Effect of flexible camshaft technology on dual-fuel engine performance using phenomenological combustion model

The objective of the current study is to investigate the effect of tunable valve timings on the performance and emissions of a dual-fuel marine engine. A simulation model was developed in MATLAB to simulate the valvetrain mechanism. The model generates different valve lift curves depending on the input conditions, and after that, it tests them against any possible collision with the piston. The valid valve lift curves were exported to a two-zone combustion model in AVL CRUISE-M platform. The combustion model depends on the fractal principle and aims to predict the in-cylinder parameters. In addition, it contains sub-models to calculate the ignition delay and emissions formation. Model results were compared against experimental data, as the latter were obtained from a heavy-duty, medium-speed, single-cylinder research engine, which employs natural gas as a main fuel. The results showed good agreement and the model was used for further investigations with other cam pairs. It has been found that the fractal combustion model can effectively represent the combustion behavior in the dual-fuel engine. Furthermore, valve timing has a significant influence on the engine performance and exhaust emissions. Results also revealed that applying Miller cycle can reduce the nitrogen oxides emissions, while the higher valve overlap period had a negative effect on methane slip.


Introduction
In the last few decades, internal combustion engines have been used as the main prime mover in many applications.They are preferred for mobility purposes since they can serve with different fuel types.Due to rapid transportation growth and increased industrial activities, more exhaust gases are emitted from internal combustion engines.Particularly in the maritime sector, diesel engines are commonly used due to their high thermal efficiency and reliability; however, they represent one of the main pollution sources [28,29].Diesel exhaust contains around 450 different compounds of gaseous emissions which are emitted during the combustion process.Some of them represent a threat for human health and environment, as well as enhancing the global warming effect.The most dangerous emissions which captured the concern in the last years are nitrogen oxides (NO x ), particulate matter (PM), carbon dioxide (CO 2 ), and methane.NO x gases have many dangerous impacts since they cause photochemical smog and acid rain.Furthermore, they participate in damaging the immune system and lungs specially for those who have respiratory disease [22].PM consists of very small solid particles with different physical and chemical properties, for instance, carbonaceous material (soot), ash, and metals.The amount of produced PM depends on the combustion characteristics and the used fuel.Exposure to PM (specially the smallest) contributes to premature cardiovascular diseases and lung cancer [11,34].In terms of CO 2 and methane, they are categorized as greenhouse gases (GHG), which increase the earth's temperature by reducing energy dissipation back to space.With the increased usage of natural gas (NG) as alternative fuel, more methane emissions are emitted to the atmosphere.Additionally, methane leakage during production and logistics represents an influential source.According to the fourth report from International Maritime Organization (IMO) in 2020, the emitted greenhouse gases from shipping activities (expressed in equivalent carbon dioxide (CO 2e )) have increased from 977 million tons in 2012 to 1076 million tons in 2018.From the global anthropogenic, this portion represents 2.76% in 2012 and 2.89% in 2018 [14].From this point of view and to fulfill the strict emissions legislations, new combustion technologies are required in order to achieve ultra-low emission target.

