基于细粒度特征的BOA-GBDT光伏出力预测PV Output Prediction Based on Gradient Boosting Decision Tree Model With Bayesian Optimization Algorithm and Fine-grained Features
谢从珍;王江储;谢心昊;刘智健;白剑锋;
摘要(Abstract):
光伏出力与天气模式密切相关,深度挖掘天气模式特征信息能有效提高光伏出力预测精度。由于信息粒度的粗细程度对光伏出力预测的精度有影响,使用粗粒度的原始特征或聚类特征的传统光伏出力预测方法在预测精度方面存在提升空间。针对以上问题,提出了一种基于细粒度特征的贝叶斯优化梯度提升树(Bayesianoptimizationalgorithm gradient boosting decision tree, BOA-GBDT)光伏出力预测方法,该方法首先对日间每条气象监测数据及光伏出力监测数据构建细粒度特征,包括瞬时天气模式特征及时窗趋势性特征,然后采用贝叶斯优化算法(Bayesianoptimization algorithm,BOA)对细粒度特征的种类进行约减,最后通过(gradient boosting decision tree,GBDT)模型拟合特征与光伏曲线的关系,建立BOA-GBDT光伏出力预测模型。对实际算例进行误差分析,结果表明相比传统支持向量机(support vector machine,SVM)方法,该方法构建的预测模型运行时间平均减少97.3%,均方根误差平均减少80.4%。验证了该方法的有效性。
关键词(KeyWords): 光伏出力预测;天气特征;特征工程;GBDT
基金项目(Foundation):
作者(Author): 谢从珍;王江储;谢心昊;刘智健;白剑锋;
Email:
DOI: 10.13335/j.1000-3673.pst.2019.0447
参考文献(References):
- [1]曾鸣.《电力发展“十三五”规划》解读[J].中国电力企业管理,2017(1):14-16.Zeng Ming.Interpretation of the 13th Five-Year Plan for power development[J]. China Academic Journal Electronic Publishing House,2017(1):14-16(in Chinese).
- [2]李建林,籍天明,孔令达,等.光伏发电数据挖掘中的跨度选取[J].电工技术学报,2015,30(14):450-456.Li Jianlin,Ji Tianming,Kong Lingda,et al.Span determining of photovoltaic generation data mining[J]. Transactions of China Electrotechnical Society,2015,30(14):450-456(in Chinese).
- [3]黎敏,林湘宁,张哲原,等.超短期光伏出力区间预测算法及其应用[J].电力系统自动化,2019,43(3):10-18.Li Min,Lin Xiangning,Zhang Zheyuan,et al.Interval prediction algorithm for ultra-short-term photovoltaic output and its application[J].Automation of Electric Power Systems,2019,43(3):10-18(in Chinese).
- [4]孙骁强,刘鑫,程松,等.光伏逆变器参与西北送端大电网快速频率响应能力实测分析[J].电网技术,2017,41(9):2792-2798.Sun Xiaoqiang,Liu Xin,Cheng Song,et al.Actual measurement and analysis of fast frequency response capability of PV-inverters in northwest power grid[J].Power System Technology,2017,41(9):2792-2798(in Chinese).
- [5]陈志宝,丁杰,周海,等.地基云图结合径向基函数人工神经网络的光伏功率超短期预测模型[J].中国电机工程学报,2015,35(3):561-567.Chen Zhibao,Ding Jie,Zhou Hai,et al.A model of very short-term photovoltaic power forecasting based on ground-based cloud images and RBF neural network[J].Proceedings of the CSEE,2015,35(3):561-567(in Chinese).
- [6]黎静华,左俊军,汪赛.大规模风电并网电力系统运行风险评估与分析[J].电网技术,2016,40(11):3503-3513.Li Jinghua,Zuo Junjun,Wang Sai.Analysis and assessment of operation risk for power system with large-scale wind power integration[J].Power System Technology,2016,40(11):3503-3513(in Chinese).
