基於微控制器之微型同步相量量測器研究

Abstract

如今越來越多元的電源及負載出現於電力系統中，諧波及雜訊的問題也越發被重視。在量測系統中的頻率及相量時也會受上述兩者之影響產生測量偏差，到目前為止已有很多的演算法被提出以期望解決此問題。DFT(Discrete Fourier Transform)為一被用於計算頻率的強大工具， 並有各式各樣的延伸應用，而Smart DFT為本文所用之DFT演算法，其優點為完全以數值解方式計算頻率，無忽略頻率變動量之參數，故並不會產生計算偏差，適用於需精確計算頻率之場合。前代微型同步相量量測器(micro synchronized Phasor Measurement Unit, µPMU)中已知有三至四處問題需改進且價格也需降低，本文使用雙核心平行運算的方式使得電壓收集的更完整、輸出資料更即時，且時間戳並不會因為長時間運行而不準確，在本文中也將驗證前代所用之演算法是否符合國際規範。本文也自行製作了一級時間源，使µPMU得以精確對時，並嘗試將其時間誤差降至最低。本文研製出一秒十筆及一秒三十筆即時輸出，並在固定頻率測量下將頻率之標準差降低至0.0007Hz以下，能夠準確測量市電之頻率，表現也比前一代版本更佳。

Nowadays, more and more diversified power source and load appear in the power system, and the problems of harmonic and noise are paid more and more attention. The value of frequency and phasor measured will also be affected by the above two terms, so far, a multiple of a variety of algorithms has been proposed to solve this problem. The Discrete Fourier Transform (DFT) is a powerful tool used to calculate frequency and has a variety of extended applications. Smart DFT is the DFT algorithm used in this study. Its advantage is that the frequency is calculated completely by numerical solution without ignoring the parameters of frequency variation, so there is no calculation deviation. The last generation of micro synchronous Phasor Measurement Unit (micro Phasor Measurement Unit, micro PMU) are known in around three or four problems need to be improved and the price also need to be reduced. This study uses multi-core and parallel computing method makes a more complete of the input voltage when collecting data, a more immediate of outputting data, and the timestamps will not incorrect after running for a long time. In this article will also verify the algorithm used by the last generation is in line with the international norms. In this study, I also made a first-level time source by myself, tried to minimize its time interval error so that PMU could match the timestamp accurately. In this study, 10 outputs per second and 30 outputs per second were produced and the standard deviation of the frequency was reduced to below 0.0007 Hz under the condition of fixed-frequency measurement.