Modern combustion technologies
In terms of conventional diesel combustion (CDC), much research has been conducted to improve its performance with lower exhaust emissions.Some of them related to exhaust after treatment, exhaust gas recirculation (EGR), and various injection strategies.Although these approaches are effective, maintenance cost and systems complexity represent an obstacle.Additionally, minimizing both PM and NO x simultaneously in diesel engines is not possible due to the trade-off characteristics [32].Modern combustion technologies involve the low-temperature combustion (LTC) concept, which allows to reduce PM and NO x together by adopting lower combustion temperature and leaner mixture.Commonly, the abbreviation "LTC" is an umbrella term which embraces a group of many technologies.The first and basic one in this group is the homogeneous charge compression ignition (HCCI), which acts as a bridge between diesel and gasoline engines.In HCCI, a homogeneous mixture of air and fuel is inducted into the cylinder during the intake stroke.Then, the mixture is autoignited around the end of the compression stroke when the temperature reaches the self-ignition limit.Considering the absence of the controlled ignition onset as it is in the diesel engines, HCCI has poor control on the combustion beginning and rate of heat release (ROHR).Furthermore, the short combustion duration in HCCI bans it from participating at the higher loads [21].In order to overcome the shortcomings of HCCI, premixed charge compression ignition (PCCI) was introduced.PCCI involves high-pressure fuel injection inside combustion chamber during the compression stroke, aiming to assure better control on the combustion initiation.Moreover, to achieve different levels of fuel stratification and ROHR control, injection event can include many small pilot pulses as shown in Fig. 1.The difference between PCCI and CDC is that the latter involves diesel injection at around 20 • before top dead center (BTDC), while PCCI employs earlier fuel injection during the air compression process [21,23].On the other hand, they both share the compression ignition concept and use some ignition control techniques such as variable valve actuation and EGR.Because of the low combustion and exhaust temperatures in HCCI and PCCI, more unburned hydrocarbon (UHC) and carbon monoxide (CO) are emitted as the high temperature is required for complete oxidation [21].Sometimes PCCI and HCCI can be categorized as single-fueled modern combustion technologies since one fuel type is included [23,24].Conversely, double-fueled modern combustion techniques utilize two fuel types as it is in the reactivity-controlled compression ignition (RCCI).In this technology two fuels with different reactivity levels are used.Low reactivity fuel (LRF) such as gasoline or natural gas is supplied into the cylinder through the intake port by port fuel injection (PFI).After that, the high reactivity fuel (HRF), which is usually diesel, is injected during the compression stroke.The start of combustion (SOC) depends on the chemical reactions between the two fuels, the air-fuel ratio, and the in-cylinder conditions.By tuning the substitution ratio of the two fuels, combustion initiation can be effectively controlled through fuel reactivity stratification [13,21,23].Due to the dual-fuel capability in RCCI, it has better combustion control in comparison with HCCI and PCCI.Developing the RCCI concept has led to the Pilot Dual-Fuel (DF) combustion, which employs two fuels also with different reactivity.The only difference is that in Pilot-DF, the HRF is injected directly at high pressure into the cylinder near the top dead center (TDC).Hence, the pilot injection is used as an igniter, which allows perfect control on the combustion process.In these types of engines, NG is preferred as LRF due to its availability and low price.Moreover, NG is a low carbon fuel with higher calorific value and has high octane rating, which allows it to sustain at the higher compression ratios without knocking.DF engines are associated with high thermal efficiency and lower NO x emissions.On the other side, they produce relatively high amount of unburned hydrocarbon or "methane slip" in the exhaust emissions [2,3,21,23].Many factors are responsible for the presence of methane slip, but the major sources are crevice volumes, incomplete combustion due to low combustion temperature, and valve timing [12,26].Usually, intake valve opens before TDC to assure high volumetric efficiency.The same considerations are applied with the exhaust valve as it closes after TDC to maintain optimal exhaust scavenging.As a result, both valves are simultaneously opened which is referred to as "overlap period" and is measured in crank angle degrees.Typically, a moderate overlap period has positive influence on engine operation, while in other cases, it can lead to poor performance and higher exhaust emissions especially in the medium-speed engines.During overlap period, two possible effects are available depending on the in-cylinder conditions.The first one is when an amount of the fresh charge escapes directly through the exhaust port without contributing to the combustion process as indicated in Fig. 2a.In other words, this case could be described as "short circuit" between the gas exchange terminals.Therefore, fuel consumption is increased and higher UHC exists in the exhaust.The second effect represents the same phenomena but in the opposite direction as illustrated in Fig. 2b.It occurs when the exhaust gases are pushed again through the intake valve instead of leaving the cylinder from the exhaust one.Consequently, they will be carried back to the cylinder with the new charge, which reduces the volumetric efficiency and behaves as internal EGR causing a dilution effect.The engine load and speed, as well as the pressure difference between intake port, exhaust port, and cylinder will determine whether the first or the second effect will be dominant.In large turbocharged and low-speed engines, the first effect will be more common since relatively long time is available for the overlap period in addition to the high intake pressure which improve the cylinder scavenging [18,20,35].
For modern combustion technologies, valve timing is considered a sophisticated approach which allows better combustion control as it governs engine breathing and affects cylinder conditions.In this regard, Miller cycle is utilized on a wide range as it can reduce NO x emissions and control combustion temperature.Miller valve timing is accomplished by advancing the intake valve closing point (IVC) with respect to bottom dead center (BDC).As a result, the charge is expanded in the cylinder which reduces pressure and temperature at the end of the intake stroke.In other words, Miller cycle provides larger expansion ratio than the compression one.Hence, pressure and temperature traces in the whole cycle can be effectively controlled [27].Many scientific literatures have discussed combustion optimization Fig. 2 Effect of valve overlap during gas exchange with the help of Miller cycle.For example, the influence of Miller valve timing and intake charge cooling was experimentally studied on a single-cylinder, heavy-duty, dual-fuel engine in Ref. [10].The engine employs methanol as a main fuel through PFI, while diesel fuel represents the ignition pulse from a common rail injection system.Using a hydraulic variable valve actuation mechanism, IVC was tested at three positions while intake valve open (IVO) was fixed.Tests were conducted at indicated mean effective pressure (IMEP) of 18 bar and 1200 rpm with variable methanol fraction ratios, while diesel injection timing was tuned for the maximum allowed pressure-rise rate.The results showed that advancing the IVC caused a lower effective compression ratio, which reduced pressure and temperature before the combustion starts.Hence, the SOC was delayed.Otherwise, the excessive reduction of the effective compression ratio resulted in less trapped mass with lower heat capacity.As a result, the averaged in-cylinder gas temperature during the combustion phase was higher.Referring to the intake air temperature, the reduction from 323 to 305 K reflected on the average in-cylinder gas temperature by a decrease of 50 K during the compression stroke.Therefore, the ignition delay was extended, and the best combustion characteristics were reobtained by controlling the pilot diesel injection.In terms of exhaust emissions, the combination of Miller cycle with the lower intake temperature led to a significant reduction in NO x emissions and increased the net thermal efficiency.In contrast, more UHC and CO were observed.

