11.4 Generator Paralleling & Emergency System Commissioning

Generator Paralleling Functional Test

Paralleling two generators — or a generator and the utility — requires matching four quantities at the instant of closing: voltage magnitude, frequency, phase angle, and phase rotation. Rotation must be verified once at commissioning and cannot be corrected by the synchronizer; the other three are what the synchronizing scheme continuously manages. The automatic synchronizing relay (device 25) and its permissive logic must be function-tested before the paralleling breaker is closed under load.

• Voltage — typically within ±5 percent of nominal.
• Frequency — typically within about 1 Hz.
• Phase angle — typically within about ±10 degrees.
• Phase rotation — must match between sources; wrong rotation will trip the 25 relay or, worse, slam motors into reverse.

Before any hot parallel, function-test the ANSI 25 synchronism-check relay, the automatic synchronizer (ANSI 15) if fitted, and the permissive logic. Then verify load sharing. Real power (kW) is controlled by the governor (frequency and droop); reactive power (var) is controlled by the exciter/AVR (voltage and reactive drop). A unit that takes unequal kW has a governor droop mismatch; a unit that takes unequal vars has an AVR/reactive-drop mismatch. Test reverse-power (ANSI 32) protection by simulating reverse power and confirming the breaker trips. Test var control by introducing a reactive load step and watching the AVRs share it within their design tolerance (typically ±5 percent of rated kW or kvar).

The Paralleling Test Sequence

1. Single-unit test first. Run each generator individually through its ATS and NFPA 110 load test before paralleling.
2. Synchronizer and 25 relay functional test. With both sources at rated voltage, verify the 25 relay closes its permissive contact only within the synchronism window; verify it blocks when phase, frequency, or voltage is out.
3. First parallel. Close the paralleling breaker on permission; monitor kW and var sharing immediately.
4. Load steps. Apply 25 percent, 50 percent, 75 percent, 100 percent load steps and confirm each unit shares within its design tolerance.
5. Reverse-power trip. Simulate reverse power; verify ANSI 32 trips the breaker and the unit isolates cleanly.
6. Transfer back and shutdown. Offload, open the paralleling breaker, run the engine cooldown, then shutdown.

Emergency System Commissioning Sequence

Commissioning an emergency power system goes from a dead, de-energized state to a fully integrated, load-tested system. The sequence:

1. Pre-energization inspection and component tests — insulation resistance, TTR, Ductor, CT polarity, relay calibration, grounding.
2. Control power on, logic check — battery and charger, DC distribution, alarm and annunciation points.
3. Individual equipment tests — generator dry-run, ATS functional test, switchgear functional test.
4. Source-to-source integration — first transfer with the ATS, in-phase or closed-transition verification.
5. NFPA 110 load test — monthly 30 percent / 30-minute exercise and the Level 1 4-hour load-bank acceptance run.
6. Integrated system test — full loss-of-normal simulation: transfer, load pick-up, load steps, retransfer, cooldown, with all alarms and event records captured.

A commissioning test per ANSI/NETA ECS of a new emergency power system typically includes a load-bank run, monthly transfer testing, and documented commissioning of source-transfer sequences.

Commissioning vs Acceptance Testing

Acceptance testing (ANSI/NETA ATS) is performed on new equipment before energization, to prove each component meets factory and design specifications. Commissioning (ANSI/NETA ECS) is the integrated, system-level process that proves the assembled system performs its intended function — the ATS, generator, switchgear, and loads all work together. You can pass acceptance on every component and still fail commissioning: a CT polarity error that only shows up under load, a 25 relay set for the wrong synchronism window, or a generator that will not share load are commissioning defects, not component defects. NETA Level 2 candidates should be able to distinguish the two scopes on exam day.

Typical Punch-List Items

• Relay settings not matching the coordination study (wrong tap, wrong curve, wrong pickup).
• CT polarity errors on differential or metering circuits.
• 25 synchronism-check window set too tight or too loose.
• Reverse-power (32) trip set too low, causing nuisance trips on load steps.
• Governor droop not matched between paralleled units — one hogging kW.
• AVR reactive drop not matched — units fight on var sharing.
• Engine cooldown timer too short — generator shuts down hot.
• ATS failure-sensing delay too short — nuisance starts on momentary voltage dips.
• Phase rotation mismatched between sources — motors reverse on retransfer.
• Alarm points not mapped at the HMI — operator blind to status.

Each punch-list item is assigned an owner and a completion date before the system is handed over.

Worked Example — First Parallel of Two 500 kW Generators

Two new 500 kW diesel generators are to parallel onto a 1,000 kW critical load. Single-unit NFPA 110 load tests pass on both. On the first parallel at 25 percent load (250 kW), Gen A carries 180 kW and Gen B carries 70 kW — a 110 kW mismatch, far outside the ±5 percent (±25 kW) tolerance. Because kW sharing is governed by the governor, the technician checks droop: Gen A is set to 2 percent droop, Gen B to 5 percent droop. Resetting both to 3 percent droop brings the sharing to 130 kW / 120 kW, within tolerance. On the reactive side, at 25 percent load the units show 180 kvar and 60 kvar — a var mismatch pointing to the AVR reactive-drop setting. Resetting both AVRs to the same reactive-drop brings var sharing within tolerance. Only then are the 50 percent, 75 percent, and 100 percent load steps applied, and only then is the reverse-power (32) trip simulated. This is the order: single-unit test, synchronizer and 25 relay check, first parallel with sharing corrected, load steps, then protection trip tests.

Commissioning Documentation

The commissioning record is more than the sum of the component test reports. It includes the integrated system test procedure, the step-by-step results (transfer times, load-sharing readings at each step, alarm and event logs), the punch list with owners and dates, and the final return-to-service sign-off. A commissioning report that contains only component test reports and omits the integrated test results has not demonstrated that the system works as a system — which is the entire point of commissioning.