A CFD Study on the Effects of Injection Timing and Spray Inclusion Angle on Performance and Emission Characteristics of a DI Diesel Engine Operating in Diffusion-Controlled and PCCI Modes of Combustion

Abstract

In three-dimensional (3D) computational fluid dynamics (CFD) simulations, the effects of injection timing and spray inclusion angle (SIA) on performance and emissions of diffusion-controlled and Premixed Charge Compression Ignition (PCCI) combustion in part load for a heavy-duty direct injection (HDDI) diesel engine are studied. The start of injection (SOI) of a 146° SIA injector is varied between −70 and −10 °crank angle (°CA) after top dead center (ATDC). For −50 °CA ATDC SOI with various SIAs between 80° and 146°, PCCI combustion reduces mono-nitrogen oxide (NO_x) emissions significantly compared to conventional diesel combustion (CDC). Due to incomplete combustion in rich zones formed by droplet–cylinder wall interaction, early wide SIA injection deteriorates combustion efficiency (CE) and Indicated Mean Effective Pressure (IMEP) and increases soot and carbon monoxide (CO) emissions. Narrow-angle sprays interacting with the piston bowl elevate CE and IMEP and decrease soot and CO emissions but increases NO_x emissions. Optimal combustion is achieved by avoiding fuel droplet–cylinder wall interaction. By spray-targeting at the stepped lip of the piston bowl, 100° SIA and −50 °CA ATDC SOI yield, respectively, the highest CE and IMEP: 97.8% and 3.37 bar and the lowest soot and CO emissions: 33.5 and 2.2 ppm, with acceptable NO_x emissions.

1. Introduction

Direct-injection (DI) diesel engines are commonly used in heavy-duty (HD) vehicles because of their high thermal efficiency, durability, and reliability. However, meeting ever-stringent emission regulations is challenging because of their significant mono-nitrogen oxides (NO_x) and soot emissions [1].

In diesel-fueled compression ignition engines, combustion occurs in both premixed and diffusion-controlled phases. For a late single injection of diesel fuel, diffusion-controlled combustion begins before injection is terminated. In this case, the fuel-rich-premixed phase of combustion precedes the main diffusion combustion, in which thick diffusion flames form in combustible fuel vapor-air mixture (FVAM) surrounding the sprays in the vicinity of stoichiometric zones. In conventional diesel combustion (CDC), the premixed phase of combustion is a standing flame in the rich FVAM zone located slightly downstream of the maximum spray tip penetration length (STPL) with an equivalence ratio of 2 to 4 at temperatures at approximately 825 K, which results in the formation of incomplete combustion products carbon monoxide (CO), and polyaromatic hydrocarbons (PAHs). As the incomplete rich premixed combustion products are advected farther downstream, a diffusion flame sheath forms around them in the vicinity of their stoichiometric zone with a temperature level of 2700 K, at which point they release most of their chemical energy, which raises their temperature to 1600 K in the zone enclosed by the diffusion flame sheath. Soot particles form from precursor PAHs via the inception, surface growth, and coagulation process in the incomplete, rich-premixed combustion products, which are sufficiently hot for soot formation via fuel pyrolysis above 1300 K. Mixing quality and temperature determine the soot-formation rate, which is elevated in fuel-rich zones. The concentration and size of soot particles, respectively, peak in the zone enclosed by the diffusion flame sheath and within the diffusion flame sheath at the periphery of the head vortex in the leading portion of the reacting jet. Then, soot particles are partly oxidized by hydroxyl radicals (OH) and dioxygen (O₂) within the diffusion flame sheath. On the other hand, thermal nitric oxide (NO) forms within the slightly lean side of the high temperature diffusion flame sheath and in hot gas that remains after the end of combustion [2].

At low loads, simultaneous reduction in NO_x and soot emissions while maintaining indicated specific fuel consumption (ISFC) at the expense of higher CO emissions is possible with low-temperature combustion (LTC) as compared to CDC. For LTC, fuel vapor should be diluted with air and exhaust gas recirculation (EGR) sufficiently in the early stages of compression stroke so that mixture homogeneity is improved [3]. Homogeneous charge compression ignition (HCCI), Premixed Charge Compression Ignition (PCCI), and Reactivity-Controlled Compression Ignition (RCCI) modes of combustion have been extensively studied within the context of LTC.

