High‐impedance fault detection in electrical power distribution systems using moving sum approach

AI in arcing‐HIF detection: a brief review

“…(iv) Asymmetry [7]: Owing to different breakdown voltages at positive and negative half cycles, there are a current peak and wave shape differences between two half cycles. (v) Build-up and shoulder [3]: After about tens of cycles, the fault current reaches a peak, at which point the accumulation will temporarily stop, keep several cycles of constant value and form the shoulder.…”

“…The voltage signal can also be used to classify arcing-HIF. Sarwagya et al [3] calculate the absolute value of three-phase voltages in several cycles. Gautam and Brahma [18] deal with voltage signals based on mathematical morphology, and takes the output difference as fault feature.…”

“…Parts of operation in distribution network share the similar characteristics with arcing-HIF signals such as non-linear load with non-linear characteristics [45], capacitor switching, feeder switching, load switching with high-frequency characteristics [3], transformer inrush current and current transformer (CT) saturation with low-frequency and asymmetry characteristics [43]. Arcing-HIF detection methods have to deal with the game between sensitivity and reliability.…”

“…Several methods use a single threshold or threshold sets to distinguish non-arcing-HIF and arcing-HIF. The threshold indexes contain input impedance variation [1], root mean square (RMS) of the current [2], the absolute value of three-phase voltage [3] etc. The key problem is how to determine the appropriate thresholds in complex operation environments.…”

HIF detection in distribution networks based on Kullback–Leibler divergence

“…There are two types of HIF faults: unbroken and broken. In the former type, the distribution conductor makes contact with a high‐impedance object such as tree limbs and wood fences while the latter type occurs when an energised conductor breaks and makes electrical contact with the ground surface such as the sidewalk, concrete, asphalt, and sand [4–6]. The HIF is mostly associated with the arcing phenomenon, and its main features are low current, randomness, non‐linearity, and asymmetry [7, 8].…”