Cross Scavenging in Legacy Two-Stroke Designs

Cross scavenging is the oldest gas-exchange method applied to two-stroke marine diesel engines. Both the scavenge inlet ports and the exhaust discharge ports are cut into the cylinder liner on opposite sides, at the same axial height near the bottom of the stroke. A raised wedge, called the deflector, is cast or forged into the piston crown on the scavenge-port side. This deflector redirects incoming charge upward instead of allowing it to flow straight across and out through the exhaust ports. There is no valve in the cylinder cover.

The scheme dominated marine two-stroke practice from Burmeister and Wain’s first commercial diesel installation in 1912 through the late 1940s, then retreated rapidly as loop scavenging and uniflow scavenging raised achievable rated outputs well beyond what the deflector-piston arrangement could sustain. Understanding why cross scavenging worked, and why it failed, clarifies the entire development arc of the slow-speed marine diesel engine. For the quantitative side, the engine scavenge pressure calculator and the engine BMEP calculator handle the numbers that cross-scavenging engineers worked through by slide rule.

Port arrangement and cylinder geometry

The cross-scavenged cylinder liner has two port belts, each occupying a substantial arc of the liner circumference. One belt, typically covering 55 to 70 percent of the liner circumference on the inlet side, connects to the scavenge receiver. The opposite belt, of roughly similar width, connects to the exhaust manifold. Both port belts sit at the same axial position, timed to open and close symmetrically as the piston passes bottom dead centre.

Port height in historical designs ran from about 80 mm in the smallest cylinders up to 200 mm in bore sizes of 700 to 800 mm. Width-to-height ratios were typically 4:1 to 6:1, producing rectangular windows stacked vertically around each side of the liner. Ports needed to be wide because the symmetrical timing meant scavenge and exhaust ports opened and closed simultaneously: the window of time during which scavenging could occur was the same for both port sets, giving the designer no means to hold the exhaust ports closed while scavenge pressure built, a freedom that uniflow scavenging gains through its separate exhaust valve.

The liner itself in cross-scavenged engines was typically cast iron, bored and honed, with ports formed by coring during casting or by machining after. Port bridges, the strips of liner material between adjacent port openings, needed to be wide enough to carry piston ring loads while the rings crossed the port belt. Narrow bridges led to ring snagging, breakage, and liner damage. This constrained the maximum achievable port area relative to bore area, which in turn limited the mass-flow rate of scavenge air and put a ceiling on achievable power density.

Port timing was fixed, set by the position of the port belt relative to the crankshaft. There was no mechanism to alter timing between part and full load, no variable-geometry option. The best any operator could do was adjust boost pressure via blower speed or turbocharger bypass, changing delivery ratio without touching the fundamental port timing.

The deflector-crown piston

The deflector is the defining mechanical feature of cross scavenging and the source of most of its problems. In its classic form it is an asymmetric ridge or ramp on the piston crown, positioned on the scavenge-port side, rising typically 80 to 150 mm above the flat plane of the crown in larger engines.

Incoming scavenge air strikes the sloped face of the deflector and is turned upward. Ideally, this upward jet sweeps across the cylinder cover, pushes combustion residuals downward on the exhaust-port side, and exits through the exhaust ports while fresh charge fills the volume below. In practice, the flow is not this clean: turbulence at the deflector face generates a recirculation zone on the leeward side of the deflector that traps residual gas, while direct short-circuiting along the scavenge-port wall bypasses the deflector entirely at higher delivery ratios.

Crown casting and materials

Early deflector pistons were cast iron. By the interwar period, alloy cast iron and carbon steel became standard. Some builders used a composite approach, with a forged steel crown and a cast iron skirt. The problem was thermal: the deflector crown received radiant and convective heat from combustion on its upper face, gas-wash heat from scavenge air on its lower edges, and conducted heat from the ring belt and skirt below. In the absence of effective internal cooling, the deflector tip ran 80 to 120 degrees Celsius hotter than the flat crown area on the exhaust-port side.

