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

Marine Sea Water Cooling Systems

Contents

What a sea water cooling system does

A sea water cooling system rejects the waste heat of a ship’s machinery into the ocean, the only heat sink a vessel carries with it everywhere. A slow-speed two-stroke main engine converts roughly half its fuel energy into shaft work; the rest leaves as exhaust, radiation, and heat carried away by cooling media. Jacket water, lubricating oil, scavenge (charge) air, and the auxiliary engines all dump their heat, in the end, into sea water drawn through the hull. Get that flow wrong and the engine derates, the bearings run hot, or the watch loses cooling at the worst moment.

The engineering sits inside a hard boundary: sea water is corrosive, it carries larvae and silt, and its temperature is set by where the ship is, not by what the plant wants. Classification rules fix the upper design point at 32 degrees C, the tropical sea water value in IACS Unified Requirement M28 (1978), so every cooler, pump, and pipe must still deliver rated duty in warm water. This article covers the two competing architectures (direct sea water cooling and central cooling), the sea chests and strainers at the inlet, the sea water circulating pumps, the heat exchangers that move the heat across, the metals that survive sea water, and the marine growth prevention systems that keep the surfaces clean. The sizing arithmetic links to the sea water pump cooling flow calculator and the central cooling pump capacity calculator.

The 32 degree design temperature and why it governs everything

Engine power ratings are not quoted in a vacuum. IACS UR M28 (1978) sets the ambient reference conditions for determining the power of main and auxiliary reciprocating internal combustion engines on ships of unrestricted service: total barometric pressure 1,000 mbar, air temperature +45 degrees C, relative humidity 60 percent, and sea water temperature 32 degrees C as the charge air coolant inlet. Those four numbers are the tropical envelope. A maker’s power table is valid up to that envelope; the major two-stroke project guides state exactly this, citing IACS M28 for the 45 degrees C blower inlet, 1,000 mbar, 60 percent humidity, and 32 degrees C sea water.

The 32 degree figure is the design sea water inlet temperature, not the test-bed reference. ISO 3046-1, the international standard for reciprocating internal combustion engine performance, defines a separate set of standard reference conditions used to declare and correct power on the test bed: 1,000 mbar total barometric pressure, 25 degrees C air temperature, and 25 degrees C charge air coolant (cooling water) temperature. So an engine declared at ISO 3046 conditions (25 degrees C) must still make its rated output when the sea reaches 32 degrees C. The cooling system carries that 7 degree gap. Size a central cooler for 32 degrees C sea water and the plant holds rated power through the Red Sea and the Gulf in August; size it for 25 degrees C and the engine derates whenever the inlet climbs.

This is why a marine engineer reads inlet temperature first. Above 32 degrees C, found in parts of the Persian Gulf in summer, the heat exchanger’s log-mean temperature difference shrinks, the fresh-water circuit runs hotter, and the engine’s charge-air and lube-oil temperatures climb toward their alarm setpoints. The defense is design margin built around the 32 degree point, not heroics on the day.

Direct sea water cooling

Direct cooling pumps sea water straight through the jacket, the lube-oil cooler, the charge-air cooler, and any condensers. It’s the older arrangement and the simpler one: one fluid, one heat-transfer step, fewer pumps, lower first cost. Small craft, harbor tugs, and many fishing vessels still use it because the plant is compact and the crew is small.

The cost shows up on the metal. Sea water corrodes cast iron and ordinary steel, so the wetted parts have to be copper alloy, bronze, or coated, which raises the price of every component it touches. Worse, the system temperature is pinned low. Run a jacket above roughly 50 degrees C on raw sea water and dissolved calcium and magnesium salts precipitate as scale on the hottest surfaces, choking heat transfer and trapping corrosion underneath. Direct cooling therefore can’t exploit the higher jacket and oil temperatures that modern engines prefer for thermal efficiency, and the marine growth that enters with the water settles on the components that are most expensive to open and clean. Through the 1960s this was standard; central cooling has displaced it on almost all ocean-going tonnage since.

