Functional Analysis for Series Arc Fault Diagnosis in Residential AC Power System

Proper monitoring of the components of a system leads to a successful control of it. Fault detection and diagnosis (FDD) is an effective methodology that uses system's information commonly measured from sensors and other devices along with mathematical and statistical theories to effectively control and improve a system's performance. In this study, fault detection and diagnosis (iFDD) was applied to analyze and detect series arcing in an AC electrical installation. The iFDD methodology presented in this article focused on pattern recognition and fingerprint analysis of voltage and current signals under induced series arcing. An AC power system, often found in houses or commercial buildings for daily use of electrical appliances and equipment, was constructed and utilized for series arcing experiments. The tests were conducted on experimental currents (2–7 A) measured on domestic loads (halogen lamp and ranger plate). All data were digitized with a sample rate of 250 kHz (oscilloscope sampling rate was 250 MHz). Fingerprints of series arcing under different working conditions were analyzed. Multiresolution wavelet decomposition of current and voltage signals were applied using MATLAB for analysis of fault transients with selection of a suitable mother wavelet. Results showed that db4 is suitable for detection of series arcing using information of the current signal