- [7]龚莺飞,鲁宗相,乔颖,等.光伏功率预测技术[J].电力系统自动化,2016,40(4):140-151.Gong Yingfei,Lu Zongxiang,Qiao Ying,et al.An overview of photovoltaic energy system output forecasting technology[J].Automation of Electric Power Systems,2016,40(4):140-151(in Chinese).
- [8]曾元静.大型光伏电站发电功率预测及其接入对电网影响的分析[D].南京:东南大学,2016.
- [9]王育飞,付玉超,孙路,等.基于混沌-RBF神经网络的光伏发电功率超短期预测模型[J].电网技术,2018,42(4):1110-1116.Wang Yufei,Fu Yuchao,Sun Lu,et al.Ultra-short term prediction model of photovoltaic output power based on chaos-RBF neural network[J].Power System Technology,2018,42(4):1110-1116(in Chinese).
- [10]管霖,赵琦,周保荣,等.基于多尺度聚类分析的光伏功率特性建模及预测应用[J].电力系统自动化,2018,42(15):24-30.Guan Lin,Zhao Qi,Zhou Baorong,et al.Multi-scale clustering analysis based modeling of photovoltaic power characteristics and its application in prediction[J].Automation of Electric Power Systems,2018,42(15):24-30(in Chinese).
- [11]黎静华,黄乾,韦善阳,等.基于S-BGD和梯度累积策略的改进深度学习方法及其在光伏出力预测中的应用[J].电网技术,2017,41(10):3292-3300.Li Jinghua,Huang Qian,Wei Shanyang,et al.Improved deep learning algorithm based on S-BGD and gradient pile strategy and its application in PV power forecasting[J].Power System Technology,2017,41(10):3292-3300(in Chinese).
- [12]葛乐,陆文伟,袁晓冬,等.基于改进相似日和ABC-SVM的光伏电站功率预测[J].太阳能学报,2018,39(3):775-782.Ge Le,Lu Wenwei,Yuan Xiaodong,et al.Power forecasting of photovoltaic plant based on improved similar day and ABC-SVM[J].Acta Energiae Solaris Sinica,2018,39(3):775-782(in Chinese).
- [13]陈中,宗鹏鹏.基于样本扩张灰色关联分析的光伏出力预测[J].太阳能学报,2017,38(11):2909-2915.Chen Zhong,Zong Pengpeng.PV output forecast based on grey correlation analysis with expanded sample[J].Acta Energiae Solaris Sinica,2017,38(11):2909-2915(in Chinese).
- [14]李因果,戴翼,何晓群.基于自适应权重的面板数据聚类方法[J].系统工程理论与实践,2013,33(2):388-395.Li Yinguo,Dai Yi,He Xiaoqun.Clustering method for panel data based on adaptive weighting[J].Systems Engineering:Theory&Practice,2013,33(2):388-395(in Chinese).
- [15] Friedman J H.Greedy function approximation:a gradient boosting machine[J].The Annals of Statistics,2001,29(5):1189-1232.
- [16] Rasmussen C E,Williams C K I.Gaussian processes for machine learning[M].Cambridge MA:MIT Press,2005.
- [17] Bergstra J,Bardenet R,Bengio Y,et al.Algorithms for hyperparameter optimization[C]//International Conference on Neural Information Processing Systems,Granda,Spain,2011:2546-2554.
- [18] Desert Knowledge Australia Centre.Download data.alice springs[EB/OL].[2018-12-17]. http://dkasolarcentrecom.au/historical-data/download.
- [19]潘超,谭启德,蔡国伟,等.基于递归量化分析的COA-SVR短期风速混合预测模型[J].电网技术,2018,42(8):2373-2381.Pan Chao,Tan Qide,Cai Guowei,et al.Hybrid short-term wind speed prediction model by COA-SVR based on recursive quantitative analysis[J].Power System Technology,2018,42(8):2373-2381(in Chinese).
- [20]王华勇,杨超,唐华.基于LightGBM改进的GBDT短期负荷预测研究[J].自动化仪表,2018,39(9):76-78+82.Wang Huayong,Yang Chao,Tang Hua.Research on the short-term load forecasting using improved GBDT based on light GBM[J].Process Automation Instrumentation,2018,39(9):76-78+82(in Chinese).