Predição das Componentes de Clarke via Interpolação Polinomial de 1º Grau com Base Newtoniana Aplicada na Estimação Digital de Frequência

Abstract

The electrical frequency is a parameter of great importance for the full functioning of the Electric Power Systems (EPS), influencing the operation of the equipment and the power quality supplied. This work presents an innovative method for estimating digital frequency in EPS. The estimation technique is based on the analysis of the network voltage waveforms, which are decomposed into their components a and ß through the Clarke Transform. The future values of the a and ß components are predicted through 1st degree polynomial interpolation with Newton’s base. From these values, the frequency of the network is then estimated as a function of the angle resulting from the product between the signal of the Clarke complex and that given by the prediction of the components a and ß. The proposed method was tested for voltage signals with step, ramp, exponential and damped sinusoidal frequency variations. Cases were also tested for signals with amplitude variation. The methodology was evaluated in terms of convergence time and minimum and maximum errors before and after convergence, showing that the proposed technique has great precision and robustness for the simulated situations.