Combustion modeling approaches
In the field of internal combustion engines research and development, experimental tests are very common since they provide precise results.But for the large engines with several sub-systems, more time and costs are consumed.Hence, modeling and simulation approach is preferred as it effectively predicts engine performance and exhaust emissions considering resources saving.Additionally, combustion simulation provides more flexibility since it allows to test all the conditions that could not be experimentally investigated.In this aspect, there are three modeling approaches.The difference between them is the model complexity, the computational efforts, and the model ability to simulate different phenomena.The first and simplest approach is the thermodynamic models (zerodimensional).Such models are only time-dependent and neglect the spatial details.Moreover, they do not involve any physical or chemical definitions and are preferred for simple and fast simulation.In the second place are the phenomenological models (Quasi-dimensional), which represent a compromising point between simplicity and high predictive capability.These models characterize the combustion chamber by dividing it into two or more zones.Further, they include formulations to describe the physical and chemical processes for combustion and exhaust formation.In the third and last place, comes the computational fluid dynamic models (CFD multi-dimensional).Compared with the other two types, CFD models involve high temporal and spatial resolution which provide a micro-scale insight during the combustion process.This is accomplished by dividing the combustion chamber into small cells, connected together forming a numerical mesh.Then, the correlations of mass, energy, and momentumconversation are applied for each cell [25].
Many studies have employed the modeling and simulation approach for engines development.For example, the authors in Ref. [33] studied the effect of Miller cycle on the performance and emissions of a medium-speed, diesel marine engine using a phenomenological simulation model.The engine involves six cylinders and is equipped with a turbocharger as well as a direct injection system.Firstly, the relevant experimental tests were performed and then the model was implemented using AVL BOOST.Model results were validated with the experimental data at different loads and showed good agreement.Results indicated that applying Miller cycle by advancing the IVC can reduce cylinder pressure and temperature.Consequently, lower NO x emissions are produced since it depends on the combustion temperature.In spite of that, extreme advancing of IVC had negative effects as the SOC was significantly delayed and ROHR increased rapidly.In addition, the NO x emissions and combustion temperature were higher.In terms of the different operating conditions, there is no one IVC position that was best for all loads and optimal IVC was earlier with the higher loads.
Considering the new combustion technologies, the effect of Miller cycle on a heavy-duty RCCI engine was investigated in Ref. [18].The engine has six cylinders and was retrofitted for the RCCI operation using NG as a primary fuel and diesel for ignition control.In addition, the engine was equipped with an electrically actuated valve train mechanism, which allows IVC tuning.In order to predict the in-cylinder pressure and exhaust emissions, a multi-zone combustion model was developed.The model is based on detailed chemical kinetics including 354 reactions through 65 species.The simulation results were validated with the experimental data, and the model was used for further analysis.It can be concluded that advancing the IVC can be used effectively to control the peak cylinder pressure and temperature which allows to exceed the engine load without heavy knocking.As the IVC was advanced from 163 • to 145 • BTDC, the IMEP can be increased from 21 to 23 bar, while a reduction of 2% was observed for the indicated thermal efficiency.Furthermore, UHC was affected negatively as it increased 3 times.Contrary, the severe retarding of IVC caused higher intake air temperature due to the exhaust back flow, while the UHC was lower with 30%.
The review shed light on the significance of the new combustion technologies, parallel with the other assistant techniques such as EGR and Miller cycle.From this perspective, the present study aims to investigate the effect of the flexible camshaft technology (FCT) on the performance and emissions of a dual-fuel marine engine.A simulation model was conducted using MATLAB to represent the different valve timing possibilities from the FCT mechanism.Subsequently, the valve timing positions were exported to a combustion model in AVL CRUISE-M platform, in order to analyze their effect on the engine.Model results were validated against experimental data, and the model was employed for further investigations.