Oil-cooled pistons, introduced in larger bore engines during the 1940s and 1950s, helped but could not cool the deflector tip as well as the flat crown because the internal oil gallery had to follow the crown profile. In cross-sectioned deflector pistons from the period, oil galleries are visible routed around and under the deflector, but the tip itself remains a conductive dead zone. Tip temperatures in 600 mm bore engines operating at full load were measured by pyrometer insert at 380 to 420 degrees Celsius in documented trials on Sulzer RD-series engines, against a design intent of 340 degrees Celsius or below for the cast iron alloys then in use.

Piston failures

Deflector cracking was the chronic failure mode. Cracks originated at the base of the deflector, where the stress concentration from the abrupt crown geometry met the highest thermal gradient. Once initiated, cracks propagated into the ring belt or down toward the oil-gallery walls. A cracked deflector in service shed fragments into the cylinder: catastrophic to piston rings, liners, and valves.

Interwar maintenance manuals from B&W and Sulzer specify deflector inspection at intervals of 4,000 to 6,000 running hours, compared to 8,000 to 10,000 hours for flat-crown piston inspection in the same builders’ later loop-scavenged engines. This doubled inspection burden imposed direct cost and introduced additional risk from the inspections themselves: pulling a deflector piston required craning out a heavy assembly through an awkward bore alignment, with significant risk of liner damage if the piston swung during extraction.

Gas-exchange physics: where the losses are

Two-stroke gas exchange in a cross-scavenged cylinder involves three measurable loss mechanisms, each reducing the fraction of the cylinder volume that contains fresh air at the start of compression.

Short-circuit flow

The most damaging loss is short-circuit flow: fresh scavenge air that enters through the scavenge ports, crosses the cylinder at low height, and exits through the exhaust ports without displacing any residual gas. The deflector does not stop this; it only redirects a portion of the inlet jet. At low delivery ratios (below about 1.1), the upward deflector jet dominates and short-circuiting is modest. At higher delivery ratios, the energy of the incoming jet overwhelms the deflector geometry: air spills around the deflector edges and along the floor of the cylinder, reaching the exhaust ports with only a short residence in the cylinder volume.

Trapping efficiency, the fraction of the delivered air charge that remains in the cylinder when the exhaust ports close, runs 0.45 to 0.65 across the operating range of well-maintained cross-scavenged engines at delivery ratios of 1.1 to 1.4. Loop-scavenged engines in the same delivery ratio range achieve 0.50 to 0.70. The improvement from loop scavenging is real but modest; the large gain comes from uniflow scavenging, where the exhaust valve closes before the scavenge ports, trapping the incoming charge column at 0.65 to 0.85.

The practical consequence is that cross-scavenged engines needed high delivery ratios to reach acceptable scavenging efficiency. High delivery ratio requires high boost pressure or high scavenge-pump work, both of which consume energy. The parasitic cost of producing excess scavenge air to compensate for short-circuiting was a direct penalty on SFOC.

Mixing within the cylinder

Scavenge air that does deflect upward and avoid short-circuiting still encounters residual combustion gas as it traverses the cylinder cover region. The mixing is not the designer’s intent but it is thermodynamically unavoidable: the scavenge jet must push into the existing gas, and at the interface it mixes. The result is dilution of the fresh charge before the exhaust ports close.

Scavenging efficiency, the mass fraction of cylinder contents that is fresh air at exhaust port closure, reaches 0.75 to 0.82 in cross-scavenged engines. The number is a ceiling, not a typical value: achieving 0.82 requires careful port geometry, deflector shaping, and operation near optimal delivery ratio. In service, with partially fouled ports, worn deflectors, and variable boost, 0.75 is more representative of a well-maintained engine.

The formula for scavenging efficiency $\eta_{sc}$ as a function of delivery ratio $\Lambda$ under perfect-mixing assumptions is:

A perfect-mixing model predicts the scavenging efficiency of a cross-scavenged engine better than a perfect-displacement model does, because the deflector-induced flow is chaotic enough that the gas contents do approach a well-mixed condition by the time the ports close. Uniflow scavenging, by contrast, approaches the displacement-flow ideal more closely, which is why its scavenging efficiency substantially exceeds what the perfect-mixing formula predicts for the same delivery ratio.

