claude code, claude code 
The relay test plan is the document that decides whether a commissioning job closes out cleanly. It states what will be tested, which settings the tests rest on, what counts as a pass, and who witnesses the result. A thin plan leaves the client’s reviewer no way to confirm that the work was right.
Most disputes on a protection job are not about the relay. They concern a settings revision nobody recorded, a test point nobody wrote down, or a trip nobody witnessed. Each is cheaper to prevent before the first injection than after energization.
What a relay test plan has to contain
A workable plan names the bays and protection functions in scope, and what is deliberately out of scope. It then points at one settings source: a dated revision of the protection settings sheet, not the latest file.
| Section | Contents | Why the reviewer asks for it |
|---|---|---|
| Scope | Bays, panels and functions covered, plus exclusions | Settles what was and was not tested |
| Settings control | The file the relay was actually loaded with | Proves the test matches the settings in service |
| Test points | Injected values, expected operation and tolerance | Lets a second engineer repeat the work |
| Acceptance criteria | The pass rule for each test, agreed in advance | Separates a failure from a judgment call |
| Record sheets | Templates for readings, timings and witnesses | The raw evidence the report is built from |
| Roles | Who tests, witnesses, signs and is notified | Avoids surprises at energization |
Keeping the settings sheet and the relay aligned
Plans fail on settings control far more often than on test technique. The relay is loaded from one revision, the plan is written against another, and the mismatch surfaces weeks later when the event record does not match the report.
The defense is simple. Freeze the settings file, record its revision in the plan, then read the settings back out of the relay after loading. A relay test plan that cannot name its revision proves nothing.
A mid-test change is normal. Approve and date it first, re-test every affected element, and write the reason beside the new readings. Unaffected points keep their original results.
Test points, acceptance criteria and record sheets
The test point schedule is where the plan becomes technical. Every element that can operate gets a row: pick-up, one or two points on the time curve, and the boundary where it must stay stable. Directional and differential elements need points on both sides of the characteristic.
Acceptance criteria belong in the plan, not the report. A tolerance chosen after the readings are known proves nothing. The plan states whether the settings sheet or the client’s specification governs.
Record sheets bridge plan and report. A useful sheet carries the date, test set identity and calibration status, injected value, measured value, expected result, verdict and initials. Worked examples of the points themselves are in overcurrent relay testing procedure.
What a test report has to prove
The report is an argument, not a data dump. It shows that the relay was tested against a known settings revision, with traceable equipment, and that somebody other than the author saw the results.
It carries the settings revision in service, the test equipment with calibration status, the raw readings, every deviation with its cause, and the retest that cleared it. Values found before adjustment are separated from values left in service, and skipped tests are listed with the reason.
Supporting evidence is cheap. An event record captured during injection shows what the relay believed it saw. On a digital substation, a network capture may be the only proof that the interface behaved, which is why digital substation testing changes the evidence list rather than the logic being proven.
Frequently asked questions
What belongs in a relay test plan?
Scope and exclusions, the settings revision the relay was loaded with, test points with tolerances, acceptance criteria, record sheet templates, and the roles for testing, witnessing and signing.
How should a settings change mid-test be handled?
Approve and date it first, re-test every element it touches, and record the reason beside the new results. Elements it does not affect keep their original readings.
What makes test records auditable?
Traceability: a named settings revision, identified equipment with valid calibration, and readings signed by the person who took them and witnessed by a second person.
Where the test set fits
Most of this discipline lives on paper, but the instrument supplies part of the evidence. The VAE-430 provides three current and four voltage channels with a built-in operating system and LCD, so tests run stand-alone and results stay with the job. The VAE-660 adds six current and six voltage channels with an embedded PC, and its binary inputs time a contact operation to 1 ms.
Both specify AC accuracy of 0.2 % or better, and that figure belongs in the report beside the measured values. Results transfer to a PC for attachment to the record sheets.
A relay test plan and the report that follows it are read together, so the plan should be written by somebody who knows what the report must contain. The instruments used to gather this evidence belong to the secondary test instruments family. To match a plan to a specific scheme, contact the team or send the scheme details with a quote request.