Dual-fuel research engine
The experimental tests were performed using a mediumspeed, dual-fuel marine engine at the Department of Piston Machines and Internal Combustion Engines, Rostock University.For flexibility considerations, the research engine consists of only one cylinder based on a MaK-M34DF engine.In order to simulate the load effect, a 1.2 MW dynamometer is coupled to the engine as shown in Fig. 3.The main engine specifications for one cylinder from serial production configuration are tabulated in Table 1.
Since the dynamometer is speed-controlled, the engine load was adopted by tuning the fuel quantity.NG was used as a main fuel through a solenoid valve (SOGAV 105) mounted on the intake port.The ignition event relies on a pilot diesel injection from a common rail injection system.The supplied energy during the combustion consists of around 95-98% NG, while the remaining portion comes from the diesel pilot.To represent the turbocharger effect, the engine is fed with pressurized intake air from two compressors.The air pressure is regulated through two pressure control valves, while air temperature is controlled with the help of heaters.
On the other hand, exhaust pressure is adjusted by two flap valves with different sizes to allow precise control.Intake and exhaust pressures are determined according to a specific map and performance data of a real-sized turbocharger.For combustion analysis, an AVL IndiSet was used to gather the pressure signals from the different sensors (intake, exhaust, and in-cylinder pressure).Exhaust emissions were measured through AVL SESAM i60 FT gas analyzer, while for smoke detection, an AVL 415s smoke meter was employed.For data acquisition and signals interface, NI hardware in conjunction with LabView software were utilized.

Flexible camshaft technology
The advanced FCT technology represents a key point in this study, as it aims to adjust both the intake and exhaust valves simultaneously during the engine operation.In addition, FCT allows to apply Miller timing with different levels according to the operation conditions.As indicated in Fig. 4, the FCT mechanism consists of specific valve drive (a), which is controlled by an actuator (b) [7].
The valve drive is integrated in an overhead valvetrain mechanism and has three arms.Every arm has an eccentric cam on its end.Using a control shaft, the eccentric cams can be rotated within the required angular position.The other ends of the arms are in contact with the engine cams through three rollers, in order to transfer the cams' displacement to the push rods as described schematically in Fig. 5.The mechanism depends on the angular motion of the control shaft.Consequently, the three arms are moved forward and backward which displaces the contact points between the cams and the rollers.As a result, the opening and closing  points of the valves are advanced or retarded depending on the rotation direction.As shown in Fig. 5, the eccentric cams at the end of the arms have an inversed angular position.Therefore, the valves' tuning occurs in an opposite manner and if one valve is advanced, the other is retarded.The third eccentric cam which triggers the fuel pump is not considered in the current study as the engine has a separate common rail injections system.In the primary design of the FCT mechanism, the hydraulic actuator in Fig. 4b was used to drive the control shaft.The hydraulic actuator had only two positions with a resolution of 180 • (two tuning positions for the valves also).For the current research project, the mechanism was developed by adopting an electric motor with an encoder (see Fig. 6), instead of the hydraulic actuator.The electric motor is combined with a worm gear and provides slow angular motion, which allows continuous and stepless control with high resolution.The characteristics of the FCT mechanism are summarized in Table 2.In this study, the second configuration is utilized aiming to accomplish different levels of Miller cycle.Additionally, the angular position of the control shaft will be used as an indicator for the Miller cycle levels, and will be referred as "FCT = angle • ".

