Selection of Numerical Weather Forecast Features for PV Power Predictions with Random Forests

被引：8
|
作者：
Wolff, Bjoern [1 ]
Kramer, Oliver [2 ]
Heinemann, Detlev [1 ]
机构：
[1] Carl von Ossietzky Univ Oldenburg, Inst Phys, Energy Meteorol, Oldenburg, Germany
[2] Carl von Ossietzky Univ Oldenburg, Dept Comp Sci, Computat Intelligence, Oldenburg, Germany
来源：
DATA ANALYTICS FOR RENEWABLE ENERGY INTEGRATION (DARE 2016) | 2017年 / 10097卷
关键词：
PV power forecasting; Random Forest; Feature importance; Support Vector Regression; SUPPORT VECTOR REGRESSION;
D O I：
10.1007/978-3-319-50947-1_8
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
The increasing volatility introduced to power grids by renewable energy sources makes it necessary that the accuracy of energy forecasts are improved. Photovoltaic (PV) power plants hold the biggest share of installed capacity of renewable energy in Germany, so that high quality PV power forecasts are vital for a cost efficient operation of the underlying electrical grid. In this paper, we evaluate multiple Numerical Weather Prediction (NWP) parameters for their ability to improve PV power forecasting features. The importance of features is decided by a Random Forest algorithm. Furthermore, the resulting top ranked features are tested by performing PV power forecasts with Support Vector Regression, Random Forest, and linear regression models.
引用
收藏
页码：78 / 91
页数：14
相关论文
共 36 条
  • [31] Data-driven upscaling methods for regional photovoltaic power estimation and forecast using satellite and numerical weather prediction data
    Pierro, Marco
    De Felice, Matteo
    Maggioni, Enrico
    Moser, David
    Perotto, Alessandro
    Spade, Francesco
    Cornaro, Cristina
    SOLAR ENERGY, 2017, 158 : 1026 - 1038
  • [32] Mathematical models to forecast temporal variations of power law shape parameters of a PV module working in real weather conditions: Prediction of maximum power and comparison with single-diode model
    Zaimi, Mhammed
    El Ainaoui, Khadija
    Assaid, El Mahdi
    SOLAR ENERGY, 2023, 266
  • [33] Comparing support vector regression for PV power forecasting to a physical modeling approach using measurement, numerical weather prediction, and cloud motion data
    Wolff, Bjoern
    Kuehnert, Jan
    Lorenz, Elke
    Kramer, Oliver
    Heinemann, Detlev
    SOLAR ENERGY, 2016, 135 : 197 - 208
  • [34] Short-term wind power prediction framework using numerical weather predictions and residual convolutional long short-term memory attention network
    Xie, Chenlei
    Yang, Xuelei
    Chen, Tao
    Fang, Qiansheng
    Wang, Jie
    Shen, Yan
    ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE, 2024, 133
  • [35] A numerical weather prediction feature selection approach based on minimal-redundancy-maximal-relevance strategy for short-term regional wind power prediction
    Zhao, Yongning
    Ye, Lin
    Zhongguo Dianji Gongcheng Xuebao/Proceedings of the Chinese Society of Electrical Engineering, 2015, 35 (23): : 5985 - 5994
  • [36] Design and Development of a Short-Term Photovoltaic Power Output Forecasting Method Based on Random Forest, Deep Neural Network and LSTM Using Readily Available Weather Features
    Rangelov, Denis
    Boerger, Michell
    Tcholtchev, Nikolay
    Laemmel, Philipp
    Hauswirth, Manfred
    IEEE ACCESS, 2023, 11 : 41578 - 41595
1234