VA-TEK Engineering There is no single correct relay test interval. DL/T 995 fixes intervals by relay technology, NERC PRC-005 sets maximum intervals by monitoring status, and NETA leaves the interval to the owner’s maintenance policy while defining what a test must cover. The defensible approach is to state which framework governs the site, then justify any deviation with monitoring coverage, criticality and service history.
Key takeaways
- China: DL/T 995-2016 gives fixed full-test and partial-test intervals for microprocessor and non-microprocessor relays.
- North America: NERC PRC-005 uses monitoring-based maximum intervals; NETA MTS does not set a universal number.
- Extensions: monitoring can justify a longer interval only where the monitored functions are defined, tested and acted upon.
What DL/T 995-2016 requires
DL/T 995 is the Chinese field standard for relay inspection. Its interval structure separates a full inspection from a partial inspection, and it treats pilot channels separately.
| Inspection type | Microprocessor relays | Non-microprocessor relays | Notes |
|---|---|---|---|
| Full inspection | Every 6 years | Every 4 years | Covers the complete 19-item field program |
| Partial inspection | Every 2 to 4 years | Every 1 year | Scope selected from the full program |
| Pilot channel | Every 6 years | Every 6 years | Channel tests follow the protection type |
| First full inspection after commissioning | Within 1 year | Within 1 year | Confirms installation and early-life stability |
The difference between 6 and 4 years is not arbitrary. Electromechanical relays wear mechanically: discs, bearings, springs and contacts change with time and operation count. Microprocessor relays have no moving parts in the measuring path, but electrolytic capacitors, memory devices and power modules age. The failure mechanisms differ, so the inspection interval differs.
The 2-to-4-year partial-inspection band is deliberately elastic. The lower end suits harsh environments, high fault duty, frequent operations or a history of defects. The upper end suits stable, lightly loaded devices with good monitoring and a clean service record. Taking the lower bound by default adds work without adding evidence; taking the upper bound without justification removes the only defence against a latent failure.
What NERC PRC-005 requires
NERC PRC-005 does not say “test every relay every X years.” It sets maximum intervals for protection system components according to how well they are monitored. The table is component-specific, so a single number cannot be applied to every relay.
| Monitoring status | Typical maximum interval pattern | What it means in practice |
|---|---|---|
| Unmonitored | Shorter interval, commonly 5 years for several protection components | No continuous supervision of the protection function; periodic proof testing carries the load |
| Partially monitored | Mid-tier, commonly 7 years | Some failures annunciate, but not every protection path is covered |
| Thoroughly monitored | Longest tier, up to 10 years for defined components | Continuous monitoring covers the defined failure modes and alarms are acted upon |
NERC’s technical reference reports survey averages of about 5 years for electromechanical or solid-state relays and about 7 years for microprocessor relays. Those averages describe practice, not a universal requirement. The PRC-005 table remains the controlling document, and its component definitions must be checked before an interval is adopted.
The important discipline is that “monitored” is not a marketing claim. It means the monitored functions are specified, the alarm path is tested, and someone responds to the alarm. A microprocessor relay with an unused self-test output is unmonitored in practice.
What NETA and manufacturers say
NETA MTS-2023 Chapter 7.9 defines the test content and the acceptance form, not a universal interval. NETA’s position is that the owner sets the maintenance policy, taking account of equipment criticality, environment, duty and history. Using the MTS checklist as a fixed annual program is a policy choice, not a NETA requirement.
Manufacturer recommendations vary. ABB, for example, recommends a 3-to-5-year test interval for the REX640 depending on environment and supervision, and a 5-year interval for the REX615 where supervision is used. These are product-specific recommendations, not replacements for the owner’s standard.
How to set an interval defensibly
| Factor | Shortens the interval | Extends the interval |
|---|---|---|
| Technology | Electromechanical or early solid-state devices | Microprocessor relay with defined self-supervision |
| Monitoring | No continuous supervision, or alarms not actioned | Thorough monitoring with tested alarm path |
| Criticality | Generator, busbar, transformer or line protection with wide impact | Non-critical auxiliary protection with redundancy |
| Environment | High temperature, humidity, vibration, corrosive atmosphere | Controlled indoor environment |
| Duty | Frequent operations or repeated faults | Rare operation, stable load |
| History | Repeat defects, drift, firmware issues | Clean test history and stable settings |
A written justification should name the governing standard, the baseline interval, the monitoring credit claimed, and the evidence for that credit. “The relay is numerical” is not evidence of monitoring. “The self-test output is wired to SCADA, the alarm is tested quarterly, and the last three tests showed no drift” is.
What resets the interval clock
A full inspection resets the full-inspection clock. A partial inspection does not. Several events should trigger an out-of-cycle test regardless of the calendar:
- Protection operation: inspect the relay, trip circuit and event record after a fault, especially where the fault current approached the CT rating.
- Firmware or software change: re-run the affected function tests and the scheme logic; a version change is a configuration change.
- Setting change: verify the new setting against the setting sheet and confirm the changed logic.
- CT or VT replacement: prove ratio, polarity and secondary wiring before returning to service.
- Repair or module replacement: repeat the tests that cover the replaced module and the interfaces it drives.
- Abnormal indication: a self-test alarm, a communication alarm or an unexplained event record deserves investigation, not deferral to the next cycle.
Frequently asked questions
How often should microprocessor relays be tested?
Under DL/T 995-2016, every 6 years for a full inspection and every 2 to 4 years for a partial inspection, with the first full inspection within one year of commissioning. Under NERC PRC-005, the interval depends on the component and its monitoring status, not on the relay being microprocessor-based.
Does NETA specify a relay test interval?
No. NETA MTS-2023 defines the test content and the acceptance criterion, which is the manufacturer’s recommended tolerance. The interval is an owner decision and should be documented in the maintenance policy.
Can continuous monitoring extend the interval?
Only where the monitoring covers the defined failure modes and the alarm is tested and actioned. A self-test that nobody reads does not reduce the risk that periodic testing addresses.
Does a partial test reset the full-test clock?
No. Partial inspection covers a selected scope; it does not substitute for the complete program. Keep the two clocks separate in the maintenance record.
Where VA-TEK test sets fit
Interval work is mostly three-phase testing: pickup, timing, characteristic points and scheme logic. The VAE-430 covers those checks with 3×0–40 A AC, 4×0–120 V AC and 8 binary inputs / 4 binary outputs. Where a periodic test includes a transformer or busbar differential scheme, the VAE-660 provides six current and six voltage channels so every end can be energized at once. The full range is listed on the relay test set hub.
For the item list behind the interval, use the commissioning checklist. For the standards that define the criteria, read relay testing standards compared. For test records that support an interval extension, see relay test planning and reporting.
Free whitepaper: the standards, test items and test-set requirements discussed in this article are consolidated in Relay Protection Testing: Standards, Test Items and Test-Set Requirements — a 29-page PDF, free to download without registration.
Standards and references
- DL/T 995-2016, Testing regulations for relaying protection and security automatic equipment, clauses 5.1.2.3 and 5.2.
- NERC, Relay Maintenance Technical Reference: nercstg.nerc.com.
- NERC PRC-005, Protection System Maintenance and Testing.
- NETA MTS-2023, Chapter 7.9; NETA: netaworld.org.
