Multi-source Heterogeneous Data Processing and Optimization Algorithm in Intelligent Distribution Network Planning System

In the process of distribution network planning (DNP), massive heterogeneous data, such as power system operation data, enterprise operation management data, non-electric energy data, etc., bring challenges to data processing because of their high dimension, large volume and strong coupling. Aiming at these problems, this paper proposes a DNP method combining multi-source heterogeneous data (MHD) processing technology and optimization algorithm. This method fuses multi-source data by Kalman filter (KF) to improve data accuracy, and extracts key features by principal component analysis (PCA) to provide a simplified data set for subsequent optimization. In the aspect of optimization algorithm, the multi-objective optimization algorithm NSGA-II is used to solve the DNP problem, taking into account the objectives of economy and reliability, and satisfying the physical constraints such as power flow balance, voltage limitation and line capacity. The experimental results show that this method is superior to the traditional planning method in terms of total investment cost, operation cost, average power outage time and power supply reliability, which effectively verifies its effectiveness and superiority. The research in this paper provides new ideas and methods for intelligent DNP, which is helpful to improve the safety, stability and economic operation level of power system.