SolarCrossFormer: Improving day-ahead Solar Irradiance Forecasting by Integrating Satellite Imagery and Ground Sensors

2025-09-19

AI Summary

To address the insufficient spatiotemporal resolution of existing day-ahead solar irradiance forecasting methods—limiting their applicability to large-scale photovoltaic grid integration—this paper proposes a cross-modal graph neural network that jointly leverages satellite remote sensing imagery and ground-based meteorological time series. The model implicitly encodes geographic coordinates as input, enabling probabilistic 15-minute-resolution forecasts up to 24 hours ahead across any location in Switzerland. It exhibits zero-shot generalization: it can seamlessly incorporate newly deployed sensors without retraining and impute irradiance values in unobserved regions. Evaluated on one year of ground-truth measurements from 127 stations, the model achieves a normalized mean absolute error of 6.1%, matching the accuracy of commercial numerical weather prediction systems while substantially improving forecast robustness and deployment flexibility.

Technology Category

• Computer Vision: Multi-modal Vision
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Graph-based Machine Learning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

Accurate day-ahead forecasts of solar irradiance are required for the large-scale integration of solar photovoltaic (PV) systems into the power grid. However, current forecasting solutions lack the temporal and spatial resolution required by system operators. In this paper, we introduce SolarCrossFormer, a novel deep learning model for day-ahead irradiance forecasting, that combines satellite images and time series from a ground-based network of meteorological stations. SolarCrossFormer uses novel graph neural networks to exploit the inter- and intra-modal correlations of the input data and improve the accuracy and resolution of the forecasts. It generates probabilistic forecasts for any location in Switzerland with a 15-minute resolution for horizons up to 24 hours ahead. One of the key advantages of SolarCrossFormer its robustness in real life operations. It can incorporate new time-series data without retraining the model and, additionally, it can produce forecasts for locations without input data by using only their coordinates. Experimental results over a dataset of one year and 127 locations across Switzerland show that SolarCrossFormer yield a normalized mean absolute error of 6.1 % over the forecasting horizon. The results are competitive with those achieved by a commercial numerical weather prediction service.

Problem

• Improving day-ahead solar irradiance forecasting accuracy
• Integrating satellite imagery with ground sensor data
• Enhancing spatiotemporal resolution for grid integration

Innovation

• Combines satellite images and ground sensor data
• Uses graph neural networks for data correlations
• Generates probabilistic forecasts without retraining model