Central cooling architecture

Central cooling inserts a closed fresh-water loop between the sea and the machinery. Sea water now does only one job: it cools a single large heat exchanger, the central cooler. Treated fresh water then circulates through every engine and auxiliary. Sea water never enters the jacket, the oil cooler, or the air cooler, so those parts see clean inhibited fresh water instead of corrosive salt water, and they can run hotter.

The fresh-water side is usually split into two circuits at different temperatures. The low-temperature (LT) circuit, leaving the central cooler at around 36 degrees C in tropical conditions (a few degrees above the 32 degree sea water), feeds the loads that need the coldest water: the charge-air cooler’s LT stage, the lube-oil cooler, and the air-conditioning condensers. The high-temperature (HT) circuit, typically running 70 to 90 degrees C, cools the cylinder jackets and the charge-air cooler’s HT stage. A controlled bleed of LT water mixes into the HT return, and a thermostatic three-way valve trims jacket outlet temperature, often held near 85 to 90 degrees C for the main engine. Splitting the circuits lets each consumer sit at its right temperature with one sea water heat rejection step. It also makes the HT jacket water hot enough to drive a fresh-water generator by flash evaporation, turning waste heat into distilled water.

The fresh water is treated with a nitrite or nitrite-borate corrosion inhibitor and held slightly alkaline, around pH 8.3 to 10, with the inhibitor concentration checked by onboard test kit. Because the loop is closed, makeup is small, oxygen ingress is low, and corrosion of the expensive jacket and cooler surfaces nearly stops. That single change (moving sea water out of the machinery and into one cooler) cut cooling-related corrosion failures sharply and is the main reason central cooling is now the default on new ships above small auxiliary vessels.

Each fresh-water circuit carries its own circulating pumps, an expansion tank, and temperature control. The expansion tank sits at the top of the circuit, vents the air that comes out of solution as the water heats, holds the static head that keeps the pump from cavitating, and provides the small makeup the closed loop loses to evaporation and packing leak-off. A steadily falling tank level is the first sign of a leak that has to be found, not just topped up, because the makeup carries fresh inhibitor and dilutes the dosing. A jacket pre-heater, steam or electric, warms the HT circuit before a cold start so the engine doesn’t take thermal shock when it fires, and the same heat keeps a standby main engine warm and ready. The LT and HT pumps follow the same one-running, one-standby rule as the sea water pumps, with automatic cut-in on a fall in discharge pressure, so a circulating-pump trip doesn’t take the cooling with it.

Sea chests, high and low

The sea chest is the recess in the hull where sea water enters. It’s a welded steel box let into the shell plating, closed outboard by a hinged or bolted grating and connected inboard through a hull valve to the sea water main. The box volume lets sand, silt, and entrained air separate from the flow before the water reaches the pumps.

Ocean-going ships carry at least two: a low sea chest near the turn of the bilge and a high sea chest higher on the side shell. The low chest draws cool, clean water at sea and in deep water; the high chest is the choice in shallow or silty harbors, where the low chest would suck mud, and in heavy ice, where the high chest can be supplied with warmed recirculated water. The watch selects between them with the cross-connect valves. Each chest sits below the lightest service waterline so it stays submerged when the ship rolls or trims by the head.

A sea chest carries more than a grating and a valve. A vent line runs to the top of the box so trapped air escapes rather than reaching the pump and breaking suction. A compressed-air or steam blow-back connection lets the crew clear the grating of weed, jellyfish, or slush ice by reversing flow, and the steam line doubles as ice clearing in freezing service. Sacrificial zinc or aluminum anodes inside the chest protect the steel and the surrounding shell from galvanic attack and are renewed each dry-docking. The grating bars, on the order of 30 to 50 mm spacing, stop rope, plastic, and large organisms while passing the silt and small larvae that the strainers and the rest of the system must then deal with.

The 2023 IMO Biofouling Guidelines treat the sea chest as a high-risk niche. Resolution MEPC.378(80), adopted 7 July 2023, says internal surfaces and inlet gratings of sea chests should be protected by an anti-fouling system suited to the flow over and through the chest, and that designers should provide the capacity to block off the sea chest for cleaning and treatment. The same resolution defines a niche area as a part of the submerged surface more susceptible to fouling than the main hull because of structural complexity or variable flow: a sea chest fits that description exactly.

