Series arc fault detection method based on statistical analysis for dc Microgrids

Arc fault detection is required for the promising dc Microgrids because they are susceptible to dc arcing which may cause fire without protection. This paper proposes a new arc fault detection method based on statistical analysis. The statistical analysis provides design insight to establish the arc detection strategy by helping to determine the key parameters and operation principle of the arc fault detector. Since it enables guaranteeing the reliability of a fault detection method, the arc fault detector can be designed to meet the system requirements. The hardware prototype verifies the effectiveness of the method. The prototype arc fault detector can achieve 99.999% of fault detection accuracy in given arcing condition and 0.027% of false detection in given non-arcing condition while avoiding malfunctions from the switching noise of dc Microgrids interfacing power circuits.