Selecting the right size generator set involves evaluating starting loads, continuous demand, fuel preferences, and ambient operating conditions. Oversizing leads to wet stacking in diesel engines, while undersizing risks system tripping during motor startup spikes.

Before selecting equipment, perform a full load audit to identify critical circuits versus non-essential building loads.

Start With the Load, Not the Rating

The most common error is beginning with a kVA figure — perhaps from a previous installation, or a neighbouring site — and working backwards. A generator is not sized by nameplate alone, but by the shape of the load it must carry. Two sites with identical monthly consumption can need very different sets if one has large motors and the other has mostly resistive heating.

Split your audit into three columns: continuous load, intermittent load, and starting load. Motors, compressors, and pumps belong in the third column, where the inrush current can be several times the running figure. That peak, not the average, is often what sets the rating.

Four Questions Worth Answering First

• 01What starts at once? A single 90 kW motor starting across the line is a different problem from four 30 kW motors staggered ten seconds apart. Sequence matters as much as total load.
• 02What is the lowest load? Diesel sets run poorly below roughly 30% of rating. If the overnight load is tiny, a single large set will spend its life underloaded — which is where wet stacking and glazed cylinders begin.
• 03How hot does it get? Ratings are quoted at standard ambient conditions. A set installed in an enclosed plant room at 45°C can lose a meaningful percentage of its output before it ever runs.
• 04How long must it run? Standby duty and prime duty are rated differently. A set expected to carry full load for days at a time should be specified as prime, not standby.

“The right question is not how big a generator you need, but how little it can be and still carry the load that matters.”

Where Oversizing Goes Wrong

An oversized set is comfortable to specify and expensive to own. Fuel consumption at light load does not fall proportionally — a 500 kVA set carrying 60 kVA can burn nearly as much fuel per hour as one carrying 200 kVA, which means the cost per usable kWh climbs sharply. Maintenance intervals are set by running hours, not by energy delivered, so an oversized set also demands more servicing for the same output.

There are legitimate reasons to oversize: planned expansion, future EV charging, or a large motor that must start across the line. The key is to name the reason explicitly in the specification, so that later readers can tell a deliberate margin from an accident.

A Note on Hybrid Options

Where the load profile is spiky, pairing a smaller generator with battery storage often beats specifying a larger set. The battery absorbs the short peaks, the generator runs steadily at its efficient point, and total fuel use falls. This is covered in more detail in our analysis of solar and battery storage working together.

Solar panels and batteries are often discussed as a single "hybrid" system, but they perform different jobs. Solar generates during daylight, with output that rises and falls with the weather. Batteries store and release energy on demand, regardless of the sun. The value of combining them lies in what each can do that the other cannot.

In an industrial setting, the combination is usually designed around three goals: reducing grid import, maintaining power through outages, and smoothing the load seen by the generator. Each goal imposes different requirements on the control system.

The Role of the Transfer Switch

In a hybrid microgrid, the power transfer switch (PTS) is the component that decides where power comes from at any moment. It sits between the grid, the generator, the solar inverter, and the battery inverter, and it must open and close in a controlled sequence to avoid back-feeding the grid or damaging equipment.

The PTS logic is usually configured around three states: grid-connected, islanded (grid down, generator and storage carrying the load), and transition. The transition state is where most commissioning problems occur, because the switch must coordinate with the inverters to match voltage and frequency before closing.

How the Inverters Talk to Each Other

• 01Frequency-watt control. The battery inverter adjusts its output based on grid frequency, providing fast response to load changes without waiting for the generator to ramp.
• 02Volt-watt control. The solar inverter reduces output when voltage rises, preventing overvoltage as generation exceeds local demand.
• 03State-of-charge limits. The battery inverter curtails charging when the battery is full and stops discharging when it is depleted, so the generator or grid picks up the difference.
• 04Generator assist. When the battery is low and solar is insufficient, the controller starts the generator and holds it at a steady load while the battery smooths the peaks.

"Solar decides when power is available. The battery decides when it is used. The controller decides which one the site actually needs."

Sizing the Two Together

Solar and battery are sized against different constraints. Solar is sized by roof or land area, by the site's daytime load, and by the export limit imposed by the utility. The battery is sized by the longest expected outage, by the peak load it must carry alone, and by how much of the solar surplus it needs to absorb.

A common mistake is to size the battery for the solar array rather than for the load. A large array paired with a small battery will export most of its generation and provide little resilience. A modest array paired with a well-sized battery will cover the site's critical loads through an outage and still reduce daytime import.

