,6(5):59-64.Tao Z Y,Fan J H, Li K Y,et al,2016. Talk about the physical meaning and forecasting application value of meteorological elements(pressure, temperature, humidity,wind)[J]. Adv Meteor Sci Technol,6(5):59-64(in Chinese).
吴启树,韩美,郭弘,等,2016. MOS温度预报中最优训练期方案[J].应用气象学报,27(4):426-434. Wu Q S,Han M,Guo H,et al,2016. Optimal training period scheme in MOS temperature prediction[J].J Appl Meteor,27(4):426-434(in Chinese).
肖擎曜,胡非,范绍佳,等,2017.风能数值预报的模式输出统计(MOS)研究[J].资源科学,39(1):116-124. Xiao Q Y, Hu F,Fan S J, et al, 2017. Model output statistics(MOS)of wind energy numerical prediction[J]. Res Sci,39(1):116-124(in Chinese).
阳坤,郭晓峰,武炳义,2010.青藏高原地表感热通量的近期变化趋势[J].中国科学:地球科学,40(7):923-932. Yang K,Guo X F,Wu B Y,2010. Recent changes in surface sensible heat flux over the Qinghai-Tibet Plateau[J]. Sci China:Earth Sci,40(7):923-932(in Chinese).
杨薛明,边继飞,朱霄珣,等,2016.基于最大熵混沌时间序列的支持向量机短期风速预测模型研究[J].太阳能学报,37(9):2173-2179. Yang X M, Bian J F, Zhu X X, et al, 2016. Research on short-term wind speed prediction model of support vector machine based on maximum entropy chaotic time series[J]. Acta Energiae Solaris Sin,37(9):2173-2179(in Chinese).
叶小岭,顾荣,邓华,等,2017.基于WRF模式和PSO-LSSVM的风电场短期风速订正[J].电力系统保护与控制,45(22):48-54. Ye X L,Gu R,Deng H,et al,2017. Short-term wind speed correc-tion of wind farm based on WRF mode and PSO-LSSVM[J].Power System Protection and Control, 45(22):48-54(in Chinese).
袁翀,戚佳金,王文霞,等,2016.采用正则化极限学习机的短期风速预测[J].电网与清洁能源,32(11):62-68. Yuan C,Qi J J,Wang W X,et al,2016. Short-term wind speed prediction using regularized limit learning machine[J]. Power System and Clean Energy,32(11):62-68(in Chinese).
张颖超,肖寅,邓华,2016.基于ELM的风电场短期风速订正技术研究.气象,42(4):466-471. Zhang Y C,Xiao Y,Deng H,2016.Research on short-term wind speed correction technology of wind farm based on ELM[J]. Meteor Mon, 42(4); 466-471(in Chinese).
张治国,崔炜,白雪涛,等,2017.第24届冬奥会海坨山赛区近两年冬季地面风场特征[J].干旱气象,35(3):433-438. Zhang Z G,Cui W,Bai X T,et al,2017. Characteristics of ground wind field in the winter of Haishu Mountain in the 24th Winter Olympic Games[J]. Arid Meteor,35(3):433-438(in Chinese).
周志华,2016.机器学习[M].北京:清华大学出版社:248-261. Zhou Z H,2016. Machine Learning[M]. Beijing:Tsinghua University Press:248-261(in Chinese).
Erdem E, Shi J,2011. ARMA based approaches for forecasting thetuple of wind speed and direction[J]. Appl Energy,88(4):1405-1414.
Li G,Shi J,2010. On comparing three artificial neural networks for wind speed forecasting[J]. Appl Energy,87(7):2313-2320.
Lin Y J,Kruger U,Zhang J P,et al,2015. Seasonal analysis and prediction of wind energy using random forests and arx model structures[J]. IEEE Trans Control Syst Technol, 23(5):1994-2002.
López E, Valle C, Allende H,et al,2018. Wind power forecasting based on echo state networks and long short-term memory[J].Energies,11(3):526.
Ren Y,Suganthan P N,Srikanth N,2016. A novel empirical mode decomposition with support vector regression for wind speed forecasting[J]. IEEE Trans Neural Netw Learn Sys, 27(8):1793-1798.
Wang J J,Wang Y F,Li Y N,2018. A novel hybrid strategy using three-phase feature extraction and a weighted regularized extreme learning machine for multi-step ahead wind speed prediction[J]. Energies,11(2):321.
Zjavka L,2015. Wind speed forecast correction models using polynomial neural networks[J]. Renew Energy,83:998-1006.
韩耀风,覃文锋,陈炜,等,2017. adaptive LASSO logistic回归模型应用于老年人养老意愿影响因素研究的探讨[J].中国卫生统计,34(1):18-22. Han Y F,Qin W F,Chen W,et al,2017. Study on the application of adaptive LASSO logistic regression model to the influencing factors of the elderly's willingness to support the elderly[J]. Chin J Health Statis,34(1):18-22(in Chinese)
胡海川,黄彬,魏晓琳,2017.我国近海洋面10 m风速集合预报客观订正方法[J].气象,43(7):856-862. Hu H C.Huang B,Wei X L, 2017. An objective correction method for the ensemble prediction of 10 m wind speed near the ocean surface in China[J]. Meteor Mon,43(7):856-862(in Chinese).
孔令彬,赵艳茹,王聚杰,等,2014.基于支持向量机风速订正方法的研究[J].西南大学学报(自然科学版),36(5):194-200.Kong L B,Zhao Y R,Wang J J,et al,2014. Research on wind speed correction method based on support vector machine[J]. Southwest Univ(Nat Sci),36(5):194-200(in Chinese).
