In short. After a trip or blackout, the useful question is what happened first and why. Sequence-of-events (SOE) lists show the order; disturbance records show waveforms. Together, with accurate time stamps, they turn an incident into a fixed cause.
Records and their use
| Record | Shows | Needs |
|---|---|---|
| Event list (SOE) | Order of status changes with time stamps | Millisecond time stamping at the source |
| Disturbance record | Waveforms of voltage and current around a trigger | Pre- and post-trigger recording |
| Trend | Slow changes over hours to months | Historian with sensible resolution |
| Alarm log | What operators saw | Operator action record |
Setup
- Time-synchronise devices. See time synchronisation.
- Define triggers for disturbance records: protection pickup, trip, voltage dip, breaker operation.
- Record enough pre-fault time.
- Name signals clearly and consistently.
- Collect records automatically after events.
- Preserve original files with metadata.
Reading an incident
Identify first event → check the protection that operated → read currents and voltages → check breaker timing → check controller actions → find the root cause. Compare actual operating time to expected from coordination curves. See TCC engineering.
Generator-specific insights
- Start time, stabilisation time and transfer time against design.
- Frequency and voltage dip during block loads.
- Speed of load shedding relative to frequency fall.
- Reverse power or loss of excitation events.
From records to action
Keep an event register with cause, corrective action, owner and closure. Repeat events show a systemic issue. Trend start time and transfer time to detect slow deterioration. See protection records.
Illustrative example
A plant loses a bus. The SOE shows a breaker failure signal 80 ms after the first trip, and the disturbance record shows current persisting. The fault is traced to a worn breaker mechanism that was slow to open. Without synchronised stamps and records, the incident would have been blamed on the relay.
Three kinds of event data
| Type | Resolution | Answers |
|---|---|---|
| Sequence of events (SOE) | 1 ms stamping at the device | What happened first? |
| Disturbance record | Waveform samples per cycle | How did current and voltage behave? |
| Trend data | Seconds to minutes | What was drifting before? |
Why 1 ms matters
In a cascading event, the first trip may precede the others by a few milliseconds. With 1 second SCADA stamping, the order is lost and the wrong device is blamed. Stamp events at the device with synchronised clocks.
Using the data
- Reconstruct the timeline from SOE across devices.
- Open waveforms at the first device to find the fault type and magnitude.
- Check whether each relay acted as designed.
- Write a short report: cause, effect, corrective action.
- Update settings or procedures.
Intelligence from patterns
Repeated small events, such as a breaker nearly tripping or a generator dipping on one particular load, show degradation before failure. Review them regularly.
Practical rules
- Retrieve records promptly before overwriting.
- Name and archive them with date and site.
- Time sync all devices. See time synchronisation.
See alarms, events and disturbance records.
The master engineering question
After any disturbance, can we reconstruct the exact order of events to the millisecond and prove what failed first?
What to send us
Send relay models and event samples. Send your application details.
Questions engineers ask
Why do time stamps matter so much?
Without a common time base, event order across devices is guesswork.
How long should records be kept?
Per policy and regulation; keep incident data long term.
Can SCADA replace relay records?
No. SCADA update rates are usually too slow for waveform analysis.
Who should review them?
Engineers, with operations, after every significant event.
This guide gives general principles. Follow site records policy.