In short. A load bank is sized from what the test must prove, not from the generator nameplate alone. To prove full engine output the resistive capacity must equal the generator’s kW rating; to prove the alternator and AVR at rated kVA a reactive section is needed to pull the power factor to 0.8. Steps should be fine enough to reproduce the load profile being checked. Cables, protection and heat rejection at site are part of the sizing.
Start with the question the test must answer
Load bank tests are run for different reasons, and each implies a different bank. A commissioning test proves that a new set delivers rated output and holds voltage and frequency within ISO 8528 limits through specified steps. An annual test on a standby set burns off carbon from light running and confirms that the cooling system can reject full heat in summer. A post-overhaul test confirms rated power without abnormal temperatures or pressures. A fault investigation reproduces the load at which a trip occurred. The first and third need the full kW rating and a step sequence; the second needs sustained output at a high percentage of rating for an hour or more; the fourth may need a modest bank with fine steps. Deciding this first avoids hiring a 1,000 kW bank to answer a 200 kW question. The load bank testing service page describes how the test plan is written.
The arithmetic
PkW = SkVA × PF and I = (SkVA × 1000) ÷ (√3 × VLL)
PkW is the real power the bank must absorb, SkVA is the generator’s apparent power rating, PF is the power factor at which the test is run, I is the line current in amperes and VLL is the line voltage in volts. A 500 kVA set rated at 0.8 power factor has a 400 kW engine (illustrative). A purely resistive bank set to 400 kW at 415 V draws 556 A at unity power factor and loads the engine fully, but the alternator sees only 400 kVA, 80 percent of its rating. To load the alternator to 500 kVA the bank must also absorb 300 kVAr, taking the current to 696 A at 0.8 power factor. That is the figure the test cables, the connection terminals and the site breaker must carry for the duration of the test. Generators are rated for continuous output at site conditions, so if the set is derated for altitude or temperature the test target should be the derated figure, not the nameplate.
Resistive only, or resistive and reactive
| Test objective | Bank required | Sizing basis |
|---|---|---|
| Prove engine output and cooling | Resistive | 100 percent of generator kW rating |
| Annual exercise of a standby set to prevent wet stacking | Resistive | At least 50 to 70 percent of kW for one to two hours; manufacturer data governs |
| Prove alternator, AVR and excitation at rated kVA | Resistive plus inductive | 100 percent kW and sufficient kVAr for 0.8 power factor |
| Prove paralleling and kVAr sharing | Resistive plus inductive | Total plant kW and kVAr at the sharing points of interest |
| Reproduce a UPS or capacitive load problem | Resistive plus capacitive | Specific to the fault being investigated |
| Check load acceptance and transient response | Resistive with fine steps | Step sizes matching the real load blocks, often 25 percent increments |
For most routine testing in India a resistive bank is sufficient and easier to hire and transport. A combined bank is needed when the alternator or excitation system is in question, when sets are to be paralleled, or when the installation feeds a genuinely low power factor load. The resistive and reactive load banks page explains what each type exercises in the machine.
Step sizes and the test profile
A bank’s total capacity matters less to the quality of the test than its step resolution. ISO 8528-5 defines load acceptance classes by the voltage and frequency excursions on sudden application of load, and the test usually applies steps of 25, 50, 75 and 100 percent, each held until readings stabilise, followed by rejections in reverse. For a 400 kW engine that means 100 kW steps, so a bank with 25 kW or 50 kW elements gives the flexibility to build each step and approach a limit gradually. Where the real load arrives in a known block, such as a 160 kW chiller, the test should include a step of that size from the standing load that will exist when it starts. The transient response guide covers how to interpret the dip and recovery figures recorded during these steps.
Cables, protection and connection
At 696 A the test cables are substantial: typically two or three single-core flexible cables per phase, sized for the current and run length, with voltage drop checked so the bank receives the voltage it needs to draw rated load. Resistive elements draw power in proportion to the square of the voltage, so a bank at the end of a long undersized cable under-loads the set and the test reads optimistically. The connection point must be a properly rated breaker or the generator’s outgoing terminals with the normal load isolated, never a tee onto a live busbar. The bank, cables and generator frame must be bonded into the site earthing, and the facility should not be depending on the generator during the test. The cable sizing guide provides the current ratings for temporary flexible cables.
Heat, noise and where to put the bank
Every kilowatt the bank absorbs is released as heat into the surrounding air. A 400 kW resistive bank discharges roughly the heat of forty domestic room heaters through a fan that needs free intake and discharge. Inside a generator room it recirculates hot air into the radiator and the set derates or overheats, which invalidates the result. The bank should stand outdoors, downwind of the generator intake, with a clear exhaust path and clearance from anything heat-sensitive. In a Hyderabad summer a midday test adds the bank’s heat to a 42 degree ambient, and the cooling margin recorded is the real worst case, which is useful. Fan noise adds to generator noise and matters near hospitals and housing.
Related photographs and videos
Common mistakes
- Hiring a bank matched to the generator kVA rather than kW, so a resistive bank is oversized or a combined bank is specified when a resistive one would do.
- Undersizing the test cables, so voltage drop reduces the power the bank actually draws and the set is never loaded to rating.
- Running the bank inside the generator room, so recirculated heat derates the set and the test reports a cooling problem that is really a test setup problem.
- Bypassing a protective shutdown to finish a test, which converts a diagnostic into an engine failure.
What to send us
Send the generator nameplate photograph with kVA, kW, voltage and power factor, the purpose of the test and any step profile you need to reproduce, the location where the bank can stand with a photograph of the connection point and the cable route length, and whether the facility must stay on supply during the test. Use the generator and load bank rental enquiry and we will size the bank, cables and test plan.
Questions engineers ask
What size load bank do I need for a 500 kVA generator?
A 400 kW resistive bank loads the engine fully, since 500 kVA at 0.8 power factor is 400 kW. To load the alternator to its full 500 kVA you need an additional 300 kVAr of inductive capacity. For a routine annual exercise a 250 to 300 kW resistive bank run for an hour is usually enough to clear light-load deposits.
How long should a load bank test last?
A commissioning test typically holds each step for 15 to 30 minutes and runs at full load for at least an hour so that coolant and oil temperatures stabilise. An annual exercise test commonly runs one to two hours at 50 percent load or more. The manufacturer’s recommendation for the specific engine should take precedence.
Can I test a generator using the building load instead of a load bank?
Only partially. Building load is rarely at rated output, cannot be stepped on command, and varies during the test. It is adequate for confirming that the set starts and transfers, but not for proving rated output, cooling capacity or load acceptance. A load bank is the only way to apply a known, repeatable load.
Why does a resistive bank not load the alternator fully?
At unity power factor the kVA equals the kW, so a 400 kW resistive load is only 400 kVA on a 500 kVA alternator. The alternator current is 80 percent of rated and the excitation system is not working at its rated duty. Reactive load is needed to bring the current and excitation up to the nameplate condition.
Technical content reviewed by Srinivas Reddy, Managing Partner, Reddy Generators. Last updated 19 October 2026.