Моделирање на паметни енергетски мрежи

Abstract

In this PhD thesis, several models are developed that cover different aspects of smart grids. One of these aspects involves the utilization of the large amount of available digital information for creating smart models for planning and forecasting based on the latest and new achievements in the field of machine learning. Specifically, a model based on deep belief network has been developed and used for short-term forecasting of the electricity consumption in Macedonia, as well as for short-term forecasting of the prices on a part of the day-ahead power exchanges in the region of Southeast Europe. It has been shown that the model yields superior results compared to the results obtained from traditional neural networks, and even more, the results for electricity consumption forecasting present better results than the forecasts obtained by the Macedonian Power Transmission System Operator (MEPSO). A second aspect of the smart grids that is analyzed in this PhD thesis is the development of optimization models for the integration of the growing number of distributed energy sources, with particular attention to renewable energy sources whose production cannot be controlled. For this purpose, an optimization model has been developed that includes models for correlation of metrological conditions with the production of electricity. The results of this model show the optimal ratio of production from these sources, which depends on their share in the total production. The third aspect of smart grids refers to the control and optimization of energy consumption. For that purpose, a model is being set up to examine the possibilities for energy savings used for heating in the Buildings sector. The model includes a non-stationary heat transfer model that is embedded in an optimization model and includes modeling of meteorological conditions (the same model as in the modeling of electricity production is used) and applied to a typical apartment with a location in Skopje, Macedonia. As a result, this model provides information on the appropriate required installed capacity, insulation, as well as the most appropriate heating technology that will allow an optimal ratio of these values and will result in maximum energy savings. Finally, as the last aspect analyzed in this PhD thesis is the electricity transmission network itself. For this purpose, a detailed model of the transmission networks of the Energy Community member countries, as well as of some of the systems of the neighboring countries, has been developed. The benefits of the results obtained from this model are twofold. Namely, on the one hand, the projects that are the most relevant for network expansion and integration of the region are obtained, on the basis of their influence on the reduction of losses and the reduction of energy not served, as well as on their potential for maximizing the net transfer capacity between the countries. On the other hand, as a methodological contribution, the developed model can be used for further analyzes of the impact of different electricity production and demand-side strategies on the electricity transmission network in the region.