WinGD low-speed Engines Licensees Conference 2015

WinGD low-speed Engines Licensees Conference 2015 Low- and high-pressure dual-fuel Technology Evaluation Process; Case Studies for LNG Carriers and Merchant Vessel ABSTRACT This document describes the low- and highpressure dual-fuel evaluation process, as Wärtsilä s/wingd s basis for selecting the best for its dual-fuel low-speed 2-stroke engine development. Differences between both technologies are revealed and analysed regarding their advantages or disadvantages. Beside the analytic approach, two case studies show the practical differences of both technologies from the customer s point of view. INTRODUCTION In the debate about the fuel of the future, natural gas is seen as the transitional fuel of the future due to its availability, environmental friendliness and cost competitiveness. In 2013 Winterthur Gas & Diesel Ltd. (WinGD), formerly Wärtsilä s 2-stroke division, introduced low-pressure low-speed engines to the emergent gas-fuelled vessel market. In 1986 WinGD successfully tested high-pressure gas diesel, applied on one cylinder, of a fullscale test engine. However, the market was not ready for LNG fuelled vessels and consequently this was not introduced at that time. In recent years it became clear that the market is now ready for LNG as ships fuel and therefore the gas had to be adapted to the latest engine generation. Instead of just adapting the already proven high-pressure, it was decided to consider also the low-pressure which had been successfully introduced on Wärtsilä s 4-stroke engines and has since become the industry standard gas supply system. Clearly, the which brings greatest added value to ship-owners and operators should be selected. Hence, among the key drivers in the selection process were outstanding environmental friendliness, simple and maintenance-friendly installation combined with the lowest possible investment costs, as well as the possibility of port-to-port operation on gas. And, the safety of the total gas installation, i.e. engine and ancillary system, including redundancy in case of gas system failures, was paramount. At the end of the process, low-pressure gas was selected by WinGD because it had proven to provide the highest benefits, especially regarding emissions, capital investment and maintenance costs. Ultimately, the lowest possible total cost of ownership (TCO) best supports the business of ship owners and operators. This paper demonstrates the evaluation process based on two case studies, one for an LNG/boil-off gas fuelled LNG carrier, and one for an LNG-fuelled merchant vessel. Licensees Conference 2015, Interlaken 1 / 14

HISTORY The first dual-fuel engine was launched as early as 1972 as the Sulzer 7RNMD90. However, at that time the market was not ready for gasfuelled vessels and, aside from that one engine, no further dual-fuel engines followed. The engine applied low-pressure gas admission via the cylinder cover, as at that time the engines were not yet uniflow scavenged, i.e. the exhaust left the engine via exhaust gas ports at the bottom end of its cylinder liners. Engine power output or, to be more precise, mean effective pressure (MEP), as well as engine efficiency were also at a much lower level than for the modern engines. Consequently, this early engine is not included in the low- and high-pressure gas evaluation process. In 1986, one cylinder of a test engine on IHI s (now Diesel United) test-bed was converted to dual-fuel with high-pressure gas injection. Gas was injected via the cylinder cover with 300 bar pressure and the test was passed successfully. At that time it also had to be accepted that the market was not yet ready for 2- stroke gas-fuelled engines. High-pressure gas 4-stroke engines were also developed. They were introduced to the market in 1987 and installed onboard seagoing vessels. Operational experience indicated satisfactory engine operation, but concerns on the operator s side regarding the high-pressure gas supply, as well as difficulties in operating and maintaining the high-pressure fuel-gas supply system, prevented a breakthrough in the market. Consequently, the next development step was to focus on low-pressure. As a first step spark-ignited pure-gas engines were introduced in 1992, but with the drawback of having no redundancy in case of possible gas supply failures, e.g. due to gas alarms, etc. Therefore, in a second step, pilot fuel ignited dual-fuel engines were launched. The introduction of the dual-fuel engines in 1995 can be seen as the historical breakthrough of gas-fuelled engines in the marine propulsion market. At the end of the first decade/beginning of the second decade of this century, it became clear that the marine market was now becoming ready for gas-fuelled low-speed 2-stroke engines. As a consequence, the decision to develop a modern dual-fuel low-speed engine was taken. However, first of all it was necessary to decide which should be selected. Following the well-known diesel combustion process with high-pressure seemed to be the easiest way, as it had already been successfully tested and only needed adapting to the latest engine generations, while the development of a modern low-pressure uniflow scavenge engine presented a real challenge. Otherwise, as history has shown, it was lowpressure dual-fuel which brought the market breakthrough for medium-speed gas-fuelled engines. Finally the evaluation described below was carried out. Based on the clear and outstanding advantages of low-pressure, together with the possibility of utilising the knowledge acquired in the development of medium-speed, low-pressure dual-fuel engines, the decision was made to develop low-pressure, low-speed 