| 引用本文: | 徐怡悦,吴在军,王洋,窦晓波,胡敏强.基于正交级数的光伏电源输出功率概率模型[J].电力自动化设备,2017,37(3): |
| XU Yiyue,WU Zaijun,WANG Yang,DOU Xiaobo,HU Minqiang.Probabilistic modeling based on orthogonal series for PV power output[J].Electric Power Automation Equipment,2017,37(3): |
|
| 摘要: |
| 目前光伏输出功率概率建模中，参数分析方法需预先假设参数分布，非参数分析方法中常用的核密度估计的带宽值选取方法不统一，在此情况下提出一种非参数方法——正交级数密度估计，对数据分布不附加任何假设，基于正交级数理论直接建立光伏电源输出功率的概率模型。利用江西南昌及浙江嘉兴两地的光伏电源输出功率历史实测数据进行仿真，结合拟合优度检验和误差分析，验证了所提模型的准确性和有效性。对于不同时段、不同地区光伏电源输出功率的随机特性，所提方法具有良好的适用性，在样本容量变化时模型保持稳定性。 |
| 关键词: 光伏功率 概率模型 正交级数密度估计 核密度估计 拟合优度检验 |
| DOI：10.16081/j.issn.1006-6047.2017.03.011 |
| 分类号: |
| 基金项目:国家自然科学基金资助项目(51677025)；教育部新世纪优秀人才支持计划资助项目(NCET-13-0129)；江苏省重点研发计划项目(BE2015012-1)；国家电网公司科技项目(SGT-YHT/14-JS-188) |
|
| Probabilistic modeling based on orthogonal series for PV power output |
|
XU Yiyue, WU Zaijun, WANG Yang, DOU Xiaobo, HU Minqiang
|
|
School of Electrical Engineering, Southeast University, Nanjing 210096, China
|
| Abstract: |
| Since the existing parameter analysis methods of probabilistic modeling for PV(PhotoVoltaic) power output require the presupposition of parameter distribution while the bandwidth selection of kernel density estimation, a commonly used method of non-parameter analysis, is not unified, another method of non-parameter analysis, orthogonal series density estimation, is proposed, based on which the probabilistic model of PV power output can be directly established without any presupposition of data distribution. Simulation is carried out based on the historical PV power output measurements of Nanchang and Jiaxing, and combined with fitting goodness test and error analysis, the accuracy and effectiveness of the proposed model are verified. The proposed method is adaptable to the randomicity of PV power output of different periods and different regions, and the established model keeps stable when the sample-size changes. |
| Key words: photovoltaic power probabilistic model orthogonal series density estimation kernel density estimation fitting goodness test |
