Protection zones

The idea

A relay does not watch the network. It watches a zone: a defined stretch of equipment, bounded by the breakers the relay can open, with current transformers at those breakers measuring what flows in. When you draw the zones for a scheme, one check comes before everything else: every cable, busbar and transformer on the drawing must sit inside at least one zone.

That requirement is why zones overlap. Suppose Zone 1 ended exactly where Zone 2 began. The boundary would be a single line on the drawing, the breaker on that line would belong to neither zone, and no relay would clear a fault at that point. So engineers cross the zones over each other at the breaker: the current transformers on each side are assigned to different zones, and the breaker itself sits inside both.

A fault in the overlap trips both zones, which costs a slightly wider outage. The trade is deliberate — a small overlap is better than a small stretch with no protection at all.

Once the zones exist, every fault follows an ordered script. Primary protection is the first layer: the relay of the faulted zone sees the fault, times out, and trips its own breaker. The clearing time is the relay's operating time plus the breaker's interrupting time, and nothing else in the network has to move.

Backup is the second layer. The upstream relay sees the same fault, because its zone reaches past the boundary, but a grading margin makes it deliberately slower. If the primary relay never acts, the backup clears the fault later, from further upstream, taking healthy equipment off with the faulted equipment. (The TCC curves page covers that margin on inverse curves; the model below uses definite times, so the margin here is a flat step rather than a gap between two curves.)

Breaker failure is a different failure, and it needs its own answer. Here the relay works perfectly and sends the trip — but the breaker does not open. A breaker-failure timer starts with the trip command and checks whether current still flows after the breaker's normal opening time. When it expires, it trips the breakers around the stuck one. This layer clears faster than graded backup, because its clock starts at the trip command, not at the fault.

Try it

Clear a fault on feeder 2 the ordinary way. Then fail the relay, and then jam the breaker. Read the three clearing times from the sequence: 0.30 s, 0.70 s, 0.55 s.

Why it matters

• A gap between zones is unprotected equipment. The overlap is not an accident of the drawing: a fault between two current transformers must still sit inside a zone, and the cheapest way to guarantee it is to put the breaker inside both.
• Every backup operation is a bigger outage. Backup does not just clear late; it clears from further upstream, so a fault on one feeder can take out the feeders beside it. That is the real cost of a relay that fails to operate, and it is why commissioning tests check the primary relay rather than assume it works.
• A stuck breaker is not a slow relay. If graded backup has to cover breaker failure, the fault stays on the network for the full margin. A dedicated timer cuts that to a fraction, and it is the only layer that acts when the relay is right and the switchgear is not.
• The deliverable is a sequence, not a setting. A protection study answers three things for each credible fault: what happens, in what order, and what the network loses. On meshed and ring networks the answer also depends on direction, the subject of directional overcurrent and of the network topology.

Related concepts