How to Size a Diesel Generator for Motor Loads

Last reviewed 2026-11-10

Sizing a generator for lights and sockets is a matter of adding up the load and allowing some headroom. Add a few induction motors, such as a borehole pump, a chiller compressor or a large fan, and the picture changes. The running load may still be modest, but starting those motors can demand several times as much from the generator for a few seconds. Diesel sets react to that demand far less gracefully than a utility supply does. This guide walks through a practical method and shows where the judgement calls lie.

Why motors are different

An induction motor at standstill behaves like a transformer with a short-circuited secondary. When it is switched straight onto the supply (direct on line, or DOL), it draws a locked-rotor current that is typically five to eight times its full-load current. Most of that current is reactive, at a power factor that may be only 0.2 to 0.4. The current falls as the motor accelerates, usually within a few seconds, but during that time it looks like a very large, very inductive load.

A utility transformer barely notices. A generator set is a different matter. The alternator has a relatively high internal impedance, so a sudden kVA demand pulls its terminal voltage down until the automatic voltage regulator (AVR) recovers it. The engine may also slow under the real-power part of the step until the governor responds. If the voltage dip is too deep, contactors can drop out, lighting flickers, other motors stall and the starting motor may not accelerate at all. Sizing for motors is therefore mostly about keeping that transient dip acceptable.

Step 1: list the running load properly

Start with a load table. For each item record the running power, the power factor and how many there are. Convert motor nameplate ratings to electrical input: nameplate kW or hp is shaft output, so divide by the motor efficiency. A 20 hp motor at 90 % efficiency draws 20 × 0.7457 / 0.9 = 16.6 kW.

Then add the loads correctly. Sum the real powers (kW) and the reactive powers (kVAr = kW × tan φ) separately, and combine them at the end as kVA = √(kW² + kVAr²). That gives the true running kVA.

The generator must satisfy two running limits. Its kVA rating must cover the running kVA, and because most sets are rated at PF 0.8, its kW rating (kVA × 0.8) must cover the running kW. Whichever requires the bigger set governs. Apply a diversity factor only where you are confident some loads never run together. For emergency and life-safety supplies, many codes expect you to assume they all do.

Step 2: find the worst starting step

For each motor, estimate the starting kVA:

The multiple depends on how the motor is started:

These are planning values. Better figures come from the motor itself. NEMA motors carry a locked-rotor kVA code letter, and IEC motor datasheets give the ratio of starting current to rated current (I_A/I_N).

Motors are normally started one at a time, so the critical moment is when the motor with the largest starting step starts while everything else is already running. The peak demand at that moment is roughly the running kVA of all other loads plus the starting kVA of that motor. Adding the two arithmetically overstates the result slightly, because the starting kVA is at a much lower power factor, but for a first estimate the error is on the safe side.

Step 3: compare with what the set can deliver

This is where generic methods reach their limit. How much starting kVA a set can absorb depends on its alternator, its excitation system and the voltage dip you are prepared to accept. Manufacturers publish motor-starting curves that relate starting kVA to percentage voltage dip for each alternator frame. A set may be able to start a motor whose starting kVA is well above its continuous rating, provided a 20–30 % dip is acceptable.

Without those curves, the conservative approach is to require the set's rated kVA to cover the peak starting kVA. That tends to oversize. It is a reasonable first estimate and an upper bound to refine once a candidate machine is chosen. In the generator sizing calculator this is the "starting capability" factor: 1.0 gives no credit for short-term capability, and you can raise it to match a manufacturer's curve at your permitted dip.

Typical acceptable dips depend on what else is connected. Around 15 % is a common figure where lighting flicker matters, and 25–35 % may be tolerated where only motors are on the bus. Always check the drop-out voltage of the contactors and relays in the system.

Step 4: margin and site derating

Add a design margin on the running load for growth and uncertainty. 20–25 % is common. Then account for site conditions. Diesel engines lose power at high altitude, where the air is thinner, and at high intake air temperature. Alternators also derate at high ambient temperature. The derating figures come from the manufacturer's data for the specific engine. Generic rules of thumb exist, but they vary too much between engines to rely on. Enter the derate as a percentage and divide the requirement by (1 − derate).

Also make sure you are comparing against the right rating. ISO 8528-1 defines standby, prime and continuous ratings with different permitted load profiles and running hours. A set used as the main power source for a remote site needs a prime or continuous rating. A standby rating suits emergency backup for limited hours only.

A worked example

A small pumping station has 10 kW of lighting and controls at PF 1.0, a 15 kW pump at PF 0.85 started DOL, and a 7.5 kW ventilation fan at PF 0.8 on a soft starter. The running load is:

Starting the pump DOL at 6 × (15 / 0.85) = 105.9 kVA on top of the other running loads (18.4 kVA) gives a peak of about 124 kVA. With no credit for short-term capability, a 125 kVA set is indicated, more than double the running requirement.

Fit the pump with a soft starter at 3× and the peak falls to about 71 kVA, so a 75 kVA set covers both running and starting. That difference in set size usually pays for the soft starter several times over, and it also cuts fuel use because the smaller set runs at a healthier load factor. You can check fuel consumption at each option with the fuel consumption calculator.

Things this method does not cover

• Non-linear loads. VFDs, UPS rectifiers and large LED installations draw harmonic currents that distort the generator voltage. Alternators feeding a large share of non-linear load are often oversized deliberately. Ask the manufacturer.
• Block load acceptance. Some applications need a large load picked up in a single step, such as a fire pump or a whole building transferring at once. Transient frequency and voltage performance classes are defined in ISO 8528-5.
• Regenerative loads. Lifts and cranes can push power back into the set, which a diesel engine cannot absorb without load banks or other measures.
• Light-load running. Oversized sets running for long periods at low load suffer wet stacking. Sizing far beyond need is not automatically "safe".

Summary

1. Tabulate loads and convert motor shaft ratings to electrical input.
2. Add kW and kVAr separately, then check both the kVA and kW limits of the set.
3. Find the motor whose start causes the highest peak, and choose starting methods deliberately.
4. Compare the peak with the set's motor-starting capability at an acceptable voltage dip, conservatively if curves are not yet available.
5. Apply margin and the manufacturer's site derating, then pick a set with the right ISO 8528 rating type.
6. Confirm the selection with the manufacturer's sizing tool before purchase.

This guide is general education, not a design procedure for any particular code or product. Life-safety and code-mandated systems need review by a qualified engineer.

Calculators for this topic

• Generator Sizing Calculator
• Diesel Generator Fuel Consumption Calculator
• kVA to kW Converter (and kVAr)