In the HCCI concept, the entire homogeneous lean mixture of fuel vapor, air, and EGR formed before it enters the cylinder or during the intake stroke auto-ignites in the compression stroke as its temperature and density increase, without the assistance of a spark plug. HCCI has diesel-engine-like efficiency with extremely low NO_x emissions, but CO emissions need after-treatment in an oxidizing catalyst to meet emissions regulations [4]. Auto-ignition depends on various parameters such as temperature, pressure, composition, homogeneity, thermophysical, and chemical properties and turbulence intensity of the combustible mixture. However, because of difficulties in the control of combustion phasing, HCCI combustion is unsuitable at high engine loads due to excessive heat release and pressure rise rates, which cause increased NO_x emissions and engine wear, respectively. Furthermore, HCCI combustion has problems in cold starting and idle operation. Therefore, HCCI combustion that is driven mainly by chemical kinetics presents drawbacks as the full range of engine operation is considered [5,6].

The PCCI concept, which is a hybrid of HCCI and CDC, can be realized by early injection of diesel fuel during the compression stroke, which allows sufficient time to partially mix the fuel vapor with air prior to auto-ignition, resulting in more homogenous combustion compared to CDC. However, the mixing quality of PCCI combustion is lower than that of HCCI combustion. In the PCCI concept, after injection ends and fuel droplets completely evaporate, lean mixture zones at various locations consecutively auto-ignite near the top dead center (TDC). PCCI combustion is governed primarily by chemical kinetics because injection–evaporation and heat release events do not overlap [7]. In PCCI combustion, soot and NO_x emissions are significantly lower than those of CDC because a more uniform mixture is formed due to longer ignition delay (ID), and flame temperatures are lower due to longer combustion duration, as it occurs at various lean mixture zones consecutively. Therefore, heat release rate (HRR) history is spread over a longer time, even though HRR peaks are relatively higher compared to HCCI combustion [8,9]. Furthermore, engine-out CO emissions of PCCI combustion are significantly lower compared to HCCI combustion, but a diesel oxidation catalyst (DOC) is still needed [10].

In PCCI combustion, pollutant formation and oxidation zones depend on the evolution of the in-cylinder equivalence ratio, pressure, and temperature fields according to EGR rate, intake valve closing (IVC), injection strategy and timing, in-cylinder flow field, spray–wall interaction, and ID. In PCCI combustion without EGR, minor amounts of NO are formed in high-temperature (above 2400 K) fuel-rich zones beneath the valves due to splashed droplets after spray impingement on either the piston bowl or the cylinder liner, depending on the spray inclusion angle (SIA) and injection timing. On the other hand, soot is formed at highly rich zones with moderately elevated temperatures of approximately 1800 K, such as crevice and squish volumes and around the spray impingement point. Moreover, soot oxidation rates are rather low due to insufficient temperature levels. In PCCI combustion, CO forms due to early incomplete combustion at fuel-rich low-temperature zones between 1000 and 1400 K, which partially oxidize after being advected to high-temperature zones. However, CO remaining in low-temperature zones cannot oxidize, which causes elevated CO emissions in PCCI combustion. Therefore, for simultaneous reduction in soot and NO_x emissions, spray type, inclusion angle, and penetration, as well as start of injection (SOI) timing, should be compatible with the chosen piston bowl geometry to ensure proper spray-targeting, and FVAM should be diluted with EGR so that premixing to leaner equivalence ratios is ensured extensively, locally fuel-rich zones are minimized, and flame temperatures are sufficiently lowered.

In a recent experimental sensitivity study on the effect of SOI timing ranging between BDC and TDC and intake charge temperature on CA50 with and without EGR, it was determined that while the intake temperature is almost the only choice to control CA50 for SOI timings between −180° and −80° ATDC, (injection during intake valve closing and classical HCCI zones), SOI timing is much more effective than the intake temperature in controlling CA50 for SOI timings between −80 °CA and −32.5 °CA ATDC (top-land/crevice effect and stratification effect zones). Furthermore, while for SOI between −80 °CA and −50 °CA ATDC, CA50 can be controlled very effectively by SOI timing in the no-EGR case, EGR causes more uniform in-cylinder equivalence ratio distribution; reduces engine-out NO_x, CO and UHC emissions; improves combustion efficiency; reduces the sensitivity of CA50 to SOI timing; and increases the influence of intake temperature on CA50 [11].

