Determinação das condições operativas de microredes baseada em redes neurais

Abstract

The increased insertion of renewable generation, mainly wind and solar, brings new challenges for the planning and operation of electric power systems due to its dependence on climatic conditions. In this perspective, this dissertation aims to propose a methodology based on Artificial Neural Networks (ANN) for pattern recognition and a self-organizing map to assist in the planning and operation of the power system. In this context, the operational conditions that can lead the system to breach of voltage limits can be identified, enabling corrective control actions to be carried out. In addition, the proposed approach is able to identify the specific unit responsible for driving the system to an unsatisfactory operating condition. For this, the methodology is tested in a microgrid represented by the modified unbalanced three-phase IEEE 34-bus system, considering the use of wind and solar generation. The data set of satisfactory and unsatisfactory operating conditions are obtained using the Monte Carlo simulation. For this purpose, the backward-forward scan power flow is employed. These data are then provided to ANNs for training, validation and testing. The results obtained indicate a robust methodology capable of assisting in decision making and determining control actions during the operation of the system with high insertion of renewables, thus avoiding overvoltages.