應用於輸電線測距電驛之新型適應性濾波演算法設計

Abstract

由於測距電驛的數位濾波演算法必須利用基頻相量準確且快速地估測故障位置，故數位濾波演算法對於測距電驛而言就顯得格外重要，特別是當輸入電驛的電流訊號中含有直流衰減偏移成分時。 為了達成高速之測距保護，本論文提出了一個新型適應性濾波演算法，其利用故障偵測器及故障型態判別演算法來偵測故障之發生與判別為何種故障，再利用可變資料視窗之相量估測演算法來估測基頻相量及視在阻抗。此新型適應性濾波演算法不但可以濾除故障電流中的直流衰減偏移成分，且對於雜訊也有很好的抑制效果。 本論文亦用了MATLAB來模擬實際電力系統，並藉此驗證所提新型適應性濾波演算法之強健性，此外也將其與全週型離散傅立葉轉換加mimic濾波器做了比較，一些模擬的結果也會在論文中呈現。由模擬的結果可知，此新型適應性濾波演算法比全週型離散傅立葉轉換加mimic濾波器有更好的性能。

Digital filtering algorithms for distance relays are very critical, because they must save fundamental frequency phasors to be used for a precise and quick estimation of fault location, even with corruption from the decaying dc offset in the current signal. For the high-speed transmission line distance protection, we present a new adaptive filtering algorithm which uses the fault detector and the fault classification algorithm to determine the inception of a fault and classify the fault types and then uses the phasor estimation algorithm with data windows of appropriate length to estimate fundamental frequency phasors and seen impedances. This new adaptive filtering algorithm which does not only remove the decaying dc offset but also attenuate the effect of the noise in the current signal. The robustness of this new adaptive filtering algorithm is tested using a real model power system by MATLAB. The performance of this new adaptive filtering algorithm is compared with the full-cycle Discrete Fourier Transform+mimic filter. Some results are reported in this thesis. These results indicate that the new adaptive filtering algorithm provides better performance than the full-cycle Discrete Fourier Transform+mimic filter.