发表机构
University of Peradeniya; MARC, University of Peradeniya（佩拉德尼亚大学; 佩拉德尼亚大学MARC机构）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究针对多模态短期太阳辐照度预测，固定多模态预测管道，仅改变视觉主干进行基准测试。比较多种主干在不同数据集上的预测表现，给出了各主干的RMSE等结果，提供了可重复的编码器比较，未确立架构优势等。
AI 中文摘要
天空图像辐照度研究通常会比较预测系统，其中图像编码器、时间模型、融合模块、目标定义和训练方法都会一起改变。我们采用了一种更狭义的协议：多模态预测管道是固定的，只有视觉主干会变化。共享设置保持预处理、晴空指数归一化、天气历史编码、融合、回归头、损失函数、优化器调度、种子和时间顺序分割策略不变。我们比较了ConvNeXt、Swin Transformer、VMamba、Spatial Mamba和MambaVision主干在福尔松进行提前10分钟预测以及严格匹配的NREL分割。预测技能是根据晴空指数智能持续性来衡量的，仅时间行作为天气历史诊断报告，而不是主要排名标准。在福尔松严格分割上，所有评估的视觉主干运行都比智能持续性有所改进。在评估的单种子严格运行中，VMamba Small和Swin Base达到了匹配的福尔松RMSE值，分别为65.39 W/m²和65.50 W/m²；仅时间诊断达到69.51 W/m²。在313样本的NREL严格分割上，智能持续性在17.48 W/m²时仍然最强，而最低的视觉RMSE由Swin Tiny获得，为23.76 W/m²。这些结果提供了在一个固定的多模态操作点下可重复的编码器比较，而不是建立架构层面的优势、统计解决的排名或完全优化的预测性能。代码可在此处获取：此https URL
英文摘要
Sky-image irradiance studies often compare forecasting systems in which the image encoder, temporal model, fusion block, target definition, and training recipe all change together. We use a narrower protocol: the multimodal forecasting pipeline is fixed, and only the visual backbone is varied. The shared setup keeps preprocessing, clear-sky-index normalization, weather-history encoding, fusion, regression head, loss, optimizer schedule, seed, and chronological split policy unchanged. We compare ConvNeXt, Swin Transformer, VMamba, Spatial Mamba, and MambaVision backbones for 10min-ahead forecasting on Folsom and a strict matched NREL split. Forecast skill is measured against clear-sky-index smart persistence, and temporal-only rows are reported as weather-history diagnostics rather than as the main ranking criterion. On the Folsom strict split, all evaluated visual-backbone runs improve over smart persistence. In the evaluated single-seed strict runs, VMamba Small and Swin Base reach matched Folsom RMSE values of 65.39 W/m^2 and 65.50 W/m^2; the temporal-only diagnostic reaches 69.51 W/m^2. On the 313-sample NREL strict split, smart persistence remains strongest at 17.48 W/m^2, while the lowest visual RMSE is obtained by Swin Tiny at 23.76 W/m^2. These results provide a reproducible encoder comparison under one fixed multimodal operating point rather than establishing architecture-level dominance, statistically resolved ranking, or fully optimized forecasting performance. Code available here: https://github.com/Oshadha345/irradiance_benchmark
Comments6 pages, 3 figures, Moratuwa Engineering Research Conference 2026 (MERCon 2026)
