TY  - EJOU
AU  - Gao, Honglian 
AU  - Zhang, Qingsong 
AU  - Chen, Zeming 
AU  - Li, anchen 
AU  - Liu, Quanhui 
AU  - Tian, Yuxiang 
AU  - Xu, Xianfeng 

TI  - Fault Reconfiguration Technology for Distribution Networks Considering Distributed Energy Output Forecasting
T2  - Energy Engineering

PY  - 
VL  - 
IS  - 
SN  - 1546-0118

AB  - With the increasing penetration rate of distributed generators (DGs) in the distribution network, DGs with independent power supply capability provide strong support for distribution network fault recovery. Traditional fault reconfiguration methods often rely on static load priorities and fail to fully consider the time-varying dynamic characteristics of outage costs, resulting in shortcomings in the economy and adaptability of restoration strategies. To address this, this paper proposes a fault reconfiguration method that integrates day-ahead prediction and a dynamic load restoration set. Firstly, a Long Short-Term Memory network optimized by Variational Mode Decomposition and the Marine Predators Algorithm is used for the day-ahead prediction of DG output, providing a data foundation for rapid response upon fault occurrence. After a fault occurs, a time-varying dynamic load priority restoration set is constructed based on load outage cost analysis, comprehensively considering the load importance level and the decay characteristics of outage costs, forming an economy-oriented restoration objective. Furthermore, the islanding process and the main network reconfiguration process are unified into a single model, constructing a Mixed-Integer Second-Order Cone Programming model. This model is validated using the CPLEX solver on the IEEE 33-node and 69-node test systems. Simulation results show that, compared to traditional static reconfiguration methods, the proposed strategy not only ensures the continuous power supply to important loads but also significantly reduces user outage costs, achieving the synergistic optimization of power supply reliability and economy.
KW  - Distribution network; intelligent optimization algorithms; LSTM neural network; islanding formation; main grid reconfiguration; second-order cone programming (SOCP); distributed generation (DG)

DO  - 10.32604/ee.2026.074052