SPIRIT: Short-term Prediction of solar IRradIance for zero-shot Transfer learning using Foundation Models

2025-02-14

AI Summary

Newly commissioned photovoltaic (PV) power plants often lack multi-year historical solar irradiance data, rendering conventional short-term solar forecasting methods ineffective. Method: This paper proposes SPIRIT—the first zero-shot transfer learning framework specifically designed for solar irradiance forecasting. Leveraging foundation models and cross-site temporal representation learning, SPIRIT achieves high-accuracy short-term forecasting without requiring any historical data from the target site. Contribution/Results: Its core innovation lies in establishing a zero-shot transfer paradigm that enables plug-and-play deployment and progressive fine-tuning as local data accumulates. Extensive experiments demonstrate that, under zero-shot settings, SPIRIT reduces mean forecasting error by 70% compared to state-of-the-art methods (p < 0.01). This substantial improvement significantly enhances prediction reliability for nascent PV sites, thereby providing a scalable forecasting infrastructure for photovoltaic–energy-storage coordinated control and flexible grid dispatch.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Multi-modal Vision
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Traditional solar forecasting models are based on several years of site-specific historical irradiance data, often spanning five or more years, which are unavailable for newer photovoltaic farms. As renewable energy is highly intermittent, building accurate solar irradiance forecasting systems is essential for efficient grid management and enabling the ongoing proliferation of solar energy, which is crucial to achieve the United Nations' net zero goals. In this work, we propose SPIRIT, a novel approach leveraging foundation models for solar irradiance forecasting, making it applicable to newer solar installations. Our approach outperforms state-of-the-art models in zero-shot transfer learning by about 70%, enabling effective performance at new locations without relying on any historical data. Further improvements in performance are achieved through fine-tuning, as more location-specific data becomes available. These findings are supported by statistical significance, further validating our approach. SPIRIT represents a pivotal step towards rapid, scalable, and adaptable solar forecasting solutions, advancing the integration of renewable energy into global power systems.

Problem

• Predict solar irradiance without historical data
• Enhance renewable energy grid management
• Enable zero-shot transfer learning for solar forecasting

Innovation

• Foundation models for solar forecasting
• Zero-shot transfer learning approach
• Fine-tuning with location-specific data