Asymmetric combustion from the deflector

The deflector creates a combustion chamber that is not axially symmetric. At top dead centre, with the deflector protruding into the clearance volume, the space between the piston crown and the cylinder cover is uneven: taller on the exhaust-port side, shallower on the deflector side. Fuel injection nozzles, typically positioned centrally in the cylinder cover, spray into this asymmetric space.

The flame front, igniting in the fuel spray zone, travels shorter distances to the cylinder walls on the deflector side and longer distances on the open side. The deflector surface itself is at 380 to 420 degrees Celsius and acts as an ignition-delay modifier: fuel striking it partially vaporizes rapidly, while fuel reaching the cooler flat-crown region behaves more uniformly. The result is non-uniform heat release, with earlier peak pressure on the deflector side and a flatter, extended burn on the exhaust side. This degrades indicated thermal efficiency by 2 to 4 percentage points compared to a uniform-chamber engine of otherwise identical geometry, as documented in engine-indicator diagram comparisons from the 1940s.

Quantitative comparison with loop and uniflow scavenging

The table below summarizes typical measured performance parameters for the three scavenging types as applied to slow-speed two-stroke marine diesel engines in their respective production generations. Values for cross scavenging reflect engines in their mature phase from the 1930s to 1950s; loop scavenging reflects the Sulzer RD/RND generation and B&W K/L series of the 1950s to 1970s; uniflow reflects the MAN B&W MC and ME series and WinGD X series from the 1980s onward.

The SFOC gap between a mature cross-scavenged engine and a modern uniflow engine runs 50 to 70 g/kWh. At a typical large slow-speed engine rating of 18,000 kW and 6,000 running hours per year, a 60 g/kWh difference corresponds to approximately 6,480 tonnes per year of additional fuel. At USD 600 per tonne for HFO, that is roughly USD 3.9 million per year per ship. This arithmetic explains why operators did not wait for engineering elegance before switching technologies.

The BMEP ceiling matters separately. Cross scavenging cannot achieve BMEP above 12 bar because port area, deflector geometry, and thermal loading interact to cap delivery ratio and fresh-charge trapping before higher BMEP levels become reachable. Loop scavenging removes the deflector and introduces angled ports that produce a more organized flow pattern, raising the ceiling to around 17 bar. Uniflow adds the exhaust valve, allowing independent control of exhaust timing, and the ceiling rises to 21 bar in modern engines. Each step required abandoning the previous arrangement rather than simply modifying it.

Engine families that used cross scavenging

Burmeister and Wain, Copenhagen

B&W built the first series-production cross-scavenged marine diesel engines. The company’s 1912 installation in the MS Selandia used a four-stroke engine, but B&W moved to two-stroke cross-scavenged designs for its larger slow-speed engines through the 1920s and 1930s. The B&W K-series, introduced in the 1930s, started in cross-scavenged configuration and transitioned to loop scavenging in later variants. Bore sizes in the K-series ranged from 500 mm to 840 mm. By the time the L-series was designed in the early 1950s, B&W had committed to loop scavenging throughout, and the K-series cross-scavenged variants were no longer offered for new installation.

B&W’s corporate history, published in its centenary documents, records that the decision to abandon cross scavenging was driven by deflector piston durability failures at higher ratings rather than by theoretical scavenging efficiency considerations. Operators were complaining about deflector cracking at 5,000 to 6,000 hours rather than at the designed 8,000-hour interval, and the repair costs were accumulating faster than the fuel savings from any alternative.

Sulzer Brothers, Winterthur

Sulzer’s early two-stroke marine engines, built from the 1910s through the 1930s, used cross scavenging. The company designated its interwar cross-scavenged type the SD (Sulzer Diesel, single-acting) series. These engines ran bore sizes from 380 mm to 700 mm with rated speeds of 100 to 140 rpm, producing unit cylinder outputs of 180 to 380 kW per cylinder in the larger sizes.

Sulzer’s transition to loop scavenging came with the RD series in the late 1930s and the RND series in the late 1940s. The Schnuerle-type angled scavenge ports of the RD/RND arrangement, with ports directed tangentially to create a rotating air column in the cylinder, produced measurably better scavenging than the older cross-scavenged SD types in Sulzer’s own comparative trials. Sulzer published performance figures showing the RND series achieving SFOC of 202 to 208 g/kWh at rated output, against 228 to 240 g/kWh for the equivalent SD series engines at the same cylinder size. Later, the RTA series introduced uniflow scavenging, and the trend continued toward the modern WinGD X series.

