Optimal Reconfiguration of Distribution Network with Solar Power Plants

Abstract

The paper is concerned with the problem of determining the optimal reconfiguration of the power grid for each hour of the day. The optimization criterion is the value of the total power losses in the power grid. The considered network operates in parallel with a large power system and includes two solar power plants. Consumers in this network are represented by electricity load curve of three types. The technique for determining the optimal configuration is based on knowledge about the features of flow distribution in grids with renewable energy sources and the fundamental principles from the graph theory. Also, the method relies on the results of forecasting the generation of solar power plants connected to the considered power grid. Solar power plants’ forecasting is carried out by a decision tree model trained using machine learning methods. To train the predictive model, data on the generation of real solar power plants are used.