Generator and ATS Restoration Test: Prove the Factory Sequence, Not Just Start-Up

A generator starting with no load does not prove that a factory will restore safely. The acceptance test must demonstrate the complete chain: source-loss detection, start command, voltage and frequency readiness, transfer, staged loading, alarms, failed conditions, return transfer and cooldown.

This guide gives owners and EPC teams a test structure for Vietnam factory projects. Example settings and loads are hypothetical. The approved design, generator and ATS manufacturer limits, life-safety strategy, utility conditions and current QCVN/TCVN govern the real test.

1. Define the test boundary

Mark on the SLD the normal source, generator, ATS/TSE, emergency or essential bus, load-shedding devices, UPS/BESS interfaces and downstream panels. State whether the test begins at a simulated voltage input, an opened normal-source breaker or an actual upstream outage.

List excluded loads and temporary operating restrictions. A safe simulated input test and a full integrated power test provide different evidence.

2. Build a sequence-of-events table

Milestone Evidence
Normal source unacceptable Measured voltage/frequency and controller timestamp
Start command ATS output and generator input
Generator ready Voltage, frequency, ready contact
Normal source opened Command and auxiliary contact
Emergency source closed Command, position and bus voltage
Load stages added kW, kvar, current, voltage, frequency
Normal source stable Return timer start and completion
Cooldown/stop Unload, timer and stop state

Use one synchronized time source or record offsets. Controller screenshots alone cannot correlate breaker, generator and power-quality events.

3. Establish the restoration envelope

Example only: source detection delay 2.0 s, generator ready in 10.0 s after start, and transfer 1.5 s after readiness. The electrical restoration is 2.0 + 10.0 + 1.5 = 13.5 s. Compare that result with each load’s permitted interruption.

Keep safety-service requirements separate from production-continuity targets. IEC 60364-5-56:2018 addresses supplies for safety services; the project must state its adopted requirements and authority interfaces.

4. Test realistic load steps

For a hypothetical 1,000 kW rated generator, a sequence of 400 kW, then 250 kW, then 150 kW reaches 40%, 65% and 80% of rating. The point is not those percentages; it is to reproduce the approved motor-start and essential-load sequence.

  • record minimum voltage and frequency after each step;
  • record recovery time to the project band;
  • confirm engine, alternator and controller limits;
  • verify load-shed priority and blocked starts;
  • repeat the most onerous credible sequence where permitted.

Apparent power, power factor, harmonics and unbalance may govern even when kW looks acceptable.

5. Test ATS/TSE functions

IEC 60947-6-1:2026 covers transfer switching equipment within its scope and includes requirements for several transfer arrangements. The project test should identify the actual TSE class, transition type, controller, bypass/isolation arrangement and OEM procedures.

  • normal-source undervoltage and phase loss;
  • source-available criteria and timers;
  • mechanical/electrical interlocks;
  • dead-source and live-source blocking;
  • manual mode and local/remote selection;
  • position indication and alarm mapping;
  • return-transfer and cooldown logic.

6. Prove failure modes

Injected condition Expected safe outcome
Failed generator start No close to unavailable source; alarm and escalation
Generator ready contact stuck Independent voltage/frequency permissive blocks unsafe close
Emergency breaker fails to close Alarm, no false “bus restored” indication
Normal breaker fails to open Interlock blocks unintended paralleling
Controller communication loss Documented local fallback state
Emergency stop/lockout Trip, alarm and controlled reset
Normal source returns briefly Stability timer prevents hunting

Use simulations or secondary injection where a live failure would create unacceptable risk. State which evidence is simulated and which is end-to-end.

7. Protect people and production

Issue a switching program, risk assessment, permit boundary and communication plan. Confirm fire systems, emergency lighting, lifts, process hazards, stored energy and restart restrictions. Assign one test director with stop authority and one person responsible for each source and process interface.

8. BOQ and bidder response

Item Required basis
FAT Controller logic, I/O, alarms, simulated sequences
Pre-commissioning Cables, settings, phase sequence, interlocks
Load test Load bank or facility load, duration and fuel
Integrated SAT Outage, transfer, steps, failures and restoration
Records SOE export, PQ trend, settings, signed sheets
Training Normal, manual, alarm and recovery scenarios

9. ITP gates

  • HP-01: cause-and-effect and load-step schedule approved.
  • HP-02: settings, protection and interlocks verified.
  • HP-03: fuel, ventilation, exhaust, cooling and starting systems accepted.
  • HP-04: no-load and simulated logic tests passed.
  • HP-05: load test passed with calibrated instruments.
  • HP-06: integrated restoration and failure matrix witnessed and closed.

Owner handover checklist

  • Do all event records use a known time base?
  • Is bus restoration distinguished from process restart?
  • Were load steps representative of the design case?
  • Were failed start and rejected close demonstrated?
  • Are settings and logic exports revision controlled?
  • Are fuel autonomy and replenishment assumptions documented?
  • Can operators recover from manual and alarm states?

10. Close findings by requirement, not by screenshot

Use a test register with requirement ID, scenario, initial state, command, expected result, measured result, raw-file reference, witness and disposition. A pass should identify the accepted tolerance and the design document that created it. If a timer, setting or wire changes after a failed test, record the change and repeat every affected scenario rather than replacing the earlier evidence.

Retain controller and relay event exports, power-quality trends, breaker operation counts, generator parameters, fuel level, ambient conditions and calibrated-instrument files. The signed summary should point to these records. This makes later maintenance tests comparable and prevents an unexplained commissioning value becoming the operating baseline.

Frequently asked questions

Is a monthly auto-start test enough?

No. It confirms a limited function and may not prove transfer, load response, downstream dependencies or failure handling.

Must the whole factory lose power during SAT?

Not always. Select safe staged tests and disclose their boundaries. Critical end-to-end claims need corresponding evidence.

Can the generator be sized only from connected kW?

No. Load steps, motor starting, kvar, harmonics, unbalance, derating and future duty matter. See our generator sizing guide.

What proves the ATS transfer time?

A synchronized event record showing source condition, commands, breaker states and bus voltage, not a stopwatch alone.

Prepare the integrated test

Quang Anh can help owners turn the SLD, essential-load schedule and cause-and-effect into a FAT/SAT matrix, switching program, ITP, record sheets and handover package.

Source register

[contact-form-7 id="7239967" title="Contact form 1"]

Related Posts