Automating the assessment of the power grid state in remote areas of Russia using smart structures

The paper discusses a method for automating the assessment of the power grid state in remote regions of Russia using smart structures. The proposed automation method is implemented as a mobile application. The smart structure underlying the described method of automating the assessment of the power grid state consists of modules for receiving and processing data from sensors, searching for patterns in the power grid characteristics and generating state classifiers, offering recommendations for repair and optimal operation of the power grid and substation. The scientific novelty of the proposed solution is in the method of analyzing and processing the power grid characteristics and their combinations. In addition, the external influence parameters in the form of natural and manmade factors are taken into account. The method of analyzing and processing information about the power grid and substation is based on the machine learning method – logical data analysis. Assessing the state of a power grid and a substation is important when studying and solving the problems of predicting changes in the power grid state, selecting recommendations and making decisions on repair and maintenance work. The method for assessing the power grid state is based on the search for patterns and the construc-tion of classifiers. It allows taking into account all the characteristics and parameters of a power grid, their totality and the relationship between them. In addition, the described method allows analyzing and obtaining patterns for incomplete and inaccurate data, which is a fairly common occurrence in real power networks. The method can be used in the design and maintenance of power grids and substations in hard-to-reach and remote regions of the Russian Federation. The proposed reduction of the characteristic regularities and their sets based on their recurrent conjunction makes it possible to obtain optimal classifiers of the states of a power grid and a substation with high interpretability and generalization. It increases the accuracy of assessing the power grid state, therefore, increases the accuracy of predicting behavior, recommendations and making decisions about repair work and the optimal mode operation.