However, due to early injection timings, attempts to exploit the low soot and NO_x emission capabilities of PCCI combustion led to deeper STPL, causing spray impingement on cylinder walls. Therefore, incomplete combustion due to the slower oxidation of fuel in proximity to the cylinder liner at lower temperatures causes high CO emissions. To improve uniformity of the mixture; eliminate cylinder liner wetting issues; and control combustion phasing, pressure rise rate (PRR), noise, and emissions while maintaining thermal efficiency, two-stage injection with judiciously selected injection timings and injected fuel quantities and EGR rate according to the load is suggested [12]. A dual-combustion-mode narrow SIA DI diesel engine concept Narrow Angle Direct Injection (NADI) with an optimized piston bowl shape facilitates fuel vapor extraction from the bowl with sufficient speed toward the squish area, thereby limiting fuel wall-wetting, promoting fuel–air mixing, and controlling combustion timing so that near-zero NO_x and particulate emissions are simultaneously achieved, while fuel efficiency is maintained for the HCCI mode of combustion for part and medium loads through effective mechanisms such as variable EGR rate, IVC, SOI timing, and intake gas temperature, and then switches to CDC with injection close to TDC at high loads [13]. For a single-cylinder DI diesel engine with a piston bowl geometry modified according to the spray-targeting requirement of the first SOI timing and decreased compression ratio (CR) to prolong ID so that dual injection with a narrow SIA strategy for which HCCI combustion of the minor portion of the fuel from early first injection after a sufficiently long ID for complete evaporation of the fuel and formation of a homogeneous mixture is followed by retarded, late injection (after top dead center (ATDC)) burning of the majority of the fuel is shown to be effective in both reducing NO_x and CO emissions and preventing the deterioration of combustion efficiency (CE) and maintaining Indicated Mean Effective Pressure (IMEP) of single-injection case if both injection timings are optimized [14]. In PCCI combustion with double injection strategy, an NADI-type modified piston bowl geometry to avoid fuel cylinder–liner interaction during the first early injection at 50–60 °crank angle (°CA) before top dead center (BTDC) with a narrow SIA of 60°, a decreased CR of 15:1 to prolong the ID for improved mixture uniformity prior to HCCI combustion, a cooled EGR rate of 30% for both increasing the first ID further and simultaneous reduction in engine-out soot and NO_x emissions, and a main injection at 5 °CA BTDC for increased IMEP due to the bulk of heat release occurring near TDC is shown to be effective [15].

In PCCI operation of a DI diesel engine with very low swirl ports and a lip-less shallow-dish-type piston cavity designed for drastically weakened in-cylinder flow to avoid start ability, and white smoke problems from a cold engine by reducing heat loss through the walls of the combustion chamber, low penetrating and highly dispersed sprays from a micro-multi injector (MMI) with numerous smaller-diameter nozzles and a wide SIA to obtain more prompt vaporization due to better atomization and to create a more uniform FVAM without wetting the cylinder liner so that the formation of fuel-rich regions around the cylinder liner is avoided during early injection, a lower CR of 14:1 to advance diffusive combustion phasing for an earlier SOI timing so that sufficient time for soot oxidation is created at full load without compromising ISFC, full-load torque, and cold condition performance are suggested. Furthermore, for high loads, combining PCCI and diffusive combustion via a multiple injection strategy, which consists of pilot injection of the minor portion of the fuel and main injection with the rest, can improve the tradeoff between NO_x and soot emissions, while combustion noise can be reduced by tuning the EGR rate [16].

In recent years, the effects of fuel SOI timing and SIA on performance and emissions of PCCI mode of operation have been studied by considering STPL, wall impingement, and in-cylinder charge motion. For PCCI mode of combustion in a relatively low CR, high-speed direct injection (HSDI) engine with an open-crater-type piston bowl, optimum spray-targeting spots around the edge of the piston bowl near the squish region result in better CO emissions compared to targeting the bottom of the piston bowl, as well as excellent soot and NO_x trade-off due to squish-flow-assisted enhancement of mixture formation in the piston bowl and prevention of local rich spots [17]. For a DI diesel engine with a reentrant piston bowl, the lowest ISFC for PCCI operation in part load can be obtained via single-injection with optimum SIA, which avoids wall impingement [18]. It is apparent that the PCCI mode of combustion in a DI diesel engine should be controlled by optimizing the SIA and SOI timing to avoid drawbacks of the concept [19,20,21].

