A Comprehensive, Zero-Setting Algorithm To Detect Secondary Arc Extinction
Single-pole tripping is used to maintain power transfer and system stability when clearing single-line-to-ground faults. While the single phase is open, the healthy phases can sustain the ionized channel created by the primary fault, causing a secondary arc. The arc must be extinguished before reclosing to successfully re-energize the open phase. This paper presents a comprehensive, zero-setting algorithm to detect secondary arc extinction for any line configuration.
Typically, open pole reclosing is delayed by a fixed, conservative dead time to allow sufficient time for the secondary arc to extinguish. This is simple, but not optimal. Secondary arc extinction depends on several deterministic and random factors and may occur well before the end of the dead time. Detecting secondary arc extinction can improve system stability by allowing faster reclosing. Conversely, for a permanent fault, the arc will not extinguish by the end of the dead time. Detecting a persistent fault on the open phase can prevent stressing the system, again, by blocking reclosing and initiating three-pole tripping to extinguish the fault.
The authors previously proposed an algorithm to detect secondary arc extinction and permanent faults. For lines with no shunt reactors, the algorithm relied on mutual coupling line parameters, which can be difficult to determine. For lines with shunt reactors, the previous algorithm focused on four-legged reactors and did not address lines with three-legged shunt reactors or series capacitors.
The algorithm presented in this paper uses first principles to identify unique signatures in the open-phase voltage for detection of arc extinction. The prior algorithm for shunt reactor compensated lines is extended to cover three-legged shunt reactors. For detecting arc extinction on lines without shunt reactor compensation, a new method is proposed, which does not require any user settings and removes reliance on mutual-coupling line parameters. Both methods can be used on any line configuration—transposed, untransposed, parallel, or series-compensated. The performance of the new algorithm is validated against field events and simulations.
