Dynamic Reconstruction of Distribution Networks with Distributed Photovoltaics Based on Weather Classification

A weather-based reconstruction method for distribution networks with distributed photovoltaics is proposed. Firstly, the historical output data of the photovoltaic power station is divided into typical weather scenarios such as sunny, cloudy, overcast, and rainy using the K-means clustering method. Subsequently, a non-parametric kernel density estimation method is applied to estimate the photovoltaic output for each weather type. Finally, the distribution network reconfiguration model for the IEEE 33-node distribution network case study is solved using the particle swarm algorithm, and optimal reconfiguration solutions for different weather types are obtained.