Method for Implementing Load Recognition Algorithm for Internet of Things Terminals Based on Distributed Storage

Due to the requirement of a significant amount of historical data, the current load recognition algorithms can only rely on the collection of real-time data from terminals and the utilization of cloud platforms for algorithmic computations. In light of this context, this paper first analyzes the existing hardware and software resources available in the terminals. Building upon this analysis, a load recognition algorithm based on distributed storage is proposed. The algorithm synergizes the power historical data from charging stations, communication base stations, motors, and other devices, employing hierarchical clustering and B-spline functions to decompose and recognize power curves of different load types. Finally, a simulation analysis of the load recognition algorithm is conducted on an actual power grid in Duyun, verifying the effectiveness of the proposed method.