One 2,000 kVA generator or four 500 kVA sets in parallel? For a growing share of data centres, factories and microgrids, the answer is parallel — better fuel economy at partial load, N+1 redundancy, and capacity that scales with the project. But paralleling only works when synchronisation and load sharing are engineered properly. Here is how it works and when it pays.

Why parallel instead of one big set

The four synchronisation conditions

Before a generator’s breaker may close onto a live bus, its output must match the bus on four parameters simultaneously:

ParameterTypical toleranceWhat happens if violated
Voltage magnitude±5%Reactive current surge between machines
Frequency±0.2 HzActive power surge; mechanical shock
Phase angle±10° (aim <5°)Severe torque transient — can shear couplings and damage windings
Phase sequenceMust be identicalCatastrophic fault on breaker close

Modern paralleling controllers (Deep Sea, ComAp, Woodward, SmartGen) automate the whole sequence: they trim engine speed and AVR excitation until conditions are met, close the breaker at the synchroscope’s zero point, then hand over to load-sharing control. Manual synchronising with lamps and a synchroscope is legacy practice — specify auto-sync on any new plant.

Load sharing: droop vs isochronous

Reactive (kVAr) sharing is handled in parallel by the AVRs, using voltage droop or cross-current compensation — unbalanced kVAr sharing overheats one alternator while the others loaf, so commissioning must verify both kW and kVAr balance under real load.

Design requirements checklist

Worked economics: 4 × 500 kVA vs 1 × 2,000 kVA

Site with 400 kW night load, 1,100 kW day peak, 16 h/day operation:

Econo Solar supplies paralleling-ready gensets across all eight engine series — Cummins (ECC), Perkins (ECP), Weichai (ECW) and more — with auto-sync controllers, motorised breakers and load-share wiring configured at the factory. Send us your load profile for a paralleled-plant proposal.