Circuit breakers are by far the largest single source of substation events, and a breaker that fails to open, or opens late, leaves everything it protects exposed. This video summarises six technical webinars on circuit breakers and switchgear: what fails, what cyclic maintenance misses and what a continuous monitor really adds.
Key points
- Between 83 and 90 % of reported problems are mechanical, at every voltage class.
- Defects develop between inspections, and several are invisible to a test with the breaker out of service.
- The trip coil current portrays the mechanism: diagnosis is a comparison with the healthy signature, not a threshold.
- The protection relay describes the fault; the monitor describes the breaker.
- With SF6, a gauge only alarms once a lot of gas has gone; in switchgear it is the insulation that fails, and partial discharge finds it.
Why the breaker matters so much
A slow or failed opening exposes the whole system to fault current. Breaker mis-operation is often the root cause of a transformer failure, so a transformer is only as reliable as the breaker protecting it. And a breaker that sits closed for years gives no sign of its condition until the day it is asked to trip.
It is almost always mechanical
The figures quoted across the sessions come from the EPRI substation life-extension guidelines:
| Origin of the problem | Share of reported problems |
|---|---|
| Mechanical | 83–90 % |
| Auxiliary and control circuits | 8–15 % |
| Main electrical circuit | 1–5 % |
The leading contributors are loss of lubrication, lack of maintenance and ageing, and the symptom that can actually be measured is a slowed interrupting time.
What cyclic maintenance misses
The problems develop between the intervals: lubrication degrading and lengthening the interrupting time, heater failure and condensation in the control cabinet, hydraulic and pneumatic leaks that drive excessive motor starts, SF6 loss and moisture ingress through failing seals, trip and close coil failures, and sticking or corroded auxiliary contacts. Timing and travel testing finds these, but it needs the breaker out of service, so it is done rarely.
Each offline test also has a blind spot. Timing and travel needs an outage; contact resistance does not see arcing contact wear, because the arcing contacts are not the ones being measured; power factor is influenced by ambient conditions and contamination. The first trip test is the one that catches what the others miss.
The coil current is the mechanism talking
Online, the central measurement is the trip coil current. Every bend in the waveform is a piece of the mechanism moving: current begins through the 52a contact, the plunger reaches speed, the latch is struck, the buffer is struck and, finally, the 52a opens and the circuit is broken. Captured at 128 samples per cycle, it is resolved to the millisecond and stored as a waveform that can be overlaid on the previous one.
That is why the diagnosis is a comparison, not a threshold. When lubrication dries out, the first half of the waveform is identical to the healthy reference and everything after the latch arrives late. Operate the breaker a second time and the interrupting time improves, which is itself the signature of a lubrication problem rather than a latch one. Left alone, it gets progressively worse.
Relay or monitor: one describes the fault, the other the breaker
The protection relay already records the trip, so why add anything? Because from the same trip the relay gives the type of fault, the distance to the fault, the interrupting time for that one event and the breaker status, while the monitor adds what the relay cannot see: arc time and open time, the coil current waveform, the cumulative I²t per pole (the contact wear that accumulates), the motor and stored energy, and the condition of the heaters and cabinet.
From each capture come the numbers a breaker is judged on: interrupting time, open and close time per contact mechanism, arc time and restrikes.
The gas: a gauge alarms when it has already gone
Gauges and density switches need someone to walk up and read them, and they only alarm after a significant loss. A sensor trends the loss, forecasts the days to alarm and to lockout, and turns cumulative mass loss into the emissions report that has to be filed. It should measure density and dew point together, not pressure alone: pressure moves with temperature, while density is what the equipment actually needs.
When the gas is tested rather than just measured, the sessions give four reasons — acceptance, safety, fault modes, component wear — and one decision at the end: uncontaminated gas is stored and reused, slightly contaminated gas is filtered and retested, and severely contaminated gas is reprocessed or destroyed.
Switchgear: here the enemy is the insulation
In switchgear the failure mode changes. Around three quarters of failures are insulation related, driven by moisture, dust, contaminants and age. The symptoms are corona, tracking on bus and insulators and, finally, flashover. Partial discharge sensors — coupling capacitors and radio-frequency CTs — detect these dielectric problems at an incipient stage. They also remove a hazard: fewer reasons to rack a breaker out for testing means less arc-flash exposure and less wear on the connectors.
What the six sessions agree on
- The failure is mechanical, and it develops between maintenance visits.
- The relay describes the fault; only the monitor describes the breaker.
- Diagnosis is a comparison against that breaker’s own baseline.
- Gas needs trending, not a gauge somebody has to walk up to.
- The value is scheduling repairs instead of testing during outages.
Related equipment
The Dynamic Ratings BPM breaker performance monitor captures the trip and close coil currents in service. For offline work, see our circuit breaker analysers, the SF6 leak detectors and the partial discharge detection range. If you would like to discuss monitoring for your own breakers, contact our engineers.