2-stroke engines. In 2013 these engines were officially introduced to the market. ENVIRONMENT Natural gas is principally a mixture of pure methane plus some amounts of ethane, propane, butane and marginally heavier hydrocarbon gases. Some other gases, mainly nitrogen, can also be present. Global warming Thanks to the higher hydrogen-to-carbon ratio of natural gas in comparison to liquid fuels, its use leads to a direct reduction in carbon dioxide (CO2) emissions. However, in order to correctly assess the global warming potential of an engine s emissions, Total Unburned Hydrocarbon (THC) emissions also need to be considered. In the Diesel Cycle combustion process on a highpressure dual-fuel engine, the fuel-gas is burned in the gas-jet flame, i.e. the combustion itself is a locally rich combustion and consequently most gas-fuel is burned. Hence, emissions of unburned hydrocarbons, mainly methane, are kept at a very low level. In the Otto Cycle combustion process on a low-pressure dual-fuel engine, a lean fuel-gas/ pre-mixture is Licensees Conference 2015, Interlaken 2 / 14

burned. Thus, it is possible to have localised areas of fuel-gas/air mixtures which are too lean and, as a result of this, some fuel-gas might not be burned completely. As a consequence, emissions of unburned hydrocarbons and methane (also known as methane slip) can be higher. The medium-speed engine development experience has shown that emissions of unburned hydrocarbons and methane slip can be significantly reduced by improving the engine design. This involves optimising the mixing process of the fuel-gas and the air, as well as avoiding dead volumes and crevices in the combustion chamber which could create areas of excessively lean fuel-gas/air mixtures. In addition, a suitable method of ignition is required for the whole fuel-gas/air mixture to maximise complete combustion. Considering that according to the ICCP report Climate Change 2013, methane has a global warming potential 28 times higher than CO2, the global warming impact of a low-pressure dual-fuel engine is still at least 15% lower than that of a conventional engine, while a high-pressure dual-fuel engine allows a reduction of up to 23%, based on Wärtsilä s measurement results. In this aspect, high-pressure currently shows some advantage, while it is expected that further engine design improvements will further reduce the gap between both technologies. Sulphur oxide emissions (SOX) Sulphur oxide emissions are significantly reduced by both dual-fuel technologies, since natural gas is practically sulphur-free and consequently the remaining source of sulphur oxide emissions is directly linked to the sulphur content of the applied pilot-fuel and the required pilot fuel amount. Less specific pilot fuel is needed for low-pressure dual-fuel engines and consequently low-pressure provides a marginal environmental advantage. However, as the difference is just marginal, this aspect is not seen as a driver for decision making. Nitrogen oxides (NOX) As just described, the combustion processes of the Diesel and Otto Cycles are very different. As a consequence of the rich diesel combustion process, local high-temperature peaks are created in the combustion chamber, and it is these higher temperatures which lead to the reaction between oxygen and nitrogen, i.e. NOX is formed. By contrast, the combustion process of the lean-burn Otto Cycle ensures a very equal combustion temperature distribution and so avoids high peak temperatures which ultimately keeps NOX emissions extremely low. These emissions are, all the time, significantly lower than defined by the IMO Tier III limit. This is seen as a clear benefit of low-pressure gas admissions. Particulate Matter (PM) The difference between lean and rich combustion processes also has an influence on PM emissions, which are significantly reduced by the Otto Cycle combustion process. The remaining PM emissions of a low-pressure dualfuel engine are mainly related to the Diesel Cycle pilot-fuel combustion. A high-pressure Diesel Cycle dual-fuel engine also helps to reduce PM emissions in comparison to a conventional diesel engine, since natural gas is, compared to liquid fuel, a significantly lighter hydrocarbon fuel and also has a higher hydrogento-carbon ratio. However, the emission levels still reach almost two-thirds of a conventional diesel engine. Besides the clear environmental impact of PM emissions, legislative limitations are expected as a challenge in the near future, and thus low-pressure gas has clear advantages. TECHNICAL ASPECTS General development challenges The development challenge for a high-pressure dual-fuel engine solution can principally be seen in the safe handling of the fuel-gas when pressurised to at least 300 bar. The 2-stroke engine test in 1986 showed that handling fuel gas at a pressure of 300 bar is, in principle, possible and field experience with medium-speed high-pressure dual-fuel engines confirmed this conclusion. The development of low-pressure dual-fuel engines is more demanding. The engine requires the correct homogeneous fuel-gas/air mixture - not too rich and not too lean which should be evenly distributed within the whole combustion chamber. Hence, the correct amount of combustion air and fuel-gas needs to be mixed in the combustion chamber, while ensuring an even mixing, resulting in a homogeneous mixture. Local areas of over-rich fuel-gas Licensees Conference 2015, Interlaken 3 / 14