In HD and LD diesel engines, a dual-fuel, or a single-fuel with additive RCCI combustion strategy enables better combustion phasing, duration, and magnitude control by adjusting the ratio of the low-reactivity fuel to the high-reactivity fuel or the additive mass fraction, respectively, with reduced EGR rates for a broader operating range compared to HCCI and PCCI combustion [22]. In RCCI concept, port or direct injection of a low-reactivity fuel is followed mostly by early cycle single or multiple DI of high-reactivity diesel fuel so that fuels are blended in the cylinder to generate fuel-reactivity gradients for controlling combustion duration. Injection timings, durations, and ratio of fuels are adapted to the operating conditions to control ignition timing and avoid high pressure rise rates and misfire [23,24,25]. In RCCI combustion, zones of varying reactivity sequentially auto-ignite [26,27]. Staged combustion of high- and low-reactivity fuels due to spatially stratified reactivity reduces heat release rate so that combustion duration extends, and pressure rise rate remains below the acceptable limit.

Innovative combustion systems and injection strategies in combination with advanced combustion concepts and alternative fuels can improve the CO₂ and pollutant emissions and increase the efficiency and performance of CI engines. Therefore, they have the potential to boost the fuel economy and to reduce engine-out emissions. Recently, direct injection of oxygenated fuel blends such as diesel–gasoline–bioethanol or diesel–biodiesel–bioethanol in CI engines has been proposed to increase the renewable content of the fuel and to reduce CO₂ and PM emissions while improving soot-NO_x trade-off characteristics. While the lower initial combustion temperatures due to the charge cooling effect of alcohol fuels reduce NO_x emissions, both the average diameter and amount of PM emissions can be decreased by using oxygenated fuel blends involving lower-viscosity alcohol fuels. Increasing the EGR rate in part loads reduces NO_x emissions without compromising PM emissions with acceptable CO and UHC emissions. However, a lower cetane number and higher autoignition temperatures of alcohol fuels limit utilization fractions of alcohols (besides corrosion and cavitation problems in the fuel-injection system) and increase CO and UHC and emissions especially at a low load. Furthermore, higher mass flow rates of alcohol fuels are needed due to their lower heating values compared to diesel fuel [28,29,30].

Conventional multi-jet injector concepts cause significant amounts of fuel impingement on the cylinder walls due to their high-momentum spraying characteristics from very small diameter nozzles. In advanced injections, spray–cylinder liner interaction can be avoided with a “Hollow Cone Nozzle” that has a larger discharge area, which significantly reduces spray penetration due to the lower momentum of its spray compared to that of a conventional multi-jet injector. PCCI combustion can be realized with a hollow cone injector at a relatively low injection pressure, and very low NO_x and soot emissions at the expense of higher CO and UHC emissions can be obtained compared to conventional diesel combustion. The operating range of PCCI combustion can be shifted to higher loads by lowering the compression ratio [31]. Recently, an innovative hybrid injector forming a hollow cone spray of low penetration and five compact jets with a wide inclusion angle that penetrate deep in the combustion chamber, HCN-5, has been suggested. At injection timings close to TDC, compared to a hollow cone nozzle injector, due to better mixing quality, HCN-5 can improve combustion stability, peak pressure rise rate, thermal and net indicated efficiencies, and engine-output CO, UHC, and soot emissions [32].

From the literature cited above, it is evident that the effects of SIA and SOI timing on the combustion performance and emissions of HDDI diesel engines operating in the PCCI mode of combustion have not been extensively studied so far. Therefore, the novelty of this study is the evaluation of the effects of single-injection timing and SIA on combustion performance and emissions of a large-bore, low-speed, HDDI diesel engine operating at a low load in PCCI mode, through 3D CFD simulations. For a single-cylinder DI diesel engine having a stepped-lip piston bowl and swirl ratio which have been optimized for CDC, operating in the PCCI mode of combustion at a low load, the present study determines the optimal single-SOI timing and SIA to enhance uniformity of lean FVAM and thereby to lower flame temperatures sufficiently by controlling the combustion phasing, so that engine-out soot and NO_x emissions are reduced simultaneously while keeping CO emissions at acceptable levels, without compromising ISFC.