The Sulzer RD and RND series remained in production and in heavy service into the 1970s on many ships. Several RND-series engines continued operating into the 1990s on bulk carriers and tankers where the fuel cost remained manageable and the mechanical condition of the engines was good. These are loop-scavenged, not cross-scavenged, designs; they are noted here only to calibrate the timeline. The cross-scavenged SD series had effectively left commercial new-build by the early 1950s.

MAN, Augsburg

Maschinenfabrik Augsburg-Nurnberg (MAN) built two-stroke marine diesel engines using cross scavenging from the 1910s through the 1930s. MAN’s G-series and early S-series used the cross-scavenged layout. The company’s 1930s double-acting two-stroke engines, where combustion occurred on both sides of the piston, were also cross-scavenged in their port arrangement, though the double-acting geometry added its own complications for the deflector: a deflector on both ends of the piston crown created a bidirectional problem that the engineers addressed with careful port phasing rather than elegant aerodynamics.

MAN’s transition to loop scavenging occurred in the 1940s and 1950s. The KSZ series and later the KSZ-C series used loop scavenging. MAN’s eventual merger with B&W (resulting in MAN B&W) produced the MC series with uniflow scavenging, which has been the standard for the company’s slow-speed marine diesels from the early 1980s onward.

Werkspoor, Amsterdam

Werkspoor, a Dutch engineering firm, produced cross-scavenged two-stroke marine engines under license from several European designers and in its own configurations from the 1910s through the 1940s. Werkspoor engines served in Dutch merchant marine applications extensively. The company’s engines were technically similar to the B&W and Sulzer cross-scavenged designs of the same period. Werkspoor ceased independent marine engine production in the 1960s following a series of corporate reorganizations.

British manufacturers

Several British shipbuilders and engine manufacturers, including Barclay Curle & Co. in Glasgow and Harland and Wolff in Belfast, built cross-scavenged two-stroke diesel engines under license from continental European designers. The British-built engines followed essentially the same port and deflector geometry as the original license designs.

Doxford of Sunderland deserves a separate note. Doxford built opposed-piston two-stroke engines in which two pistons per cylinder move toward each other, the scavenge ports are uncovered by the lower piston and the exhaust ports by the upper piston. This arrangement is inherently uniflow (charge enters from one end, exits from the other) and is not cross-scavenged. Doxford engines remained in production from 1920 through 1980 and continued in commercial service into the 2000s on some vessels; they belong to the opposed-piston engine lineage, not the cross-scavenging one.

Why cross scavenging persisted as long as it did

The transition from cross to loop scavenging was not abrupt. Loop scavenging, particularly the Schnuerle-patent angled-port design, was already understood in the 1930s and demonstrated superior scavenging efficiency in research engines. Yet major OEMs continued building cross-scavenged engines into the late 1940s and some operators continued buying them. Four factors account for the lag.

Valve-free simplicity at the power levels of the era

The cylinder cover of a cross-scavenged engine carries no exhaust valve. In the 1920s and 1930s, high-temperature exhaust valves with adequate service lives were genuinely difficult to manufacture reliably. Valve seat materials, valve stem sealing, and hydraulic actuators all required capabilities that weren’t routine. An engine without any valve in the combustion gas path was simpler to maintain at sea with limited workshops. This advantage eroded rapidly as materials technology advanced: by 1960, exhaust valves in uniflow engines were achieving 12,000-hour service intervals, and the maintenance comparison had reversed.

The power ceiling was not yet the constraint

Ships of the 1920s and 1930s were generally smaller and slower than the postwar fleet. A large cargo ship of 1930 might require 3,000 to 5,000 kW of propulsion power. A cross-scavenged engine delivering 7 to 9 bar BMEP at moderate bore sizes could meet that demand. The power ceiling of cross scavenging became a problem only when postwar containerization and the growth of the tanker fleet pushed demand toward 15,000 to 30,000 kW per ship. At those power levels, cross scavenging was simply not in the game.

