By ShipCalculators.com · Published 26 April 2026 · last updated 28 May 2026

Marine Auxiliary Engines and Generators

Contents

What the auxiliary plant does and how it splits from propulsion

A ship carries two separate power chains. The main engine turns the propeller and does little else; the auxiliary plant makes the electricity that runs the rest of the ship. On a conventional motorship the two are mechanically independent: a slow-speed two-stroke crosshead drives the shaft, while three or four four-stroke gensets feed the main switchboard. The split matters because the loads behave differently. Propulsion is a single large steady demand set by ship speed; electrical demand is the sum of dozens of smaller intermittent loads that swing with the operating mode.

The auxiliary engine is the prime mover for a generator. It’s a four-stroke trunk-piston diesel, directly coupled to a synchronous AC machine, with its own cooling, fuel, lube oil, governing and protection. The set is called a genset. A medium ship runs 2 to 4 of them plus an emergency generator. The fleet term “auxiliary” is a little misleading on diesel-electric tonnage, where the same gensets supply propulsion through electric motors, so the “auxiliary” power becomes the primary power. The distinction is functional, not about size.

Why four-stroke for the gensets when the main engine is two-stroke? A two-stroke fires every revolution and runs at 70 to 250 rpm, which suits a propeller turning at shaft speed but makes for a tall, heavy machine that’s awkward to couple to a 720 or 900 rpm generator. A four-stroke fires every second revolution, runs at 500 to 1,000 rpm for the medium-speed class, and packs more power into a given footprint. The penalty is fuel efficiency: a large two-stroke main engine clears 165 to 175 g/kWh, while a medium-speed auxiliary sits around 180 to 195 g/kWh at its best point. For the propeller you chase efficiency over decades of running; for ship service you accept a few extra grams per kilowatt-hour in exchange for compactness, fuel flexibility, and a machine that takes load swings without complaint.

What draws the electrical load

The auxiliary plant feeds the ship’s services through the main switchboard. The big consumers fall into clear groups, and naming them is the first step in sizing the plant.

Main engine ancillaries: jacket and central cooling water pumps, fuel oil supply and booster pumps, fuel and lube oil purifiers, lube oil pumps, control air compressors.
Navigation and bridge: radar, ECDIS, GPS, AIS, gyro, autopilot, and the GMDSS communications suite.
Accommodation and hotel: HVAC, lighting, galley, laundry, refrigeration, and the fresh-water generator.
Cargo-related: cargo pumps and inert-gas blowers on tankers, reefer-container sockets on container ships, ballast pumps, cargo cranes on geared bulkers.
Deck machinery: mooring winches, windlasses, and bow thrusters during berthing.
Safety: fire pumps, emergency lighting, and the bilge and alarm systems.

Total demand spans a wide band by ship type. A dry bulk carrier or product tanker at sea sits between 1 and 4 MW. A large container ship with a heavy reefer count can reach 15 to 20 MW. A cruise ship, where propulsion is also electrical, runs 40 to 80 MW. An offshore vessel holding station on dynamic positioning can pull 20 to 40 MW into its thrusters. The plant is built to the peak it must hold, not the average, and that peak is rarely at sea.

Sizing the plant: the load cases and the running number

Genset sizing starts from an electrical load balance, sometimes called a load table. The naval architect lists every consumer, assigns each a connected rating, then applies a load factor (the fraction of the rating actually drawn) and a diversity factor (the fraction of consumers running at once) for each operating condition. The conditions that drive the design are at sea, in port, manoeuvring, and cargo handling. Each produces a different total, and the plant must cover the worst of them with margin.

Sea load is usually the lowest. The propeller carries the ship, so the only electrical demand is the main engine’s pumps, the bridge, and the hotel. One genset at 50 to 85 percent load typically covers it. Manoeuvring is heavier: bow and stern thrusters come online during berthing, steering gear works hard, and the engine room runs at full readiness, so two or three sets parallel up. Cargo handling is the swing case. On a crude tanker the cargo pumps during discharge can pull more than the rest of the ship combined; on a reefer-heavy container ship the refrigerated boxes can be 50 to 70 percent of the total electrical load on a fully laden tropical voyage. In port without cargo work, demand falls back toward the hotel and standby figure, and a single set carries it.

