Automatic retrieval downloads COMTRADE oscillography and event records from each relay as soon as they are created, and stores them on a central server with a name that identifies substation, device, date and time. It uses IEC 61850 file transfer over MMS, FTP or SFTP, or the manufacturer's software. It keeps records from being lost when the relay overwrites its memory and lets a trip be analyzed in hours, not days.
The problem: the record was there, but it is gone
After a trip, the relay’s oscillography record is the main evidence of what happened: what the protection saw, when it picked up, when it tripped and how long the breaker took. The guide on how to read an oscillography record shows everything it contains. The problem is getting to it in time.
In many substations, downloading is manual: a technician goes out with a laptop, connects to the relay and copies the files. Meanwhile, the relay keeps recording. Its memory holds a limited number of records and, when it is full, overwrites the oldest ones. A reclose, a second fault or restoration switching can erase the record that mattered. In addition, files copied by hand end up with generic names, in personal folders, with no clear link to the event.
Automatic retrieval solves both: it downloads each record as soon as it exists and stores it in a central repository with a name that identifies it.
What is retrieved
- Oscillography in COMTRADE format, defined by IEC/IEEE 60255-24:2013 (IEEE C37.111-2013): the configuration and data files, or the single .cff file.
- Event records and sequence of events (SOE): pickups, trips, status changes, alarms, with their time stamps.
- Fault reports generated by some relays: fault type, phases, estimated distance, currents.
- Records from digital fault recorders and PMUs, where present, with longer duration and more channels than the relay’s.
How the files reach the server
There are three main paths, which can coexist in the same substation:
| Method | How it works | Comment |
|---|---|---|
| IEC 61850 file transfer over MMS | The server reads the IED’s record directory and requests each file with the file services of the communication model | Vendor-independent in theory; in practice each IED organizes its directories its own way |
| FTP or SFTP | The server connects to the IED as to a file server | SFTP encrypts the session; FTP does not, and must be confined to closed networks |
| Vendor software | A manufacturer’s application polls its relays with its proprietary protocol and downloads the records | Covers its own brand well; substations with several brands need several |
How the server finds out. In IEC 61850, the disturbance recorder logical node (RDRE) flags that a new record was made, and that data can arrive in an MMS report. The server reacts by requesting the file. With the other methods, the server polls each device periodically. The guide on IEC 61850 MMS describes reports and services.
- 01EventThe relay records the disturbance and creates the COMTRADE file
- 02Notice or pollThe server receives notice of the new record or detects it in its periodic poll
- 03DownloadThe server copies the file over MMS, SFTP or the vendor software
- 04Name and storeRenamed with substation, device, date and time and stored in the repository
- 05AnalysisThe engineer reaches the record through a secure channel and analyzes it
Naming and organization
A file named with a sequential number says nothing when there are hundreds. IEEE C37.232 (COMNAME) proposes a naming convention for time sequence data files that includes the event date and time and the names of the substation, circuit and recording device. IEEE lists the 2011 edition as inactive since 2022 and has a revision project open, but the idea is still the basis of good practice: the name must let you find the record without opening it.
The rest is organization: one folder per substation and per device, original files kept unmodified, and a retention policy that defines how long they are kept.
Synchronized time
An oscillography record without a reliable time stamp shows the waveform, but cannot be correlated with the record from the other end of the line, with the SCADA SOE or with the breaker’s record. Automatic retrieval only delivers its full value if the relays are synchronized, as described in time synchronization in substations.
Secure access
The retrieval server needs to reach every relay, and analysts need to reach the server. That combination is sensitive. The practical rules are:
- The server sits in its own network zone, separate from the control network and the office network, with an industrial router and firewall at the boundary.
- It accesses the relays with a read-only account that cannot change settings.
- Analysts, in-house or external, access the server and not the relays, through a controlled access point, with individual accounts and session logging.
- Unencrypted FTP and default passwords are ruled out.
What to specify
- Which devices are retrieved: relays, recorders, power quality meters.
- Which method and which account are used for each brand.
- How often each device is polled and how new records are detected.
- The naming convention and folder structure.
- Where the server sits, who has access and with what session logging.
- Which alarm is raised if a device stops responding to retrieval.
A useful record starts in the relay
No retrieval system fixes a badly configured record. What gets retrieved is what the relay recorded, and that depends on its recording settings:
- What triggers it: protection pickup, in addition to trip, so that events where the protection saw the fault and did not operate are also kept.
- How much before and after: enough pre-fault to see the normal condition and enough post-fault to see the breaker opening and the reclose.
- Which signals it includes: the currents and voltages of the three phases and neutral, and the key binary signals (trips, breaker position, reclosing, teleprotection).
- How many records fit and what happens when the memory is full.
- The time, with the relay synchronized to the substation’s common source.
These settings vary from one brand to another and should be reviewed at commissioning and after every settings change.
What you gain
- No record lost to overwriting.
- Faster analysis: the engineer has the records from both ends of the line shortly after the event.
- History: repeated events on the same circuit can be compared and trends detected, such as a breaker that takes longer each time.
- Support for event reports to the utility or system operator when an event must be reported.
To review a downloaded record, you can open it in our free COMTRADE viewer, which runs in the browser.
How we help
Having the records is half the job; the other half is interpreting them. If a trip is not clear, send us the COMTRADE files and events from both ends: at Tiza Group we carry out relay trip analysis, quote the analysis for each case and deliver the cause, the performance of each protection and the recommended actions.
If your relays record short records, without the signals that matter or without a reliable time stamp, we review and adjust their disturbance recording with our remote relay configuration service, for the main brands, in medium, high and extra-high voltage substations and in industry.
Related guides
- COMTRADE files
- How to read an oscillography record
- Electrical fault analysis
- Digital fault recorder and PMU
- IEC 61850 MMS
- IEC 61850 to the control center: TR 61850-90-2
- WAN networks for substations: IEC TR 61850-90-12
- Configuration management of relay settings and the SCD
- Substation communications by voltage level
References
- IEC/IEEE 60255-24:2013, Common format for transient data exchange (COMTRADE) for power systems (IEEE Std C37.111-2013).
- IEEE C37.232-2011, Standard for Common Format for Naming Time Sequence Data Files (COMNAME).
- IEC 61850-7-2, Abstract communication service interface (ACSI) (file transfer model).
- IEC 61850-7-4, Compatible logical node classes and data object classes (RDRE).
- IEC 61850-8-1, Mappings to MMS and to ISO/IEC 8802-3.
Frequently asked questions
Why are relay oscillography records lost?
Because each relay stores a limited number of records and, when its memory is full, overwrites the oldest ones. If an event is followed by several switching operations, reclosures or further faults, the important record may disappear before anyone goes to download it.
How does the server know a relay has a new oscillography record?
In IEC 61850, the disturbance recorder logical node (RDRE) flags that a new record was made, and the server can receive it in a report and then request the file. With other protocols or with vendor software, the server polls each device's record directory periodically.
Is it safe to give remote access to oscillography records?
Yes, if the retrieval server sits in its own network zone, accesses the relays with a read-only account, and analysts reach it through a controlled access point, with individual accounts and logging. What should be avoided is exposing the relays directly outside the substation.