電力訊號頻率偵測演算法研究與應用

Abstract

頻率偵測在對於電力訊號系統上來說，一直都是一重要標的，經由觀測頻率訊號，可使吾人得知有關負載與系統電源間的動態關係，對於越來越多的分散式電源應用在電力系統上，在分散式電源與市電系統併聯運轉時的訊號頻率偵測就顯得愈發重要，因為唯有掌握到確切的頻率訊號，才能夠使得在進行同步併聯操作時可能會產生的一些危害降到最低。另外，在頻率訊號上，吾人亦須特別關注有關諧波汙染方面的問題，目前因為諧波汙染所造成的非自然傷害時有耳聞，因此本文希望能在對頻率訊號做分析時，能最迅速得到最正確的相關參數，並且能利用濾波方法來將諧波及雜訊等訊號加以濾除以應用於分散式電源系統之併聯上。 本文首先介紹常見的分散式電源系統，並針對與分散式系統息息相關的孤島偵測技術，探討其優缺點，以及孤島偵測失效之原因。對於孤島發生後再次與市電系統併聯時的頻率偵測，本文將引用一以離散傅立葉轉換法為基礎的新型離散傅立葉轉換法，此演算法不僅保有原本離散傅立葉轉換法對於頻率偵測方面計算快速且能消除直流偏壓所帶來的影響以外，更因結合了FIR (Finite Impulse response)濾波器的使用而改善離散傅立葉轉換法的計算誤差使濾除諧波及雜訊頻率偵測有更高的準確率，如此可使同步併聯操作時更為精準以減少相關傷害。

The frequency estimation is an important issue for power signal system. The dynamic relationship between the load and the generation can be detected by measuring the power system frequency. Because more and more distributed power source units are used in generation, it is more important for us to detect the frequency signals when distributed power source units are connected with synchronous utility system. In order to minimize the damage caused by the operation of synchronizing connection, we must have the exact frequency signals. Furthermore, we must pay attention to the harmonic pollution in frequency signals because the ruin of harmonic pollution is heard usually. We hope to get the correct parameters fast from the frequency signals analysis, and to use the filtering methods for the harmonics and noise from signal, for the operation of synchronizing connection of distributed power source units. Firstly we will introduce the common distributed pwer source units and probe into the advantage and disadvantage of islanding detection that is closely related to distributed power source units and the reason why the detection will fail. After islanding operation, we have to estimate the frequency on both sides of utility system and island system for the reconnection of both systems. So we will present the Extended Discrete Fourier Transform which is based on Discrete Fourier Transform (DFT) and combines FIR (Finite Impulse Response) filter then the algorithm not only can keep advantage of fast computation of frequency detection and eliminate the DC offset but also can enhance the resistance of the computation error caused by Discrete Fourier Transform. The algorithm can improve the accuracy for islanding detection and have a good performance for filtering harmonics and noise. Finally, we can reduce the damage when distributed resources units and utility operate in parallel synchronization.