Duplex strainers

Between the sea chest and the pumps sit the strainers, the last barrier before the rotating and heat-transfer equipment. A strainer is a perforated basket in a housing that catches the weed fragments, shell, and grit that pass the grating. Essential cooling lines use a duplex (twin) arrangement: two baskets in parallel with a changeover valve, so the watch can isolate and clean one basket while the other carries full flow. The plant never loses cooling for a routine clean.

Strainers are watched by differential pressure. As a basket loads up, the pressure drop across it rises; a gauge or transmitter across the strainer tells the crew when to switch and clean. In fouling-prone water that can be daily; in clean open ocean, weekly. Some installations add automatic self-cleaning back-flushing strainers that scavenge the dirty fraction overboard without manual changeover. Whatever the type, the strainer protects the pump impellers from impact damage and keeps debris out of the narrow channels of a plate heat exchanger, where a single lodged shell can block a port and starve a plate column.

Sea water circulating pumps and class redundancy

Sea water cooling pumps move the cooling flow from the sea chest, through the central cooler (or through the equipment in a direct system), and overboard. They are almost always single-stage centrifugal pumps: simple, tolerant of a little entrained grit, and well matched to high flow at modest head. Heads are low, typically 1.5 to 3 bar, because the resistance is mostly friction in large-bore pipe, the cooler, and the overboard discharge; pushing high flow against high head would be wasteful and the pipework would not justify it. Wetted parts are corrosion-resistant alloys: gunmetal or bronze casings, aluminum-bronze or nickel-aluminum-bronze impellers, and duplex stainless shafts in the hardest service.

Redundancy here is a class requirement, not a design choice. SOLAS Chapter II-1, Regulation 26, requires that means be provided so the normal operation of propulsion machinery can be sustained or restored even though one of the essential auxiliaries becomes inoperative, and it lists sources of cooling water for essential systems among the auxiliaries that need this treatment. In practice classification societies translate that into at least two sea water cooling pumps per essential service, each sized for the full required flow, so one pump can be opened for overhaul while the other holds the plant at full power. Larger ships fit three or four pumps with combined capacity above demand. On a vessel certified for unattended machinery spaces, the standby pump starts automatically on low discharge pressure with an alarm, a safeguard the IACS machinery requirements call out for pumps whose automatic starting is required. The same logic extends to the fresh-water circulating pumps in a central system: one running, one standby, automatic cut-in.

Pump speed control has become the main efficiency lever. Cooling demand falls steeply when a ship slow-steams or sits at anchor, yet a fixed-speed pump keeps delivering full flow against a throttling valve, wasting power. A variable-frequency drive trims pump speed to the actual heat load, and because a centrifugal pump’s shaft power varies with roughly the cube of speed, modest speed cuts give large energy savings. Holding the central-cooler outlet at a fixed setpoint by varying sea water pump speed is now common new-build practice and a documented saving on the cooling-water account.

Plate versus shell-and-tube heat exchangers

The heat exchanger is where sea water and fresh water meet without mixing. Two types dominate marine cooling, chosen on duty, space, and the cleanliness of the water.

The gasketed plate heat exchanger (PHE) is the standard central cooler on modern ships. A stack of thin corrugated plates, clamped in a frame between two end pieces, forms alternating narrow channels: sea water in one set, fresh water in the next, counter-flowing across each plate. The corrugations force turbulence at low velocity, so a PHE achieves a high heat-transfer coefficient in a small footprint, often a quarter the volume of an equivalent tube unit. It opens for service: slacken the tie bolts, slide the plates apart, inspect and clean each one, replace a gasket, re-clamp. Duty can be raised later by adding plates to the same frame. The trade-offs are the gaskets, which set a temperature and pressure ceiling and need periodic renewal, and the narrow channels, which clog if a strainer lets debris through.

