Detection of Arc Fault on Low Voltage Power Circuit by Using Fuzzy Theory and Neural Network

• Main Authors: Ming-Ghe Shih, 史明哲
• Other Authors: Chi-Jui Wu
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/60375659724229303630

碩士 === 國立臺灣科技大學 === 電機工程系 === 103 === The main purpose of this paper is to find the difference of current characteristics between arc fault and normal operation by using the arc fault experimental platform to gather characteristics of the load. This study proposed two detecting methods. First of all, the experimental data are analyzed by Fast Fourier Transforms (FFT) for the signal processing to capture feature of current. Thereafter, these extracted features are applied to learning vector quantization (LVQ) and fuzzy system, respectively. Therefore, the two detecting methods are developed. At last, the two detecting methods are used to test experimental data, including normal operation, on/off switching, series arc fault and branch series arc fault. The results are compared with the commercial Arc-Fault Circuit Interrupter (AFCI). According to this research, the detecting methods can effectively detect series arc fault and show little malfunction. They have better ability than commercial AFCIs.