Where Hybrid Designs Go Wrong

The failures we see most often are not in the equipment but in the control logic. A transfer switch configured for the wrong islanding condition will trip the site instead of protecting it. A battery inverter set to the wrong state-of-charge window will either refuse to discharge when needed or fail to recharge before the next outage.

The remedy is commissioning: run the system through each transition state deliberately, log the behaviour, and adjust the setpoints before handover. A hybrid system that has never been islanded on purpose is a system whose behaviour is unknown.

A Note on Generators

Where a generator is part of the hybrid, the battery can absorb the peaks that would otherwise force the generator to run at low load. This is covered in more detail in our guide to generator sizing, and in the analysis of fuel consumption.

Backup power is often planned around a single question: how big a generator do we need? That question is the last one to answer, not the first. Business continuity depends on knowing which loads must stay on, for how long, and what happens when the fuel runs low — none of which is a generator sizing problem.

A continuity plan starts with the business, not the plant room. What does the site do, what does it lose per hour of downtime, and which of those losses can be avoided with power and which cannot?

Step One: Audit the Emergency Circuits

The first task is to walk the site and list every circuit that must remain energised. This is not the same as listing every circuit that is currently connected to the backup panel. Panels accumulate loads over time, and a circuit that was critical five years ago may now be redundant — or a new critical load may have been added without updating the plan.

Split the audit into three tiers:

• 01Life safety. Fire alarms, emergency lighting, sprinkler pumps, and evacuation systems. These must run regardless of cost or duration.
• 02Business critical. IT servers, cold storage, process controls, and security systems. These determine how long the site can operate without grid power.
• 03Discretionary. Comfort cooling, non-essential lighting, and general office loads. These can be shed first when fuel or capacity is limited.

Step Two: Establish the Duration Target

The duration target is a business decision, not an engineering one. A site that can tolerate a four-hour outage needs a very different system from one that must run for three days. The target should be written down and agreed with the people who own the business risk, because it drives every downstream choice: fuel storage, generator redundancy, and battery capacity.

"A backup power plan is only as good as the fuel plan behind it. Most sites know their generator rating and not their hours of autonomy."

Step Three: The Fuel Management Plan

Fuel is where continuity plans most often fail. A generator rated for the load is useless if the tank is empty, the fuel has degraded, or the delivery route is blocked. The fuel plan should cover:

• 01Autonomy. How many hours of run time does the on-site tank provide at the expected load? This should be calculated, not assumed from tank size.
• 02Fuel quality. Diesel degrades over time. Stored fuel should be tested periodically and polished or replaced on a schedule.
• 03Resupply. Who delivers, how quickly, and what happens if the roads are closed? A resupply contract is not a resupply plan.
• 04Priority allocation. If fuel is limited, which loads get it? This should be decided in advance, not during the outage.

Step Four: Test the Plan, Not Just the Generator

Most sites test the generator. Fewer test the plan. A generator test confirms that the set starts and carries load. A plan test confirms that the transfer switch operates, that the critical circuits are actually energised, that the fuel log is current, and that the people who need to act know what to do.

Run the plan test at least annually, and after any significant change to the site's loads or infrastructure. Record what happened, including what went wrong, and update the plan accordingly.

Where Battery Storage Fits

Battery storage can extend autonomy without adding fuel, and can bridge the gap between grid failure and generator start. It is not a replacement for a generator on long outages, but it is often the cheapest way to cover short outages and to reduce the load the generator must carry. This is covered in more detail in our analysis of solar and battery storage working together.

A standby generator that has never failed is not necessarily a reliable generator — it is a generator that has never been asked to carry load for long enough to reveal a fault. The purpose of a maintenance calendar is to make sure the set is tested in the conditions it will actually face, not just the conditions that are convenient to test it in.

The calendar below is written for a diesel standby set in an industrial or commercial plant room. The intervals assume the manufacturer's recommendations are the baseline; the notes explain where sites commonly deviate and why.

Weekly Checks

• 01Visual inspection. Look for leaks, loose fittings, and signs of rodent or water ingress. Check that the enclosure is secure and the exhaust is clear.
• 02Fluid levels. Oil, coolant, and fuel. A slowly falling level is a problem before it becomes a failure.
• 03Battery condition. Check the charger is operating and the battery voltage is within spec. Most no-start failures are battery failures.
• 04Fuel log. Confirm the tank level matches the log. Discrepancies indicate a leak or a theft — both are continuity risks.