FCT model
The first part in the simulation framework was developed in MATLAB, aiming to predict the behavior of the FCT as shown in Fig. 7. Depending on the mechanism geometry, the FCT model calculates the valve timing according to the angle of the control shaft.Additionally, different cam profiles can be used as an input.After that, the model generates valve lift curves and tests them against any possible collision with the piston.If no collision is found, the curves are exported to the combustion model in AVL CRUISE-M environment.In this study, two cam pairs were tested as illustrated in Figs. 8 and 9.The first pair (Fig. 8) is the standard one, which is specified for the engine and was used during the experimental tests.The second one was considered in the simulation only for the analysis purpose.The difference between them is the nose of the cams, which controls the time period in which the valve is held open.

Engine model
The engine model represents the second part in the simulation framework.It aims to analyze the engine performance and emissions under the effect of the FCT, taking into account different valve timings.The model was established in AVL CRUISE-M platform by connecting the required blocks together as shown in Fig. 10.Some of them represent the relevant engine components (intake manifold, runners, valves, combustion chamber, etc.).The other blocks were used to control the simulation conditions and observe Fig. 6 FCT actuator with electric motor the results.Each block contains the required parameters to define its thermodynamic state and geometric properties.
As in the engine test bench, only one cylinder is considered in the simulation model with the full thermodynamic cycle (four strokes).Furthermore, the model is based on the phenomenological approach, in which a number of sub-models were included to simulate the thermodynamic processes in the cylinder.

Combustion model
The dual-fuel engine in the current study depends mainly on the NG as the main fuel, while the diesel pilot injection is employed only to ignite the mixture.In this context, the  combustion process can be treated as premixed combustion and diesel fuel is considered only in the fuel composition.In order to predict the in-cylinder pressure and ROHR, a quasidimensional two-zone combustion model based on the fractal combustion concept was implemented.In this concept, the mixture burning rate is calculated depending on the turbulent flame propagation.In addition, the effect of the combustion chamber shape and cylinder composition is considered.The fractal principle treats the flame front as a thin curved layer separating between burned and unburned zone as illustrated in Fig. 11.Then during the flame propagation, the laminar flame surface (A L ) is highly wrinkled under the effect of turbulent eddies with different length scales.Consequently, the burning rate is enhanced, and the total area of the flame front is increased leading to a turbulent flame surface (A T ).
The ratio between the turbulent and laminar flame surface (A T /A L ) represents the wrinkling factor and is responsible for the increase in the turbulent flame speed (S T ) with respect to the laminar one (S L ) [4,9].According to Damköhler [8], the mass burning rate can be expressed as where m b is the mass of burned charge, u is the density of unburned charge, and t represents the time.The wrinkling factor (A T /A L ) can be characterized by the ratio between the maximum turbulent length scale ( L max ) to the minimum one ( L min ).The fractal principle is then applied as it is a math- ematical method to represent the irregular geometries with repeated patterns [9].Hence, the correlation is formulated as where D 3 is the fractal dimension, which describes the roughness of the flame wrinkling and can be defined in terms of laminar flame speed and turbulent intensity.According to Ref. [19], it is defined as the following: where u ′ is the turbulence intensity.Substituting Eq. (3) in Eq. ( 1) gives the mass burning rate as L min is assumed to be equal to the Kolmogorov length scale ( L k ).In addition, L max is equal to the integral length scale ( L I ).Based on the fractal concept, the computation of L min , L max , and D 3 depends on the turbulent flow inside the cyl- inder.The modified K-k model according to Refs.[5,6] is used to characterize the turbulence inside the cylinder as the following: The kinetic energy of the mean flow K: The kinetic energy of the turbulent flow k: Fig. 11 The principle of the fractal combustion model 1 3 ṁin and ṁex are the mas flow rates into and out of the cylin- der, u in is the velocity of the charge entering the cylinder, ρu is the change rate of the unburned charge density, m is the mass of the charge, P is the turbulence production rate, is the rate of the kinetic energy dissipation for unit mass, U f is the mean velocity of the charge, and R e represents Reynolds number.The two factors c t and c l are the turbulent produc- tion constant and the turbulent length scale, respectively, which are used together for the calibration purpose.The main boundary conditions for the simulation model are summarized in Table 3 [1].