The number of sets running at any moment follows the load, not the clock. A power management system (PMS) starts and stops gensets so the bus stays inside a target band, usually keeping running load between roughly 50 and 85 percent of the online capacity. Below the lower threshold it sheds a set to save fuel and keep the remaining engines off the low-load region where fouling and SFOC both rise. Above the upper threshold it auto-starts a standby set, synchronises it, and shares the load on. The hysteresis between start and stop thresholds stops the plant from cycling a set on and off every few minutes around a steady demand.

Sizing also has to survive a failure. SOLAS Chapter II-1 Part D and class rules require that essential services stay powered with the largest generating set out of action. In practice that translates to the N-1 rule: the remaining sets must carry the essential load with one set, the largest, removed. The N-1 case is why an owner fits three or four medium gensets rather than two large ones. With two equal sets, losing one halves capacity; with four, losing one leaves three-quarters, and the surviving plant covers the at-sea and manoeuvring essentials without a blackout. The N-1 redundancy check calculator works this case against a stated load table, and the medium-speed four-stroke system calculator sizes the individual set.

Number and rating by ship type

The arrangement reflects the demand profile, not a universal recipe.

Bulk carrier or product tanker: 3 to 4 sets of 1,200 to 2,500 kW each, total installed 4 to 10 MW, plus a 200 to 500 kW emergency set.
Container ship: 4 sets of 2,500 to 4,500 kW each, sized around the reefer peak, total 10 to 20 MW.
Crude or product tanker with high pump demand: 3 to 4 sets of 1,500 to 3,000 kW each.
Cruise ship (diesel-electric): 4 to 6 sets of 8 to 14 MW each, total 40 to 80 MW; the gensets are both propulsion and hotel power.
Offshore DP vessel: 4 to 6 sets of 1,500 to 4,000 kW each, split across redundant buses for DP-2 or DP-3 class.

The cruise and offshore numbers run high because propulsion lives on the same bus as ship service, and because DP class demands that no single failure can take down station-keeping. That redundancy requirement, not raw power, dictates the multi-bus, multi-set architecture on those hulls.

The generator and how it shares load

The machine bolted to the engine is a brushless synchronous AC generator. It turns at the engine’s rated speed and produces three-phase AC. Frequency follows engine speed: 60 Hz on US-flag and much of Asian tonnage, 50 Hz on European and most other flags. Distribution voltage is 440 V or 450 V on smaller ships and 690 V, 3.3 kV or 6.6 kV on high-power installations, where higher voltage cuts cable cross-section and copper losses for the same power. Excitation is brushless, fed by a rotating exciter with a permanent-magnet pilot exciter so the machine can build voltage from rest without an external supply.

Two control loops govern every genset. An automatic voltage regulator (AVR) holds terminal voltage at its set point by adjusting field current, and a speed governor holds frequency by adjusting the fuel rack. On a single set running alone the governor runs isochronous: it keeps frequency dead flat regardless of load. The interesting behaviour starts when two or more sets parallel onto the same bus.

The electrical relations are worth stating exactly. For a balanced three-phase load the active power a set delivers is $P = \sqrt{3}\, V I \cos\phi$ , where $V$ is line voltage, $I$ line current, and $\cos\phi$ the power factor. The reactive power is $Q = \sqrt{3}\, V I \sin\phi$ , and the apparent power is $S = \sqrt{3}\, V I$ , with $S^2 = P^2 + Q^2$ . The active part $P$ is the real work done by motors, lighting and heating; the reactive part $Q$ is the magnetising power that flows back and forth in inductive loads and never does net work. Ship loads are mostly induction motors, so the power factor runs lagging, typically 0.8 to 0.95. The generator active and reactive power calculator resolves a load into these components.