concentrations could lead to self-ignition, while areas of over-lean fuel-gas concentrations might not ignite and, consequently, the unburned gas would be lost in the exhaust. The fuel-gas injection solution as applied on the Sulzer 7RNMD90 was not seen as a suitable solution since modern engines are uniflow scavenged. Besides finding good solutions for controlling the combustion air quantity and fuelgas mixing, the compression ratio needs to be reduced in the Otto Cycle gas operation mode but kept as high as possible for the Diesel Cycle operation mode. WinGD s modern, fully-electronically controlled engines provide the basis for ingenious solutions to these challenges. Most conventional diesel engines are selected in the lower range of their rating fields due to the fuel-saving potential of de-rating. Low-pressure dual-fuel engines have quite similar performance within their whole rating field, and are even slightly better in upper area and, thus, the rating point can be freely selected within the whole rating field. The theoretical drawback of the reduced maximum engine output is, in most cases, practically no issue, as can be seen in Figure 1, which shows the selected rating points of current X62 engine projects in comparison to the X62DF s rating field. Safety Irrespective of which is applied, safety onboard the ship has to be on the same level as with conventional diesel engines. Therefore, double-walled pipes are required for both solutions, so that in case of any fuel-gas leakage, the gas cannot enter the gas-safe engine room, and the leakage can be detected immediately. The amount of fuel-gas in the engine room piping system is generally similar for both solutions, as the fuel-gas is either transported in larger diameter pipes at lower pressure/density or in smaller pipes at higher pressure/density. However, in case of a fracture in a pipe or a pipe connection, high-pressure gas has a higher expansion energy potential. However, it can be concluded that both solutions can be made safe. Engine performance A high-pressure Diesel Cycle dual-fuel engine applies the same combustion process as a conventional diesel engine. Thus, in general no basic performance differences are expected. By contrast, with an Otto Cycle low-pressure dual-fuel engine, some performance differences need to be considered due to the different combustion process applied. Power output: The maximum power output of a dual-fuel engine employing the Otto Cycle is lower than that of an engine employing the Diesel Cycle. This can be seen as a drawback of the low-pressure and a deeper evaluation is needed in order to establish if a power density deficit is acceptable or not. Figure 1: Rating field comparison: conventional diesel engine versus dual-fuel engine. Methane number dependency: As the fuel-gas/air mixture is being compressed in the combustion chamber before controlled ignition, uncontrolled self-ignition needs to be avoided. Since the fuel-gas is a mixture of different gases, its resistance to self-ignition, indicated by the methane number (MN), can vary. For fuel-gas with a lower stability against self-ignition, maximum engine output needs to be reduced. This could be another drawback, and therefore deeper investigation is needed. The engine output which can be reached under all conditions will be at least as indicated in Figure 2. For conditions colder than the tropical design condition, more engine output can be attained. Licensees Conference 2015, Interlaken 4 / 14

Engine Power 105% 100% 95% 90% 85% 80% 75% 70% 65% rating on R1 to R3 line rating on R2 to R4 line specific energy consumption Diesel cycle specific energyconsumption Otto cycle 60% 55% valid for W-XDF engines 50% 60 65 70 75 80 85 90 95 100 MN Figure 2: Engine power output in gas operation Considering that the majority of available fuelgas has a methane number in the range of 70 to 90, an LNG- fuelled vessel can always be operated up to its service speed, which is typically reached at 85 to 90 % of the contracted maximum power output. For LNG carriers, using LNG as both their cargo and their fuel, the situation is somewhat different. At about MN100, the natural boil-off gas (NBOG) has a high methane number, since the lighter fraction, i.e. methane, evaporates first. For forced boil-off of gas, a controlled evaporation is usually carried out, which allows the heavier fractions to be returned to the cargo tank, while delivering the light methane to the engine. Thus, the methane number of the gas supplied to the engine is usually above MN 80, even though the methane number of the remaining LNG in the tank during a voyage in ballast can be much lower. It can be concluded that the dependency on methane number needs to be considered, for example in the design of the engine, but for practical vessel operation it is only of minor relevance. Efficiency: As well known, a Diesel Cycle engine has generally the best efficiency of all internal combustion engine types, taking into account the entire load range. However, at high engine loads the efficiency of an Otto Cycle engine is similar or even better. Figure 3 shows a comparison between the specific energy consumption of a Diesel Cycle and an Otto Cycle engine. Engine power Figure 3: Specific energy consumption. Considering the entire load range, a drawback of low-pressure versus high-pressure is to be expected. However, high-pressure requires a high-pressure gas supply which consumes more electrical power than a low-pressure gas supply. In addition, high-pressure requires an exhaust gas treatment system in order to operate at low NOX emission levels. Depending on the selected treatment system, additional chemicals will be consumed, additional electric power is required and engine performance is reduced. Thus, it is advisable not only to consider engine efficiency as a stand-alone, but to consider overall system performance, in order to come to a final conclusion on the subject. Cases studies, as detailed