Ignition delay model
In dual-fuel engines, ignition delay is considered a critical parameter since it includes interaction between two fuels.In this study, the ignition delay was predicted using a developed correlation from chemical reaction kinetics with mechanism [30,31].It considers the diesel substitution ratio and the air-fuel ratio of the NG mixture.The formula is expressed as where ID is the ignition delay, C 1 and C 2 are calibration factors, is the AFR Actual to AFR Stoich., P Cyl. is the cylinder pressure, W Diesel is the mass fraction of diesel fuel, T Cyl. is the cylinder temperature.The correlation does not exist in ( 13) T Cyl. , the software and was implemented in the model by a userdefined function.

NO x formation model
The reduction of NO x emissions represents one of the main targets in this study.Using the extended Zeldovich mechanism, the NO x formation was predicted [36].The model is widely used and depends on three reactions as the following: where K 1 , K 2 , K 3 are the reaction rates.

Methane slip formation model
Methane slip is a crucial challenge in dual-fuel engines because of the gaseous nature of the utilized fuel.Furthermore, the lean mixture operation increases the misfiring probability.As mentioned in the literature review, the main sources of the methane slip are: • The crevice volumes between the rings and the piston, where the unburned mixture is pushed, and the flame cannot propagate.• The misfiring or the partial burning when the combustion quality is poor.• The valve overlap period, depending on the operating conditions.
( In this study, the three previous sources are considered, and the sum of them will be mentioned as "Total methane slip".Since the methane solubility in oil is low, the unburned methane due to the absorption in the lubrication oil was neglected.Methane slip from the crevice volumes according to Ref. [17] was calculated as the following: where m crevice is the mass of unburned charge in the crev- ices, V crevice is the total crevice volumes, M is the molecular weight of unburned charge, R is the gas constant, and T Piston is the piston temperature.The model assumes that the pressure inside the crevice volumes and in the cylinder is the same.
In dual-fuel engines, partial burning can be a result of too lean mixture or high charge dilution.It was described with a semi-empirical correlation which was proposed by Lavoie in Ref. [16] as follows: where F P is the fraction of unburned charge due to partial burning, C P1 a tunable constant, EVO is the crank angle at exhaust valve opening, 90 is the crank angle at which 90% of the fuel is burned, C P2 is a constant depends on the equivalence ratio, and 0 is the crank angle at which 0% of the fuel is burned.
Methane slip from the valve overlap was calculated during the simulation, as all the chemical species are transported through the blocks during the engine cycle.For the three mentioned sources, it was assumed that no post-oxidation occurs for the unburned mixture.

Simulation setup
Simulation was carried out in AVL CRUISE-M software using the filling and emptying method.During the simulation process, the principles of mass and energy conversion were applied for each time step.The two cam pairs in Figs. 8 and  9 were investigated at three engine loads (50, 75, 100%).For every load, brake power, lambda, and center of combustion (COC) were held constant by utilizing three PID controllers during the entire simulation.The first PID controller modifies the intake air pressure to attain the desired power and keeps it constant according to the selected load.In order to maintain constant lambda with the different valve timing positions, the second PID controller was used.It regulates the NG injection rate through the SOGAV valve.The third PID controller adopts the ignition timing to hold the COC constant with certain degrees after top dead center (ATDC).The previous (19) procedures were applied with the same manner during the experimental tests.Table 4 indicates the target values for the three PID controllers [1].

Model validation
In order to use the model for further investigations, simulation results were compared against the experimental data.The validation of the FCT model is shown in Fig. 12 using the standard cam pair.In Fig. 13, in-cylinder pressure, ROHR, and exhaust Fig. 13 Validation of in-cylinder pressure, ROHR, and exhaust emissions using cam pair (1) emissions are compared against the results from the test bench using cam pair (1).