When sets run in parallel, the governors decide who carries what share of $P$ and the AVRs decide who carries what share of $Q$ . Two governing modes exist. In droop, the governor lets speed (and so frequency) fall slightly as load rises, on a fixed slope, usually 3 to 5 percent from no-load to full-load. Two droop sets on the same bus settle at a common frequency, and each picks up the load that its droop line assigns at that frequency. Droop is stable and needs no communication between sets, but the bus frequency sags as total load climbs, so an operator trims the governors to keep it near nominal. In isochronous mode with load sharing, a controller holds the bus at exactly nominal frequency and actively splits real load in proportion to set rating, so a 2,000 kW set and a 1,000 kW set carry load in a 2:1 ratio. Reactive load is shared the same way through the AVRs, in a cross-current or voltage-droop scheme that stops one machine from hogging the kVAR while another runs underexcited. The load sharing droop calculator maps governor droop to load split, and the generator parallel operation calculator covers the synchronising window.

Synchronising before the breaker closes

A set can’t be paralleled onto a live bus until its output matches the bus. Four conditions must line up: voltage, frequency, phase sequence, and phase angle. The operator or the auto-synchroniser trims the AVR until the incoming voltage matches the running bus, trims the governor until frequency matches with the incoming set running fractionally fast so it picks up load rather than motoring on connection, and waits for the phase angle to close on zero, read off a synchroscope or a set of synchronising lamps. The breaker closes at the synchroscope’s slow-rotating null. Close it out of phase and the two machines snap into step with a torque transient that can shear couplings and trip protection. Auto-synchronisers do this in a few seconds and more repeatably than a hand on the governor, but the manual capability stays as a backup because a synchroniser fault must not strand a set offline.

The power management system

Modern plants run under a PMS that automates the load-following and protection that an engineer once did by hand. Its core functions are auto-start of a standby set on rising load, auto-synchronise and auto-load-on, auto-stop on falling load, and load-dependent set selection. It also runs the heavy-consumer interlocks: before a bow thruster or a large cargo pump can start, the PMS confirms enough spinning capacity is online, and starts a standby set first if not. If a running set trips while a heavy consumer is drawing, the PMS sheds non-essential load in priority order to hold the bus rather than let a cascade trip the lot. Black-start sequencing, bus-tie management, and the protective trips for reverse power, over-current, over- and under-voltage, and over- and under-frequency all sit inside the same system. The protections matter as much as the automation: a generator that loses its prime mover but stays connected will motor on power drawn from the bus, and the reverse-power trip is what disconnects it before it drags the whole plant down.

The shaft generator

A shaft generator (often a power take-off, PTO) draws electrical power off the main engine instead of off a genset. The generator is geared to the propeller shaft or to the engine’s free end, so while the main engine runs the ship can make part or all of its sea-load electricity from the propulsion fuel. The economic case is real on long sea passages: the two-stroke main engine burns fuel at 165 to 175 g/kWh against the four-stroke genset’s 185 to 195 g/kWh, so taking ship-service power off the shaft saves the difference, and one genset can be shut down entirely at sea.

The catch is frequency. A simple shaft-driven alternator makes frequency proportional to shaft speed, so it only delivers constant 50 or 60 Hz when the engine holds constant rpm. Older installations restricted the shaft generator to a narrow speed band or used a constant-speed coupling. Modern arrangements put a power-electronic frequency converter between the alternator and the bus, so the shaft generator stays online across the engine’s full operating speed and even feeds power into the propeller as a power take-in (PTI) booster on hybrid plants. The shaft generator can’t cover manoeuvring or port loads, where the main engine is stopped or running erratically, so it supplements rather than replaces the genset plant. The shaft generator output calculator sizes the available take-off against engine margin, and the shaft generator credit calculator covers the efficiency-index credit it earns.