Monthly Checks

• 01Test run under load. A no-load run confirms the set starts. A load run confirms it can carry the site. Run at a meaningful load for at least 30 minutes.
• 02Transfer switch test. Confirm the switch operates cleanly and the critical circuits energise. This is the step most sites skip.
• 03Alarm and fault log review. Read the controller log and investigate any recurring fault, even if the set still starts.

"A generator that has only ever been tested at no load has only ever proven that it starts. It has not proven that it can carry the site."

Quarterly Checks

• 01Fuel quality test. Diesel degrades. Test for water, microbial growth, and cetane. Polish or replace if results are out of spec.
• 02Air and fuel filter inspection. Replace on condition or on schedule, whichever comes first.
• 03Cooling system check. Inspect hoses, clamps, and the radiator for blockage. A set that overheats under load is a set that will trip.
• 04Battery load test. A voltage check is not enough. Load-test the battery to confirm it can deliver cranking current.

Annual Checks

• 01Full load bank test. Run the set at or near rated load for a sustained period. This is the only way to confirm it can carry the site at full demand.
• 02Oil and filter change. Even on low-hour sets, oil degrades with time and contamination.
• 03Fuel system inspection. Check injectors, fuel lines, and the tank for signs of degradation.
• 04Control system review. Confirm setpoints, alarm thresholds, and auto-start logic are correct for the current site configuration.
• 05Documentation update. Record what was done, what was found, and what changed. An unrecorded maintenance history is a maintenance history that will be repeated.

The Two Items Most Sites Skip

The first is the transfer switch test. Sites test the generator and assume the switch will work. When the grid fails, the switch is the component that determines whether the site actually receives power — and it is the component most likely to have seized or drifted out of adjustment.

The second is the fuel quality test. Sites check the fuel level and assume the fuel is usable. Diesel that has sat for a year may still be at the right level and still be unfit to burn. The failure mode is a generator that starts, runs briefly, and then stops — which is worse than a generator that never started, because it delays the response.

A Note on Low-Hour Sets

Standby sets often accumulate very few running hours. This does not mean they need less maintenance — it means they need different maintenance. Time-based items (fuel quality, battery condition, coolant) matter more than hour-based items. A set that runs 20 hours a year still needs its fuel tested, its battery load-tested, and its transfer switch exercised on schedule.

This case note describes a distribution centre in a region with a single grid feed and a history of outages that ranged from a few minutes to half a day. The site had a standby generator, but a capacity review showed that it could not carry the full refrigeration load during a long outage. The brief was to fix that without buying a larger generator.

The Starting Point

The site's critical load was refrigeration: cold rooms, blast freezers, and the compressors that serve them. The existing generator was rated at 250 kVA, and the combined refrigeration load at the height of summer was around 380 kVA. During a long outage, the generator would be overloaded within minutes of the transfer switch closing.

The obvious response was to specify a 500 kVA set. The site had the space, and the budget could stretch to it. But a 500 kVA set would spend almost all of its life carrying a small load — the site's normal backup demand was closer to 120 kVA — and would be inefficient and expensive to maintain. The engineering team proposed a different approach.

The Hybrid Design

The design combined the existing 250 kVA generator with a battery storage system and a roof-mounted solar array. The roles were divided as follows:

• 01Solar. A 120 kWp array on the warehouse roof, which reduced daytime import and provided some generation during outages that occurred in daylight.
• 02Battery storage. A 200 kWh battery with a 150 kW inverter, sized to cover the peak refrigeration load that the generator could not carry alone, and to bridge the gap between grid failure and generator start.
• 03Generator. The existing 250 kVA set, retained and controlled so that it ran at a steady load while the battery handled the peaks.
• 04Control system. A single controller that coordinated the transfer switch, the battery inverter, the solar inverter, and the generator, with the ability to shed non-critical loads if the battery state of charge fell below a set threshold.

"The right answer was not a bigger generator. It was a smaller generator, used better."

The Outage

The system was commissioned in the spring. The first significant test came in the summer, when a grid failure lasted twelve hours. The sequence was as follows:

• 010–10 seconds. The battery inverter detected the grid failure and islanded the critical panel. Refrigeration compressors continued to run without interruption.
• 0210–60 seconds. The generator started and synchronised. The battery absorbed the compressor inrush as the generator came online.
• 031–12 hours. The generator carried a steady load of around 200 kVA. The battery covered the peaks above that, and the solar array contributed during the daylight hours. Non-critical loads were shed automatically when the battery state of charge fell below 30%.
• 04Recovery. When the grid returned, the controller synchronised and transferred the load back, then recharged the battery from the grid and solar.

