On site, across the UAE
Load bank testing & load test reports
A monthly off-load test start proves your generator will start. It does not prove it will carry the building. This is how you find out — and how you get a document that says so.
What it is
A device that turns your generator's output into heat
So the set can be fully loaded without the building's actual load being involved. Which type you need depends entirely on what you are trying to prove.
Resistive
A purely resistive load at unity power factor. It loads the engine fully, which proves fuel, air, cooling and the governor. This is the type used for the great majority of standby generator testing.
Reactive
Adds an inductive component so the set can be tested at a lagging power factor, typically 0.8 — which is what actually loads the alternator and its excitation, because real building loads are inductive.
Resistive/reactive combined
Does both. What you want when the acceptance criterion is written in kVA at 0.8 power factor, because a resistive-only test at full kW does not prove the alternator can deliver kVAr.
Matched to your criterion
The rig is specified against your set's rating and the standard you are testing to — not against what happens to be free. Where a criterion is one the equipment cannot honestly meet, we say so before the visit rather than quietly testing to something easier and reporting a pass.
The connection is planned before anyone arrives. How the load bank lands on your set — at the alternator terminals or the outgoing breaker, with the cable route and the hot-air discharge made safe — is a site question, and getting it wrong on the day turns a half-day test into a wasted one. It is agreed in advance from photographs of your terminal box and the space around the set. And the readings are recorded independently of the controller. A set's own panel is one instrument, calibrated at some unknown point in its life; a test that transcribes it has proved only that the panel is self-consistent.
Why standby sets need this
Wet stacking
A diesel engine is designed to run hot enough to burn all the fuel it injects. Run it at a small fraction of its rating for long enough and cylinder and exhaust temperatures never reach complete combustion. Unburnt fuel and soot then accumulate — on the cylinder walls, past the rings, around the valves, on the turbocharger and through the exhaust. You can usually see it before you measure it: black oily weeping at exhaust joints, carbon at the stack outlet, and a dirty, smoky run-up.
NFPA 110 addresses it directly: if a set does not reach at least 30 % of nameplate kW — or the manufacturer's minimum exhaust gas temperature — during its routine monthly exercise, an annual load bank test is required. Most standby sets in commercial buildings do not reach 30 %, which is why most standby sets need one.
A load bank test catches deposits early. It does not repair an engine light-loaded for a decade — at that point the set needs an overhaul, and the load test becomes the proof afterwards rather than the fix.
- What it costs you
Glazed bores and stuck rings, so lost compression and rising oil consumption. Carboned injectors and poor spray pattern. Turbocharger deposits and reduced boost.
- And then the irony
An engine that is progressively less able to take the very load it exists to carry.
- It is not unusual
It is the normal outcome for a standby set that only ever sees a small office load during its monthly exercise.
Step-loading is the point
The test procedure
A single jump to full load tells you far less than a staged climb with readings at every step, each held long enough for temperatures to stabilise.
- Pre-test inspectionOil, coolant, fuel quantity for the full duration, belts, filters, battery, exhaust condition, and the alternator's insulation resistance. A set that is not fit to run is not loaded — that is how load bank tests turn into breakdowns.
- Safety and isolationThe set separated from the site load and the changeover, the load bank connected at the alternator terminals or the outgoing breaker, cable run and exhaust discharge made safe, permit-to-work raised where the site requires it.
- Start and warm throughTo normal operating temperature before load is applied.
- Step loadA typical proving profile steps 25 %, 50 %, 75 % and 100 % of the set's rating, holding each step to stabilise. The current NFPA 110 annual test for standby sets is 50 % for 30 minutes then 75 % for one hour; earlier editions used 25/50/75 %, and the edition your authority enforces governs. Manufacturer and client criteria often specify their own profile, including 110 % overload for a defined period on sets rated for it.