below, are the appropriate way to analyse overall system performance. Required fuel-gas handling system Engine power The main difference between the two dual-fuel solutions is obviously the fuel-gas pressure, i.e. the design of components suited to the low- or high-pressures and the need to achieve that pressure level in the fuel-gas system. In general, designing for higher pressures can be seen as more challenging, even though it is manageable. The system component costs are on a clearly higher level for high-pressure, more energy is needed to reach the required pressures and, accordingly, maintenance costs will also be on a clearly higher level. Cryogenic centrifugal pumps are sufficient for the pressurisation of LNG in the low-pressure concept, but for the high-pressure concept the more demanding maintenance of reciprocating (piston) compressors has been taken into account. Looking at the handling of boil-off gas, many different types of gas compressors are available for the low-pressure concept, such as centrifugal, screw or piston compressors. By contrast, Licensees Conference 2015, Interlaken 5 / 14

for the high-pressure concept, only piston compressors are currently available from a limited number of makers. Moreover, the electrical power demand for operating the compressors is significantly different when comparing the two concepts, since the level of compression work, i.e. work done in compressing the fuel-gas, needs to be taken into account. The selected rating points are shown in Figure 4. 18,500 18,000 17,500 17,000 16,500 16,000 15,500 15,000 Engine Power [kw] W5X72DF 5G70ME-C9.5-GI Design point Thus, in terms of fuel-gas handling, the lowpressure concept exhibits clear advantages against the high-pressure concept. 14,500 14,000 13,500 13,000 12,500 12,000 11,500 CSR CMCR CASE STUDIES Up to this point, most evaluated topics show clear advantages, or at least similar results, for the low-pressure concept. For detailed investigation of the impact on operating costs and investments costs, case studies are seen as being the most helpful tool. The following case studies are based on the latest performance data (as of July 2015) available for the X-DF engines as representative of lowpressure, and the MAN ME-GI engines as representative of high-pressure. However, during the initial selection process, the performance data needed for assessments was not yet available. Case study for LNG carrier An LNG carrier in the size range of 170,000 m 3 to 180,000 m 3 is assumed. Configuration: The following main configurations have been selected for the low-pressure (LP) and the highpressure (HP) solution case study: - LP: 2 x 5X72DF HP: 2 x 5G70ME-GI CMCR: 2 x 12,500 kw @ 69.0 rpm CSR: 90% CMCR - LP: 2 x 8L34DF + 2x 6L34DF HP: 2 x 9L34DF + 2x 6L34DF - LP: 2 x Centrifugal compressor HP: 2 x Piston compressor, HP: 1 x HP LNG piston pump 11,000 10,500 Engine Speed [rpm] 10,000 55.0 60.0 65.0 70.0 75.0 80.0 85.0 90.0 95.0 Figure 4: Selected rating points - LNGC LNG is loaded at a terminal in the Gulf of Mexico and needs to be transported via the Panama Canal to Asia, e.g. Japan. Keel-laying is considered to be after 2016, but before introduction of any of the new emission control areas (ECAs) which are currently under discussion. Consequently emission control is only required in US waters. Apart from transit through the Panama Canal at an average speed of 10.0 kn (without waiting times), service speed operation at 19.5 kn is assumed. The power demand for part-load operation is estimated by the mathematical power-speed prediction formula with an assumed beta-factor (exponent) of 3.5.The same power demand is assumed for both laden and ballast voyages. The contracted maximum continuous rating (CMCR) is assumed to be 2 x 12,500 kw at an engine speed of 69 rpm. The service speed can be attained at a continuous service rating (CSR) power of 90% of the CMCR. Whenever possible, the engines operate in gas mode. In the US ECA, the 5G70ME-GI needs to operate an exhaust gas treatment system. Exhaust gas recirculation (EGR) is the selected solution. Outside the ECA, HFO is assumed as the pilot fuel for the 5G70ME-GI engine, while MDO is assumed for the 5X72DF. No switching to HFO is taken into account, as the pilot-fuel quantity is so minor. Consequently the operating costs for the HFO system, as well as the reduced component maintenance intervals, would exceed the savings of about 100 USD per day, i.e. HFO system operating costs would need to be added Licensees Conference 2015, Interlaken 6 / 14

to the 5G70ME-GI operating costs, but for reasons of simplicity these costs are not taken into account in the case study. The fuel prices are considered to be those detailed in Table 1. In comparison to the LNGfuelled vessel case study, lower costs are assumed for the gas. Different ship owners/charterers assume gas prices in their own calculations in different ways, e.g. different prices for natural and forced boil-off gases or different prices for laden and ballast voyages. This case study is based on an average gas price of 450 USD/ton, equivalent to 9.5 USD/mmBTU. However, the influence of the LNG price on the final result is minor, as shown in Figure 7. Fuel type Fuel price LHV Max sulphur content [USD/ton] [$/mmbtu] kj/kg [%] LNG 450 9.5 50,000 0.0% MGO 850 21.0 42,800 0.1% MDO 650 16.1 42,707 0.5% HFO 450 11.7 40,500 3.5% Table 1: Assumed fuel prices for LNG carrier. In laden voyages only NBOG consumption is assumed, but depending on the tank insulation quality, which affects the boil-off rate (BOR), some forced boil-off might be needed in order to reach the service speed. The 5X72DF engines are fed by a centrifugal compressor, which is currently the