The Outcome

The site ran through the twelve-hour outage without losing refrigeration. The generator ran at its efficient point for the duration, and fuel consumption was lower than it would have been with a larger set running at light load. The battery was recharged by the following morning.

The capital cost of the hybrid system was approximately 60% of the cost of a new 500 kVA generator and its associated switchgear and fuel storage. The site retained its existing generator, added resilience, and reduced its daytime grid import through solar generation.

What Made It Work

Three things made the design succeed. First, the load audit was done properly, and the peak load was measured rather than estimated. Second, the battery was sized for the peak that the generator could not carry, not for the solar array. Third, the control system was commissioned through each transition state deliberately, including a planned islanding test before handover.

The lesson is not that hybrid systems are always the right answer. It is that the right answer depends on the load profile, and that a smaller generator used well can outperform a larger one used poorly.

Related Notes

The load audit approach used here is described in more detail in our guide to generator sizing. The control logic is covered in our analysis of solar and battery storage working together.

Oversizing a standby generator is the easiest decision in the plant room. It removes the risk of being caught short, it leaves room for expansion, and it is rarely challenged because the cost is paid over years rather than at the point of purchase. The problem is that the cost is real, and it accumulates quietly in fuel, maintenance, and capital that never delivers a return.

This analysis looks at where the money actually goes when a generator is larger than the load it serves — and why the cheapest set to buy is rarely the cheapest to own.

Where the Money Goes

• 01Fuel at light load. Fuel consumption does not fall in proportion to load. A 500 kVA set carrying 60 kVA can burn nearly as much fuel per hour as one carrying 200 kVA, because the engine must overcome its own friction and cooling losses regardless of output. The cost per usable kWh climbs sharply.
• 02Maintenance on running hours. Service intervals are set by running hours, not by energy delivered. An oversized set that runs for the same number of hours delivers less useful energy per service, so the maintenance cost per kWh is higher.
• 03Capital tied up. The difference between a 250 kVA and a 500 kVA set is not just the generator — it is the switchgear, the fuel tank, the foundation, and the plant room space. That capital delivers no return if the capacity is never used.
• 04Wear from underloading. Diesel sets run poorly below roughly 30% of rating. Wet stacking, glazed cylinders, and degraded oil are the long-term consequences of a set that never gets warm enough to burn cleanly.

"The cost of idle capacity is not paid at the point of purchase. It is paid every hour the set runs below its efficient point."

The Comfort Trap

Oversizing is comfortable because the downside is invisible. A set that is too large still starts, still carries the load, and still passes its test runs. The fuel bill is higher, the maintenance cost is higher, and the engine wears faster — but none of these show up as a failure, so none of them trigger a review.

The trap is reinforced by the way specifications are written. A designer who specifies a 500 kVA set for a 200 kVA load is unlikely to be criticised. A designer who specifies a 250 kVA set and is later asked to carry an unplanned load is likely to be blamed. The asymmetry pushes specifications upward, one safety margin at a time.

When Oversizing Is Justified

There are legitimate reasons to specify a larger set, and they should be named explicitly in the specification so that later readers can tell a deliberate margin from an accident:

• 01Planned expansion. A documented plan to add load within the set's life, with a date and a load estimate.
• 02Large motor starting. A single motor that must start across the line can set the rating even when the running load is small.
• 03Future EV charging. A known future load with a known profile.
• 04Redundancy. Two sets each rated for the full load, so that one can carry the site while the other is serviced.

The Alternatives

Where the load profile is spiky or the peak is short, a smaller generator paired with battery storage often beats a larger set. The battery absorbs the peaks, the generator runs steadily at its efficient point, and total fuel use falls. This is covered in more detail in our analysis of solar and battery storage working together, and in the case note on a cold storage site that replaced a planned 500 kVA set with a hybrid design.

Where the load is genuinely variable but not spiky, load consolidation — moving loads onto fewer sets and shutting the rest down — can achieve the same effect at lower cost.

What to Measure

The first step is to measure the actual load profile, not the design load. Log the load for at least a week, including the peak, the average, and the lowest sustained load. Compare the measured profile against the set's rating. If the average load is below 40% of rating, the set is a candidate for right-sizing or for a hybrid design.

The second step is to measure fuel consumption against the manufacturer's part-load curve. If the measured consumption is significantly higher than the curve predicts, the set is either poorly matched to the load or in need of service.