The objective of this study is to determine the root causes of combustion characteristics, engine performance, in-cylinder pollutant formation, and engine-out emissions of a HDDI diesel engine operating in the PCCI mode of combustion at a low load. In this study, firstly, for a single-cylinder research engine, depending on the injection timing, operating at a low load either in conventional diesel or PCCI mode, the interaction of fuel spray with combustion chamber walls, FVAM formation, combustion efficiency, and IMEP and engine-out NO_x, soot, CO, and UHC emissions were investigated. Then, for the injection timing determined to be optimal for PCCI combustion at a low load, the SIA enabling optimum engine performance and engine-out emissions depending on the combustion characteristics developed according to the FVAM equivalence ratio distribution resulting from the spray–wall interactions was determined. The novelty of this study is that it reveals successive combustion and quenching zones at low temperature in PCCI combustion in a diesel engine operating at a low load. It also presents the interactions of the spray with the piston and cylinder walls, as well as the RER, temperature, NO_x, and soot distributions for a single-SOI timing–SIA pair at which optimum PCCI combustion is obtained. As a result of this study, the design of an innovative injection system that can dynamically increase both the spray inclusion angle and spray tip penetration along with the injected fuel amount as the load increases is suggested due to delayed SOI timing.

2. Materials and Methods

2.1. Single-Cylinder Research Engine

In this study, the combustion performance and emissions of a turbo-charged, liquid-cooled, four-stroke, diesel-fueled, HDDI, single-cylinder research engine based on Ford Otosan 13 L Ecotorq engine are tested on the dyno installation shown in Figure 1 and simulated by 3D CFD software CONVERGE. The engine test bed setup is shown in Figure 2. Technical specifications of the engine are listed in

Table 1. Engine technical specifications.

Bore × stroke [mm]	130 × 160
CR	17.0:1
Swept volume [L]	2.1
Nozzle diameter [mm]	0.18
Number of injector nozzles	8
SIA [°]	146
Maximum injection pressure [bar]	2500
Intake valve opening (IVO) [°CA BTDC]	4
IVC [°CA ABDC]	13
Exhaust valve opening (EVO) [°CA BBDC]	40
Exhaust valve closing (EVC) [°CA ATDC]	4

. The engine, which has a CR of 17:1, is equipped with an 8-nozzle diesel fuel injector with a nozzle diameter of 180 μm and an SIA of 146°. The engine is tested at a low load operation of 3 bar IMEP and an engine speed of 1100 rpm, for which a simultaneous reduction in NO_x and soot emissions is possible without compromising engine torque and experiencing excessive noise in PCCI mode via proper control of combustion. All measurements and simulations are performed in steady state operation. The operating condition of the engine is given in

Table 2. Engine operating conditions.

Engine speed [rpm]	1100
Injected fuel mass [mg/cycle]	39
Injection pressure [bar]	800
Absolute intake pressure [bar]	1.36
Fuel (diesel surrogate)	n-heptane
Air inlet temperature [K]	298
Air to fuel ratio [-]	71:1
Total residual gas rate [%] (no-EGR)	3.4%
Swirl ratio [Tippelmann]	1.2

.

2.2. CFD Model of the Research Engine and Solution Procedure

In this study, 3D CFD simulations were performed by CONVERGE [33], a compressible, reactive, transient flow solver for complete internal combustion engine (ICE) cycle analyses. Turbulent motions of intake charge and fuel sprays, spray combustion, and emissions of a single-cylinder HDDI diesel research engine are calculated. Statistical behaviors of diesel fuel sprays are determined by solving the Williams spray equation for the droplet distribution function coupled with equations of motion of the droplets [34,35,36], Reynolds-averaged Navier–Stokes (RANS) equations and re-normalization group (RNG) k-ε turbulence model equations [37,38], together with droplet breakup [39,40] collision and coalescence [41], vaporization [36,42], and wall interaction sub-models [43,44].

The RNG k-ε turbulence model in which turbulent kinetic energy production is modeled by isotropic eddy-viscosity assumption is utilized to characterize rapidly compressed swirling in-cylinder flow. The rapid distortion RNG k-ε turbulence model with temperature wall functions for variable density flows, which considers the increase in the turbulent Prandtl number in the boundary layer according to Han and Reitz, is preferred for accurate and computationally efficient evaluation of wall heat flux distribution in engine combustion simulations [38].

