Atom AI

AI built inside your CMMS to make your FM operations smarter and easier

Explore Atom Assistants
Preventive Maintenance

Generator Maintenance Checklist

A generator maintenance checklist template built around the test most standby sets never get. An off-load weekly run proves the engine starts and proves nothing at all about whether the set will carry the building, and running a diesel lightly loaded for years causes wet stacking, which is a failure mode in its own right. This records the load actually tested to, the changeover actually timed, and the fuel stock actually held.

  • Maximum load actually tested to, not the nameplate rating
  • Changeover timed and compared with what the loads survive
  • Fuel stock measured against the stated autonomy
  • Set confirmed back in automatic before anybody leaves
generator-maintenance-checklist.xlsx

Generator Maintenance Checklist

Tested on load

Per test
Load applied, kW achieved
Per transfer step
Time taken against required
Per service
Load tested to, as a percent
Checked
Fuel against stated autonomy
#LOAD TEST RECORDDURATIONLOADKW
1Off-load run
2Building load test
3Load bank test, 25 percent
plus load bank tests at 50, 75 and 100 percent, a black building test and an extended run, then ten timed changeover steps, twelve engine checks, ten on the fuel system and stock, twelve on the alternator, controls and protection, and ten on the installation and its environment.

The document you will get. Download for the full, editable file.

Who this generator maintenance checklist is for

One set, four people, and three of them only think about it twice a year.

  • The service engineer

    Engine, fuel, alternator and the environment, with a reading against each rather than a tick. The load test record is the part that makes the visit worth the money.

  • Whoever owns business continuity

    Maximum load tested to, changeover time and fuel autonomy. Three numbers that say whether the building actually has standby power or merely owns a generator.

  • The facilities manager

    Fuel stock against the stated autonomy, and the date of the last full load bank test. Both tend to drift quietly and neither announces itself.

  • Whoever answers for life safety

    What the set supports, and whether life safety loads are among them. That single field changes what a Critical defect on this document means.

Where standby power is not optional

The structure suits any standby set. What changes is how long the building can be without power and who notices first. If you run one of these, the sector page goes further than the template does.

What a generator service should cover

A generator maintenance checklist is the service record for a standby generating set. It records the load actually tested to rather than the nameplate rating, times every step of the changeover against what the critical loads can survive, checks the engine, the fuel stock against the stated autonomy, the alternator, controls and protection, and the installation around it.

A. The tests first, then the machine

FieldWhat goes in itWhy it earns its place
Generator detailsWhat it supports, the required autonomy, and the transfer arrangementWhat the set supports and for how long it must run decide the fuel stock, the service interval, and whether a failure is an inconvenience or a life safety event.
Load test recordEight tests with duration, load applied, kilowatts achieved and the exhaust conditionOff-load run, building load, load bank at four levels, a black building test and an extended run. The exhaust column is where wet stacking shows up.
Maximum load tested toAs a percentage, reported separately from the nameplateThe honest measure. A set serviced every quarter and never taken past thirty percent has an untested two thirds, and that is exactly where the failures hide.
Transfer and changeoverTen steps, each timed and compared with a required figureMains failure simulated through to reset to auto. Timing each step is what lets the result be compared against what the critical loads can actually survive.
EngineTwelve checks covering oil, fuel filtration, coolant, belts, exhaust and mountsIncludes the jacket water heater, because a cold engine starts slowly and the changeover time is measured from a cold start on the day it matters.
Fuel system and stockTen checks, including stock recorded against the required autonomyAutonomy is a stated number of hours at full load. Fuel held is a measured quantity. The template asks for both so the shortfall is arithmetic rather than opinion.
Fuel conditionSampled and tested for water, sediment and microbial growth, with the age recordedDiesel degrades. A tank filled three years ago and never turned over holds water and growth, and the first time it matters the engine runs for twenty minutes and stops.
Alternator, controls and protectionTwelve checks through to alarms proved to the monitoring pointProtection settings verified rather than assumed: overspeed, overtemperature and low oil pressure are the three that have to operate on the one occasion they are needed.
Battery and chargerCondition, electrolyte, terminals, charger output and float voltageBatteries cause more standby generator failures than engines do. The panel indication tells you the charger thinks it is working, which is a different claim.
Installation and environmentTen checks from ventilation and combustion air to housekeepingVentilation louvres that have to open on start are on this list because a set that runs fine for ten minutes and overheats at twenty is usually an airflow problem, not an engine one.
Defects and severityCritical, Major or Minor, spelled exactly that wayCritical on a standby set means the building currently has no backup power, and somebody senior needs to know today rather than at the next review.
Set returned to automaticConfirmed on the panel at sign-offThe commonest way a generator fails to start is that it was left in manual after a test. It takes five seconds to confirm and it is the last thing anybody thinks about.