The shell-and-tube exchanger is a bundle of tubes inside a cylindrical shell, one fluid through the tubes and the other across them. It handles higher pressures and dirtier water than a PHE, tolerates a wider temperature range, and can be cleaned mechanically by rodding or hydro-blasting straight through the tubes. It’s bulkier and heavier for the same duty, which is why it has given ground to the PHE for the main central cooler, but it remains common where pressure is high, where the water is silty, or where a sturdy bundle that takes a brush is worth the space. Sea water is usually put through the tubes so the inside, not the shell, takes the corrosive and erosive duty and so cleaning is a straight pass.

Sizing balances first cost against running cost and fouling margin. A tighter exchanger costs less to buy but runs at higher velocity and higher pressure drop, demanding more pump power and leaving less reserve as surfaces foul. Marine central coolers are typically specified with a fouling margin so they still meet duty at 32 degrees C sea water near the end of a cleaning interval, not just clean and cold on the test floor.

The duty itself is a simple heat balance: the heat carried away equals mass flow times specific heat times temperature rise, $Q = \dot{m}\,c_p\,\Delta T$ . For sea water, $c_p$ is close to 3.99 kJ/kg-K and density near 1,025 kg/m3, so a chosen sea water flow and an acceptable temperature rise across the central cooler (commonly 6 to 10 degrees C) fix the heat that can be rejected. Turn the relation around and the required sea water flow follows from the total cooling load and the rise the design allows. The sea water pump cooling flow calculator runs that arithmetic for a given duty.

Sizing the cooling load and the sea water flow

The cooling-water plant is sized from the heat every consumer rejects at maximum continuous rating, plus a fouling margin. A slow-speed two-stroke main engine sends a known split of its fuel energy to the cooling media rather than to the shaft or the exhaust: the cylinder jacket, the lubricating oil, and the charge-air (scavenge air) cooler are the three large loads, and the charge-air cooler is usually the biggest single one at high rating because it pulls the heat of compression back out of the boost air. The auxiliary engines add their own jacket, oil, and charge-air loads, and a string of smaller consumers (the main air compressors, the fuel-oil cooler, the stern-tube cooler, the air-conditioning and provisions-refrigeration condensers, and the fresh-water generator’s distillate cooler) round out the total.

That total heat load, all of it, leaves through the central cooler into the sea water. So the sea water flow is fixed by the same heat balance read from the sea water side: take the sum of all rejected heat, divide by sea water $c_p$ near 3.99 kJ/kg-K and by the temperature rise the design allows across the cooler. A larger allowed rise means less sea water flow and a smaller pump, but a hotter overboard discharge and a tighter temperature approach in the cooler; a smaller rise means more flow and a bigger pump but more margin. Most designs land the sea water rise in the 6 to 10 degree band, which keeps the central cooler’s approach reasonable at the 32 degree inlet while keeping pump size and power sensible. Run the sums at 32 degrees C inlet, not at a temperate average, or the plant will be short of cooling exactly when the sea is hottest. The central cooling pump capacity calculator sizes the pump from the load, the rise, and the inlet temperature.

Materials for sea water service and galvanic considerations

Material choice on the sea water side is corrosion engineering. Bare carbon steel and cast iron pit and waste in sea water; the parts that touch it are copper alloys, titanium, or lined.

Copper-nickel alloys are the long-running workhorses of sea water piping and tube bundles. The 90-10 grade (90 percent copper, 10 percent nickel) and the more resistant 70-30 grade form a thin, adherent, protective oxide film in clean sea water and resist the impingement attack that destroys plain copper. They are velocity limited, because too-fast flow strips the protective film and erosion-corrosion sets in at bends and inlets. Published guidance from the copper-nickel materials bodies puts the maximum design velocity for heat exchanger and condenser tubes at about 2.5 m/s for 90-10 and 3 m/s for 70-30, with higher limits, near 3.5 m/s for 90-10 and 4 m/s for larger 70-30 pipe (100 mm and above), allowed in straight pipe runs. Stay under the limit and a Cu-Ni system runs for decades; exceed it at a sharp bend and the wall thins fast.