Low fuel prices through most of the cross-scavenging era

Through the 1920s, 1930s, and 1940s, fuel oil prices were low enough that a 30 to 40 g/kWh SFOC advantage for loop scavenging over cross scavenging did not translate into a compelling commercial argument. The arithmetic changed after the 1973 oil crisis, when bunker prices quadrupled. By that point cross scavenging had already been displaced, but the oil price shock reinforced the direction of travel and accelerated the adoption of the uniflow engines with their further SFOC improvements.

Manufacturing inertia and installed-base economics

Shipowners who had purchased cross-scavenged engines in the 1930s and 1940s had spare-parts inventories, trained crews, and maintenance procedures calibrated to those engines. Transitioning to a new scavenging type mid-ship-life was not a natural action. OEMs also had tooling, casting patterns, and supply chains built around cross-scavenged designs. Switching required investment in new tooling and new training for service engineers. Each factor individually was manageable; together they produced a roughly ten-year lag between when loop scavenging was proven in research and when it dominated new-build orders.

The transition to loop scavenging

Loop scavenging, specifically the Schnuerle-type angled port arrangement, removed the deflector and replaced the cross-flow path with a U-shaped loop. Scavenge ports are angled, typically 20 to 40 degrees from the cylinder axis tangentially and 15 to 25 degrees toward the cylinder cover axially. This angling produces a rotating, rising air column inside the cylinder. The column rises, turns over at the cylinder cover, and descends on the exhaust-port side, pushing residuals downward and out without requiring a deflector to redirect the flow.

The loop-scavenged cylinder cover is still valve-free, which preserved the mechanical simplicity of cross scavenging’s most valued feature while eliminating its most problematic one. The piston crown became flat, or nearly flat, removing the thermal stress concentration of the deflector. Piston crown cooling became more uniform and predictable. Cylinder liner wear became more symmetric, as the loading on the ring pack no longer varied with the deflector-side combustion asymmetry.

Sulzer’s transition from the cross-scavenged SD series to the loop-scavenged RD series in the late 1930s and the RND series in the late 1940s is the clearest documented OEM transition. Sulzer’s own technical records from Winterthur, referenced in Woodyard’s Pounder’s Marine Diesel Engines (9th edition, 2009), describe the RD series as achieving scavenging efficiency of 0.83 to 0.87 under the delivery ratios achievable with the turbocharging systems of the period, compared to 0.75 to 0.80 for the SD series. The improvement in BMEP ceiling was equally significant: the RD series was rated to 14 bar, the SD series to 9 bar.

The detail of why the Schnuerle design works better than cross scavenging is that the loop flow’s organized rotation generates a more uniform displacement front. Fresh air does not arrive as a concentrated jet that must be redirected by an obstacle; it arrives as a rotating annulus that sweeps upward from all around the port periphery simultaneously. Short-circuit flow still occurs, particularly at high delivery ratios, but it is less severe and less sensitive to exact port geometry than in cross scavenging. For a more detailed treatment of loop scavenging compared to uniflow, the comparison article covers the Schnuerle port geometry, the angled-port design variants, and the further step to the exhaust valve.

The further transition to uniflow scavenging

Loop scavenging solved the deflector problem and raised the BMEP ceiling but left one structural limitation intact: both scavenge and exhaust ports are still controlled by the same piston, which means they open and close at the same crank position (with minor adjustments for port height differences). This symmetric timing means the exhaust port must already be open when the scavenge port opens, so fresh air always arrives with an open exhaust path available for short-circuiting.

Uniflow scavenging breaks this constraint by separating exhaust timing from scavenge timing. In the uniflow arrangement, scavenge ports near the bottom of the liner are uncovered by the piston as usual, but the exhaust is handled by a valve in the cylinder cover, operated hydraulically on a separate cam or electronic control schedule. The exhaust valve can close before the scavenge ports open on the upstroke, trapping the fresh charge. It can open earlier on the downstroke to blow down residual gas pressure before scavenging begins. These two degrees of freedom are simply not available to any port-only design, whether cross- or loop-scavenged.

