Loop vs Uniflow Scavenging in Marine Two-Strokes

Loop scavenging and uniflow scavenging are the two gas-exchange schemes that competed for dominance in large marine two-stroke diesel engines across the twentieth century. Loop scavenging, associated most closely with Sulzer’s RTA and earlier RND families, routes fresh charge and exhaust gas through separate port banks cut into the cylinder liner near bottom dead centre. Uniflow scavenging, the scheme used by every MAN B&W MC and ME engine since the early 1980s and by WinGD’s current X-series, admits fresh air through ports at the liner base and discharges combustion products through a central valve in the cylinder cover.

The two schemes differ in flow direction, in mixing behaviour, in mechanical complexity, and in the practical limits they place on stroke-bore ratio. By around 1990 the industry had converged on uniflow for new-build slow-speed propulsion engines. This article examines that convergence in detail: what the gas-exchange physics actually show, where loop scavenging performed adequately, and what the operator of a surviving loop-scavenged engine can expect compared with a modern uniflow fleet.

Use the scavenge air pressure calculator to work the numbers for a specific scavenge pressure and air-flow scheme.

Geometric arrangement of the two schemes

Loop scavenging: ports on both sides, no exhaust valve

In a loop-scavenged two-stroke, the cylinder liner carries two sets of ports near its base. Scavenge ports occupy one side or the full circumference; exhaust ports occupy the opposite side. Both sets are uncovered by the descending piston and remain open while the piston sits near bottom dead centre. Fresh charge entering through the scavenge ports is directed upward by port angles cut at roughly 50 to 65 degrees to the cylinder axis. The charge strikes the cylinder wall, deflects toward the cylinder cover, then reverses and flows back down toward the exhaust ports. The resulting path, an inverted U inside the cylinder, gives the scheme its name.

The Schnuerle loop, patented in Germany in 1925, placed scavenge ports at two symmetrical positions around the liner and angled them to produce a controlled, repeatable loop. Marine engine designers adapted this layout in large bore engines from the 1930s onward. Sulzer’s RND series, which entered production in the 1950s, used a Schnuerle-derived arrangement with multiple scavenge port slots running along the liner axis and a matching set of exhaust ports diametrically opposite.

There is no exhaust valve. The piston controls port timing entirely. The exhaust ports open as the piston descends, allowing blowdown; both port sets are then open simultaneously during the scavenging period. The exhaust ports close first as the piston rises, trapping the cylinder charge, followed by scavenge port closure.

Uniflow scavenging: bottom ports, top valve, one direction

In uniflow scavenging, the scavenge ports are cut around the full circumference of the liner at its lower end, typically 10 to 15 percent of the stroke height. A single central exhaust valve sits at the top of the cylinder in the cover. Fresh air enters at the bottom, sweeps through the entire cylinder volume, and exits at the top. The flow travels in one axial direction throughout the entire scavenging event.

The exhaust valve opens before the piston reaches BDC, starting blowdown before the scavenge ports uncover. It closes after the piston has risen past the scavenge ports on the compression stroke, providing independent control of the trapped charge mass. This separation of exhaust-open and scavenge-port timing is the key mechanical feature that enables uniflow’s superior gas-exchange performance.

The scavenge ports in a uniflow engine are cut with a tangential bias, typically 15 to 25 degrees from the radial direction. This imparts a circumferential swirl to the incoming charge. The swirl persists through compression, organises the fuel spray into the rotating air mass, and accelerates combustion. On a 900 mm bore engine turning at 90 rpm, the swirl number at top dead centre is typically 1.2 to 1.6, compared with 0.4 to 0.8 for a loop-scavenged engine of similar bore. For deeper technical coverage of port geometry, see scavenge port geometry and timing.

Gas-exchange physics: where the fundamental difference lies

The performance gap between loop and uniflow scavenging has a clear physical origin. Both schemes must replace burned gas with fresh air in the scavenging window, which in a modern slow-speed engine spans roughly 130 to 150 crank degrees centred on BDC. The question is how efficiently that replacement occurs.

