Application of hybrid intellectual systems in energy

Another impetus of development in the field of automation of designing and monitoring parameters of electric networks was given by achievements in the creation of artificial intelligence systems and, first of all, the theory of expert systems and knowledge bases, as well as artificial neural networks (ANNs). This is explained by the fact that expert systems and ANNs allow working with insufficiently formalized methods and models, which are the main part of the methods of structural and parametric synthesis of energy systems. Currently, one of the classes of expert systems is hybrid expert systems (HES), which allow not only combining different models for representing knowledge in knowledge bases, but also using several technologies for processing them, which makes this type of system flexible enough when setting up for a specific subject area. In the process of designing and operating urban electric networks, many problems arise related to solving poorly formalized tasks and storing a large amount of engineering data and knowledge. One of these tasks is the task of predicting electrical loads for a certain time period, due to technological and economic reasons. On the other hand, the development of the Industry 4 technology requires energy engineers to have the skills to use intelligent and knowledge-based systems. The paper presents an approach to solving the problem of predicting the electrical loads of a city electric grid based on hybrid intelligent systems. The composition of such systems includes both expert systems and artificial neural networks. The main directions of the application of the methodology of neural networks in the energy sector are considered. The functional and infological models of the automation system for designing power systems have been developed, containing a bank of engineering knowledge in this subject area and based on methods and algorithms of hybrid intelligent systems adapted to such poorly formalized processes as functional and parametric synthesis.