Short-Term Forecasting-Based Network Reconfiguration for Unbalanced Distribution Systems With Distributed Generators

摘要

This article proposes a network reconfiguration methodology using repository-based constrained nondominated sorting genetic algorithm with preference order ranking for an unbalanced distribution system. The algorithm can accommodate the variable nature of load demand and distributed generator output. A mathematical multiobjective model is formulated to obtain the optimal topology for a whole day considering minimization of daily energy loss, energy not supplied, and cumulative current unbalance factor under the constraint of minimum switching action. A wavelet transform-based ARIMA model is used for wind speed forecasting and is proposed for solar irradiance and load forecasting as well. Hourly network reconfiguration (NR) is also performed, and a comparison is performed between hourly and whole-day NR. The proposed approach has been implemented on IEEE 34-bus and IEEE 123-bus systems to evaluate the effectiveness of the developed methodologies.