Testing the control system starts from the substation signal list and verifies each signal end to end: the command from the screen to the coil, the status from the contact to the screen, and every alarm with its text and priority. A worked example with 148 assumed signals adds up to 422 min (7.03 h) and, with 5 % rejects, about 23 min of retesting.
Short answer
Before energization, the control-system test proves that what the operator sees and commands on the screen (HMI) matches what happens at the equipment. It starts from the substation’s signal list, which enumerates every command, status and alarm, and verifies each one end to end. The interlocks that allow or block each operation are tested too. It is a long, repetitive test, and its value is that an error found here, with the yard de-energized, puts no one at risk.
The signal list as the contract of the test
Each row says what the signal is, which device it comes from, what text it carries on the screen, which protocol carries it, where it is sent (for example to the control center) and its priority. It is the document the test is run against, so it must be approved before starting; if it changes during testing, the affected signals are tested again. The link to the control center is covered in SCADA and PMU signals for the grid operator.
Three types of test
Status. The equipment is taken to each position (open, closed, and the intermediate or fault position if there is one) and the screen, the event log and the transmission to the control center are checked to show the same thing, with the right time.
Commands. The command is executed from every point it can be given from (local screen, panel, control center) and the equipment is checked to operate and the status to confirm. A command has a chain: the order, the condition check, the output to the equipment and the position confirmation. Two-step select-before-operate commands are covered in commands in IEC 101 and 104 and IEC 61850 MMS.
Alarms. The condition is produced at its source and the alarm is checked to appear with text, priority and time, to be acknowledgeable and to be logged. The design and hierarchy of alarms are in HMI: screen design and alarm management. Testing the alarms of a conventional panel is described in annunciator and panel alarm test.
Interlocks: what the command must not allow
A correct command, executed under the wrong condition, is dangerous. So, for each operation, the test covers the condition in which it must be allowed and the condition in which it must be refused. Examples of what is verified:
- A disconnector neither opens nor closes while the breaker of its bay is closed.
- An earthing switch does not close if voltage is present.
- A breaker close is refused if the device’s operating mode does not allow it (for example, in maintenance or local mode).
- If communication is lost with the device that supplies the condition, the interlock goes to the defined safe state.
An interlock test has two halves: it permits when it should and it refuses when it should. One half alone leaves a risk unverified.
Worked example: time of the point-to-point test
Assumptions. A substation with 148 signals on the list: 40 commands, 62 statuses and 46 alarms. Average times per signal (illustrative assumptions): 4 min per command, 2 min per status and 3 min per alarm. Assume 5 % of the signals fail the first test and must be repeated.
- Commands: 40 × 4 = 160 min.
- Statuses: 62 × 2 = 124 min.
- Alarms: 46 × 3 = 138 min.
- First pass total: 160 + 124 + 138 = 422 min = 7.03 h.
- Average time per signal: 422 / 148 = 2.85 min.
- Signals that fail: 5 % of 148 = 7.4, that is, 8 signals rounding up.
- Retest: 8 × 2.85 = 22.8 min, about 23 more minutes, not counting the time to fix the cause.
Fixing the cause (a crossed wire, a misspelled text, a wrong address in the configuration) usually takes longer than repeating the test. That is why the list and the settings should be reviewed before the crew arrives.
Recommended order of the test
It is best to test in this order. First the statuses, because everything else is confirmed through them. Second the commands from the lowest level (the bay panel) up to the highest (the substation screen and the control center): if one fails, it is clear at which level the problem lies. Third the interlocks, with the equipment in the conditions that trigger them. Fourth the alarms, grouped by origin so as not to repeat operations. Fifth, the test of behavior under system failures: loss of auxiliary supply, a dropped link and a device restart, which must leave the system in a safe state and with the right information on return.
What evidence is left
Each row of the signal list ends with a result: passed, passed with remark or rejected, with date, responsible person and, if there is a remark, the agreed action. An event log exported from the screen, with the times of each operation, serves as backup. The evidence for rejected and repeated signals must keep the first result, not replace it: that shows what was corrected. These documents are part of the commissioning dossier and are what the client or the operator reviews before energization.
Common mistakes
- Testing the command but not its status confirmation.
- Testing only from the local screen and forgetting the control center.
- Accepting an alarm because it “shows up”, without checking text, priority and time.
- Testing interlocks in one direction only.
- Leaving no signal-by-signal evidence: without it there is no test, only a claim.
- Changing the signal list during the test without repeating the affected signals.
Checklist
- Approved signal list, with its version noted in the protocol.
- Each status tested in all its positions, with the right time.
- Each command tested from all command points, with status confirmation.
- Interlocks tested in the permitting condition and in the refusing one.
- Safe behavior on loss of communication verified.
- Each alarm verified for text, priority, acknowledgement and log.
- Signals sent to the control center confirmed with its operator.
- Result recorded signal by signal, with date and responsible person.
- Failed signals corrected and repeated.
Related guides
- HMI: screen design and alarm management
- Annunciator and panel alarm test
- SCADA, SAS, DCS and HMI: differences in the power sector
- Commands in IEC 101 and 104: direct or select-before-operate
- SCADA and PMU signals for the grid operator: input tests
- What is checked before energizing a substation
References
- IEC 61850, communication networks and systems for power utility automation (control model with select-before-operate, and messages between devices).
Need these tests?
We design the signal list, the interlock logic and the acceptance criteria for the control-system tests as part of our engineering design and studies service, for projects in any country. Contact us.
Frequently asked questions
What is tested first, the command or the status signal?
The status. Without a reliable position on the screen there is no way to know whether a command worked. The breaker or disconnector status is verified in both positions first, and then the commands are tested, confirmed by that status.
Must commands be sent to the real yard equipment?
It depends on the phase. With the primary equipment still de-energized it can be really operated: each command opens and closes the breaker. Once equipment is energized, the test is done with the breaker in test position or with the trip circuit isolated, as the site safety procedure requires.
How are alarms tested without causing faults?
By simulating the condition at its source: opening the contact, lowering a level, removing a supply or forcing the signal with the test set. Each alarm must appear with its text, priority and time, and must be acknowledgeable and logged.