Method of Automation of Power Supply Restoration Processes in Distribution Networks Using Neural Network Models and Graph

The issues of improving the reliability of power supply by automating the processes of power restoration in distribution networks are being considered. A combined approach combining graph analysis algorithms and artificial neural networks is proposed. The architecture of the system implementing this approach and the results of its tests on modelled network structures are described. The efficiency of the methodology is evaluated and a comparative characterisation of existing solutions is given. The results obtained confirm the prospects of integration of intelligent algorithms into power supply control systems