Fuel sprays are tracked via a Lagrangian approach. Blobs with a diameter equal to that of the injector nozzles and a velocity determined by the injection rate are introduced stochastically in parcels. Because wall temperatures are below the fuel boiling point for DI diesel engines, fuel droplets impinging on a dry wall spread or splash, forming an evaporating liquid film on the wall. Droplets impinging on the wetted wall either rebound, spread, or splash back to the chamber, depending on their Weber numbers, which are based on their normal incident velocities. In IC engine dense spray simulations of this study, spray primary and secondary breakup, dynamically changing droplet drag, droplet collision and coalescence, and heat and mass transfer among the droplets and the surrounding gas mixture during vaporization of the fuel and spray–wall interaction sub-models are included. The effects of a droplet’s Reynolds number based on its velocity relative to the gas, as well as its oscillation and distortion, as estimated by Taylor’s analogy based on a forced spring–mass–damper harmonic oscillator system, on its dynamically varying drag coefficient are considered [45]. In primary breakup calculations of liquid fuel jets near nozzle exits, the formation of child droplets from injected liquid blobs due to Kelvin–Helmholtz (KH) instability is considered. For secondary breakup calculations, the Kelvin–Helmholtz–Rayleigh–Taylor (KH-RT) instabilities competition model is applied [39,40]. In this study, a no time counter (NTC) collision model based on Monte Carlo simulations of gas dynamics and composite post-collision outcomes model are applied to estimate the number of droplet collisions and their outcomes [46]. Bounce, grazing, stretching, and reflexive separations and permanent coalescence cases are considered as post-droplet collision outcomes of hydrocarbon droplets according to droplet collision Weber number, impact parameter, and droplet size ratio based on the composite post-collision outcomes model [41]. The dispersion of spray droplets via turbulent velocity fluctuations and the depletion of gas-phase turbulent kinetic energy due to work performed by turbulent eddies on droplets are calculated by including the effect of fluctuating velocity on the drag coefficients of droplets in droplet equations of motion and adding spray source terms to gas-phase turbulent transport equations based on the O’Rourke turbulent dispersion model, which assumes a Gaussian distribution for each fluctuating velocity component [36]. A DI diesel engine’s thermodynamic efficiency and emissions are significantly affected by evaporation rates of fuel spray droplets since only vaporized fuel mixed with air at a combustible ratio can react with oxygen. Mediocre air–fuel vapor mixing because of an insufficient fuel evaporation rate due to decreased specific surface area of spray droplets associated with poor atomization, and/or unfavorable in-cylinder charge motion, or fuel film formation on combustion chamber walls caused by the adhesion of impinging droplets increase soot emissions. On the other hand, extremely rapid evaporation of diesel fuel before auto-ignition increases NO_x emissions due to fast premixed combustion at elevated temperatures [47]. Various evaporation models suitable for calculating heat and mass transfer rates of a single-component fuel droplet in DI diesel engine combustion simulations exist in the literature [36,42,48,49,50]. In this study, the temperature of each droplet is assumed to be uniform, and its temperature and size change histories are calculated by solving two ODEs for heat and mass transfer rates on its surface by utilizing Frossling correlations for the forced convection of heat and mass. [36,42]. Finally, a rebound/slide model is implemented to simulate the interaction of sprays with walls [43,44].

In CFD simulations, the number of cells is kept below three million by the Adaptive Mesh Refinement (AMR) algorithm. AMR automatically embeds finer cells at locations where absolute values of estimated sub-grid scales of temperature and velocity fields are above dynamically adjusted threshold values to resolve flow features that have high gradients away from boundaries and removes the finest layer of excessively refined cells adjacent to the walls to comply with targeted y+ values [51]. The cell size of the adaptive mesh varies between 0.25 mm and 4 mm depending on temperature and velocity gradients as the in-cylinder charge motion and chemical reactions evolve. Satisfactory grid-convergence of the flame liftoff length and reaction-zone thickness of the reacting diesel sprays is assured by the selected minimum cell size.

In this study, direct numerical time integration of production rate equations of chemical species constituting an optimally reduced multi-step chemical reaction mechanism of the diesel fuel surrogate n-heptane (n-C7H16) is performed. Since thermophysical properties of n-heptane are unlike those of diesel fuel, those of a heavier hydrocarbon, such as dodecane, tetradecane, hexadecane, etc., can be substituted. The thermophysical properties of tetradecane (C14H30) such as normal boiling point, critical properties, vapor pressure, latent heat of vaporization, liquid density, viscosity, thermal conductivity, gas diffusion coefficients, and surface tension are utilized in simulating non-chemical processes such as fuel injection, spray break-up, droplet interactions, and evaporation.