Copper-nickel brings a second benefit: it resists fouling on its own. The slow release of copper ions from the surface is toxic to settling larvae, so barnacles and mussels struggle to colonize Cu-Ni piping and tube sheets, a property that reduces the fouling burden inside the system without any active dosing. Aluminum-brass appears in older condenser tubing; nickel-aluminum-bronze is favored for pump impellers and valve internals for its strength and erosion resistance.

Titanium is the material for the most aggressive duty, and it’s why modern PHE plates are commonly titanium. Titanium carries no general corrosion rate in sea water, doesn’t pit or suffer crevice attack in clean sea water at normal temperatures, and isn’t velocity limited the way copper alloys are, so it tolerates the high local velocities inside a plate channel and the chlorination used for fouling control. The catch is galvanic: titanium is noble, so coupling it to a less noble metal in the same sea water can drive heavy corrosion of the active partner. The defenses are the standard ones for dissimilar metals in an electrolyte: isolating gaskets and dielectric unions to break the electrical path, generous sacrificial anodes to feed the galvanic current instead of the steel, and material sequencing so the small, expensive, noble part isn’t sacrificing a large active one. Where copper alloy pipe meets a titanium plate cooler or a steel hull penetration, the joint detail and the anode plan decide whether the system lasts a docking cycle or eats itself. The same cathodic-protection thinking that protects the hull applies here; see marine cathodic protection and hull coatings.

The area ratio between the noble and the active metal sets how fast the galvanic cell runs. A small steel fitting bolted into a large titanium or Cu-Ni assembly is the worst case: the whole noble area drives current into a tiny anode, and that fitting wastes quickly. Reverse the ratio, a small noble part among large active metal, and the attack spreads thin and slows. Designers therefore keep the rarely-changed noble components large and the consumable active parts (anodes, isolating spool pieces) the ones that wear, so maintenance replaces cheap parts on schedule rather than discovering a wasted hull penetration in dry dock. Pipe internals get a corrosion allowance on top of the velocity rule, so a Cu-Ni sea water main is specified thicker than a fresh-water line of the same bore to carry decades of slow film loss and the occasional local impingement at a fitting.

Marine growth prevention systems against biofouling

Sea water carries the larvae of barnacles, mussels, tubeworms, and hydroids, plus the bacteria that lay down the biofilm those larvae settle on. Inside a cooling system, that growth narrows pipe bores, blocks strainer baskets, fouls tube and plate surfaces, and cuts heat transfer; left alone it can choke a sea chest. A marine growth prevention system (MGPS) is the active defense. IMO Resolution MEPC.378(80) defines an MGPS as an anti-fouling system used to prevent biofouling accumulation in niche areas or other surfaces, and it specifically asks that internal sea water cooling systems be designed with a minimum number of bends and flanges, of appropriate material to minimize biofouling, compatible with any MGPS fitted, with dead ends and stagnant cross-overs avoided and standby pumps and piping fully integrated so water doesn’t sit still and breed growth.

The impressed-current copper-and-aluminum system is the most common shipboard MGPS. A pair of anodes, one copper and one aluminum (or ferrous), is mounted in the sea chest or a dedicated treatment tank and fed a controlled direct current. The copper anode releases copper ions at a concentration toxic to larvae before they can settle, while the aluminum (or iron) ions lay a protective film through the downstream pipework. The dosing is set to maintain a few parts per billion of copper at the points to be protected, enough to stop settlement without wasting anode.

Electrochlorination is the other main method, especially where flow rates are large. An electrolytic cell passes current through sea water, generating sodium hypochlorite in situ from the chloride already present; the hypochlorite is dosed into the cooling inlet to keep a low free-chlorine residual that kills planktonic organisms before they attach. It needs control, because the chlorine residual in the overboard discharge is an environmental limit, and many ports cap it. Both methods attack the larvae stage, which is far cheaper than removing established growth. The 2023 Biofouling Guidelines push management upstream: a ship-specific Biofouling Management Plan and a Biofouling Record Book, with sea chests treated as the niche they are, and proactive cleaning of microfouling before macrofouling can take hold. MEPC.378(80) is voluntary at the IMO level, but Australia, New Zealand, and California enforce biofouling rules on arriving ships, so the management plan is in practice a trading requirement on those routes.