李丽辉,朱建生,强丽霞,等,2017.基于随机森林回归算法的高速铁路短期客流预测研究[J].铁道运输与经济,39(9):12-16. Li L H,Zhu J S,Qiang L X,et al,2017. Research on short-term passenger flow forecasting of high-speed railway based on random forest regression algorithm[J].Rail Way Transport and Economy,39(9):12-16(in Chinese).
戚双斌,王维庆,张新燕,2009.基于支持向量机的风速与风功率预测方法研究[J].华东电力,37(9):1600-1603. Qi S B,Wang W Q,Zhang X Y,et al,2009. Research on wind speed and wind power prediction based on support vector machine[J]. East China Electric Power,37(9):1600-1603(in Chinese).
孙军波,钱燕珍,陈佩燕,等,2010.登陆台风站点大风预报的人工神经网络方法[J].气象,36(9):81-86.Sun J B,Qian Y Z,Chen P Y,et al,2010. Artificial neural network method for landing typhoon site gale forecast[J]. Meteor Mon,36(9):81-86(in Chinese).
陶祖钰,范俊红,李开元,等,2016.谈谈气象要素(压、温、湿、风)的物理意义和预报应用价值[J].气象科技进展,6(5):59-64.Tao Z Y,Fan J H, Li K Y,et al,2016. Talk about the physical meaning and forecasting application value of meteorological elements(pressure, temperature, humidity,wind)[J]. Adv Meteor Sci Technol,6(5):59-64(in Chinese).
吴启树,韩美,郭弘,等,2016. MOS温度预报中最优训练期方案[J].应用气象学报,27(4):426-434. Wu Q S,Han M,Guo H,et al,2016. Optimal training period scheme in MOS temperature prediction[J].J Appl Meteor,27(4):426-434(in Chinese).
肖擎曜,胡非,范绍佳,等,2017.风能数值预报的模式输出统计(MOS)研究[J].资源科学,39(1):116-124. Xiao Q Y, Hu F,Fan S J, et al, 2017. Model output statistics(MOS)of wind energy numerical prediction[J]. Res Sci,39(1):116-124(in Chinese).
阳坤,郭晓峰,武炳义,2010.青藏高原地表感热通量的近期变化趋势[J].中国科学:地球科学,40(7):923-932. Yang K,Guo X F,Wu B Y,2010. Recent changes in surface sensible heat flux over the Qinghai-Tibet Plateau[J]. Sci China:Earth Sci,40(7):923-932(in Chinese).
杨薛明,边继飞,朱霄珣,等,2016.基于最大熵混沌时间序列的支持向量机短期风速预测模型研究[J].太阳能学报,37(9):2173-2179. Yang X M, Bian J F, Zhu X X, et al, 2016. Research on short-term wind speed prediction model of support vector machine based on maximum entropy chaotic time series[J]. Acta Energiae Solaris Sin,37(9):2173-2179(in Chinese).
叶小岭,顾荣,邓华,等,2017.基于WRF模式和PSO-LSSVM的风电场短期风速订正[J].电力系统保护与控制,45(22):48-54. Ye X L,Gu R,Deng H,et al,2017. Short-term wind speed correc-tion of wind farm based on WRF mode and PSO-LSSVM[J].Power System Protection and Control, 45(22):48-54(in Chinese).
袁翀,戚佳金,王文霞,等,2016.采用正则化极限学习机的短期风速预测[J].电网与清洁能源,32(11):62-68. Yuan C,Qi J J,Wang W X,et al,2016. Short-term wind speed prediction using regularized limit learning machine[J]. Power System and Clean Energy,32(11):62-68(in Chinese).
张颖超,肖寅,邓华,2016.基于ELM的风电场短期风速订正技术研究.气象,42(4):466-471. Zhang Y C,Xiao Y,Deng H,2016.Research on short-term wind speed correction technology of wind farm based on ELM[J]. Meteor Mon, 42(4); 466-471(in Chinese).
张治国,崔炜,白雪涛,等,2017.第24届冬奥会海坨山赛区近两年冬季地面风场特征[J].干旱气象,35(3):433-438. Zhang Z G,Cui W,Bai X T,et al,2017. Characteristics of ground wind field in the winter of Haishu Mountain in the 24th Winter Olympic Games[J]. Arid Meteor,35(3):433-438(in Chinese).
周志华,2016.机器学习[M].北京:清华大学出版社:248-261. Zhou Z H,2016. Machine Learning[M]. Beijing:Tsinghua University Press:248-261(in Chinese).
Erdem E, Shi J,2011. ARMA based approaches for forecasting thetuple of wind speed and direction[J]. Appl Energy,88(4):1405-1414.
Li G,Shi J,2010. On comparing three artificial neural networks for wind speed forecasting[J]. Appl Energy,87(7):2313-2320.
Lin Y J,Kruger U,Zhang J P,et al,2015. Seasonal analysis and prediction of wind energy using random forests and arx model structures[J]. IEEE Trans Control Syst Technol, 23(5):1994-2002.
López E, Valle C, Allende H,et al,2018. Wind power forecasting based on echo state networks and long short-term memory[J].Energies,11(3):526.
Ren Y,Suganthan P N,Srikanth N,2016. A novel empirical mode decomposition with support vector regression for wind speed forecasting[J]. IEEE Trans Neural Netw Learn Sys, 27(8):1793-1798.
Wang J J,Wang Y F,Li Y N,2018. A novel hybrid strategy using three-phase feature extraction and a weighted regularized extreme learning machine for multi-step ahead wind speed prediction[J]. Energies,11(2):321.
Zjavka L,2015. Wind speed forecast correction models using polynomial neural networks[J]. Renew Energy,83:998-1006.