The reason load testing sits at the top of this document is that it is the only part of a generator service that tests the thing the generator is for. Everything else on the page tests the machine: the oil is clean, the coolant is correct, the belts are sound, the battery holds charge, the alternator insulation is good. All of that is necessary and none of it answers the question the building is actually asking, which is whether this set will pick up the load when the mains fails and hold it for as long as it takes. An off-load weekly run, which is what most sets get, demonstrates that the engine starts and runs. That is worth having and it is a small fraction of the claim being made. A diesel engine run lightly loaded for long periods does not simply fail to prove itself; it actively deteriorates, because unburnt fuel and soot accumulate in the cylinders, the exhaust and the turbocharger, which is wet stacking, and a set that has been exercised off load for five years can be in materially worse condition than one that was left alone. The second thing worth copying is that the changeover is timed step by step rather than recorded as a pass. The useful number is not that the set started but how many seconds elapsed between the mains failing and the load being carried, because that figure has to be read against something: the ride-through of the uninterruptible supplies, the tolerance of the control systems, the restart behavior of the chillers. A set that accepts load in eight seconds and a set that takes forty both pass a test that only asks whether the transfer happened.

B. What it looks like filled in

One quarterly service on a set that was tested properly. Sixty two checks, and the finding is a battery.

LineValue
Checks completed62: engine, fuel, alternator, installation, plus the tests
Maximum load tested to100 percent, 412 kW achieved on a load bank
Changeover time8 seconds, against a required 15
Fuel held against autonomy4,200 against 3,800 required, no shortfall
Defect raisedBattery bank, cell voltage low, charger float out of range
SeverityCritical

This is close to a model service: tested to full load on a bank, a changeover comfortably inside its requirement, and fuel stock above the stated autonomy. Sixty two checks were completed across the engine, the fuel system, the alternator and the installation, plus the eight load tests and the ten timed transfer steps, and almost all of them passed. The single defect is a battery bank with a low cell voltage and a charger float voltage out of range, and it is graded Critical, which is the correct grading and the part worth dwelling on. Nothing about that set is going to fail while it is running. It will fail to start, which on a standby machine is the same outcome and arrives with no warning at all, because the panel will show the charger working right up to the morning the engine will not turn over. Batteries cause more standby generator failures than engines do, which is why the template asks for the charger output and the float voltage as measurements rather than accepting the indication on the panel. Critical on this document has a specific meaning: the building currently has no backup power, and that is a conversation to have today rather than at the next review. The fixed distribution this set feeds has its own document, the electrical maintenance checklist, and the planned intervals both sit inside a PPM schedule alongside the equipment maintenance checklist that covers the rest of the site's machinery.

Download the generator maintenance checklist

Word for the version an engineer completes at the set, Excel for the one that works out the fuel shortfall against stated autonomy and holds a Service History showing the maximum load tested across a year, and PDF for the copy that goes in the plant room file. Free, and yours to rebrand.

Get the template

How do you service a standby generator?

Test it the way it will be used, time the transfer, and check the fuel is still fuel. Six steps.

  1. Write down what the set supports and for how long

    Which loads, whether life safety is among them, and the required autonomy in hours at full load. Those two answers set everything else on the document.

  2. Load test to the manufacturer's schedule

    A bank if the building cannot take it. Record the kilowatts achieved and the exhaust condition, because that column is where wet stacking appears.

  3. Time every step of the changeover

    Mains failure to start signal, to speed, to load accepted, to transfer, and back again. Then compare the total with what the critical loads can actually survive.

  4. Sample the fuel and count the stock

    Water, sediment and microbial growth, with the age of the fuel recorded. Then measure what is held against the autonomy that was stated, and report the shortfall.

  5. Measure the battery and the charger

    Cell condition, terminals, charger output and float voltage. Batteries cause more standby failures than engines do and the panel indication is not a measurement.

  6. Report the load tested to, then put it back in automatic

    The maximum percentage actually reached, not the nameplate. Then confirm on the panel that the set is in automatic before leaving, every time.

A set that has only ever run off load is a set nobody has tested. See how Atom AI assistants hold load test results, fuel stock and changeover times against the asset.
See Atom AI assistants

An off-load run versus a load test

Two things that both look like exercising the generator, one of which tests almost nothing and is what most sets get.

AspectThe off-load runThe load test
What it provesThe engine startsThe set carries the building
What it costsAlmost nothingA load bank, or an outage
How often it happensWeekly, usually automaticRarely, and often never
What it does to the engineWet stacking, over yearsBurns the cylinders clean
What it says about the alternatorNothing at allIt holds voltage under load
What it leaves untestedMost of the set's capacityVery little

The uncomfortable part of this comparison is that the cheap test is not merely less informative than the expensive one, it is actively harmful if it is the only test the set ever gets. A diesel engine is designed to run under load. Run it at a small fraction of its rating for ten minutes a week, year after year, and fuel that does not fully burn accumulates as soot and unburnt diesel in the cylinders, past the rings, into the exhaust and through the turbocharger. That is wet stacking, and its symptoms are exactly what a nervous operator does not want: a set that smokes heavily, loses power and may not reach its rating when it is finally asked to. So a building that has diligently exercised its generator every Monday morning for five years can be in worse condition than one that has left it alone, and the weekly run is the thing that hid the problem, because it ran every time. The resolution is not to stop exercising the set but to load it properly on a schedule, which usually means a load bank, which usually means a line item somebody has to approve. That is the real reason full load tests are rare, and it is why this document reports the maximum load actually tested to as a headline figure rather than burying it: it makes the gap between what was paid for and what was proved visible to the person who signs the contract. The fixed distribution downstream of the set has its own regime in the electrical maintenance checklist, and a test that requires taking the building off supply belongs in a plant shutdown checklist.