- Readings at every stepVoltage and current on all three phases, frequency, power factor, kW and kVA, oil pressure, coolant temperature, exhaust temperature where a probe is fitted, battery and charging voltage, and ambient — plus what instruments do not capture: smoke, noise, leaks, vibration.
- Load acceptance and rejectionHow far voltage and frequency dip when a block of load is applied, how quickly they recover, and how far they overshoot when load is removed. ISO 8528-5 classifies this as performance classes G1 to G4, G3 being the tight class specified for critical loads such as hospitals and data centres. If your set was bought to a G-class, this is the test that shows whether it still meets it.
- Controlled unload and cool-downLoad off in steps, run down at no load to bring temperatures down, then stop.
- Reconnect and restore to AutoAnd confirm it is in Auto before leaving. A set left in Manual after a test is a standby set that will not start.
The deliverable
The load test report
It should be usable by someone who was not there.
- Set identification: make, model, rating, engine and alternator serial numbers, running hours
- Test date, ambient temperature, personnel, load bank used
- A table of readings at every load step, and the hold time at each
- Load acceptance and rejection behaviour if tested
- Observations — smoke, leaks, noise, alarms raised
- Pass or fail against the criterion used, with the criterion named
- Defects found, with recommendations and priorities
Our reporting sits inside an ISO 9001:2015 certified quality system. For a customer that is not a badge on a letterhead: it means the report is produced to a controlled procedure, that the readings on it are recorded rather than remembered, and that the document can be traced back to the test that produced it. If you are handing it to an insurer, an auditor or an FM client, that is the property that matters.
Ambient temperature and derating
Generator ratings are quoted at reference conditions. Real UAE conditions are hotter, and every manufacturer publishes a derating that reduces available output as inlet air temperature rises. So test at the temperature you actually operate at — a load test done at 06:00 in January flatters the cooling system — and record the ambient in the report, so the numbers can be compared honestly year on year. Derating is taken from the engine and alternator makers' own published curves for that machine, not from a general rule. A load test report without an ambient temperature on it is not comparable with anything.
Triggers
When you need a load bank test
Common questions
Load testing, answered
How long does a generator load bank test take?
The test itself follows the profile — an NFPA 110 style annual test is 90 minutes under load, plus warm-up and cool-down. Allow additional time for rigging, connecting cables and pre-test checks. A straightforward single-set test is normally a half-day on site. Tell us the rating, the site and the standard you are testing to and we give you the window to book before the visit, so shutdown coordination can be arranged around it.
Does the building have to be shut down?
No — that is the whole point. The load bank replaces the building's load, so the site stays on the mains throughout and the test is completely independent of what the building is doing. The only coordination needed is around the changeover and the cable route.
What percentage load should a generator be tested at?
To prove capacity, up to 100 % of the set's rating. To satisfy the current NFPA 110 annual requirement for a standby set, 50 % for 30 minutes then 75 % for one hour. To prevent wet stacking, the working threshold is at least 30 % of nameplate kW during routine running. If your specification or your insurer names a different profile, we test to that and record which one was used.
What is wet stacking, in one sentence?
Unburnt fuel and carbon accumulating in the cylinders and exhaust because the engine has been run too lightly for too long to reach proper combustion temperature.
Can load bank testing damage my generator?
Not if the set is inspected first. Loading a machine that is already faulty — low on coolant, restricted on fuel, with a slipping belt or a marginal winding — can turn a latent fault into a failure. That is why the pre-test inspection exists, and why we will decline to load a set that should not be loaded until it is put right. Everything found is documented either way.
Do you do load testing as part of maintenance?
Yes. It is the C-level item on generator maintenance and preventive servicing, and it is included at a defined frequency on a generator AMC.
Arrange a load test
Send the set's make, model, rating, location and the standard or acceptance criterion you need to meet. Photographs of the alternator terminal box and the space around the set let us plan the connection before the visit rather than on the day. You get back the rig proposed, the profile we will test to, and the window to book.