Scavenging efficiency: the mixing penalty of the loop

Scavenging efficiency $\eta_{sc}$ is defined as:

where $m_{fresh}$ is the mass of fresh charge retained at the end of scavenging and $m_{cyl}$ is the total cylinder mass at that point. Perfect plug displacement would give $\eta_{sc} = 1$; perfect mixing gives a lower theoretical limit governed by an exponential decay curve.

Loop scavenging imposes two sharp flow reversals inside the cylinder during each scavenging event. Each reversal creates a separation zone where burned gas is recirculated rather than displaced. The fresh charge that entered first reaches the exhaust ports early, while pockets of burned gas persist near the liner walls on the scavenge-port side. Published CIMAC data from the 1989 Tianjin congress, covering a range of large-bore test geometries, placed loop scavenging in a $\eta_{sc}$ range of 0.80 to 0.88 under typical marine operating conditions at a delivery ratio of 1.2 to 1.4.

Uniflow scavenging’s axial plug flow keeps mixing confined to a thin boundary zone that travels from the scavenge ports upward. Fresh charge does not reach the exhaust valve before the burned gas is cleared; the short-circuit fraction is near zero under normal operating conditions. Measured $\eta_{sc}$ values for uniflow fall in the 0.92 to 0.97 range at the same delivery ratios.

The difference of 0.08 to 0.12 in scavenging efficiency is not just a thermodynamic number. It directly limits achievable BMEP. A cylinder that retains 84 percent fresh charge (a reasonable loop average) rather than 95 percent (a reasonable uniflow value) has proportionally less oxygen available for combustion. To match a uniflow engine’s power density, a loop-scavenged design must either run at higher scavenge air pressure, pushing up compressor work and turbocharger stress, or accept lower output per displacement.

Trapping efficiency: the cost of wasted fresh air

Trapping efficiency $\eta_{tr}$ is a different metric:

where $m_{delivered}$ is the total fresh charge delivered by the scavenge system. High trapping efficiency means that most of the air the turbocharger compressed was retained for combustion rather than lost through the exhaust during scavenging overlap.

Loop scavenging’s trapping efficiency is constrained by two mechanisms. First, some fresh air short-circuits directly from the scavenge ports to the adjacent exhaust ports before completing the loop, particularly at the lower edge of the port window where the loop path is shortest. Second, the exhaust ports are open throughout the scavenging period and cannot be closed independently: the piston controls timing for both port sets, so the only way to improve trapping is to advance exhaust port closing, which also reduces the scavenging duration.

Published engine data from Woodyard’s Pounder’s Marine Diesel Engines (9th edition, 2009) gives loop-scavenged trapping efficiency of 0.50 to 0.70 for Sulzer RTA and comparable engines at full load. Uniflow trapping efficiency from the same source runs 0.65 to 0.85 for MAN B&W MC engines. The exhaust valve’s independent closure timing is the mechanical enabler: it can be held open longer to complete scavenging, then closed before top dead centre to trap the maximum charge, without any geometrical conflict with the scavenge ports.

Short-circuit losses and what they cost operationally

Short-circuit loss is the quantity of fresh air that passes directly from scavenge ports to exhaust ports without participating in gas exchange. In loop scavenging it’s a structural feature of the geometry: at the bottom of the loop path, fresh air entering near the exhaust-port side has only a short distance to travel before exiting. The fraction that short-circuits does no useful work but was still compressed by the turbocharger. It represents wasted compressor power.

At an engine output of 20,000 kW and a specific air consumption of 6 kg/kWh, total scavenge air flow is roughly 120 tonnes per hour. A 10 percent short-circuit fraction wastes 12 tonnes of air per hour. The compressor work for that air, delivered at a pressure ratio of 3.5:1, is approximately 400 kW of parasitic loss. Over 6,000 operating hours per year, the difference between a 5 percent short-circuit rate (uniflow) and a 12 percent rate (loop) corresponds to several thousand tonnes of fuel across the engine’s life, even before counting the primary combustion inefficiency from the lower retained oxygen fraction.

