Privacy protection methods of power grid communication data aggregation based on federated learning

Abstract

In recent years, as more and more countries attach importance to the next generation grid, a large amount of human and financial resources have been invested in the research and development of smart grid. In order to carry out real-time monitoring and intelligent control of the power grid, the control center of the smart power grid needs to continuously collect users' electricity consumption data, which, if leaked, will reveal individual users' electricity consumption rules and living habits. How to monitor regional power consumption without revealing how much individual users use has become a challenge. Therefore, the power grid communication data aggregation privacy protection method is designed based on federated learning. Firstly, the power grid communication data privacy types and encryption tools are analyzed. Secondly, the power grid communication data aggregation privacy protection scheme is designed based on federated learning, so as to realize the power grid communication data aggregation privacy protection. The results show that the privacy protection method designed for power grid communication data aggregation is effective and has certain application value.