The emergency generator and the SOLAS endurance rule

The emergency generator is independent safety equipment, not part of the normal supply, and SOLAS Chapter II-1 Part D sets its rules precisely. On a cargo ship of 500 GT and above, Regulation 43 requires an emergency source of electrical power. On a passenger ship, Regulation 42 applies. The set sits above the uppermost continuous deck, outside and clear of the main machinery space, so that a fire or flood in the engine room can’t take it out with the main plant. It carries its own fuel, cooling, and starting, and the fuel must have a closed-cup flashpoint of not less than 43 degrees Celsius. On loss of main power it starts automatically and connects to the emergency switchboard “as quickly as is safe and practicable subject to a maximum of 45 s”, per Regulation 43.

The endurance differs by ship type because the survival problem differs. A cargo ship’s emergency source must feed its essential services for 18 hours: emergency lighting in alleyways, stairways, machinery spaces, control stations, the steering gear flat and cargo pump-rooms; navigation lights and the radio installations; internal communications, navigational equipment, fire detection and the general alarm; one fire pump; and the steering gear as required. Lighting at the muster and embarkation stations and over the ship’s side must run for 3 hours. A passenger ship, where a casualty may mean a long evacuation of thousands of people, must hold its emergency power for 36 hours under Regulation 42, covering the same classes of service plus the sprinkler and bilge pumps and the watertight-door system. Ships on short voyages can be approved for a reduced period, but not below 12 hours.

The emergency generator is tested under load on a routine schedule and surveyed by class. Its dedicated fuel tank typically holds 6 to 12 hours at full load with bunkering to top it up, and the auto-start sequence is proven at every periodic test because the 45-second figure is a hard limit, not a target. A high-speed diesel, 1,500 or 1,800 rpm, is usual for the emergency set: it’s compact, starts fast on its batteries or hydraulic accumulator, and the higher SFOC doesn’t matter for a machine that runs only in drills and casualties. The emergency genset high-speed diesel calculator sizes this set against the emergency load list.

Fuel grades and the auxiliary fuel system

Auxiliary engines burn the same fuel families as the main engine, and on most motorships they share the bunkers. Residual heavy fuel oil (HFO), up to 700 cSt at 50 degrees Celsius (RMK 700 under ISO 8217), needs heated storage, settling and service tanks, centrifugal purification, and a viscosity controller that heats the fuel to roughly 12 to 14 cSt at the injector. Distillates are simpler: marine gas oil (MGO, grades DMA and DMZ) and marine diesel oil (MDO, grade DMB) need no heating and far less treatment, which is why they’re the default for emergency and harbour use. Since the IMO 2020 sulphur cap, very low sulphur fuel oil (VLSFO) at 0.50 percent sulphur covers global running, with 0.10 percent distillate or VLSFO inside the emission control areas. Dual-fuel auxiliaries add a gas mode, burning LNG or boil-off with a small pilot injection of distillate, and methanol-capable variants are entering service. The MAN L23/30H GenSet, for instance, is type-approved for HFO up to RMK 700, MDO (DMB), and MGO (DMA, DMZ) to ISO 8217.

Fuel changeover between HFO and a low-sulphur distillate before an ECA boundary is a managed procedure, not a valve flip. The two fuels differ in viscosity and temperature by a wide margin, so the changeover is ramped over 30 to 60 minutes to avoid thermal shock to the fuel pumps and injector needles, and the time and position are logged for MARPOL Annex VI compliance with a retained fuel sample. Each set carries dual fuel filters with switchover, a service tank with several hours of capacity, and on common-rail engines an accumulator rail that holds injection pressure independent of engine speed.

NOx limits, the test cycle, and the ECA application

Auxiliary engines fall under MARPOL Annex VI Regulation 13, which limits nitrogen oxides from any marine diesel engine of more than 130 kW output, with narrow exemptions for engines used only in an emergency. The limit is a tiered curve set by the engine’s rated speed $n$ in rpm, and the tier a given engine must meet is fixed by the ship’s construction date. Within each tier the allowable NOx, weighted over a test cycle and expressed in grams per kilowatt-hour, follows three speed bands.

