Energy and AI, Volume 26, 2026

GateFormer: Forecast-Space Decomposition for Intra-Hour Solar Irradiance Prediction from Sky Images

David Hamlyn, Sunny Chaudhary, Jonathon Hare, Tasmiat Rahman

Article 100917, arXiv version submitted 22 October 2026

Abstract

Sub-hourly dispatch, balancing, and storage control increasingly depend on forecasts that anticipate rapid variability in solar generation. Accurate sky-image-based irradiance forecasting requires a combination of visual cloud dynamics with recent irradiance state while adapting to different short-term irradiance-change regimes.

To address this, we introduce GateFormer, a GHI-conditioned dual-decoder ViT for intra-hour global horizontal irradiance (GHI) forecasting through forecast-space decomposition. GateFormer embeds observed GHI into the visual token stream, allowing interaction with sky-image patch features through self-attention. From this shared representation, a Trajectory Head predicts future GHI sequence, while a Delta Head predicts final-horizon change from the latest observed GHI.

The two final-horizon forecasts are adaptively weighted in prediction space, with magnitude- and direction-supervision targeting the remaining error. On the NREL-TSI benchmark, GateFormer attains 125.01 W/m2 RMSE at +30 min, improving over a state-of-the-art transformer baseline by 9.6% under paired-timestamp evaluation. Across NREL-TSI, Chilbolton, and SIRTA, GateFormer achieves the lowest final-horizon RMSE among the evaluated baselines, with forecast skill of 22.97-28.86% relative to smart persistence.

Component ablations show adaptive weighting outperforms both individual decoders and fixed averaging, distinguishing GateFormer from conventional forecast combination, while the supervised correction further reduces RMSE by 1.68 W/m2 and MAE by 4.86 W/m2. Ramp-regime diagnostics show larger gains during operationally important positive ramps, reducing RMSE by 26.27 W/m2.

These results identify forecast-space decomposition as an alternative forecasting formulation, with potential relevance to other multimodal short-horizon forecasting tasks and deployment using only sky images and measured GHI.

Author abstract from arXiv, by David Hamlyn, Sunny Chaudhary, Jonathon Hare, Tasmiat Rahman. Licensed under CC BY 4.0. Formatting adapted for this page.

PDF mirrored unchanged from arXiv under CC BY 4.0.

GateFormer research illustration
GateFormer, Energy and AI, 2026

Cite this paper

David Hamlyn, Sunny Chaudhary, Jonathon Hare, Tasmiat Rahman (2026). GateFormer: Forecast-Space Decomposition for Intra-Hour Solar Irradiance Prediction from Sky Images. Energy and AI, 26, 100917. https://doi.org/10.1016/j.egyai.2026.100917

View BibTeX
@article{hamlyn2026gateformer,
  title = {{GateFormer: Forecast-Space Decomposition for Intra-Hour Solar Irradiance Prediction from Sky Images}},
  author = {David Hamlyn and Sunny Chaudhary and Jonathon Hare and Tasmiat Rahman},
  journal = {Energy and AI},
  volume = {26},
  pages = {100917},
  year = {2026},
  doi = {10.1016/j.egyai.2026.100917},
  eprint = {2610.30942},
  archivePrefix = {arXiv},
  primaryClass = {eess.IV},
  url = {https://doi.org/10.1016/j.egyai.2026.100917}
}