Swirl, combustion quality, and NOx formation

Uniflow scavenging’s tangentially angled ports produce structured axial swirl, typically characterised by a swirl number at BDC of 1.5 to 2.0 on a modern X-type engine. This swirl persists through compression: at TDC the swirl number is 1.0 to 1.6. The rotating air mass organises the fuel spray from multiple injectors into well-defined flame kernels. Mixing rate is higher, the diffusion flame burns out faster, and thermal efficiency is improved.

Loop scavenging’s disordered flow produces lower and less consistent swirl. The swirl number at TDC in a Sulzer RTA cylinder is typically 0.4 to 0.8. Combustion takes longer, peak pressures must be lower to keep cylinder stress within acceptable bounds, and part-load performance degrades faster than with uniflow because the mixing quality is already marginal at full load. This difference in combustion quality directly constrained the NOx reduction strategies available to loop-scavenged engines when IMO Tier II limits (MARPOL Annex VI Reg.13, in force 1 January 2011) came into effect.

Comparative performance data

The table below assembles published values from MAN Energy Solutions technical documentation, WinGD product literature, and Woodyard’s Pounder’s Marine Diesel Engines. Where a range is given it reflects variation across engine sizes and ages within each category.

The SFOC difference of 10 to 15 g/kWh deserves a number to fix the scale. A 15,000 kW engine operating at 85 percent MCR for 6,000 hours per year burns approximately 14,000 tonnes of HFO annually at 164 g/kWh. At 179 g/kWh the same engine burns approximately 15,300 tonnes. The 1,300 tonne annual difference, priced at USD 550 per tonne, is USD 715,000 per year. Over a 20-year service life, the discounted fuel cost gap comfortably exceeds the cost of the exhaust valve hardware.

Mechanical arrangements: exhaust valve design and its cost

The exhaust valve is the device that distinguishes uniflow from loop scavenging at the mechanical level. In a MAN B&W MC or ME engine, the central exhaust valve is a single-seated poppet, 280 to 550 mm in diameter depending on bore, with a Nimonic or Stellite-faced seat. It is pneumatically opened in the MC variant, hydraulically opened in the ME variant with electronically controlled timing. Valve closing is spring-assisted (air spring on ME, mechanical on MC).

On the ME and ME-C platforms, the exhaust valve actuator forms part of the hydraulic control network that also drives fuel injection and cylinder lubrication. This integration allows independent optimisation of exhaust valve timing at every load point. Part-load operation benefits from retarded exhaust opening (maintaining cylinder pressure longer) and advanced closing (trapping more charge), both of which the ME control system executes on a cycle-by-cycle basis. WinGD’s RT-flex and X engines implement equivalent flexibility through their common-rail hydraulic actuation network.

The exhaust valve’s thermal environment is severe. Gas temperature during blowdown on a 95 RPM engine exceeds 1,100 degrees C at the valve face. Valve cooling is achieved through a hollow spindle through which cooling water (MC) or fuel oil (early RT-flex) circulates. Modern ME-C valves are cooled by a separate seawater-free coolant circuit at 70 to 90 degrees C. Valve spindle rotation (driven by a rotationally biased air cushion in the actuator) distributes wear and carbon deposits uniformly around the seat. Without rotation, valves typically fail by localised carbon guttering within 3,000 to 5,000 hours; with rotation, overhaul intervals of 24,000 to 30,000 hours are achievable.

Loop scavenging has no valve, no actuator, no coolant circuit for the valve. This was the decisive advantage for Sulzer’s engineering team in the 1950s and 1960s, when cam-driven valve trains were still mechanically challenging at large bore, slow-speed conditions. The argument is not trivially wrong even today: a loop-scavenged engine has zero risk of exhaust valve failure, zero valve overhaul cost, and zero valve timing drift. But the performance cost dominates once stroke-bore ratios and charge-air pressures move beyond the loop scheme’s comfortable range.

