Spatio-Temporal Graph Neural Networks for Multi-Site PV Power Forecasting

• Published in: IEEE transactions on sustainable energy Vol. 13; no. 2; pp. 1210 - 1220
• Main Authors: Simeunovic, Jelena, Schubnel, Baptiste, Alet, Pierre-Jean, Carrillo, Rafael E.
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.04.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Convolution; Correlation; Data models; Electric power grids; Forecasting; Graph neural networks; graph signal processing; Information processing; Long short-term memory; machine learning; Mathematical models; Neural networks; Numerical prediction; Photovoltaic cells; Photovoltaic systems; Photovoltaics; Predictive models; Production; Production methods; Signal processing; Solar power; Solar power generation; Spatial discrimination; Spatial resolution; Temporal resolution; Weather forecasting; Weather stations
• ISSN: 1949-3029; 1949-3037
• DOI: 10.1109/TSTE.2021.3125200

Accurate forecasting of solar power generation with fine temporal and spatial resolution is vital for the operation of the power grid. However, state-of-the-art approaches that combine machinelearning with numerical weather predictions (NWP) have coarse resolution. In this paper, we take a graph signal processing perspective and model multi-site photovoltaic (PV) production time series as signals on a graph to capture their spatio-temporal dependencies and achieve higher spatial and temporal resolution forecasts. We present two novel graph neural network models for deterministic multi-site PV forecasting dubbed the graph-convolutional long short term memory (GCLSTM) and the graph-convolutional transformer (GCTrafo) models. These methods rely solely on production data and exploit the intuition that PV systems provide a dense network of virtual weather stations. The proposed methods were evaluated in two data sets for an entire year: 1) production data from 304 real PV systems, and 2) simulated production of 1000 PV systems, both distributed over Switzerland. The proposed models outperform state-of-the-art multi-site forecasting methods for prediction horizons of six hours ahead. Furthermore, the proposed models outperform state-of-the-art single-site methods with NWP as inputs on horizons up to four hours ahead.