The MAN B&W MC series, introduced in 1982 with the first production engines delivered in 1983, used uniflow scavenging throughout. It achieved BMEP values of 18 to 19 bar and SFOC figures of 168 to 172 g/kWh at rated output in its initial production configuration. The ME series, introduced in 2001 with electronically controlled fuel injection and exhaust valve actuation, pushed SFOC below 165 g/kWh on the larger bore variants while allowing independent exhaust valve timing adjustment across the load range. These figures are not achievable under any loop-scavenged or cross-scavenged design, because they depend on exhaust valve timing optimization that port-only designs cannot perform.

WinGD’s X-DF and X-series engines, the successor lineage to the Sulzer RTA series (which itself was uniflow), continue in the same direction: SFOC values below 160 g/kWh on the largest bore variants, BMEP above 20 bar, and exhaust valve timing adjusted electronically for each firing event.

Thermal loading on the deflector crown: a quantitative look

The deflector crown thermal problem is worth examining with numbers, because it defined the practical BMEP ceiling of cross scavenging in a way that no amount of port-geometry optimization could fully overcome.

Heat flux into the piston crown from combustion in a slow-speed diesel engine runs approximately 1.0 to 2.5 MW/m² averaged over the crown area. This flux is not uniform: the zone directly under the injector spray cone receives 3 to 4 MW/m² at peak, while the crown periphery receives 0.5 to 1.0 MW/m². In a flat-crown piston with a well-designed internal oil gallery, this heat is conducted away rapidly enough to keep crown surface temperatures below 300 degrees Celsius, which is within the safe operating range for nodular cast iron and for the alloy steels used in later designs.

A deflector piston changes this balance. The deflector protrudes 100 to 150 mm above the flat crown plane. The protruding surface area exposed to combustion radiation is larger than in a flat crown, but the internal gallery area available for cooling directly under the deflector is smaller, because the gallery must avoid the mechanical section needed to support the deflector. The deflector tip has a surface-area-to-cooling-gallery-area ratio roughly three to four times worse than the flat crown zone. At the same heat flux per unit area, the tip temperature runs proportionally higher.

Measured deflector tip temperatures of 380 to 420 degrees Celsius, recorded in Sulzer RD series trials, were already at the limit for the cast iron alloys in service. BMEP increases drive higher peak cylinder pressures, which in turn drive higher peak heat flux into the piston crown. Attempting to run a cross-scavenged engine at 14 or 15 bar BMEP would push deflector tip temperatures into the 450 to 480 degrees Celsius range, beyond the safe material limits of the time. Loop scavenging’s flat-crown piston, even in the same bore size, had no such ceiling: flat-crown piston tip temperatures at 14 bar BMEP ran 280 to 310 degrees Celsius, with substantial margin.

Cylinder liner interaction with the deflector

Cylinder liner design in cross-scavenged engines had to accommodate the asymmetric load imposed by the deflector piston. The deflector crown, heavier on the scavenge-port side, shifted the piston’s centre of mass laterally from the bore axis. This produced a rocking couple on the piston at each firing event: the piston tilted slightly, loading the ring pack unevenly around the bore circumference. On the scavenge-port side, ring-to-liner contact pressure increased, driving higher wear rates. On the exhaust-port side, contact pressure decreased, leading to blow-by if the rings lost conformity.

The practical result was that cylinder liner wear in cross-scavenged engines was directionally asymmetric: wear was reliably higher on the scavenge-port side of the liner. Experienced engineers expected this and calibrated their bore-measurement protocols to check the scavenge-side diameter more frequently. Bore gauges were taken at eight radial positions rather than four when inspecting cross-scavenged liners, and the maximum out-of-round limit that triggered re-boring was set more conservatively than for uniflow liner assessments.

Liner honing patterns in cross-scavenged engines also had to account for the directional wear. Plateau honing with directional tool strokes was sometimes used to produce a slightly harder surface on the scavenge side of the bore, though this was more empirical art than documented engineering in most yards of the interwar period.