generally accepted and proven solution, although more efficient screw or piston compressors could be applied. The 5G70ME-GI is fed via a high-pressure piston compressor. In ballast voyage additional forced boil-off is considered. For the 5X72DF solution, LNG is supplied by pumps to the evaporator and then pressurised by the compressor. For the 5G70ME-GI solution, the gas is supplied in two parallel streams 1 : the natural boil-off gas is supplied via the piston compressor while the LNG for the forced boil-off supply is pressurised by high-pressure piston pumps before the gas is evaporated in a high-pressure evaporator. This arrangement allows electrical power savings for the 5G70ME-GI solution, since the part loadpower savings of the compressor are more significant than the additional power demand from the high-pressure LNG pump. The same principle could be applied for the 5X72DF solution. However, the part load saving with a centrifugal compressor, as selected in the study in order to keep the system simple and as similar to widelyused installations as possible, is limited. Namely, start-stop operation would be required for significant power savings or, alternatively, a displacement-type compressor, such as a screw or piston compressor would need to be applied, together with some minor gas-system adaptations. As an electric baseload for ship and engine room operation 2,000 kw is assumed. Depending on the selected, additional electrical power is required to operate the gas compressor and/or the LNG pump, as well as the EGR system for the ME-GI. The EGR power consumption of the ME-GI only takes into account the blowers and no water treatment system, since clean fuel-gas is burned. During the laden voyage through the Panama Canal, power demand for operation of the gas combustion unit (GCU) is assumed to be due to its own air blower as well as to the pressurised gas supply. Costs of gas burned by the GCU are not considered in this study. However, the costs are lower for the low-pressure 5X72DF solution, since at least some gas is utilised by the slowsteaming engines and consequently does not need to be burned by the GCU, while the 5G70ME-GI cannot utilise any gas as a pure diesel operating mode is required. The detailed load profile as well as the assumed electrical power demand is shown in Figure 14. The electrical power is provided by the abovementioned generator sets. Fuel consumption: In calculating the total fuel consumption of the main engine and gen-sets per round-trip, as shown in Figure 5, it can be seen that the 5X72DF solution has marginally higher gas consumption compared to the 5G70ME-GI, but lower liquid fuel consumption in about the same marginal range. 1 Start-stop operation of the compressor/pump might be possible. Licensees Conference 2015, Interlaken 7 / 14

Fuel consumption Fuel costs: different LNG price 3,500 t 2,500 kusd 3,000 t 2,500 t 2,000 t 2,000 kusd 1,500 kusd 1,500 t 1,000 t 500 t 0 t 2x 5X72DF 2x 5G70ME-GI 1,000 kusd 500 kusd saving 24 kusd 0 kusd 2x5X72DF: 300 $/t 2x5G70ME-GI: 300 $/t saving 9 kusd 2x5X72DF: 450 $/t 2x5G70ME-GI: 450 $/t loss 7 kusd 2x5X72DF: 600 $/t 2x5G70ME-GI: 600 $/t gas MGO MDO HFO Main engines Gensets Figure 5: Consumption of different fuel types - LNGC. Comparing the distribution of fuel costs between main engines and gen-sets, as presented by Figure 6, it can be seen that the 5X72DF itself causes marginally higher fuel costs than the 5G70ME-GI, but these higher costs are, in fact, slightly overcompensated by the lower fuel costs for operating the auxiliary engines, i.e. the 5X72DF solution facilitates some, if only minor savings. The detailed numbers can be taken from Figure 16 in the. 1,750 kusd 1,500 kusd 1,250 kusd 1,000 kusd 750 kusd 500 kusd 250 kusd 0 kusd 2x 5X72DF Fuel costs Main engines Gen-sets 2x 5G70ME-GI Figure 6: Fuel costs, main engine and gen-sets, LNGC. The influence of the assumed LNG price in relation to fixed liquid fuel prices can be seen in figure 7. As long as the LNG price per energy content is not higher than that of HFO, the 5X72DF helps to save fuel costs. Figure 7: Influence of LNG price on fuel costs - LNGC. Installation costs: The main differences are, as expected, caused by the differences in the fuel-gas handling system, which needs either to be designed for 16 bar (g) pressure or for at least 300 bar (g). In addition, an exhaust gas treatment system is needed for the 5G70ME-GI solution in order to comply with the IMO Tier III NOX emission limitation, while the 5X72DF complies without any exhaust gas treatment as long as it is intentionally operated on gas. Cost comparisons are generally very difficult and can only provide a first indication since costs depend on market conditions, like supply and demand; the selected detailed solution: e.g. piston, screw or centrifugal compressors for the 5X72DF solution; the selected makers and the makers location; political pricing, etc. Taking the above into account, the same indicative costs are assumed for both main engine solutions. The gen-sets for the 5G70ME-GI solution are slightly more expensive, since in total two cylinders more need to be installed. In addition, an exhaust gas treatment system is required for the 5G70ME-GI solution - As currently promoted by MAN Diesel & Turbo, EGR systems are the chosen solution. The main difference in investment costs is caused by the gas system as such. Low- or high-pressure heat exchangers, pipes, instruments and with the highest cost impact lowor high-pressure compressors are needed. In this study two sets of compressors, i.e. one in operation and one on stand-by are assumed for redundancy reasons. Mean values received from different makers are used for indicative prices. Licensees Conference 2015, Interlaken 8 / 14