When the template starts to feel limiting

The file holds one service well. Four things get difficult between visits, which is where a standby set spends its life.

  • Maximum load tested is per visit

    The useful figure is the highest load reached across a year, and on paper that means four service sheets lined up and compared by hand.

  • Fuel ages quietly

    The age of the stock matters as much as the quantity, and nothing counts the months since the tank was last turned over.

  • Nothing confirms it went back to automatic

    The sign-off asks the question. Nothing checks the answer, and a set left in manual is the commonest reason one fails to start.

  • The changeover time has nothing to compare against

    It is recorded here and the UPS runtime is recorded somewhere else, and the gap between them only exists when somebody puts the two documents side by side.

What running this in Facilio looks like

The service does not change. What changes is that the three numbers that matter stop being annual facts.

  • Work Order Intelligence

    A Critical defect is escalated, not filed

    A battery bank out of range means the building has no backup power, and that becomes an immediate job rather than a line in a report read next week.

  • Ops Performance Intelligence

    Load tested builds across the year

    The highest load reached across four services is the honest measure of whether the set is proven or merely serviced, and it assembles itself.

  • Asset Intelligence

    Fuel age and autonomy are tracked

    Stock held against the stated autonomy, and the months since the fuel was last turned over, held against the set rather than recalculated at each visit.

  • Audit Report Intelligence

    Test records stay retrievable

    Load test results and changeover times against the asset, which is the form the question takes after an outage when somebody asks what was proved and when.

Hallucination-free by design. Atom AI answers from the records in your tenant rather than generating plausible text, so an empty field reads as empty rather than filled in for you.

Frequently asked questions

What is a generator maintenance checklist?

It is the service record for a standby generating set. This one carries eight load tests, ten timed changeover steps, twelve engine checks, ten on the fuel system, twelve on the alternator and controls, and ten on the installation.

It reports the maximum load actually tested to rather than the nameplate rating, because those are frequently very different numbers.

Why is an off-load run not enough?

Because it proves the engine starts and says nothing about whether the set will carry the building. The alternator is never asked to hold voltage under load and most of the capacity is never reached.

Worse, running a diesel lightly loaded for years causes wet stacking, where unburnt fuel and soot accumulate in the cylinders and exhaust. The weekly run that was meant to prove the set is what degrades it.

What is wet stacking?

The accumulation of unburnt fuel and soot in the cylinders, exhaust and turbocharger of a diesel engine that is habitually run at a small fraction of its rating.

It shows as heavy smoke and lost power, and it is why the load test record has a column for exhaust condition rather than just a pass mark.

How should the changeover be tested?

Step by step and timed: mains failure simulated, start signal received, time to reach speed, time to accept load, transfer, running on load, mains restored, transfer back, cooldown and reset to auto.

Then compare the total with what the critical loads can survive. A UPS sized for thirty seconds and a set that takes forty to accept load is a gap nobody finds until the mains actually fails.

Why does generator fuel need testing?

Because diesel degrades. A tank filled three years ago and never turned over will hold water and microbial growth, and the first time it matters the engine runs for about twenty minutes and stops.

The template asks for the fuel to be sampled, the age recorded, and polishing or rotation carried out, alongside the stock held against the stated autonomy.

What causes most standby generator failures?

Batteries, more often than engines. A flat or failing battery bank means the set will not crank, which on a standby machine is indistinguishable from total failure.

That is why the template asks for charger output and float voltage as measurements rather than accepting the healthy indication on the control panel.

What load should a generator be tested to?

To the manufacturer's schedule, which the template records across four load bank levels plus a building load test, a black building test and an extended run.

What matters for reporting is the maximum actually reached. A set serviced every quarter and never taken past thirty percent has an untested two thirds, and the failures hide there.

What is the last thing to check before leaving?

That the set has been returned to automatic, confirmed on the panel rather than assumed. It is on the sign-off for that reason.

A generator left in manual after a test is the commonest way one fails to start, and it is the easiest failure in this entire document to avoid.

In one paragraph

A generator maintenance checklist works when it records the load actually tested to rather than the nameplate rating, because an off-load run proves only that the engine starts and years of light running cause wet stacking. Time every step of the changeover and read the total against what the critical loads can survive, sample the fuel and count the stock against the stated autonomy, measure the battery and charger rather than reading the panel, and confirm the set is back in automatic before leaving.

Starting is not the same as carrying

Load test to the schedule, time the transfer, sample the fuel, measure the battery and put it back in automatic. Once the maximum load tested only exists on four separate sheets, or nothing counts the months since the tank was turned over, the paper version has reached its limit.

Generator Maintenance Checklist Word · Excel · PDF