Exploring the Capabilities and Limitations of Recommender System Models in the Electric Power Sector

As the modern bulk power system (BPS) expands and electricity users increase, an ever-growing amount of power data needs to be processed. The recommender system (RS), as an artificial intelligence (AI) algorithm, leverage big data to suggest additional products to consumers, making them naturally suited for BPS analysis. This paper summarizes published works on RS applications in the electric power sector, explores the goals and objectives of each proposed use case, and analyzes the advantages and limitations of their capabilities. Furthermore, this paper evaluates the performance of three large language models (LLMs) on power system domain challenges to compare the differences between the models. Finally, the paper summarizes the current applications of RS models in the electric power sector and explores prospects for future development.