Figure 8 gives an indication regarding the average investment cost saving of applying lowpressure. 10,500 10,000 9,500 9,000 Engine Power [kw] W6X52DF 6G50ME-C9.5-GI Design point Indicative investment costs 8,500 160% 8,000 CMCR 140% 120% 30 musd 7,500 7,000 CSR 100% 80% 60% 40% 20 musd 10 musd Gas system EGR Gen-sets Main engines 6,500 6,000 Engine Speed [rpm] 5,500 70.0 75.0 80.0 85.0 90.0 95.0 100.0 105.0 110.0 115.0 20% 0% Alt. 1 2 x 5X72DF Alt. 2 2x5G70ME-C9.5-GI 0 musd Figure 9: Selected rating points LNG fuelled vessel. Figure 8: Indicative machinery solution investment costs - LNGC. Case study for LNG-fuelled vessel An LNG-fuelled product tanker of around 55,000 dwt is assumed. Configuration: The following main configurations have been selected for the low-pressure (LP) and the highpressure (HP) solution case study: - LP: 1 x 6X52DF HP: 1 x 6G50ME-GI CMCR: 7,800 kw @ 95.0 rpm CSR: 90% CMCR - LP: 3 x 6L20DF HP: 3 x 6L20DF - LP: 2 x LNG centrifugal pumps HP: 2 x LNG centrifugal supply pumps and 2 x LNG piston high-pressure pumps - Both: 2 x 700 m 3 single shell LNG Both: tanks with 10 bar(a) design Both: pressure The selected rating points are shown in Figure 9. It is assumed that all NBOG can be taken by the gen-sets. However, depending on the operating profile and the installed tank size, which might in reality deviate from the size assumed in this study, the gas consumption of the gen-sets might not be sufficient to avoid excessive tank pressure. If such conditions need to be covered, either a small reliquefaction plant would need to be added, or a compressor with a relatively small capacity but with an output pressure matching the main engines required gas admission pressure level, i.e. low- or high-pressure. Fuel prices are assumed as detailed in Table 2. The liquid fuel prices are the same as in the LNG carrier case study, but the LNG price is considered to be at a higher level, since the LNG has already passed through additional handling steps which have added costs. Currently, LNG prices are substantially different in different parts of the world. While being cheapest in the US, medium level prices can be found in Europe and high level prices in Asia. Even though US IMO Tier III operation is mentioned in the operating profile in Figure 17, an average price level, e.g. as could found in Europe, is assumed. The influence of different LNG prices on the total fuel costs is shown in Figure 12. Fuel type Fuel price LHV Max sulphur content [USD/ton] [$/mmbtu] kj/kg [%] LNG 700 14.8 50,000 0.0% MGO 850 21.0 42,800 0.1% MDO 650 16.1 42,707 0.5% HFO 450 11.7 40,500 3.5% Table 2: Assumed fuel prices for LNG fuelled vessel. Licensees Conference 2015, Interlaken 9 / 14

Fuel consumption: Similar to the LNG carrier comparison, the gas consumption of the LNG-fuelled vessel is marginally higher with the 6X52DF solution, while the liquid fuel consumption is marginally lower. 7,000 t 6,000 t 5,000 t 4,000 t 3,000 t Fuel consumption 7,000 kusd 6,000 kusd 5,000 kusd 4,000 kusd 3,000 kusd 2,000 kusd saving 48 kusd 1,000 kusd 0 kusd 6X52DF: 500 $/t Fuel costs: different LNG price 6G50ME-GI: 500 $/t saving 4 kusd 6X52DF: 700 $/t Main engines 6X50ME-GI: 700 $/t Gen-sets loss 40 kusd 6X52DF: 900 $/t 6X50ME-GI: 900 $/t Figure 12: Influence of LNG price on fuel costs LNG-fuelled vessel. 2,000 t 1,000 t 0 t 6X52DF gas MGO MDO HFO 6G50ME-GI Installation costs: The general problem of investment cost comparisons, as explained for the LNG carrier case, is also valid for an LNG-fuelled vessel. Figure 10: Consumption of different fuel types LNG fuelled vessel. Comparing the distribution of fuel costs between the main engine and the gen-sets, as presented in Figure 11 Fuel costs, main engine and gen-sets LNG fuelled vessel.it can be seen that the 6X52DF itself causes marginally higher fuel costs than the 6G50ME-GI, but these higher costs are actually slightly overcompensated by the lower fuel costs for operating the auxiliary engines, i.e. the 6X52DF solution enables some savings, albeit only minor. The detailed numbers can be taken from Figure 19 in the Appendix. 5,000 kusd 4,500 kusd 4,000 kusd 3,500 kusd 3,000 kusd 2,500 kusd 2,000 kusd 1,500 kusd 1,000 kusd 500 kusd 0 kusd 6X52DF Fuel costs Main engines Gen-sets 6G50ME-GI Figure 11 Fuel costs, main engine and gen-sets LNG fuelled vessel. The influence of the assumed LNG price in relation to fixed liquid fuel prices can be seen in Figure 12. The highest investment costs for operating a vessel on LNG as its fuel are incurred by the cryogenic LNG tanks, i.e. in that case study about 40% of the installation costs for the lowpressure solution are caused by the tanks. Depending on the required endurance between bunkering, the tank size could be increased or decreased, which can have a significant influence on investment costs. The selected tank configuration of 2 x 700 m 3 allows service speed operation on gas-fuel for more than three weeks. If assuming that no gas compressor is required, at least for gas supply to the main engine, the differences in investment costs are much lower compared to the LNG carrier case. However, the high-pressure machinery solution still indicates an additional investment effort of about 15%. 120% 100% 80% 60% 40% 20% 0% Indicative investment cost Alt. 1 6X52DF Alt. 2 6G50ME-GI 15 musd 10 musd 5 musd 0 musd Gas system LNG tank(s) Figure 13: Indicative machinery solution investment costs LNG-fuelled vessel. EGR Gen-sets Main engine Licensees Conference 2015, Interlaken 10 / 14