Results and discussion
Figure 14 shows the engine characteristics with the full of the FCT angles using cam pair (1).It can be seen that as the FCT angle is increased, the maximum cylinder pressure is reduced.The reason is due to the earlier closing of the intake valve which leads to lower effective compression ratio (strong Miller effect).Hence, higher intake air pressure is required to attain the same output power, which is also noticeable by the slight increase in the intake air pressure.In terms of the NO x emissions, they show higher values with the increasing of the FCT angle.
The earlier IVC at the higher FCT values resulted in less trapped mass in the cylinder.Consequently, less heat capacity is gained, and higher NO x are produced since the average in-cylinder temperature is increased.The same effect was reported in Ref. [10], as more NO x emissions are observed with the lower effective compression ratios.The total methane slip is higher as the FCT angle is increased, which reflects the effect of the longer overlap period.Therefore, more fresh charge can escape through the exhaust valve.The partially increasing in the intake air pressure supports this cause as demonstrated before in Fig. 2a.
The effect of cam pair (2) is shown in Fig. 15.Generally, in comparison with cam pair (1), the results show the same behavior.But with cam pair (2), the effect of the FCT angle is more obvious especially for the intake air pressure and the maximum cylinder pressure.As the intake cam in cam pair (2) is smaller than that in cam pair (1), higher intake air pressure has to be exerted to produce the same power.The NO x emissions show the same influence as cam pair (1), but with less sharp effect.In addition, the average NO x values with cam pair (2) at the three loads are less than that with cam pair (1) which indicates lower average in-cylinder temperatures.With cam pair (2), the total methane slip shows slightly higher values with the lower FCT angles.The advanced opening of the exhaust valve at the low FCT angles reduces the gases exposure to the high temperature and shortens the combustion stroke.Subsequently, more unburned mixture is emitted.For the same reason lower NO x emissions were observed with the lower FCT values for both cam pairs.Furthermore, the earlier closing of the exhaust valve (low FCT angles) negatively affects the exhaust scavenging efficiency, which causes dilution effect and reduces the NO x emissions.In comparison with cam pair (1), the shorter overlap period in cam pair (2) prevented the methane slip from sharp

Conclusion and outlook
In this study, a simulation model was developed aiming to analyze the performance of a dual-fuel marine engine using the FCT technology.The model includes two parts, the first one for the FCT was developed in MATLAB.It aims to represent all the available valve timings considering different cam profiles.The resulting valve lift curves were exported to the phenomenological combustion model in AVL CRUISE-M.The combustion model represents the second part and is based on the fractal combustion principle aiming to predict in-cylinder parameters and emissions formation.Simulation results were compared with experimental data from a singlecylinder marine research engine, which employs NG as main fuel, and diesel pilot injection as igniter.Two cam pairs were tested in this research.The first one is the standard pair (cam pair 1), which was used also for the validation purpose.The second pair (cam pair 2) was considered only in the simulation for investigations and analysis.It can be concluded that fractal combustion model can successfully predict ROHR and cylinder pressure in the dual-fuel engine.Additionally, the FCT has a significant effect on the engine performance and can be used effectively to control the exhaust emission.
As the FCT angle was increased, the intake valve closed earlier which reduced the effective compression ratio and was reflected on the maximum cylinder pressure.Furthermore, the longer overlap period with the higher FCT values allowed more fresh mixture to escape through the exhaust valve which increases the total methane slip in the exhaust gases.In terms of NO x emissions, they also showed higher values as the FCT angle was increased.In this study, the higher FCT values not only includes earlier intake valve closing (Miller timing), but also later exhaust valve closing.Since the latter increases, the exhaust gases exposure to the high temperature and allows for longer combustion stroke, more NO x emissions were observed with the higher FCT angles.In other words, the influence of the later exhaust valve opening was more dominant than the influence of the earlier intake valve closing.Additionally, with the low FCT values, the exhaust valve closes earlier which reduces the exhaust gases scavenging efficiency.As a result, the dilution effect increases and contributes to lower NO x emissions.Regarding the intake air pressure, it was higher with cam pair (2), as higher pressure is required with the smaller intake cam to get the same power.

Fig. 1
Fig. 1 Fuel injection pattern in LTC technologies and CDC

Fig. 3
Fig. 3 Test bench of dual-fuel single-cylinder marine engine

Fig. 14
Fig. 14 Effect of cam pair (1) on engine performance

Fig. 15
Fig. 15 Effect of cam pair (2) on engine performance

Table 1
Main engine specifications

Table 2
• CA Fig. 7 Flow chart for the FCT model

Table 4
Desired values in the three PID controllers Fig. 12 Validation of FCT model with cam pair (1)