Fuel injection in the asymmetric chamber

Fuel injection into the cross-scavenged cylinder chamber faced constraints that flat-crown designs did not. With the deflector occupying a portion of the clearance volume at top dead centre, the injector or injectors had to target fuel away from the deflector surface. Direct impingement of spray on the hot deflector caused partial fuel decomposition, coke deposits on the deflector, and reduced combustion efficiency.

The standard practice was to position the injector off-centre, toward the exhaust-port side of the cylinder cover, directing the spray into the larger open portion of the clearance volume. Some engines used two injectors, one on each side of the centre, with the deflector-side injector angled to spray past the deflector rather than at it. Neither arrangement was as efficient as the centrally positioned, symmetric-spray injection that is standard in uniflow cylinder covers.

Injection pressure and nozzle hole count in cross-scavenged engines were also limited compared to later designs. Injection pressures of 300 to 400 bar were typical in the 1940s and 1950s; modern ME-series engines inject at 1,000 to 1,200 bar. Higher injection pressure enables finer atomization and faster mixing, both of which improve combustion efficiency. The asymmetric chamber of a cross-scavenged engine was poorly suited to exploiting high injection pressures even if they had been available, because the asymmetric flame travel distances meant that even perfect atomization could not produce a spatially uniform heat release.

Operational considerations for surviving cross-scavenged installations

A small number of cross-scavenged engines remain in commercial service on coastal and river vessels in various parts of the world. Operators of these engines face a specific set of challenges that operators of modern uniflow engines do not.

Spare parts sourcing

No major OEM lists deflector pistons, cross-scavenged liner sections, or cross-scavenged ring sets as standard catalogue items. Parts must be sourced from specialist marine engine parts brokers, from cannibalized engines on decommissioned vessels, or machined to order from drawings by a qualifying machine shop. Lead times of 6 to 16 weeks for machined parts are common. Operators who maintain cross-scavenged engines in service should carry at least one complete spare deflector piston assembly per engine and verify the availability of ring sets before each operating season.

Modern fuel compatibility

Cross-scavenged engines were designed for fuels with sulphur content of 1.0 to 3.5 percent by mass, high viscosity, and high lubricity. Modern very low sulphur fuel oil (VLSFO) has sulphur content below 0.50 percent and different combustion characteristics. Lower sulphur content reduces the lubricating film available to cylinder liners from cylinder oil’s sulphated compounds. Operators converting cross-scavenged engines to VLSFO should use a cylinder oil with a base number (BN) of 40 to 70, calibrated to the new fuel’s sulphur content, and monitor liner wear rates closely during the first 1,000 hours after conversion. Marine gas oil (MGO) introduces a further viscosity reduction that can affect injection equipment on engines designed for heavy fuel oil.

NOx emissions and regulatory position

Cross-scavenged engines pre-date the IMO NOx Technical Code. Engines built before 1 January 2000 are not subject to Tier I, II, or III limits under MARPOL Annex VI Regulation 13, provided they have not undergone a “major conversion” as defined by that regulation. Most surviving cross-scavenged engines were built before 1965 and fall outside the Tier regulations. However, operation in Emission Control Areas (ECAs) designated for NOx under MARPOL Annex VI Regulation 13.6 (specific ECAs where Tier III applies) remains subject to flag-state and port-state interpretation. Operators should verify their vessel’s regulatory position with the flag state before entering NOx ECAs.

SOx compliance is a separate matter. MARPOL Annex VI Regulation 14’s 0.50 percent global sulphur cap applies to all vessels regardless of engine build date. Cross-scavenged engines burning compliant fuel meet this requirement on the fuel side, but the engine’s injection and combustion systems may need adjustment for optimal performance on low-sulphur fuels, as noted above.

Re-engining decisions

The re-engining threshold for a vessel with a cross-scavenged main engine is reached when the combination of rising maintenance cost, spare-parts sourcing difficulty, and fuel overconsumption exceeds the annualized capital cost of replacement. For a coastal vessel with an engine rated below 2,000 kW, this threshold may not be reached if the maintenance program is competent and the vessel’s trading profile keeps bunker consumption low. For a vessel trading internationally on heavy ocean passages, the fuel cost differential alone typically justifies replacement within 5 to 8 years from the decision point. Modern IMO-compliant engines in the 1,500 to 5,000 kW range from MAN or WinGD achieve SFOC values 60 to 80 g/kWh below any cross-scavenged engine, which at 3,000 hours per year and USD 600 per tonne can recover the re-engining cost within 4 to 6 years.

