Óptima ubicación de generación distribuida en redes de distribución con el uso de redes neuronales artificiales

Abstract

In the present work, artificial neural network training is developed by means of the MATLAB simulink software, in order to find the optimal location of distributed reactive power generation units in the IEEE 34-bus distribution system. For the creation and training of artificial neural networks, the IEEE 33-bar test system will be used. Subsequently, to this system one or two generation units are added or added with values specified in the problem statement; we will have the variables of interest (voltages in the bars) which will be used as data for the training of the network. Once the ANN training has been carried out, the IEEE 34-bar distribution system is evaluated, allowing the identification of the optimal location of the UGD for the improvement of voltage profiles. Finally, some results of the proposed case studies are presented, such as the time it takes to train the ANN, the best validation, the effectiveness of the network training, among other parameters.