Intelligent control and optimization of power system facilities using artificial neural networks

Abstract

As electric power systems modernize, the need to develop intelligent decision support methods for the technical refurbishment and reconstruction of electrical grid facilities is increasing. This study aims to develop an intelligent approach to selecting priority alternatives for modernizing power grid facilities based on artificial neural networks (ANN). This paper proposes a two-level decision-making structure: a tactical level for selecting preferred alternatives for individual power grid facilities, and a strategic level for setting modernization priorities at the regional power grid level. To evaluate alternatives, a system of specific criteria was developed, encompassing economic, technical, reliability, and socio-environmental indicators. A new criterion for unifying equipment nomenclature is proposed, enabling the classification of technical solutions and the clustering of alternatives using Kohonen self-organizing maps (SOM). A multilayer neural network trained with backpropagation was employed to evaluate alternative preferences. Experimental studies have shown that the best results are achieved with the Levenberg-Marquardt (LM) algorithm and a three-layer network architecture, yielding an F-score of up to 0.979 and a classification accuracy of 99.04%. The results confirm the effectiveness of neural network methods for intelligent decision support in modernizing the electric power infrastructure.