Scavenging efficiency analysis and modelling

Cross-scavenging dynamics can be analysed using the same one-dimensional gas-exchange modelling tools applied to loop and uniflow designs, though the asymmetric deflector geometry means that three-dimensional CFD is needed for accurate prediction of short-circuit fraction and delivery-ratio sensitivity.

The key relationships for any scavenging analysis begin with delivery ratio $\Lambda$ (the mass of fresh air delivered to the cylinder divided by the swept-volume mass at standard conditions), scavenging efficiency $\eta_{sc}$ (mass of fresh air retained at port closure divided by total cylinder mass), and trapping efficiency $\eta_{tr}$ (mass of fresh air retained divided by mass delivered):

For a perfect-mixing scavenging model, which approximates cross scavenging better than perfect-displacement models:

At a delivery ratio of 1.2, the perfect-mixing model predicts $\eta_{sc} = 0.699$ and $\eta_{tr} = 0.583$. Measured values in actual cross-scavenged engines at $\Lambda = 1.2$ run $\eta_{sc} = 0.76$ to $0.80$ and $\eta_{tr} = 0.53$ to $0.60$: somewhat better than perfect mixing on scavenging efficiency because the deflector does organize the flow somewhat, but close to perfect-mixing on trapping efficiency because the short-circuit losses work in the opposite direction on that metric.

Research-grade modelling of cross-scavenged cylinders uses moving-mesh CFD with the piston crown geometry (including the deflector as a dynamic boundary) and simultaneous port opening/closing conditions. The engine BMEP calculator and mean piston speed calculator provide the boundary conditions for such analyses from engine specification data.

Limitations of this treatment

This article covers cross scavenging as applied to slow-speed marine diesel engines in commercial propulsion service from approximately 1910 to 1970. Several important adjacent topics are outside its scope.

Medium-speed four-stroke diesel engines use different scavenging principles (they are not two-stroke) and are not addressed here. Two-stroke gasoline engines, which also use various scavenging schemes including cross scavenging, have different geometry and operating conditions than marine diesel engines; the quantitative values in this article do not transfer directly to small gasoline engines.

The Doxford opposed-piston engine, frequently mentioned in the same historical context as cross-scavenged engines, uses uniflow scavenging by its geometry and is covered in the Doxford opposed-piston engines article. The two lineages should not be conflated.

Performance figures cited in this article (BMEP, SFOC, scavenging efficiency) are representative of production engines in their mature operational phase. Individual engines may have fallen outside these ranges due to condition, bore wear, injection equipment state, or operating profile. The figures are consistent with published values in Woodyard’s Pounder’s Marine Diesel Engines and Gas Turbines (9th edition, Butterworth-Heinemann, 2009), Heywood’s Internal Combustion Engine Fundamentals (2nd edition, McGraw-Hill, 2018), and manufacturers’ historical product documentation.

Dates for specific OEM product transitions are based on published historical accounts and OEM documentation available through technical references. Individual ship commissioning dates, exact first-production-unit delivery dates, and trial-result publications were not independently verified against primary shipyard records for this article.

See also

• Loop Scavenging Versus Uniflow Scavenging in Two-Stroke Engines
• Uniflow Scavenging in Two-Stroke Marine Engines
• Two-Stroke Marine Diesel Engine Fundamentals
• Crosshead Diesel Engine Architecture Overview
• Scavenge Port Geometry and Timing in Two-Stroke Engines
• Sulzer Marine Diesel Engines: History and Engine Families
• Doxford Opposed-Piston Engines
• Piston Ring Pack Design for Two-Stroke Engines
• Cylinder Liner Design for Two-Stroke Engines
• Cylinder Liner Wear Monitoring
• Piston Crown Cooling in Slow-Speed Engines

Related calculators:

• Engine Scavenge Pressure Calculator
• Engine BMEP Calculator
• Engine Mean Piston Speed Calculator
• Engine BTE from SFOC
• Turbocharger Surge Margin Calculator