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

📄 arXiv: 2610.30942v1 📥 PDF

作者: David Hamlyn, Sunny Chaudhary, Jonathon Hare, Tasmiat Rahman

分类: eess.IV, cs.AI, cs.CV, cs.LG

发布日期: 2026-10-22

期刊: Energy and AI 26 (2026), 100917

DOI: 10.1016/j.egyai.2026.100917


💡 一句话要点

提出GateFormer以解决基于天空图像的太阳辐射预测问题

🎯 匹配领域: 支柱二:RL算法与架构 (RL & Architecture) 支柱九:具身大模型 (Embodied Foundation Models)

关键词: 太阳辐射预测 天空图像 深度学习 自注意力机制 多模态预测 ViT 预测空间分解 智能电网

📋 核心要点

  1. 现有方法在快速变化的太阳辐射预测中难以有效结合视觉云动态与近期辐射状态,导致预测精度不足。
  2. GateFormer通过将观察到的GHI嵌入视觉令牌流,利用自注意力机制实现对天空图像特征的有效交互,提出了一种新的预测空间分解方法。
  3. 在多个基准测试中,GateFormer的最终预测RMSE最低,相较于基线模型提升了22.97-28.86%的预测技能,表现出色。

📝 摘要(中文)

随着对快速变化的太阳能发电的调度、平衡和存储控制的需求增加,准确的基于天空图像的辐射预测变得至关重要。本文提出GateFormer,一个基于GHI条件的双解码器ViT,通过预测空间分解实现对短时间内全球水平辐射(GHI)的预测。GateFormer将观察到的GHI嵌入视觉令牌流中,利用自注意力机制与天空图像特征进行交互。通过这种共享表示,Trajectory Head预测未来GHI序列,而Delta Head则预测从最新观察到的GHI的最终变化。实验结果表明,GateFormer在NREL-TSI基准测试中实现了125.01 W/m²的RMSE,相较于最先进的变换器基线提高了9.6%。

🔬 方法详解

问题定义:本文旨在解决基于天空图像的全球水平辐射(GHI)预测中的快速变化问题。现有方法未能有效结合视觉云动态与近期辐射状态,导致预测精度不足。

核心思路:GateFormer通过将观察到的GHI嵌入视觉令牌流,利用自注意力机制与天空图像特征进行交互,提出了一种新的预测空间分解方法,能够适应不同的短期辐射变化模式。

技术框架:GateFormer的整体架构包括两个主要模块:Trajectory Head用于预测未来GHI序列,Delta Head用于预测从最新观察到的GHI的最终变化。这两个模块的输出在预测空间中进行自适应加权。

关键创新:GateFormer的核心创新在于其预测空间分解方法,通过自适应加权输出,超越了传统的预测组合方式,显著提高了预测精度。

关键设计:在设计中,GateFormer采用了自注意力机制来增强视觉特征的交互,使用了针对剩余误差的幅度和方向监督,进一步降低了RMSE和MAE。

📊 实验亮点

在NREL-TSI基准测试中,GateFormer实现了125.01 W/m²的RMSE,相较于最先进的变换器基线提高了9.6%。在多个基准测试中,GateFormer的最终预测RMSE最低,预测技能提升幅度达到22.97-28.86%。

🎯 应用场景

该研究的潜在应用领域包括太阳能发电的调度与管理、智能电网的负荷预测以及气象预报等。通过提高短期辐射预测的准确性,GateFormer能够帮助相关行业更好地应对快速变化的能源需求,优化资源配置,提升可再生能源的利用效率。

📄 摘要(原文)

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/m^2 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/m^2 and MAE by 4.86 W/m^2. Ramp-regime diagnostics show larger gains during operationally important positive ramps, reducing RMSE by 26.27 W/m^2. 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.