Research on Distribution Network Load Classification and Identification Technology Based on Edge Computing

In response to the needs of distribution network load classification and identification, this study proposes an efficient classification method based on edge computing, aiming to solve the problem of high real-time and accuracy requirements in the power system. Through in-depth analysis of the application potential of edge computing in power systems, a lightweight neural network model was designed and optimized through the momentum gradient descent method, which significantly improved the model's convergence speed and computational efficiency. In the experimental verification, through field testing in multiple scenarios, the results showed that the method performed well in terms of classification accuracy, latency and resource utilization, especially in the classification of residential, commercial and industrial loads, showing high accuracy and Lower processing latency. In addition, through in-depth analysis of actual application cases, this study further proves the wide applicability and application value of edge computing in power load classification. The results of this research provide a new technical path for the optimization and management of power systems, and lay a theoretical and practical foundation for the development of future smart grids.