Passive measures back up the active ones. Copper-nickel surfaces resist settlement on their own, copper-alloy or titanium tube sheets paired with the MGPS slow regrowth, and the mechanical clean at dry-docking removes whatever the active systems missed: sea chests, the sea water main, condenser tubes, and the central cooler plates are opened, descaled, and inspected against the same fouling-margin assumption used when the cooler was sized.

Operation and condition monitoring

Running the system well is mostly watching the right numbers. The sea water inlet temperature is read first, because it sets how much margin the plant has that day; sensors at the sea chest and pump suctions give it. The central-cooler fresh-water outlet temperature, the jacket outlet, and the lube-oil and charge-air temperatures are trended against their setpoints, and the cooling-water pump speed (on a variable-speed plant) or the bypass valve position (on a fixed-speed plant) is the control that holds them.

Fouling is caught by pressure drop and temperature approach. A rising differential across a strainer means a basket needs cleaning; a rising pressure drop across the central cooler, or a shrinking gap between the sea water outlet and the fresh-water inlet temperatures, means the cooler is fouling and a clean is due. On a well-run plant with effective MGPS, the central cooler is opened on a multi-year interval aligned with dry-docking; heavy fouling between dockings points to an MGPS fault, an undersized clean margin, or unusually dirty trading water, and is investigated rather than just cleaned away. In freezing service the watch keeps water moving and uses the high sea chest with recirculation and steam to stop ice forming in the chest, the inlet pipe, or a stopped pump. Recirculating warm overboard water back to the chest, a standard cold-climate arrangement, keeps the inlet above freezing without stopping the cooling flow. Cooling demand and pump energy both fall in port and at slow steaming, so trimming pump speed or stopping a parallel pump is routine economy, not a special measure. Logging the pump speed, the differential pressures, and the temperature approaches every watch turns slow drift into a trend the engineer can act on before an alarm, which is the whole point of condition monitoring on a system that has to run without interruption for the length of a voyage.

The cooling plant doesn’t stand alone. It rejects the heat that the marine lubricating oil systems and the charge-air side of the marine diesel engine shed, it carries the condenser load from marine refrigeration and cargo cooling and the air-conditioning of marine HVAC systems, and it serves the marine auxiliary engines and generators. Engine-room ventilation and the uptakes handle the rest of the heat the water doesn’t take; see marine engine room ventilation and uptakes.

Limitations

The figures here are design references, not promises about a specific ship. The 32 degree sea water value is the IACS UR M28 unrestricted-service envelope; a vessel notated for restricted or tropical-only service, or one trading where the sea exceeds 32 degrees C, needs the cooler and pump margins set to its own worst case, not to the rule’s default. Sea water properties shift with temperature and salinity: $c_p$ near 3.99 kJ/kg-K and density near 1,025 kg/m3 are typical of open-ocean water and drift in brackish or hot water, so heat-balance numbers should use values for the actual service water.

Velocity limits for copper alloys are guidance for clean, aerated sea water; polluted, sulfide-bearing, or sediment-laden water lowers them, and a single sharp bend or a partly throttled valve creates local velocities well above the pipe average. Treat the published 2.5, 3, 3.5, and 4 m/s figures as ceilings for well-designed straight runs, not licenses to run the whole system at the limit. PHE gasket ratings, fouling margins, and MGPS dosing are equipment-specific; use the maker’s data and the class-approved settings, not generic numbers, for any real installation.

Redundancy rules vary by classification society and by the service the pump feeds. SOLAS II-1 Reg.26 sets the principle that loss of one essential auxiliary must not stop propulsion, but the exact pump count, standby-start logic, and automation requirements come from the chosen society’s rules and the ship’s automation notation. MEPC.378(80) is a voluntary IMO guideline; the binding requirements are whatever the flag state and the ports of call enforce, and those differ. Verify every threshold against the current class rules, the maker’s manuals, and the applicable national regulation before relying on it.