摘要
本文对实际风力发电机在线检测装置记录的历史数据进行挖掘,确定影响风机输出功率及随机分布的主要因素;建立风机输出功率均值预测模型,求出风机输出功率偏差;按风力等级将风机输出功率偏差样本划分成多个子集,分别采用Bootstrap方法和参数化概率方法求解风机输出功率偏差子集和全集的置信区间,进而对风机输出功率可能出现的范围进行概率预测。利用实际运行数据对方法进行了验证和对比,结果表明:按风速划分风机输出功率偏差子集能显著地提高风机输出功率置信区间预测的准确率;Bootstrap方法预测准确率比参数概率方法高10%以上,但置信区间平均宽度约是后者的1.3倍。
The historical data recorded by the actual wind turbine on-line detection device is mined in this paper,and the main factors which affect the output power of the fan and the random distribution are determined.Thus,the fan output power mean prediction model is established,and the fan output power deviation is obtained.According to the wind level,the fan output power deviation sample is divided into multiple subsets,the Bootstrap method and the parameterized probability method are used to solve the confidence interval of the fan output power deviation subset and the complete set,and then the probability prediction of the range in which the fan output power may occur is predicted.The method is verified and compared with the actual operation data.The results show that the subset of fan output power deviation according to wind speed can significantly improve the accuracy of fan output power confidence interval prediction.The Bootstrap method has a prediction accuracy that is more than 10% higher than the parameter probability method,but the average width of confidence interval is about 1.3 times that of the latter.
引文
[1] 王琦,关添升,秦本双.基于MRMR的ORELM的短期风速预测[J].可再生能源,2018,36(1):85-90.
[2] 刘红柳,杨茂,于宁,等.风向空间分散性及其对全场风电功率预测误差的影响[J].电测与仪表,2017,54(12):54-59.
[3] Liu H,Shi J,Qu X.Empirical investigation on using wind speed volatility to estimate the operation probability and power output of wind turbines[J].Energy Conversion & Management,2013,67(67):8-17.
[4] Ruili Y E,Guo Z,Liu R,et al.Reliable power output based on confidence interval estimation and optimal ESS configuration of wind farm[J].Electric Power Automation Equipment,2017.
[5] 梁海峰,曹大卫,刘博,等.风电场概率分布模型建模及误差分析[J].华北电力大学学报(自然科学版),2017,44(3):8-14.
[6] Wan C,Xu Z,?stergaard J,et al.Discussion of “Combined Nonparametric Prediction Intervals for Wind Power Generation”[J].IEEE Transactions on Sustainable Energy,2017,5(3):1021-1021.
[7] Qin X,Li Y,Shen C,et al.The Correlation Analysis of Clean Energy Output Based on Nonparametric Kernel Density Estimation Probability Models[C]// International Conference on Artificial Intelligence and Industrial Engineering.2016.
[8] Liu Y,Yan J,Han S,et al.Uncertainty Analysis of Wind Power Prediction Based on Quantile Regression[C]// Power and Energy Engineering Conference.2012:1-4.
[9] 刘晓楠,周介圭,贾宏杰,等.基于非参数核密度估计与数值天气预报的风速预测修正方法[J].电力自动化设备,2017,37(10):15-20.
[10] Johnson R W.An Introduction to the Bootstrap[J].Teaching Statistics,2001,23(2):49-54.
[11] 杨宏,苑津莎,张铁峰.一种基于Beta分布的风电功率预测误差最小概率区间的模型和算法[J].中国电机工程学报,2015,35(9):2135-2142.
[12] 叶瑞丽,刘建楠,苗峰显,等.风电场风电功率预测误差分析及置信区间估计研究[J].陕西电力,2017,45(2):21-25.
[13] YE Ruili,LIU Jiannan,MIAO Fengxian,SONG Honglei,et al.Analysis of Wind Power Prediction Error and Its Confidence Interval Estimation in Wind Farms[J].SHAANXI ELECTRIC POWER,2017,45(2):21-25.