Tier I, for ships built on or after 1 January 2000, allows $17.0$ g/kWh below 130 rpm, $45\,n^{-0.2}$ across 130 to under 2000 rpm, and $9.8$ g/kWh at 2000 rpm and above. Tier II, for ships built on or after 1 January 2011, tightens this to $14.4$ g/kWh, $44\,n^{-0.23}$ , and $7.7$ g/kWh across the same bands, a cut of roughly 15 to 20 percent achieved through combustion tuning: injection timing and pressure, rate shaping, valve timing, and compression ratio, with no after-treatment. Tier III, for ships built on or after 1 January 2016 and operating inside a designated NOx emission control area, drops to $3.4$ g/kWh below 130 rpm, $9\,n^{-0.2}$ across the middle band, and $2.0$ g/kWh at 2000 rpm and above, about 80 percent below Tier I. A medium-speed auxiliary at 720 to 900 rpm lands in the middle band, so its Tier II limit works out near $44 \times 750^{-0.23} \approx 9.8$ g/kWh and its Tier III limit near $9 \times 750^{-0.2} \approx 2.4$ g/kWh.

The figures aren’t a single-point measurement. The NOx Technical Code 2008 (IMO Resolution MEPC.177(58)) sets the test cycles. A constant-speed auxiliary, the normal genset case, is certified on the D2 cycle: five load points at 100 percent rated speed, run at 100, 75, 50, 25 and 10 percent power, weighted 0.05, 0.25, 0.30, 0.30 and 0.10. The weighting puts most of the test emphasis on the 25 to 75 percent band where a genset actually spends its life, not at full power. A variable-speed auxiliary is certified on the C1 cycle instead. Every certified engine carries a NOx Technical File documenting the as-tested configuration; surveyors check that the in-service engine still matches it.

Tier III’s gap below Tier II is too wide for combustion tuning alone, so it needs after-treatment or an alternative fuel. Selective catalytic reduction (SCR) injects a urea solution into the exhaust ahead of a catalyst, where ammonia reduces NOx to nitrogen and water; it’s the dominant Tier III route and can reach the 80 percent cut on its own. Exhaust gas recirculation (EGR) routes part of the exhaust back into the cylinder to lower the peak combustion temperature where NOx forms, and is more common on two-stroke main engines than on four-stroke auxiliaries. Burning LNG in a dual-fuel set drops NOx far enough to meet Tier III in gas mode without after-treatment, because the lean premixed gas combustion runs cooler. Tier III binds only inside an ECA: the North American and US Caribbean areas from 1 January 2016, and the Baltic and North Sea areas from 1 January 2021. Outside an ECA, a Tier III-equipped ship runs to Tier II, which is why SCR systems have a bypass.

Specific fuel oil consumption and the load curve

A modern medium-speed four-stroke auxiliary burns 180 to 195 g/kWh at its best point, which sits around 75 to 90 percent of rated load. The curve is a shallow bowl: SFOC rises steeply below 50 percent load, falls to the minimum near 85 percent, and ticks up slightly past it. The shape is why the PMS tries to keep running sets in the 50 to 85 percent band and sheds a set rather than let two engines idle at 30 percent each, where both burn fuel inefficiently and run cold enough to glaze liners and foul exhaust passages. Low-load running for extended periods is a maintenance problem as much as a fuel one. The specific fuel oil consumption article treats the curve and its corrections in detail, and the medium-speed four-stroke marine engines article covers the engine class itself.

Engine families and where they sit

The auxiliary engine market is split between a handful of OEMs and the licensed builders at the major yards. MAN Energy Solutions sells its Holeby-origin four-stroke gensets across the cargo fleet: the L23/30H, a 225 mm bore, 300 mm stroke engine in 5 to 8 cylinders, runs 720, 750 or 900 rpm and produces 130, 135 or 160 kW per cylinder at those speeds, for a range around 650 to 1,280 kW; MAN reports more than 12,000 units installed. Larger MAN sets run up through the L21/31, L27/38, and the common-rail 32/44CR. Wartsila’s auxiliary range centres on the Wartsila 20, 26 and 32, with the 32 the most widely fitted, and the 46 crossing into propulsion duty on diesel-electric ships. Caterpillar covers the field with its 3500 and 3600 series, Cummins and Yanmar serve smaller tonnage, and Mitsubishi, Hyundai HiMSEN and Doosan supply the Japanese, Korean and Chinese yards. MTU high-speed engines handle the compact, fast-start niche, including many emergency sets. The choice usually follows the yard’s preferred supplier and the owner’s fleet standardisation, both filtered through class-society type approval. Two-stroke main engines and the broader engine background sit in the marine diesel engine article.

