The Correction of Wind Speed Forecasted of Wind Farms under Complex Topographic Conditions using Kalman Filtering

  • LI Zhaorong ,
  • DA Xuanfang ,
  • WANG Yousheng ,
  • ZHAO Wenjing ,
  • YAN Xiaomin
Expand
  • Gansu Meteorological Service Centre, Lanzhou 730020, Gansu, China

Received date: 2017-10-27

  Online published: 2018-10-28

Abstract

The distribution characteristics of wind speed and direction at 70 m and prediction errors of forecast wind speed of three typical wind farms (i. e., Nanqiu, Heiyazi, and Ganhekou) with complex topographic conditions in Gansu were analyzed. A correction model based on Kalman filtering was established to correct wind speed forecasted by Beijing Rapid Updated Cycling Analysis and Forecast System (BJ-RUC). Results show that:The percentage of effective wind speed hours of the total year is 90.9% in Nanqiu, 85.06% in Heiyazi, and 82.93% in Ganhekou, respectively. Significant differences of the percentage of effective wind speed hours exist in different seasons for each wind farm. Specifically, percentages of effective wind speed hours are higher in summer and autumn, and can reach up to 29.65%, 27.19%, 23.24% in Nanqiu, Heiyazi and Ganhekou, respectively. In summer, the diurnal variation of wind speed in Nanqiu has two or three peaks while in Heiyazi is characterized by a single peak and in Ganhekou is performed as double peaks and single valley. From summer to autumn, the prevailing wind direction in Nanqiu is southeast, the leading wind direction in Heiyazi shifts from east to west and the dominant wind direction in Ganhekou persistently keeps easterly persistently. The gustiness of wind speed varies from season to season. The Weibull distribution indicates that the wind speed in september is more volatile than wind speed in June in Heiyazi and Ganhekou. On the contrary, wind speed in June is more volatile than that in september in Nanqiu. Limitations in forecasting wind speed by BJ-RUC are mainly two aspects:one is that the gustiness of prediction wind speed is relatively weaker; The other one is that the wind speed forecasted by model is relatively larger than measured one. However, with the correction by Kalman Filtering method, the capability of forecasting gustiness of wind speed based on numerical weather model has been improved; Weibull distribution shape and scale parameters of the corrected wind speed are approximation to those of measured wind speed. The correlation coefficient of observed speed and corrected speed can be increased by 15%; the absolute error and root mean square error have also been improved significantly, which drop to 1.30 m·s-1 and 1.66 m·s-1.

Cite this article

LI Zhaorong , DA Xuanfang , WANG Yousheng , ZHAO Wenjing , YAN Xiaomin . The Correction of Wind Speed Forecasted of Wind Farms under Complex Topographic Conditions using Kalman Filtering[J]. Plateau Meteorology, 2018 , 37(5) : 1402 -1412 . DOI: 10.7522/j.issn.1000-0534.2018.00025

