Digital twin for smart electricity distribution networks

Abstract

This paper presents the results of composing a digital model of critical states of the power distribution network with intelligent control. A specific example of a power network with thirty transformer substations connected by main and reserve cables is discussed. The analysis of critical events occurring in the power system was carried out, and a set of decision rules was formed to counter these events. A practical need was noted for regular updating and expansion of the list of critical events and corresponding parrying rules. It is shown that it is expedient to use faceted active neural networks with a structure similar to that of the controlled object as a digital twin model of intelligent control of the system under consideration. The use of such neural networks based on the rule “one event - one ensemble of neurons” provides the digital duplicate model with the ability to unlimited cumulative expansion, similar to the evolutionary property of natural biological systems.