Aplicação de redes neurais artificiais à análise de estabilidade de tensão no monitoramento de sistemas elétricos de potência

• Autor(a) principal: Fabris, João Vitor
• Data de Publicação: 2022
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Biblioteca Digital de Monografias da UFMT
• Texto Completo: http://bdm.ufmt.br/handle/1/3759
• topic: CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA; Estabilidade de tensão; Sistema de energia elétrica; Redes neurais artificiais; Voltage stability; Electric power system; Artificial neural networks

In recent years, the large increase in demand for active power, associated with restrictions on the electrical system, makes it operate in conditions close to the operational limit, which may result in the inability to deal with voltage disturbances, making it instructive to the point of causing partial or total Blackouts. In this context, several methods have been combined to assess the voltage safety level in an electrical system in order to prevent trouble-free operation. However, these methods traditionally require a high computational effort. In view of this, the present work aims to collaborate with the development of an Artificial Neural Network (ANN) to assist in the operation of an electrical system, indicating to the operator, voltage stability indexes of all the system's load bus, with the great advantage of using the low computational effort. In order for the ANN to have good forecasting capacity, data such as inputs of system parameters, a methodology was also proposed in this work for the creation of operational points with the application of contingencies and calculation of their respective voltage stability indices in each load bus. The voltage stability index used in this work is based on the Jacobian matrix resulting from the Newton-Raphson Power Flow method, the determinant of the D' matrix. The results appreciated the efficiency of the standard model in estimating indices having as input only the voltage module and angle of the bars, presenting a fast execution. The results show that the ANN is able to deal with predictions of voltage stability indices with the aid of measuring devices, providing preventive actions to the operator.