SolarNet: A Deep Convolutional Neural Network for Solar Forecasting via Sky Images

The exponential growth of solar energy poses challenges to power systems, mostly due to its uncertain and variable characteristics. Hence, solar forecasting, especially short-term solar forecasting (STSF), has been adopted to assist power system operations. The STSF takes inputs from various sources, among which sky image-based STSF is not yet well-studied. In this paper, a deep convolutional neural network (CNN) model, called the SolarNet, is developed to forecast the 10-min ahead GHI by only using sky images without numerical measurements and extra feature engineering. The SolarNet, composed of 21 convolutional, max-pooling, and fully-connected layers, learns latent patterns between sky images and global horizontal irradiance (GHI) in an end-to-end manner. Numerical results based on ten years data show that the developed SolarNet outperforms the benchmarking persistence of cloudiness model and machine learning models by up to 11.88% under various weather conditions.

[1]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[2]  Soteris A. Kalogirou,et al.  Machine learning methods for solar radiation forecasting: A review , 2017 .

[3]  Hai Wang,et al.  3D-CNN-based feature extraction of ground-based cloud images for direct normal irradiance prediction , 2019, Solar Energy.

[4]  T. Soubdhan,et al.  A benchmarking of machine learning techniques for solar radiation forecasting in an insular context , 2015 .

[5]  Xiaoxia Qi,et al.  A comparison of day-ahead photovoltaic power forecasting models based on deep learning neural network , 2019, Applied Energy.

[6]  Joshua S. Stein,et al.  PVLIB: Open source photovoltaic performance modeling functions for Matlab and Python , 2016, 2016 IEEE 43rd Photovoltaic Specialists Conference (PVSC).

[7]  Marc Pollefeys,et al.  KloudNet: Deep Learning for Sky Image Analysis and Irradiance Forecasting , 2018, GCPR.

[8]  Yugang Niu,et al.  Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM , 2018 .

[9]  Jie Zhang,et al.  Hourly-Similarity Based Solar Forecasting Using Multi-Model Machine Learning Blending , 2018, 2018 IEEE Power & Energy Society General Meeting (PESGM).

[10]  B. Hodge,et al.  The value of day-ahead solar power forecasting improvement , 2016 .

[11]  Jie Zhang,et al.  Short-term global horizontal irradiance forecasting based on sky imaging and pattern recognition , 2017, 2017 IEEE Power & Energy Society General Meeting.

[12]  Bri-Mathias Hodge,et al.  Unsupervised Clustering-Based Short-Term Solar Forecasting , 2019, IEEE Transactions on Sustainable Energy.

[13]  Dazhi Yang,et al.  OpenSolar: Promoting the openness and accessibility of diverse public solar datasets , 2019, Solar Energy.

[14]  Bri-Mathias Hodge,et al.  A suite of metrics for assessing the performance of solar power forecasting , 2015 .

[15]  Merlinde Kay,et al.  Calculating the financial value of a concentrated solar thermal plant operated using direct normal irradiance forecasts , 2016 .