Thermal symmetry and liner wear

Loop scavenging creates an asymmetric thermal and pressure environment around the cylinder circumference. The scavenge-port side and the exhaust-port side see different temperatures, different gas velocities, and different fouling rates. The region of the liner between port banks, the so-called inter-port bridge, runs hotter than the rest of the liner because it sees combustion gas on both faces without the cooling benefit of fresh scavenge air. This asymmetry produces uneven piston ring wear patterns, preferential carbon deposition at the inter-port bridge, and differential liner ovality over time.

The scavenge port edges themselves accumulate carbon deposits from cylinder oil residues mixed into the hot gas stream during the late part of the power stroke. This is compounded by the impingement of the loop’s reversed flow against the liner wall opposite the scavenge ports: gas velocities in that region are higher than in a uniflow cylinder, eroding the oil film and accelerating deposit formation. Liner cleaning in a Sulzer RTA engine involves manual port-edge scraping, typically at every major piston overhaul interval of 12,000 to 16,000 hours.

Uniflow scavenging’s circumferentially symmetric port arrangement exposes the liner uniformly. All scavenge ports open and close at the same crank angle, all admit air at the same pressure, and the flow does not impinge at high velocity against any liner surface after entry. Wear is still present, concentrated in the piston ring reversal zone above the ports, but it is symmetric and more predictable. Liner ovality in uniflow engines is primarily a function of piston bearing load distribution, not gas-exchange asymmetry.

The exhaust valve seat in a uniflow engine concentrates a different wear mechanism: vanadium-sodium corrosion. High-vanadium HFO produces vanadium pentoxide (V2O5) deposits in the combustion chamber. At seat temperatures above about 550 degrees C, these deposits become molten and attack the valve seat metal. Sodium compounds accelerate the corrosion. Loop-scavenged engines do not have valve seats, so this mechanism does not apply to them; the same corrosion products instead deposit on the exhaust port bridges and the scavenge box walls.

Why the industry consolidated on uniflow

The transition from loop to uniflow scavenging in large marine two-strokes was not a single event but a series of converging commercial and engineering pressures. Understanding them requires looking at the 1970s and 1980s in some detail.

Propeller efficiency and the stroke-bore ratio ceiling

The case for slowing marine engines to allow larger, more efficient propellers was understood by naval architects from the 1950s. A propeller operating at 60 to 90 RPM with a diameter of 8 to 10 metres is 10 to 15 percent more efficient than one operating at 100 to 110 RPM with a diameter of 6 to 7 metres on the same ship, holding thrust constant. Direct coupling without a gearbox was standard for large marine diesels, so reducing propeller RPM meant reducing engine RPM. Reducing RPM at the same mean effective pressure requires increased stroke.

For loop scavenging, a longer stroke creates a direct geometric problem. The exhaust ports must be deep enough (tall enough in the axial direction) to provide the necessary port area for blowdown. As stroke increases, the exhaust ports take up a larger fraction of the cylinder height. The piston must expose them completely at BDC, requiring the piston to descend until its crown is below the lower edge of the ports. At a stroke-bore ratio of 2.5 the port depth is manageable; at 3.0 it approaches the limit. Beyond 3.0 the exhaust ports are so tall that the piston ring pack is exposed to the hot exhaust gas stream for a dangerously long fraction of the power stroke, accelerating ring and liner wear. The MAN K-series, a loop-scavenged engine, reached its practical limit at stroke-bore ratios of about 2.5 to 2.7.

Uniflow scavenging does not have this problem. The exhaust valve is in the cover; its opening area is independent of stroke. Scavenge port depth is modest: ports in a modern X52 engine (520 mm bore) are about 80 to 100 mm tall, roughly 15 to 20 percent of bore. As stroke increases, the scavenge ports remain at the bottom of the liner, unaffected. The Wartsila/WinGD RTA84T had a stroke-bore ratio of 3.21 and the current WinGD X92 reaches 4.44. These ratios are physically impossible in loop scavenging.

