Foundation-Style Pretraining for Scalable and Transferable Distributed PV Forecasting
作者: Dayin Chen, Xiaodan Shi, Xin Xu, Zhongyue Su, Shibo Zhu, Ruizhe Deng, Dongxiao Zhang, Yuntian Chen, Jinyue Yan
分类: cs.LG
发布日期: 2026-10-20
💡 一句话要点
提出基础风格预训练框架以解决分布式光伏预测的可扩展性问题
🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)
关键词: 分布式光伏 预测模型 基础风格预训练 多站点学习 气象预测 序列补全 可扩展性 零-shot学习
📋 核心要点
- 现有的光伏预测方法通常针对单个站点,缺乏可扩展性,难以适应新站点的集成与重用。
- 本文提出了一种基础风格的预训练框架,能够从多站点和多时间段的数据中学习共享的预测模型。
- 在479个分布式光伏站点的实验中,提出的模型在多个预测时间段上均优于现有的基线模型,展示了显著的性能提升。
📝 摘要(中文)
随着分布式光伏发电的快速增长,对地理分散的多个站点的准确预测需求日益增加。现有方法通常针对单个站点或特定预测时间段开发,导致在新站点集成时需要重复训练和维护,限制了在未见过站点的直接重用。为了解决这一可扩展性问题,本文提出了一种领域特定的基础风格预训练框架,从异构的多站点和多时间段光伏数据中学习共享预测模型。通过将气象预测辅助的预测形式化为序列补全问题,统一了历史光伏功率、气象观测、未来数值天气预测(NWP)和时间编码。设计并系统评估了三种预训练目标,其中混合目标在预测准确性、降级的NWP条件和下游适应性方面表现出最佳的平衡性能。
🔬 方法详解
问题定义:本文旨在解决现有光伏预测方法在新站点集成时的可扩展性问题。现有方法通常针对单个站点或特定时间段,导致重复训练和维护,限制了模型的重用性。
核心思路:提出的基础风格预训练框架通过学习来自多个站点和时间段的异构数据,构建一个共享的预测模型,以应对不同站点的预测需求。
技术框架:该框架将气象预测辅助的预测问题形式化为序列补全问题,整合历史光伏功率、气象观测、未来NWP预测和时间编码。设计了三种预训练目标:预测对齐重建、通用重建和混合重建。
关键创新:最重要的创新在于混合重建目标的提出,它结合了光伏预测的操作结构与随机序列重建,提供了在多种条件下的最佳性能平衡。
关键设计:在模型设计中,采用了多站点数据进行预训练,并在损失函数中引入了不同的重建目标,以优化模型在不同预测条件下的适应性和准确性。实验中使用了340个站点进行模型开发,139个站点进行零-shot评估。
🖼️ 关键图片
📊 实验亮点
在479个分布式光伏站点的实验中,提出的多站点预训练模型在6小时到7天的预测范围内,显著优于代表性的站点特定和共享基线模型,以及最近的大规模预训练基础模型,展示了在预测准确性上的显著提升。
🎯 应用场景
该研究的潜在应用领域包括大规模分布式光伏电站的能量管理和调度。通过提供可扩展的预测模型,能够在新站点快速部署并根据当地数据进行适应,提升光伏发电的经济性和可靠性,推动可再生能源的广泛应用。
📄 摘要(原文)
The rapid growth of distributed photovoltaic (PV) generation is increasing the demand for accurate forecasting across large numbers of geographically dispersed stations. Existing approaches are commonly developed for individual sites or specific forecasting horizons, requiring repeated training and maintenance as new stations are integrated and limiting direct reuse at previously unseen sites. To address this scalability gap, this paper presents a domain-specific foundation-style pretraining framework that learns a shared forecasting model from heterogeneous multi-site and multi-horizon PV data. NWP-assisted forecasting is formulated as a sequence-completion problem that unifies historical PV power, meteorological observations, future numerical weather prediction (NWP) forecasts, and time encodings. Three pretraining objectives are designed and systematically evaluated, including forecast-aligned reconstruction, generic reconstruction, and mixed reconstruction. The mixed objective combines the operational structure of PV forecasting with random sequence reconstruction and provides the most balanced performance across forecasting accuracy, degraded NWP conditions, and downstream adaptation. Experiments on 479 distributed PV stations use 340 stations for model development and 139 spatially held-out stations for zero-shot evaluation. Across horizons from 6 hours to 7 days, the multi-site pretrained MR model outperforms representative station-specific and shared baselines, as well as generalist time-series models, including a recent large-scale pretrained foundation model. The pretrained model can also be directly deployed at new PV stations and subsequently adapted to region-specific operational objectives when local data become available, providing a scalable forecasting paradigm for large distributed PV fleets.