References

[1]Conradsen K, Nielsen L B, Prahm L P, 1984. Review of Weibull statistics for estimation of wind speed distributions[J]. J Appl Meteor, 23(23):1173-1183.
[2]Marzban C, Stumpf G J, 1998. A neural network for damaging wind prediction[J]. Wea Forecasting, 13(1):151-163.
[3]Wang C H, Feng S L, 2014. Error features and their possible causes in simulated low-level winds by WRF at a wind farm[J]. Wind Energy, 17(9):1315-1325.
[4]Cao Y, Chen H B, Wang P C, 2017. Analysis of the data reliability and wind field characteristics near surface boundary layer with doppler sodar observations[J]. Plateau Meteor, 36(5):1315-1324. DOI:10.7522/j.issn. 1000-0534.2016.00100.<br/>曹杨, 陈洪滨, 王普才, 2017.声雷达资料可靠性及近地面边界层风场特征分析[J].高原气象, 36(5):1315-1324.
[5]Gong W J, Li W X, Zhang G M, et al, 2011. The estimation algorithm on the probabilistic distribution parameters of wind speed based on Weibull distribution[J]. Renew Energy Res, 29(6):20-23.<br/>龚伟俊, 李为相, 张广明, 等, 2011.基于威布尔分布的风速概率分布参数估计方法[J].可再生能源, 29(6):20-23.
[6]Huang F X, 2013. The WRF model forecast wind speed correction under complex terrain based on Kalman Filter[D]. Nanjing: University of Information Science and Technology, 1-68.<br/>黄凤新, 2013.基于卡尔曼滤波的复杂地形WRF模式预报风速订正[D].南京: 南京信息工程大学, 1-68.
[7]Jiang G X, Pan J W, Tian J K, et al, 2015. Energy distribution research based on the parameter K analysis of two-parameter Weibull distribution wind conditions[J]. Electric Power Construction, 36(3):105-108.<br/>姜广绪, 潘晶雯, 田景奎, 等, 2015.双参数威布尔分布风况中基于k值分析的能量分布研究[J].电力建设, 36(3):105-108.
[8]Kong Y S, Qian J M, Zang Z L, et al, 2010. Statistical weather prediction principles and methods[M]. Beijing:China Meteorological Press, 441-447.<br/>孔玉寿, 钱建明, 臧增亮, 等, 2010.统计天气预报原理与方法[M].北京:气象出版社, 441-447.
[9]Li X X, Huang T, Wang X, et al, 2017. Analysis of characters of wind field in surface layer in Lanzhou New District[J]. Plateau Meteor, 36(4):1001-1009. DOI:10.7522/j.issn. 1000-0534.2016.00092.<br/>李晓霞, 黄涛, 王兴, 等, 2017.兰州新区近地层风场时空特征分析[J].高原气象, 36(4):1001-1009.
[10]Lu R H, Xu C Y, Zhang L, et al, 1997. Calculation method for initial value of Kalman Filter and its application[J]. J Appl Meteor, 14(1):34-43.<br/>陆如华, 徐传玉, 张玲, 等, 1997.卡尔曼滤波的初值计算方法及其应用[J].应用气象学报, 14 (1):34-43.
[11]Sheng D C, 2015. The development of large-scale wind turbines in China[J]. Solar Energy (2):11-14.<br/>沈德昌, 2015.我国大型风电机组技术发展情况[J].太阳能(2):11-14.
[12]Shi L, Xu L N, Hao Y Z, 2017. Application research on the multi-model fusion forecast of wind speed[J]. Plateau Meteor, 36(4):1022-1028. DOI:10.7522/j.issn. 1000-0534.2017.00021.<br/>石岚, 徐丽娜, 郝玉珠, 2017.多模式风速融合预报应用研究[J].高原气象, 36(4):1022-1028.
[13]Sun D, 2015. Analysis and research on transient stability of power system integrated with wind power generation[D]. Beijing: North China Electric Power University, 1-51.<br/>孙丹, 2015.风力发电系统对电网暂态稳定性影响的分析与研究[D].北京: 华北电力大学, 1-51.
[14]Sun X J, Li Y, Zhang Y H, et al, 2017. Near-Surface wind simulation over acrid lakeshore area and sensitivity studies using the WRF-LES[J]. Plateau Meteor, 36(3):835-844. DOI:10.7522/j.issn. 1000-0534.2016.00058.<br/>孙学金, 李岩, 张燕鸿, 等, 2017.基于WRF-LES的干旱湖区近地面风场模拟与敏感性研究[J].高原气象, 36(3):835-844.
[15]Yan Y, Xu C, Liu D Y, et al, 2011. Study on probabilistic distribution parameters of wind speed influenced by different anemometer time intervals[J]. Renew Energy Res, 29(6):24-28.<br/>严彦, 许昌, 刘德有, 等, 2011.测风数据的时间间隔对风速概率分布参数计算的影响[J].可再生能源, 29(6):24-28.
[16]Yu J, Jiang Z H, Yu W, et al, 2015. Error analysis and correction of wind speed numerical forecast at wind farm[J]. J Meteor Sci, 35(5):587-592.<br/>余江, 江志红, 俞卫, 等, 2015.风电场风速数值预报的误差分析及订正[J].气象科学, 35(5):587-592.
[17]Zhang F M, Wang C H, 2014. Experiment of surface-layer wind forecast improvement by assimilating conventional data with WRF-3DVAR[J]. Plateau Meteor, 33 (3), 675-685. DOI:10.7522/j.issn. 1000-0534.2012.00198.<br/>张飞民, 王澄海, 2014.利用WRF-3DVAR同化常规观测资料对近地层风速预报改进的试验[J].高原气象, 33(3), 675-685.
[18]Zhang L H, Shang K Z, Cheng Y F, et al, 2011. Study on the correction of numerical prediction products[J]. Journal of Lanzhou University (Natural Sciences), 47(3):44-49.<br/>张兰慧, 尚可政, 程一帆, 等, 2011.数值预报产品的误差订正方法[J].兰州大学学报(自然科学版), 47(3):44-49.
[19]Zhang S Y, Hu F, 2017. Review on study of atmospheric boundary layer and wind power generation interaction[J]. Plateau Meteor, 36(4):1127-1137. DOI:10.7522/j.issn. 1000-0534.2016.00095.<br/>张双益, 胡非, 2017.大气边界层与风力发电的相互作用研究综述[J].高原气象, 36(4):1127-1137.
[20]Zhu Y, Liu Y X, Cheng X H, et al, 2013. A research on revision of simulated wind speed based on linear rolling and extremum procession[J]. J Trop Meteor, 29 (4):681-686.<br/>祝赢, 柳艳香, 程兴宏, 等, 2013.线性滚动极值处理方法对数值模拟风速的订正研究[J].热带气象学报, 29(4):681-686.
Outlines

/