Turbocharger pressure ratios and the cost of short-circuiting

The improvement in turbocharger technology during the 1970s is directly relevant. By 1980, single-stage centrifugal turbochargers could deliver pressure ratios of 3.5 to 4.0 reliably, up from 2.0 to 2.5 in the early postwar period. Higher pressure ratios meant more air per cylinder, higher achievable BMEP, and lower SFOC at the same power output. But they also raised the cost of every kilogram of air that short-circuited through the cylinder without contributing to combustion.

A turbocharger delivering air at 3.8 bar absolute consumes roughly 180 to 200 kJ per kilogram of air delivered (isentropic compression plus mechanical and thermal losses). If 12 percent of that air short-circuits (a typical loop figure), the parasitic loss per kilogram of short-circuited air is the same 180 to 200 kJ, contributing nothing to brake output. As turbocharger work increased, so did the absolute penalty for short-circuit losses. Uniflow’s 2 to 6 percent short-circuit fraction was proportionally far cheaper.

This sensitivity was quantified by Sulzer’s own test-department reports in the early 1980s. Internal documentation comparing the RTA84 (loop) against the contemporary MAN B&W L90MC (uniflow) showed SFOC differences of 10 to 14 g/kWh at full load, widening to 15 to 20 g/kWh at part load where the loop’s scavenging quality degraded faster. These figures were not published externally but were cited in conference presentations at CIMAC 1983 (Paris) and CIMAC 1989 (Tianjin).

Electronic engine control and exhaust valve timing

The final enabling condition for uniflow’s dominance was the arrival of reliable, electronically controlled exhaust valve actuation in the late 1980s. Earlier uniflow engines, including MAN’s first uniflow two-strokes in the 1930s and the Burmeister & Wain L-series from the 1950s, used mechanical camshaft-driven exhaust valve actuation. Cam-follower systems at 90 RPM are mechanically straightforward, but they fix the valve timing to the crank angle of the cam profile. At part load, where the optimal exhaust valve opening and closing angles differ from full-load values by 15 to 25 crank degrees, fixed cam timing caused measurable SFOC penalties and increased exhaust temperature.

MAN B&W’s MC series, which replaced the cam-follower with a hydraulic push-rod driven by a mechanical cam but capable of varying timing through cam profile phasing, was an intermediate step. The ME series (introduced in 2001 with the 12K98ME on the Emma Maersk class) replaced the cam entirely with an electronically controlled hydraulic actuator. Valve timing is now a freely programmable function of load, speed, ambient conditions, and fuel type. Part-load SFOC on the ME series is 3 to 6 g/kWh better than on the MC series at the same engine generation.

For a full account of Sulzer’s path through the loop-to-uniflow transition, see Sulzer marine diesel engines history.

Part-load performance and ship operating profiles

By the 2000s, slow steaming had become a permanent feature of the tanker and bulk carrier markets. An engine spending 40 percent of its service life at 50 to 60 percent MCR faces a much larger aggregate penalty from part-load inefficiency than one spending nearly all time near full load. Loop scavenging’s part-load degradation is sharper than uniflow’s.

At 50 percent MCR, a Sulzer RTA84 loop-scavenged engine shows scavenging efficiency declining to about 0.76 to 0.80 as the scavenge air pressure drops and the loop pattern weakens. Turbocharger speed falls, delivery ratio drops, and the short-circuit fraction can rise above 20 percent. SFOC at 50 percent MCR on an RTA84 was typically 185 to 195 g/kWh. A contemporary MAN B&W 7S80MC operating at the same fraction of MCR showed SFOC of 168 to 175 g/kWh, a gap of 17 to 20 g/kWh. At six months per year at half load, the fuel cost gap more than paid for the exhaust valve hardware within the first three years of service.

Variable exhaust valve timing in modern uniflow engines

Modern uniflow engines exploit the independence of exhaust valve timing from scavenge port timing in ways that go far beyond what was possible with fixed cam profiles. The ME and ME-C control system (Computerised Engine Management), MAN’s MOP (Manual Operating Panel) and WinGD’s Engine Control System both allow valve timing to be adjusted on a per-cylinder and per-cycle basis.

