AUTHOR=Dai Qiangsheng , Huo Xuesong , Hao Yuchen , Yu Ruiji 

TITLE=Spatio-temporal prediction for distributed PV generation system based on deep learning neural network model

JOURNAL=Frontiers in Energy Research

VOLUME=Volume 11 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2023.1204032

DOI=10.3389/fenrg.2023.1204032

ISSN=2296-598X

ABSTRACT=Over the past century, global electricity consumption has increased due to factors such as population growth and technological advancements, leading to a rise in the use of fossil fuels and consequent environmental degradation. To combat this, many countries are promoting the use of green energy sources, including solar power. However, due to the uncertainty of photovoltaic (PV) power generation systems, voltage fluctuations can occur in the power grid, resulting in losses. Therefore, in order to obtain more accurate PV power prediction. This paper proposes a spatio-temporal prediction method based on a deep learning neural network model specifically designed for Distributed PV power generation systems, it can take the spatial dimension into account, and can capture the strong spatial correlation of the Oahu Island site, improving the model prediction accuracy. Firstly, spatio-temporal correlation analysis is performed for 17 PV sites located on Oahu Island. Secondly, to confirm the effectiveness of the proposed hybrid model, we compare it with a single CNN or LSTM model trained on the same dataset. From the evaluation indexes such as loss map, regression map, RMSE, and MAE, the CNN-LSTM model that considers the strong correlation of spatio-temporal correlation among the 17 sites can fully explore the before-and-after features of the time series data and has higher prediction accuracy.