Algorithm Development and Performance Evaluation for Series Arc Fault Detection in Linear Loads

Statistical data from the last eight years-suggest that electrical fires account for approximately 20% of all fire incidents in Korea, with over 80% of these being related to arc faults. In this study, the detection and interruption performance of commercial earth leakage breakers and arc fault circuit interrupters were experimentally evaluated under serial arc- fault conditions for linear loads. Under load conditions of 5, 10, 15 and 20 A, the existing breakers did not trip despite observable characteristics in the current waveform, such as shoulder formation and flattening. To address the limitations of conventional circuit breakers in mitigating arc faults, fundamental data on the electrical characteristics of the voltage and current variations under serial arc conditions were obtained through experimental measurements for different load currents. On the basis of the results, analytical tools were implemented in Google Colab using Python to develop algorithms with enhanced reliability. Furthermore, the criteria for arc threshold determination were established using the total harmonic distortion and root mean square values, thereby enabling verification of the detection and interruption performance for serial disconnection arc faults.