A load bank test applies a controlled artificial load to a diesel generator so the set can be exercised at or near its rated output without depending on the building load. It is done to prove the machine can carry its rating, to verify cooling, governing and exhaust behaviour under real thermal conditions, and to burn off the unburned fuel and carbon deposits that accumulate when a set runs persistently at light load. The test is as much a diagnostic as a compliance exercise — the readings taken during it usually reveal more than a no-load run ever will.
Why Load Bank Testing Is Done
Standby generators spend most of their life idle, and when they do run it is often at a small fraction of rating. That operating pattern causes its own problems.
- Wet stacking. Prolonged light-load running leaves unburned fuel and soot in the exhaust tract and can cause oil dilution. Loading the set properly raises exhaust temperature and clears the deposits.
- Verification of rating. A commissioning test proves the set delivers its stated kW and kVA before the owner accepts it.
- Post-repair confirmation. After an engine, injector or alternator repair, a load test confirms the work under thermal load rather than at idle.
- Periodic assurance. Cooling systems degrade, radiators foul, and fuel quality drifts. A periodic test detects the decline before it causes a failure.
- Compliance. Emergency power systems are subject to a defined testing regime in many jurisdictions, and records of those tests are part of the evidence.
Resistive vs Reactive Load Banks
The two types stress different parts of the machine, and the choice depends on what you need to prove.
For correcting wet stacking, a resistive bank is sufficient — the objective is exhaust temperature and engine loading. For proving an installation, a combined bank is the honest test, because it loads the alternator and the distribution equipment as well as the engine.
Test Sequence and What to Log
A typical sequence steps the load up, holds at full load long enough for temperatures to stabilise, and then unloads. Specific durations depend on the standard applied, the manufacturer's recommendation and the purpose of the test, so treat the following as a structure rather than a prescription.
| Step | Load level | Purpose | Readings to record |
|---|---|---|---|
| 1 | 0–25% | Warm-up and stabilisation | Oil pressure, coolant temperature, voltage, frequency |
| 2 | 50% | Intermediate check | Exhaust temperature, current per phase, smoke observation |
| 3 | 75% | Approach to rated load | Coolant and oil temperature trend, fuel pressure |
| 4 | 100% | Rated-load hold | All temperatures and pressures, voltage and frequency under steady load, current balance |
| 5 | Step application and rejection | Transient response | Voltage dip and recovery, frequency dip and recovery, recovery time |
| 6 | Cool-down run | Turbo and exhaust protection | Temperatures falling, no alarms |
Where a set is rated for a short-duration overload, a brief period above 100% may be included, but only if the manufacturer's rating and the test standard allow it. Do not improvise an overload step.
Acceptance Criteria: What "Pass" Looks Like
A test passes when the set holds its rated output within stated tolerances and nothing degrades or trips during the hold. The checks below are the usual ones:
- Voltage and frequency stay within the regulation band stated for the machine and remain steady, not merely within limits momentarily.
- Current per phase is balanced within the alternator's stated limit, and no phase exceeds rating.
- Coolant temperature, oil temperature and oil pressure remain within the manufacturer's design range at full load.
- Exhaust temperature stabilises rather than continuing to climb, and visible smoke clears after warm-up.
- The governor and AVR recover acceptably from step load application and rejection.
- No protective device trips during the hold, other than one deliberately being tested.
- No fuel, oil or coolant leaks appear under thermal load and vibration.
- Exhaust backpressure and cooling air temperature rise are within design limits.
The most useful single indicator is often the trend rather than the absolute value. A coolant temperature that is legal but has risen several degrees since the previous test is telling you something about the radiator or the coolant condition.
Common Findings and What They Usually Mean
| Finding | Likely cause | Next check |
|---|---|---|
| Coolant temperature climbs through the hold | Radiator fouling, low coolant level, degraded coolant, or a thermostat fault | Airflow obstruction, core condition, coolant concentration |
| Exhaust temperature high at moderate load | Restricted air intake, turbocharger issue, or over-fuelling | Filter condition, intake ducting, boost pressure |
| Voltage sags progressively | AVR or excitation problem, or alternator heating | Excitation type, AVR sensing, winding temperature |
| Frequency unstable under load | Governor adjustment, fuel supply restriction, or air in the fuel | Fuel filters, governor settings, fuel line integrity |
| Heavy smoke at load that persists after warm-up | Injector condition, air restriction, or incorrect fuel timing | Injector test, air filter, timing |
| One phase current noticeably higher | Unbalanced load or a connection problem | Load distribution and terminal tightness |
| Cannot reach rated load | Fuel supply limitation, derating for site conditions, or an engine issue | Fuel flow, altitude and temperature derating |
Site Preparation, Safety and Records
A load bank test puts the machine under real thermal and electrical stress, so the site has to be ready for it.
- Confirm ventilation and that radiator or remote cooling discharge air has somewhere to go, since recirculation will invalidate the test.
- Route exhaust safely and check that no combustible material is near hot surfaces.
- Size load bank cables and connectors for full-load current and verify terminations before energising.
- Position the load bank so its own heat discharge does not affect the generator or nearby equipment.
- Confirm earthing and that the changeover or isolation arrangement prevents back-feeding the building or the utility.
- Brief personnel and keep a fire extinguisher to hand — the set is running hot and under load.
Isolation is the item that causes accidents. Before testing, confirm the set is positively isolated from the normal supply and that no parallel path exists through the transfer equipment.
The record of the test is only useful if the next one can be compared against it. Record the date, running hours, ambient temperature and altitude, load steps and durations, all readings at each step, any alarms, and the corrective actions taken. Photographs of the installation and copies of the load bank calibration certificate are worth keeping alongside the log.
Over successive tests, trend data supports maintenance decisions in a way that a single pass/fail result cannot. Our diesel generator maintenance guide sets out how periodic testing fits into the wider maintenance plan, and the distinction between prime and standby ratings determines which rating the test should be measured against.
FAQ
How often should a load bank test be done?
The interval depends on the applicable standard, the manufacturer's guidance and how the set is used. Emergency power systems have defined testing requirements in many jurisdictions. Confirm the interval against the edition adopted in your jurisdiction and record the basis for the decision.
Is running the building load an acceptable substitute?
It can be, if the building load genuinely brings the set to a sufficient level and can be held there for the required period. Most sites cannot guarantee that, which is why a load bank is specified.
Can load bank testing damage the generator?
Not when it is done correctly and within the machine's rating. The risks come from improvised overload steps, poor ventilation, undersized cabling and inadequate isolation.
Do I need a reactive load bank?
For wet-stacking correction, no — resistive loading is what raises exhaust temperature. For proving an installation at a stated power factor, including the alternator and switchgear, yes.
What load level is enough to prevent wet stacking?
Manufacturers publish guidance on minimum sustained loading, and the figure varies by engine design and rating. Follow the guidance for your specific engine rather than a general rule of thumb.
Reference Sources
- NFPA 110 — Standard for Emergency and Standby Power Systems
- ISO 8528-5:2025 — Generating sets, design and performance criteria
- ISO 8528-1:2018 — Application, ratings and performance of generating sets
- EGSA — Electrical Generating Systems Association
Testing intervals, load levels and acceptance limits must be confirmed against the standard edition adopted in your jurisdiction and the generator manufacturer's published guidance.