CONCLUSION The low-pressure solution provides the greatest benefits in the majority of aspects. It is the most environmentally-friendly solution, always complies with IMO Tier III emissions rules when operated on gas, even when outside ECAs. The associated investment costs are lower and for LNG carriers even drastically lower compared to the high-pressure solution. The overall fuel costs for operating a vessel are similar for both solutions. Using the assumed fuel prices and operating profiles, the low-pressure solution even results in some savings. Maintenance costs have not been investigated in the case studies, since calculating these costs is very complex and needs to be the subject of a separate study. However, since fewer components are required for the low-pressure solution, and only lower-loaded gas equipment is required, clear cost savings can be expected for the low-pressure solution. An overview of the low- and high-pressure comparison is provided in Table 3. It is clear that the weighting different topics into five different levels from great benefit to significant drawback is always somewhat subjective and consequently different people might come to deviating results. However, as is the intention of this Table, it indicates a trend. Finally it can be concluded that it is the right decision to apply the low-pressure industry standard used on medium-speed 4-stroke engines to WinGD s low-speed 2-stroke engines. Environment Technical aspects Financial Summary Low-pressure High-pressure GHG + ++ SOX ++ + (++) 2 NOX ++ ü (+) 3 PM ++ + Power output FGHS selection ü ++ ü ++ ++ ü OPEX ++ 4 + CAPEX ++ -- (-) 5 MN dependency Conclusion ++ ü (+) 5 Table 3: Overview of a comparison of low- and high-pressure gas-fuelling solutions. 2 Depending on sulphur content of pilot fuel. 3 If NOX treatment system (EGR or SCR) is in operation. 4 If considering: - savings for HFO treatment system operation, - potential gas system efficiency improvements, e.g. optimised compressor layout, displacement type compressor installation, etc. - gas burned by the GCU of a LNGC, - lower maintenance costs. 5 Less significant CAPEX disadvantage for LNG fuelled vessels. Licensees Conference 2015, Interlaken 11 / 14

Legend table 3: GHG: greenhouse gas, i.e. contribution to global warming PM: particulate matter FGHS: Fuel-gas Handling System, i.e. number of available component suppliers REFERENCES WINTERTHUR GAS & DIESEL: GTD 1, General Technical Data, 2015 MAN DIESEL & TURBO: CEAS Engine Calculations v1.8.54, 2015 ++: large benefit +: benefit AUTHOR ü: meets requirements -: drawback --: significant drawback Daniel Strödecke Manager Application Development Business & Application Development daniel.strodecke@wingd.com Licensees Conference 2015, Interlaken 12 / 14

APPENDIX Operating mode Ship speed Voyage type Engine power predicted with β = 3.5 (2 engines) Relative engine power Engine speed Figure 14 Operating profile LNGC. Electric hotel load Total non ECA ECA Non ECA electric zone load ECA zone Main fuel non ECA Pilot fuel non ECA Main fuel in ECA Pilot fuel in ECA Extra electric load due to selected specification of 2x W5X72DF kn [kw] [%] [rpm] [kwe] [kwe] h h Service speed US waters 19.5 laden 22500 90% 66.6 2000 2825 15 DF LNG MGO 825 kwe compressor Service speed passage 19.5 laden 22500 90% 66.6 2000 2825 455 DF LNG MDO 825 kwe compressor Slow steaming Panama Canal 10.0 laden 2173 9% 30.6 2000 3275 5 DF LNG MDO 1275 kwe compressor (incl. GCU supply) + GCU (part load) Service speed passage 19.5 ballast 22500 90% 66.6 2000 2835 455 DF LNG MDO 835 kwe compressor + LNG supply pump Service speed US waters 19.5 ballast 22500 90% 66.6 2000 2835 15 DF LNG MGO 835 kwe compressor + LNG supply pump Slow steaming Panama Canal 10.0 ballast 2173 9% 30.6 2000 2835 5 DF LNG MDO 835 kwe compressor (incl. GCU supply) 2x 5G70ME-GI kn [kw] [%] [rpm] [kwe] [kwe] h h Service speed US waters 19.5 laden 22500 90% 66.6 2000 3450 15 DF EGR T.III LNG MGO 1450 kwe compressor + EGR (blower only) Service speed passage 19.5 laden 22500 90% 66.6 2000 3300 455 DF LNG HFO 1300 kwe compressor Slow steaming Panama Canal 10.0 laden 2173 9% 30.6 2000 2900 5 DF HFO 900 kwe compressor GCU supply + GCU Service speed passage 19.5 ballast 22500 90% 66.6 2000 3025 455 DF LNG HFO 1025 kwe compressor (part load) + HP pump Service speed US waters 19.5 ballast 22500 90% 66.6 2000 3175 15 DF EGR T.III LNG MGO 1175 kwe compressor (part load) + HP pump + EGR (blower only) Slow steaming Panama Canal 10.0 ballast 2173 9% 30.6 2000 2000 5 DF HFO - 0 kwe pure diesel operation (GCU operation not needed) total distance 9215 nm Operating mode Operating Ship Enviro. Voyage Total electric Gen-set type gas MDO/MGO time [RH] speed area type load consumption [t] consumption [t] 2x 5X72DF Wärtsilä 8L34DF Wärtsilä 8L34DF Wärtsilä 6L34DF Wärtsilä 6L34DF kn 3690 kwe 3690 kwe 2770 kwe 2770 kwe Service speed US waters 15 19.5 Tier III laden 2825 kwe 77% - - - 6.8 0.1 Service speed passage 455 19.5 Tier II laden 2825 kwe 77% - - - 206.7 3.5 Slow steaming Panama Canal 5 10.0 Tier II laden 3275 kwe 89% - - - 2.6 0.0 Service speed passage 455 19.5 Tier II ballast 2835 kwe 77% - - - 207.3 3.5 Service speed US waters 15 19.5 Tier III ballast 2835 kwe 77% - - - 6.8 0.1 Slow steaming Panama Canal 5 10.0 Tier II ballast 2835 kwe 77% - - - 2.3 0.0 2x 5G70ME-GI Wärtsilä 9L34DF Wärtsilä 9L34DF Wärtsilä 6L34DF Wärtsilä 6L34DF kn 4150 kwe 4150 kwe 2770 kwe 2770 kwe Service speed US waters 15 19.5 Tier III laden 3450 kwe 83% - - - 8.2 0.1 Service speed passage 455 19.5 Tier II laden 3300 kwe 80% - - - 239.9 4.0 Slow steaming Panama Canal 5 10.0 Tier II laden 2900 kwe 70% - - - 2.4 0.0 Service speed passage 455 19.5 Tier II ballast 3025 kwe 73% - - - 223.5 4.1 Service speed US waters 15 19.5 Tier III ballast 3175 kwe 77% - - - 7.7 0.1 Slow steaming Panama Canal 5 10.0 Tier II ballast 2000 kwe - - 