The surrogate fuel n-heptane is selected to generate the reduced kinetic mechanism for CFD simulations due to the similarities of its cetane number, auto-ignition temperature, ID, and partially premixed and diffusion-controlled combustion to those of the diesel fuel [52]. However, because the lower heat value of n-heptane is higher than that of the diesel fuel, heat release, peak pressure, and temperature during its combustion is higher than those of the diesel fuel, resulting in less smoke and CO and slightly higher NO_x emissions. In this study, a reduced mechanism of n-heptane combustion which involves 42 species and 168 elementary reactions is implemented to simulate the combustion kinetics [53]. In this study, sub-grid scale turbulence–chemistry interaction is not explicitly modeled. Instead, species concentrations are calculated via direct, fully implicit time integration of the stiff system of ODEs resulting from the multistep chemical kinetics mechanism within each CFD time step. The transient multistep chemical kinetics solver of CONVERGE CFD software (SAGE) solves ignition, premixed, and diffusion-controlled combustion regimes [54]. Furthermore, in simulations of both low- and high-temperature combustion occurring after either early or late or multiple injections in diesel-fueled compression ignition engines, direct time integration of a reduced n-heptane mechanism for each computational cell, which is assumed to be filled with a homogeneous mixture, without invoking a sub-grid scale turbulence–chemistry interaction model, is a validated method on the KIVA-CHEMKIN platform. This approach accurately predicts the ID, in-cylinder pressure, and AHRR histories, as well as in-cylinder phenomena such as the locations of both cool and hot flame ignition and second-stage combustion sites and the spatial distribution of soot and OH [55,56].

The extended Zeldovich mechanism, which consists of three elementary reactions of seven species, is implemented to calculate NO emissions [57]. NO is converted to NO_x by utilizing a mass scaling factor of 1.533.

Furthermore, prompt NO_x formation in fuel-rich, low-temperature conditions at the initial stages of ignition, which is a small contribution, less than 5%, to NO_x emissions in the lean combustion of diesel fuel, is calculated via the De Soete model in the CFD simulations [58].

A semi-empirical model which consists of Hiroyasu soot formation and Nagle and Strickland-Constable (NSC) soot oxidation kinetics equations is applied to calculate soot emissions [59,60].

The solution domain is the full 3D single-cylinder research engine with a moving piston and valves including intake and exhaust ports and part of the intake manifold, as shown in Figure 3. Simulations start at 28 °CA before exhaust valve opening (BEVO) and run for 720° rotation of the crankshaft to complete the four-stroke cycle so that flow in each stroke is accurately calculated. Cycle-averaged wall-temperature distributions of the cylinder liner and head as well as the piston, which are determined via conjugate heat transfer calculations of the engine-cooling jacket and piston-cooling jet, are applied as wall thermal boundary conditions for the fully 3D CFD simulations of in-cylinder reactive flow [61]. Wall temperatures and inlet conditions of the charge and fuel spray are presented in

Table 3. Initial and boundary conditions for CFD simulations.

Absolute inlet manifold pressure [bar]	1.36
Air inlet temperature [K]	298
Average piston temperature [K]	543
Average liner temperature [K]	398
Average cylinder head temperature [K]	521

.

3. Results

3.1. Validation of Simulations

The results of 3D CFD simulations of in-cylinder reactive flow are validated with experimental data obtained from the single-cylinder DI diesel research engine runs. To validate the performance and emissions of the diesel-fueled compression ignition engine for both diffusion-controlled and PCCI modes of combustion, retarded and advanced single-injection cases, respectively, are considered. A piezoelectric pressure transducer (AVL GU21C) is utilized to measure in-cylinder pressure history according to °CA, which is recorded and displayed by an indicated pressure measurement system (AVL IndiCom—advanced combustion analysis software). NO_x and CO emissions are determined in a dynamic time-synchronous way via the Fourier-Transform Infrared Spectroscopy (FTIR) exhaust gas composition measurement system (AVL SESAM i60 FT). Soot emissions are measured by an AVL smoke meter based on the paper filter method, by which filter smoke number (FSN) is determined. The range and accuracy of the exhaust gas measurement system are listed in

Table 4. Range and accuracy of exhaust gas measurement system.

Measured Gas	Units	Range	Linearity	Reproducibility
CO	ppm	0–2000	<2.0% of measured value	<0.5 of range full scale
NOx	ppm	0–1000	<2.0% of measured value	<0.5 of range full scale
Soot	FSN	0–10	-	+/−0.005 +6% measured value

.

