Yizhao Wang

Existing protection schemes often fail to detect cross-line successive single-phase-to-ground (SSPG) faults in resonant-grounded active distribution networks, due to the severely weakened transient signature that results when a secondary fault (f₂) superimposes on an uncleared primary fault (f₁). To address this, a novel detection method based on inner-product analysis is proposed. First, a transient equivalent circuit for the f₂ fault is established. Expressions for the transient zero-sequence currents in healthy, primary, and secondary fault feeders, as well as for the bus zero-sequence voltage, are subsequently derived and analyzed. Then, the inner-product is applied to reveal the inherent correlation between healthy and faulty feeders during an f₁ fault. For the f₂ fault, distinct dynamic patterns, specifically in the amplitude and polarity of the inner products, are revealed for the primary, and secondary fault feeders. This analysis provides the theoretical foundation for SSPG fault detection. Furthermore, a differential preprocessing algorithm is introduced to enhance transient signatures, thereby improving the reliability of the inner-product based detection. Simulation and field test results confirm that the proposed method accurately detects both the f₁ and f₂ faults under challenging conditions, including short fault intervals, diverse grounding resistance combinations, distributed generation integration, arc faults, and strong noise interference.

In non-effectively grounded system, the rise in phase voltage during single-phase grounding (SPG) fault poses a significant threat to the system’s insulation integrity. The increasing cabling rate in distribution network has led to a rising proportion of cross-line double successive single-phase grounding (SSPG) faults triggered by initial fault. To address the challenges of weak characteristic and difficult detection of such successive fault, this paper proposes a SSPG detection method based on zero-sequence capacitance variation. First, the dynamic zero-sequence equivalent network during successive faults is simplified, and derive the zero-sequence current (ZSC) expressions for each feeder under both overdamped and underdamped conditions. Subsequently, the singular value decomposition-total least squares extended Prony algorithm is employed to fit the zero-sequence voltage (ZSV) and ZSC, enabling the calculation of zero-sequence capacitance for both initial and successive faults in the complex frequency domain. A detection criterion is then constructed using the variation of capacitance. Finally, digital simulations and recorded field data validate the correctness and feasibility of the theoretical analysis.

When cross-line same-phase successive faults (CSSF) occur in a distribution network, the second fault makes the zero-sequence voltage consistently exceed 15% of the phase voltage, resulting in the existing detection devices not being able to return, which makes the fault persist for a long time and endangers the stable operation of the power grid. To address this problem, we propose a fault detection method based on vertical and horizontal double criteria. Firstly, zero-sequence current and voltage features are analyzed for a flexible grounding system when the grounding mode changes during the CSSF period. Secondly, constructing two boundary cases calculates the zero-sequence admittance modulus of the first and second faulty feeders. Finally, vertical criterion 1 is constructed because the fault feeder’s zero-sequence admittance modulus is greater than the healthy feeder. Horizontal criterion 2 is constructed because the faulty feeder’s zero-sequence admittance modulus ratio is greater than the healthy feeder. PSCAD simulations and field tests show that the method can accurately detect CSSF in flexible grounding systems without being affected by current transformer reversal, missing data, strong noise and intermittent arcing.