72% - 1.6 0.0 Port operation not considered Figure 15 Gen-set load distribution. Operating mode Operating time [RH] Ship Enviro. Voyage speed area type Main engines Gen-sets Total gas MGO MDO HFO gas MGO MDO gas MGO MDO HFO consumption consumption consumption consumption consumption consumption consumption consumption consumption consumption consumption 2x 5X72DF kn [t] [t] [t] [t] [t] [t] [t] [t] [t] [t] [t] Service speed US waters 15 19.5 Tier III laden 46.9 0.4 - - 6.8 0.1-53.8 0.5 0.0 - Service speed passage 455 19.5 Tier II laden 1424.0-11.3-206.7-3.5 1630.7 0.0 14.8 - Slow steaming Panama Canal 5 10.0 Tier II laden 1.7-0.0-2.6-0.0 4.3 0.0 0.1 - Service speed passage 455 19.5 Tier II ballast 1424.0-11.3-207.3-3.5 1631.3 0.0 14.8 - Service speed US waters 15 19.5 Tier III ballast 46.9 0.4 - - 6.8 0.1-53.8 0.5 0.0 - Slow steaming Panama Canal 5 10.0 Tier II ballast 1.7-0.0-2.3-0.0 4.0 0.0 0.1 - total distance 2x 9215 nm 2945.5 0.7 22.6 0.0 432.5 0.2 7.1 3377.9 1.0 29.8 0.0 1325.5 kusd 0.6 kusd 14.7 kusd 0.0 kusd 194.6 kusd 0.2 kusd 4.6 kusd 1520.1 kusd 0.8 kusd 19.3 kusd 0.0 kusd 1341 kusd 199 kusd 1540 kusd 2x 5G70ME-GI kn [t] [t] [t] [t] [t] [t] [t] [t] [t] [t] [t] Service speed US waters 15 19.5 Tier III laden 45.5 2.3 - - 8.2 0.1-53.7 2.4-0.0 Service speed passage 455 19.5 Tier II laden 1350.3 - - 71.2 239.9-4.0 1590.2 0.0-71.2 Slow steaming Panama Canal 5 10.0 Tier II laden 0.0 - - 2.1 2.4-0.0 2.4 0.0-2.1 Service speed passage 455 19.5 Tier II ballast 1350.3 - - 71.2 223.5-4.1 1573.8 0.0-71.2 Service speed US waters 15 19.5 Tier III ballast 45.5 2.3 - - 7.7 0.1-53.2 2.4-0.0 Slow steaming Panama Canal 5 10.0 Tier II ballast 0.0 - - 2.1 1.6-0.0 1.6 0.0-2.1 total distance 2x 9215 nm 2791.6 4.5 0.0 146.6 483.2 0.3 8.2 3274.9 4.8 0.0 146.6 1256.2 kusd 3.8 kusd 0.0 kusd 66.0 kusd 217.4 kusd 0.2 kusd 5.3 kusd 1473.7 kusd 4.1 kusd 0.0 kusd 66.0 kusd 1326 kusd 223 kusd 1544 kusd total distance Figure 16 Fuel costs LNGC costs for gas burned by the GCU is not included. Licensees Conference 2015, Interlaken 13 / 14

Operating mode Ship speed Engine power Relative engine Engine predicted with β = 3.5 (2 engines) power speed Electric hotel load Total non ECA ECA Non ECA electric zone load ECA zone Main fuel non ECA Pilot fuel non ECA Main fuel in ECA Pilot fuel in ECA Extra electric load due to selected specification of 6X52DF kn [kw] [%] [rpm] [kwe] [kwe] h h Service speed US waters 14.5 7020 90% 66.6 500 505 4200 DF LNG MGO 5 kwe LNG pump Service speed non US waters 14.5 7020 90% 66.6 500 505 1500 DF LNG MDO 5 kwe LNG pump Slow steaming US waters 10.0 1912 25% 43.2 500 505 150 DF LNG MGO 5 kwe LNG pump Slow steaming non US waters 10.0 1912 25% 43.2 500 505 150 DF LNG MDO 5 kwe LNG pump 6G50ME-GI kn [kw] [%] [rpm] [kwe] [kwe] h h Service speed US waters 14.5 7020 90% 66.6 500 590 0 4200 DF EGR T.III LNG MGO 90 kwe LNG pumps + EGR (blower only) Service speed non US waters 14.5 7020 90% 66.6 500 540 1500 0 DF LNG HFO 40 kwe LNG pumps Slow steaming US waters 10.0 1912 25% 43.2 500 555 0 150 DF EGR T.III LNG MGO 55 kwe LNG pumps + EGR (blower only) Slow steaming non US waters 10.0 1912 25% 43.2 500 540 150 0 DF LNG HFO 40 kwe LNG pumps Calculated for one year operation - in total 6000 RH Figure 17 Operating profile LNG fuelled vessel. Operating mode Operating time [RH] Ship speed Enviro. area Total electric load Gen-set type gas consumption [t] MDO/MGO consumption [t] 6X52DF Wärtsilä 6L20DF Wärtsilä 6L20DF Wärtsilä 6L20DF kn 1065 kwe 1065 kwe 1065 kwe Service speed US waters 4200 14.5 Tier III 505 kwe 47% - - 451.1 20.9 Service speed non US waters 1500 14.5 Tier II 505 kwe 47% - - 161.1 7.4 Service speed US waters 150 10.0 Tier III 505 kwe 47% - - 16.1 0.7 Slow steaming non US waters 150 10.0 Tier II 505 kwe 47% - - 16.1 0.7 6G50ME-GI Wärtsilä 6L20DF Wärtsilä 6L20DF Wärtsilä 6L20DF kn 1065 kwe 1065 kwe 1065 kwe Service speed US waters 4200 14.5 Tier III 590 kwe 55% - - 508.8 21.8 Service speed non US waters 1500 14.5 Tier II 540 kwe 51% - - 169.8 7.6 Service speed US waters 150 10.0 Tier III 555 kwe 52% - - 17.3 0.8 Slow steaming non US waters 150 10.0 Tier II 540 kwe 51% - - 17.0 0.8 Port operation not considered Figure 18 Gen-set load distribution Operating mode Operating time [RH] Ship speed Enviro. area Figure 19 Fuel costs LNG fuelled vessel Main engines Gen-sets Total gas MGO MDO HFO gas MGO MDO gas MGO MDO HFO consumption consumption consumption consumption consumption consumption consumption consumption consumption consumption consumption 6X52DF kn [t] [t] [t] [t] [t] [t] [t] [t] [t] [t] [t] Service speed US waters 4200 14.5 Tier III 4154.3 53.1 - - 451.1 20.9-4605.4 73.9 0.0 - Service speed non US waters 1500 14.5 Tier II 1483.7-20.0-161.1-7.4 1644.8 0.0 27.4 - Service speed US waters 150 10.0 Tier III 41.8 1.1 - - 16.1 0.7-57.9 1.8 0.0 - Slow steaming non US waters 150 10.0 Tier II 41.8-1.1-16.1-0.7 57.9 0.0 1.9-5721.5 54.1 21.1 0.0 644.5 21.6 8.2 6366.0 75.7 29.3 0.0 4005.1 kusd 46.0 kusd 13.7 kusd 0.0 kusd 451.1 kusd 18.4 kusd 5.3 kusd 4456.2 kusd 64.4 kusd 19.0 kusd 13.2 kusd 4065 kusd 475 kusd 4540 kusd 6G50ME-GI kn [t] [t] [t] [t] [t] [t] [t] [t] [t] [t] [t] Service speed US waters 4200 19.5 Tier III 3971.5 206.4 - - 508.8 21.8-4480.3 228.2-0.0 Service speed non US waters 1500 19.5 Tier II 1387.9 - - 76.6 169.8-7.6 1557.7 0.0-76.6 Service speed US waters 150 19.5 Tier III 37.6 4.7 - - 17.3 0.8-55.0 5.5-0.0 Slow steaming non US waters 150 10.0 Tier II 36.5 - - 4.9 17.0-0.8 53.4 0.0-4.9 5433.4 211.1 0.0 81.5 713.0 22.6 8.4 6146.4 233.7 0.0 81.5 3803.4 kusd 179.4 kusd 0.0 kusd 36.7 kusd 499.1 kusd 19.2 kusd 5.4 kusd 4302.5 kusd 198.6 kusd 0.0 kusd 36.7 kusd 4019 kusd 524 kusd 4543 kusd Licensees Conference 2015, Interlaken 14 / 14