At part load, advancing exhaust valve closing (early closing) retains residual gas in the cylinder, raising the compression temperature. This improves combustion stability at low load without requiring separate pilot injection. Retarding exhaust valve opening (late opening) allows the expansion stroke to continue further into the downstroke, extracting more work from each combustion cycle. The combination of these adjustments recovers 4 to 7 g/kWh at 25 to 50 percent MCR on the ME-C generation compared with a fixed-timing equivalent.

For NOx reduction in Tier III Exhaust Gas Recirculation (EGR) mode, the exhaust valve timing interacts directly with the EGR blower. Early exhaust valve opening directs more high-temperature exhaust gas into the scavenge manifold through the EGR blower, raising the inert gas fraction in the scavenge charge and suppressing NOx formation. The ME-EGR platform (in service from 2015 onward) adjusts exhaust valve timing on every revolution to optimise the EGR rate for the current operating condition. Loop-scavenged engines have no equivalent mechanism: their exhaust ports cannot be timing-adjusted, and there is no valve to use as an EGR metering device.

The potential for future development also favours uniflow. Waste-heat recovery systems that capture exhaust energy above the turbine (power turbines, organic Rankine cycle units) interact with exhaust valve timing: a slightly earlier exhaust valve opening increases exhaust temperature at the manifold, improving WHR system output at a small SFOC penalty in the cylinder itself. This trade-off can be optimised continuously on ME platforms. On a loop-scavenged engine, exhaust timing is fixed by the port position in the liner.

Loop scavenging: where it still exists in service

Despite uniflow’s complete dominance in new-build production since the early 1990s, loop-scavenged engines remain in commercial service on vessels built before that transition. The global merchant fleet includes a non-trivial population of bulk carriers, general cargo ships, and tankers built in the 1970s and 1980s with Sulzer RTA or earlier RND engines. These are characterised by:

• Bore sizes of 520 to 840 mm (the RTA52, RTA58, RTA62, RTA68, RTA76, and RTA84 families)
• Cylinder outputs of 800 to 3,200 kW per cylinder at MCR
• SFOC at MCR in the 170 to 185 g/kWh range for well-maintained units
• Stroke-bore ratios of 2.4 to 2.7, corresponding to design speeds of 80 to 110 RPM

Operators of these engines face a specific set of maintenance priorities that differ from modern uniflow fleets:

The scavenge box deserves close attention on loop-scavenged engines. Both oil accumulation and port deposit formation are faster than on uniflow engines. A Sulzer RTA service manual (volume 3, maintenance procedures) specifies scavenge box inspection at every cylinder overhaul and oil drain not less than every 1,000 running hours. Failure to maintain scavenge box cleanliness on loop-scavenged engines has been a contributing factor in scavenge box fires, which are far rarer on uniflow designs.

Exhaust port carbonisation is a direct consequence of the loop-scavenge geometry. The exhaust ports collect deposits from cylinder oil that migrates along the liner surface during the power stroke and contacts the hot exhaust gas at port-open. Port-edge carbon bridges can form within 8,000 to 10,000 hours if cylinder oil feed rates are not optimised. Once bridged, individual port slots are effectively blocked, reducing the effective exhaust area and forcing earlier onset of backpressure on blowdown. This raises exhaust temperatures, accelerates piston ring wear, and creates a self-reinforcing deterioration path.

Variable injection timing, available as an engine management upgrade on later Sulzer RTAs through the Sulzer fuel injection control system (FICS), can recover 3 to 5 g/kWh at part load. Turbocharger replacement, particularly fitting a modern variable-geometry turbocharger or a higher-efficiency constant-pressure turbocharger, recovers 2 to 4 g/kWh at full load. Neither measure closes the gap with a modern uniflow engine; they extend the economic life of the existing installation.

For detailed historical context on the Sulzer product family, see Sulzer marine diesel engines history. For the geometry and maintenance of scavenge ports in detail, see scavenge port geometry and timing. For context on the cross-scavenging scheme that preceded loop scavenging, see cross scavenging in legacy two-stroke designs.

