Avaliação de técnicas para ajuste de parâmetros de linhas de transmissão no cálculo de tensão induzida

Abstract

Lightning strikes are a frequently occurring phenomena that influence the dynamics of the electrical system, especially in overhead power transmission lines, promoting induced overvoltages, resulting from the effects of electromagnetic fields and with the ability to compromise the operation of the electrical network. Thus, the present work aims to evaluate different techniques for adjusting transmission line parameters in the calculation of the induced voltage, in order to quantify the effects inherent to the phenomenon. The main line parameters considered are: line and segment propagation function and characteristic impedance. Therefore, with a robust adjustment technique, it is possible to predict different overvoltage scenarios in numerous regions of lightning incidence, and consequently minimize the effects, by carrying out a study and adopting preventive measures, such as the adequacy of the dimensioning of the protection system. At first, a theoretical foundation was developed, based on the main procedures and processes involved in the study of the effects caused by an atmospheric discharge in an electrical network, such as the modeling of the return channel, the electromagnetic fields, and the coupling models, responsible for the integration between the interaction of the electromagnetic fields with the line. Subsequently, to obtain the responses resulting from the adjustment techniques, three of the most commonly found network structures (single-phase, two-phase and conventional) in the Brazilian electrical system were defined, whose induced voltage results are determined, in the time domain, for each conductor constituent, through the MATLAB software. Therefore, based on the behavior of the acquired response, it appears that most of the techniques provide an adequate adjustment of the line parameters, in the frequency domain, with a small mean square error. However, the time domain response, in certain study scenarios, results in a divergent curve in relation to the reference, which depends on the configuration and technique considered, since the set of poles and residuals provided by the fit can result in unstable models, making it unusable.