Localização de faltas em linhas transmissão HVDC utilizando a transformada wavelet e redes neurais

Abstract

The development and growth of Electric Power Systems (EPS), in particular the High Voltage Direct Current (HVDC), has increased interest from companies, manufacturers and researchers regarding the application and development of new tools in order to maintain the quality, safety and reliability in the planning, operation and protection of these systems. In this context, we highlight the importance of protection and especially, the point location of occurrence of a fault in the transmission lines of HVDC systems, due to its great extent and the environment that are exposed. Therefore, the immediate location of a fault point is extremely important, because it allows shorter time on line maintenance and repair, which increase the availability and continuous supply, ensuring economical and reliable operation of the system. Regarding this trouble, this paper aims to present a study for fault location in HVDC transmission lines, using voltage and current data recorded in a single system terminal. This is based on the application of the Wavelet Packet Transform (WPT) for the extraction of the energies of certain coefficients, which will be used as input to the Artificial Neural Network (ANN) via Multilayer Perceptron Network (MPN), in order to estimate the distance of occurrence transmission line fault.