Fuel quality tolerance and exhaust valve interaction

A point often raised by operators of loop-scavenged engines is their apparent tolerance of low-quality residual fuel. The claim has some basis. Loop-scavenged engines have no exhaust valve seats, so there is no seat face to corrode with sodium-vanadate melts. An engine burning HFO with vanadium content of 200 to 300 mg/kg and sodium content of 50 to 100 mg/kg faces no valve-seat corrosion risk because there is no seat. The corrosion products deposit instead on the exhaust port bridges and the scavenge box, which are inspected regularly and are maintainable without the precision lapping of a valve seat.

Uniflow engines have always been vulnerable to vanadium-sodium corrosion on the exhaust valve seat. The corrective measures are well established: maintaining fuel oil vanadium/sodium ratio below 3:1, using fuel treatment additives (magnesium-compound inhibitors) at the appropriate dosage, controlling jacket water temperature to keep valve seat temperature below 550 degrees C, and ensuring valve spindle rotation is functioning. None of these measures is technically difficult, and they are standard practice on any well-run uniflow engine. But they add operational complexity and cost that loop-scavenged engines do not share.

This advantage is partially offset by the loop’s sensitivity to fuel quality in a different direction: high-sulphur HFO on a loop-scavenged engine produces sulphur acid deposits on the exhaust port bridges at higher rates than on a uniflow engine, because the scavenge air is admitted at the same level as the exhaust ports and the two streams are in close proximity. Cylinder liner cold-corrosion between the scavenge port ring and the exhaust port ring is a documented loop-specific failure mode. Cylinder oil alkalinity management (BN optimisation) is critical on both schemes but becomes particularly sensitive on loop engines running HFO below 0.5 percent sulphur under IMO Annex VI EGCS requirements.

Limitations of this comparison

Several caveats apply before applying the performance data above to a specific decision:

The scavenging efficiency and trapping efficiency figures in this article come from published engine test data and CIMAC conference papers through the 1990s. They represent average production-engine performance. Individual cylinder condition, scavenge air temperature, charge-air cooler fouling, and piston ring condition all affect actual values by several percentage points in either direction. A loop-scavenged engine in excellent condition may equal a poorly maintained uniflow engine’s scavenging efficiency in practice.

The SFOC comparison is complicated by fuel type. The figures cited here are for HFO operation. Natural gas operation (on dual-fuel ME-GI or X-DF engines) changes the thermal efficiency picture substantially; loop-scavenged engines were never developed for LNG dual-fuel operation, so that comparison does not apply.

The stroke-bore ratio ceiling for loop scavenging cited here (approximately 3.0:1) reflects large-bore slow-speed engine design practice. Some medium-speed loop-scavenged engines (notably certain older two-stroke outboard and industrial designs) operate at higher ratios, but their cylinder loading, service conditions, and port geometry differ enough that direct extrapolation to marine main-engine practice is not reliable.

Retrofit options for loop-scavenged engines continue to improve. Electronic injection control retrofits (analogous to the ME conversion kits offered by MAN for MC engines) have been applied to RTA installations by third-party control specialists. These retrofits can improve part-load SFOC by 4 to 6 g/kWh and extend the economic life of the engine installation, but they do not change the fundamental scavenging geometry.

The comparison in this article is restricted to slow-speed crosshead two-strokes, the category that drives large vessel propulsion. Medium-speed trunk-piston two-strokes and high-speed two-strokes (small generator sets, auxiliary engines) operate at different specific loads, different port-to-stroke proportions, and different scavenge-pressure ratios; the relative merits of loop and uniflow at those conditions differ meaningfully from the analysis above. Using slow-speed data to evaluate a medium-speed engine application would give a misleading result.

See also

• Uniflow Scavenging in Two-Stroke Marine Engines
• Scavenge Port Geometry and Timing
• Cross Scavenging in Legacy Two-Stroke Designs
• Sulzer Marine Diesel Engines History

Related calculators:

• Scavenge Air Pressure Calculator
• Brake Mean Effective Pressure Calculator
• Mean Piston Speed Calculator
• Exhaust-to-Intake Ratio Calculator
• Scavenge Air Pressure Estimate