Research on Privacy Protection Technology of Smart Grid Based on Federated Learning

With the rapid development of smart grid technology, data privacy protection and sharing has become an urgent problem to be solved. This paper discusses the privacy protection technology of smart grid based on federated learning, and reviews the latest research progress in the field of privacy protection. This paper deeply analyzes the root cause of privacy threats, and proposes a differential privacy method to adaptively adjust the privacy budget to ensure the security of data transmission and storage based on the characteristics of smart grid terminal data interaction. Experimental results show that reasonable allocation of privacy budget and increasing the value of Dirichlet distribution parameters can effectively improve the robustness of privacy-preserving model training in federated learning for a single smart meter data.