摘要 電力系統中變壓器是很重要的設備,其激磁諧波電流亦是電力諧波的來源之一,對電力品質有不利影響,激磁諧波電流主宰於變壓器鐵心的磁滯迴線,故要掌握激磁電流的諧波特性,必須對磁滯迴線與激磁諧波電流之關係加以探討,本論文提出兩個磁滯迴線的特性參數,來描繪單相變壓器磁滯迴線特性,並可據以進一步分析其激磁電流的諧波特性。文中首先以傅立葉級數為基礎發展一套程式以計算激磁電流的諧波成分,然後再發展一套類神經網路可用以獲得磁滯迴線的特性參數。最後將舉一實例,先利用類神經網路以求得磁滯迴線參數,然後再計算激磁電流的諧波成分,並與實測做比較,證實本文方法可行。
Abstract The power transformers are very important apparatuses in the power system; their magnetizing current harmonics are also the harmonic sources of power system, which have detrimental effects to power quality. The magnetizing current harmonics are mainly characterized by the hysteresis loop of core of transformer. So, the relationship between the hysteresis loop and the magnetizing current harmonics should be studied for analyzing the magnetizing current harmonics. This thesis presents two parameters of hysteresis loop to describe the hysteresis loop characteristic of single-phase transformer and to analyze the magnetizing current harmonics. A program based on Fouries-series for calculating the harmonic components of magnetizing current is first developed. Then, a neural network used to find the parameters of hysteresis loop is also developed. Finally, the practical case of using neural network to finding the parameters of hysteresis loop and Fourier-series program to calculating magnetizing current harmonics is presented. The comparisons between the calculation results and test data of harmonic components of magnetizing current for a single-phase transformer are presented to certify the availability of the proposed method.