Distance Relay Three-Pole Trip Caused by a Blown CVT Fuse
A 220 kV, 150 km single-circuit line is protected by a numerical distance relay with quadrilateral zones: Zone 1 at 80% of the line, Zone 2 at 120%, Zone 3 at 200%, all set on an 85° line angle; the positive-sequence line impedance is 0.42 Ω/km. A VT-fuse-failure supervision function exists in the relay but has never been enabled, and the trip output is three-pole. During the evening peak the line carries 620 A at 0.92 power factor lagging. At 19:42 the breaker tripped with no fault current recorded; the only plant event was a blown fuse in the Y-phase of the capacitor voltage transformer, found later during a routine check. The disturbance record shows the three line currents unchanged at 620 A with the same phase angles throughout, the Y-phase secondary voltage falling to about 3 kV primary-equivalent, the other two phase voltages essentially unchanged at 127 kV, and no measurable zero- or negative-sequence current. The relay event log shows the Y-phase impedance falling to 4.8 Ω and a Zone 1 trip 18 ms later. The protection engineer proposes increasing the Zone 1 reach and adding a 100 ms delay to the Zone 1 trip.
1. Reconstruct why the relay tripped, using the zone settings and the recorded quantities.
Read the worked answer
Normal running condition. Phase voltage = 220/√3 = 127 kV. With 620 A at 0.92 pf lagging, the load impedance seen by the relay is |Z| = 127 000/620 = 205 Ω at an angle of about 23°. Zone 1 reaches 0.8 × 0.42 × 150 = 50.4 Ω, Zone 2 reaches 75.6 Ω, Zone 3 reaches 126 Ω. So the healthy load point sits far outside Zone 3 - which is exactly as it should be.
The moment the fuse blew. The Y-phase voltage measured by the relay collapses to about 3 kV while the Y-phase current remains 620 A. The relay's Y-phase impedance becomes 3000/620 ≈ 4.8 Ω, which is inside Zone 1 (50.4 Ω) and inside the resistive reach as well. The relay sees what it interprets as a close-in fault on phase Y and trips in Zone 1 time, about one cycle later - hence the 18 ms.
The important point is that this is a measurement failure being processed as a genuine fault. The relay behaved exactly as designed for the numbers it was given. No setting change fixes that, because any reach large enough to clear real faults will also be entered by a collapsed voltage.
2. Which recorded quantities distinguish a VT failure from a real Y-phase-to-earth fault, and what would a real fault have looked like?
Read the worked answer
Real single-phase-to-earth fault on phase Y would show all of the following:
- Current change. The Y-phase current would rise sharply and change in phase relative to the pre-fault load current; here all three currents sat unchanged at 620 A with identical angles.
- Zero and negative sequence. A solid Y–earth fault produces I0 ≈ I2 ≈ I1/3. The record shows no measurable I0 or I2. This is the single most reliable discriminator.
- Healthy-phase voltages. In an effectively earthed 220 kV system, the two healthy phase voltages would rise (approaching line-to-line value depending on the earthing coefficient) rather than sit at 127 kV.
- Fault path. Both ends would feed the fault; the local relay would see a current contribution and the remote end would report a corresponding event. No fault current was recorded anywhere.
VT failure signature: one voltage collapses, all three currents unchanged in magnitude and angle, sequence currents negligible, and the total three-phase power computed by the relay becomes nonsensical because V_Y is not a real measurement. The relay sees a physically impossible combination - a fault-level impedance with no fault-level current - and that inconsistency is precisely what fuse-failure supervision is designed to detect.
3. Assess the proposed remedy and specify what should actually be done, including the logic settings and how they would be proven.
Read the worked answer
The proposal is wrong on both counts. Increasing Zone 1 reach makes the relay more likely to see load and to overreach; a 100 ms Zone 1 delay slows clearing of genuine close-in faults on a 220 kV line and buys nothing against a voltage collapse that lasts until the fuse is replaced. Neither change addresses a measurement failure.
What should be done:
- Enable VT fuse-failure supervision (VTS). The standard logic combines: a voltage-change element (one or more phase voltages drop by more than, say, 10–15% with no corresponding current change), a negative-sequence comparison (V2 rises while I2 stays below threshold - a real fault produces both), an undervoltage-with-no-current criterion, and a zero-sequence criterion to block during genuine earth faults. On operation the element should alarm, block the distance zones, and in some schemes transfer to a current-only backup mode.
- Coordinate the CVT secondary protection. Fuse and MCB ratings and their time-current curves must be checked against the relay's voltage input burden, and the fuse should be monitored so that a single blown fuse raises an alarm before it trips anything.
- Keep the three-pole trip, but make it depend on validated voltages. Single-pole tripping on a 150 km line with a fused CVT would introduce more risk than it removes.
Proof: secondary injection with a balanced three-phase voltage and current set, then collapse one phase voltage while holding the currents constant. The VTS element must pick up and the distance zones must block; a genuine single-phase fault case (current rise plus V2 and I2) must still trip. Record both cases in the relay's test file, since the whole failure was one of a function that existed but was never commissioned.