Before energizing, the settings file that the relay actually has loaded is read and compared, parameter by parameter, with the approved calculation memo. It is not enough that the relay passed its functional tests: it can work perfectly with wrong settings. Each difference is explained, corrected or approved in writing, and the result is recorded before energization.
Short answer
A relay that passes its functional tests only shows that it operates with the settings it has loaded. It does not show that they are the study’s. Before energizing, the settings file the relay really holds is read, compared parameter by parameter with the approved calculation memo, and every difference is resolved: corrected, or approved in writing with its reason. The result goes into the settings record with the as-found and the final state.
Why it is a separate test
There are three moments when settings can drift away from the study:
- When transcribing. Someone copies the values from the memo into the relay, by hand or with a template, and a value is changed or left out.
- When testing. To inject, settings are modified (a pickup is lowered, a function disabled) and not all of them are put back.
- When the study changes. The memo is updated and the relay keeps the earlier version, or the other way round.
None of the three is detected by a functional test. They are detected by comparing, which is the only test that catches all three, and it takes far less time than finding the error after a real fault.
What is compared
Everything the relay holds is compared, not only what the study calculated:
- Installation data: CT and VT ratios, nominal voltage, frequency, phase sequence. They condition what the relay measures and displays.
- Settings of the study’s functions: pickup, dial or time, curve, directionality, blocks, each in its settings group.
- Logic and outputs: which contact or message each trip goes to, which inputs block what.
- Communication and recording parameters that affect operation (addresses, GOOSE used).
- Enabled and disabled functions. A function disabled by mistake is not noticed until it is needed.
How it is done
- The file is read from the relay with the manufacturer’s software, before changing anything. This is the as-found state.
- The calculation memo of that version is prepared, with each parameter and its expected value. It can be a table.
- The two are compared, parameter by parameter. For long lists a semi-automatic process that flags differences is convenient, with a person reviewing the result.
- The differences are classified: loading error, admissible difference with a reason, or a change to the study that was not updated.
- What has to be corrected is corrected, the file is read again and saved as the final state.
- The comparison is signed: who, when, with which version of the memo and of the file.
This step does not include the device programming procedure, which depends on the manufacturer and model.
Which differences are most dangerous
- The CT ratio, because it changes the relay reading and every function referred to primary values.
- Dials and times, because they can remove the coordination margin with the upstream relay. See coordination time interval.
- A disabled function, for the reason above. A relay with no documentation at all is the extreme case: see relay without documented settings.
- The active settings group: the relay can have the correct value loaded in a group that is not the one operating.
Worked example
Example data (assumed): relay 51 on a feeder with a real 600/5 A CT (ratio 120). The calculation memo defines 40 parameters. Pickup 4.00 A secondary (480 A primary), dial 0.30, normal inverse curve. Upstream relay: 1.35 s at the same current. Minimum margin of the example: 0.30 s. Evaluated at 10 times pickup.
The IEC 60255-151 normal inverse curve (see inverse-time curves) is t = dial × 0.14 / ((I / Is)^0.02 − 1). At 10 times pickup, 10^0.02 = 1.0471, so the denominator is 0.0471 and the time is dial × 2.971.
- Comparison result: of 40 parameters, 38 match (38 / 40 = 95 %) and 2 differ.
- Difference 1, CT ratio: the memo says 600/5 (120) and the relay has 400/5 (80). At 480 A actual, the relay shows 4.0 × 80 = 320 A. If the pickup were given in primary amperes, the relay would translate it with 80: 480 / 80 = 6.0 A secondary, which is 6.0 × 120 = 720 A actual.
- Difference 2, dial: memo 0.30; relay 0.35.
- Time with the memo: 0.30 × 2.971 = 0.891 s.
- Time as loaded: 0.35 × 2.971 = 1.040 s. Difference: 0.149 s, about 0.15 s.
- Margin with the upstream relay: with the memo, 1.35 − 0.891 = 0.459 s (0.46 s); as loaded, 1.35 − 1.040 = 0.310 s (0.31 s). The example minimum is 0.30 s: the loaded margin is almost lost.
Both differences are corrected and the comparison is repeated: 40 of 40. The as-found and final states are archived.
How the memo is prepared so it can be compared
A calculation memo written to convince a reader is no use for comparing with a file. To compare, each parameter carries the name as the relay software shows it, the unit (stating whether primary or secondary), the value and where it was calculated. If the memo uses primary amperes and the relay secondary amperes, the conversion is written in the same table, with the CT ratio. Without this, every apparent difference forces an argument about whether it is an error or a matter of units.
Parameters the study does not calculate
A modern relay has dozens of parameters the study does not touch: language, texts, filters, factory values. The memo must say what is done with them: they are left at the factory value, or set to a company standard. If it does not say, the comparison fills up with unimportant differences and the important ones get lost among them.
After energizing
The final reading is repeated after energization and after any later intervention, and archived with the date, together with the backup of the settings of the device. Comparing the final state of the commissioning with that of a future maintenance is the quickest way to know whether someone changed something.
Who signs
The comparison is signed by whoever did it and whoever reviewed it, and approved by whoever signed the memo. Three signatures seem a lot for a table, but each covers a different error: transcription, review and judgment. In small projects they can be two people, never just one.
What not to do
- Do not compare against a screenshot: use the full file.
- Do not accept a difference without a written reason.
- Do not accept a memo without version and date.
- Do not forget to restore settings that were touched for testing.
- Do not close commissioning without the signed final state.
Checklist
- Approved calculation memo, with version and date.
- File read from the relay before changing anything and saved as the as-found state.
- Complete comparison of the parameters, not only those of the study.
- Differences classified: error, admissible with reason, change not updated.
- Errors corrected and file read again.
- Active settings group verified.
- Enabled and disabled functions reviewed.
- Comparison signed and archived in the dossier.
Related guides
- Protection settings memo
- Settings record: as-left and as-found
- Backup and recovery of IED settings
- IEC and IEEE inverse-time curves
- Coordination time interval (CTI) between protections
- Relay without documented settings
References
- IEC 60255-151, Measuring relays and protection equipment, Part 151: Functional requirements for over/under current protection (normal inverse curve equation, constants 0.14 and 0.02).
Need these tests?
The calculation memo and the comparison with the loaded settings are part of our engineering design and studies service, for projects in any country. Contact us.
Frequently asked questions
If the relay passed the functional tests, why compare the settings?
Because functional tests check that the relay operates with the settings it has, not that those settings are the study's. A wrong dial or a badly loaded CT ratio passes a functional test and fails at the first real fault. That is why comparison with the calculation memo is a separate test.
What is compared: the relay screen or the file?
The file read from the relay with the manufacturer's software, because it covers all parameters, including those not visible on the screen. It is saved as the as-found state and, if corrections are made, saved again as the final state.
Who approves a difference that is not an error?
Whoever signed the calculation memo, or whoever is responsible for the study, in writing. Admissible differences, for example a value rounded to the step the relay allows, are documented with their reason; the rest are corrected.