Estimação de fasores baseada na transformada wavelet para uso na proteção de distância de linhas de transmissão.

Abstract

A novel phasor estimation algorithm is proposed in this thesis, which uses full cycle orthogonal filters combined with a an adaptive digital mimic filter. The orthogonal filters are designed offline in advance using an innovative filter design technique, which takes into account the filtering characteristics of the maximum overlap discrete wavelet transform and the ones of the least error square algorithm. The proposed adaptive digital mimic filter is obtained using an adaptive strategy to compute the traditional digital mimic filter parameters, in order to eliminate the decaying DC component effect on phasor estimation. The time and frequency responses of the proposed algorithm are compared to the ones of the traditional least error square and the full cycle and half cycle Fourier algorithms, which are combined with the traditional digital mimic filter. In order to do that, actual fault data are used, as well as simulated fault data using an electromagnetic transients program and a power system real time digital simulator. The obtained results reveal that the proposed algorithm better rejects the decaying DC component in the majority of cases. In addition to that, the proposed algorithm rejects all harmonics, presents better transient responses and may be faster than the traditional full cycle algorithms in some cases. These characteristics and the real tim implementation of the proposed algorithm reveal the feasibility of its application in manufactured protective relays. As a byproduct, despite the adaptive digital mimic filter has been evaluated in conjunction with the designed orthogonal filters, another ones may be used too, such as the full cycle Fourier filters, in order to provide better decaying DC rejection in phasor estimation using these filters.