Operating practice across the voyage

The plant’s job changes through a voyage, and the running configuration changes with it. At sea on a long passage, one genset carries the load (or the shaft generator does, with all gensets stopped), and the engineers run the deepest maintenance the open ocean allows: a top-end overhaul on an offline set, filter rotations, oil sampling. Approaching an ECA, the watch runs the fuel changeover an hour or two before the boundary and logs it. Through restricted waters and into the berth, two or three sets parallel up to cover the thrusters and steering load, then shed back to one or two for cargo work depending on whether pumps or reefers dominate. The emergency generator sits in auto-standby throughout, proven at the periodic test.

Maintenance is interval-based and tracked against running hours. Top-end work (cylinder heads, valves, injectors) falls due every 6,000 to 12,000 hours; bottom-end work (pistons, rings, main and big-end bearings) every 12,000 to 24,000 hours; a major overhaul touching the crankshaft and all bearings every 24,000 to 60,000 hours. Lube oil for a medium-speed trunk-piston engine is an SAE 30 or 40 grade, sampled monthly for viscosity, base number, water and wear metals, and changed every 4,000 to 8,000 hours. Jacket water runs 70 to 95 degrees Celsius at the outlet under 1 to 3 bar, dosed with a corrosion inhibitor and watched for pH and conductivity. Class surveys run on the continuous machinery survey cycle, with each set examined in rotation and certified at delivery and renewal.

Limitations

The figures in this article are nominal, and three categories of caveat matter to anyone sizing or operating the plant. First, every rating is conditional on reference ambient conditions. Engine makers quote power and SFOC at ISO 3046 reference conditions (around 25 to 45 degrees Celsius air, defined humidity and barometric pressure, defined charge-air coolant temperature). A genset in the Persian Gulf in summer, with 45-degree engine-room air and warm seawater, will derate from its plate rating, and a load balance built on plate figures can leave the plant short on the hottest day. Always check the derating curve in the project guide against the worst design ambient, not the nameplate.

Second, the load-table factors are estimates, and the gap between design and reality is where blackouts live. Load factors and diversity factors are the designer’s judgement about how hard each consumer runs and how many run together. They’re conservative on paper, but a ship that adds reefer plugs, retrofits a scrubber’s pumps, or operates a cargo cycle the designer didn’t anticipate can outgrow its plant. The N-1 case in particular assumes the load table is current; an out-of-date table makes the redundancy claim optimistic. Re-run the balance after any major electrical retrofit.

Third, the regulatory figures are version-dependent and jurisdiction-dependent. The NOx tier limits and ECA boundaries quoted here are MARPOL Annex VI as it stands, and both the limits and the ECA list have changed over time and continue to evolve through MEPC resolutions. The tier an engine must meet is set by the ship’s keel-laying or construction date, not the engine’s manufacture date, and the Tier III obligation depends on the ship being inside a designated NOx ECA at the time, which means an SCR or dual-fuel system needs a documented mode-switch and bypass. The emergency-power endurance figures (18 hours cargo, 36 hours passenger, 3 hours muster lighting, 45-second auto-start, 43-degree flashpoint) are SOLAS Part D as currently in force; administrations can grant the short-voyage reduction down to 12 hours, and flag-specific interpretations exist. Treat the numbers here as the starting point for a project, and confirm against the controlling regulation edition and the vessel’s class rules before committing to a design or a compliance claim. The article describes typical practice; it isn’t a substitute for the engine project guide, the class rules, or the certified NOx Technical File for a specific installation.