Integrating intelligence and security: A review of large language models in smart grids with renewable energy

The growing complexity and data intensity of modern renewable power systems demand advanced artificial intelligence solutions capable of supporting intelligent, adaptive, and resilient grid operations. Large language models based on transformers can deliver substantial support in multiple areas of renewable power systems. At the same time, their deployment introduces new challenges and threats that must be addressed proactively. This paper comprehensively reviews large language models applications in renewable power systems and identifies major challenge dimensions. Based on the integrated consideration of deployment applications and security defence, we propose a defence framework with endogenous security to promote the development of smarter and safer smart grids.