In-cylinder pressure and AHRR calculated via 3D CFD simulations are compared with measurements for SOI of −10 °CA ATDC and −50 °CA ATDC in Figure 4 and Figure 5, respectively, to validate CDC and PCCI combustion cases.

In CDC case, auto-ignition of rich FVAM evolving at the tips of the fuel sprays during the ID period initiates the premixed phase of combustion at a low temperature which lasts a few °CA, causing high AHRR accompanied with chemiluminescence emissions resulting in PAH and CO formation, followed by soot formation in the higher temperature region downstream and release of the majority of chemical energy due to the oxidation of rich-premixed combustion products in the diffusion-controlled high-temperature combustion zone around the spray in the vicinity of the stoichiometric mixture where CO₂, H₂O, and NO_x are produced [57]. As shown in Figure 4, 3D CFD simulations predict in-cylinder pressure and AHRR histories satisfactorily for the diffusion-controlled phase of diesel combustion. However, the initial peak in AHRR, 264 J/°CA, which was calculated during the premixed phase of combustion, is not observed in the experimental result due to the smoothing effect of the software processing the test data.

On the other hand, in the PCCI combustion CFD simulation for an SOI of −50 °CA ATDC, after a much longer ID of 24.7 °CA due to a low-temperature environment in which fuel sprays evaporate and form an inhomogeneous FVAM, five AHRR peaks with various magnitudes are observed depending on the mixture Reaction Equivalence Ratio (RER) and temperature distribution at each low temperature premixed lean combustion event, as shown in Figure 5. In the Early Single-Injection (ESI) case, after the first auto-ignition, a sequence of three rapid premixed lean combustion-flame quenching events of very short durations at various locations occur. First, extremely lean FVAM packets with an RER slightly above 0.04 auto-ignite forming cool flames near the valves, which are quenched rapidly by convective heat losses to still-evaporating fuel droplets and richer and colder mixtures surrounding them, as well as to colder walls of the combustion chamber. Heat released during each event and further compression of the reactants–products mixture by the piston increase the mixture temperature further, causing auto-ignition of higher RER regions in the next event. These flames are also rapidly quenched via the above-mentioned mechanism. The remaining fuel is consumed by two major heat release events of longer durations without flame quenching due to a significantly decreased amount of fuel droplets and completion of evaporation, respectively. In the final lean combustion event, AHRR gradually diminishes until the fuel is completely depleted near TDC. Thus, the peak combustion chamber temperature, which is 750 K at the start of the first auto-ignition, increases to 1050 K for the first cool flames, and then reaches 2800 K during the final phase of combustion, due to the sequence of auto-ignition-quenching and final major heat-releasing lean combustion events. The final major heat release lasts longer and subsides as the remaining fuel vapor is depleted near TDC.

For the ESI case with SOI of −50 °CA ATDC, the evolution of RER, temperature, and OH radical mass fraction contour plots in a plane passing through the axes of both the cylinder and a fuel spray, as well as the OH iso mass fraction surface of 1 × 10⁻⁷ in top view, during the first heat release peak are presented according to °CA BTDC in Figure 6. As shown in Figure 6a, just before the first auto-ignition, fuel vapor accumulates near the cylinder liner, while some fuel vapor spots are formed in the piston bowl because spray droplets with limited momentum are incapable of penetrating long distances as injection pressure decreases during injector closing. Cool flames with lower temperatures, above 800 K, are formed at the lean mixture spots near the liner with an RER of approximately 0.2 due to heat losses to fuel droplets accumulated nearby. On the other hand, an extremely lean mixture with an RER slightly above 0.04 formed under the valves auto-ignites, such that the cool flame temperature is approximately 1000 K, causing local fast combustion. These flames are rapidly quenched by the surrounding mixture, which has a lower temperature and higher RER. Zones with increased temperature under the valves due to fast combustion of extremely lean FVAM are shown in Figure 6b. The formation and destruction of OH in extremely lean rapid combustion zones can be seen in its mass fraction contour plot history presented in Figure 6c. The evolution of the iso surface of the OH mass fraction of 1 × 10⁻⁷ during the first heat release peak according to °CA BTDC is presented in Figure 6d in the top view.

Comparisons of calculated engine-out NO_x, soot, and CO emissions with measurements for both injection timings are presented in Figure 7. NO_x and CO emission predictions agree with the measurements well for both injection timings. However, there are significant differences between the calculated and measured soot emissions for both injection timings, which are rather amplified in PCCI combustion. Soot emissions predicted by the Hiroyasu–NSC semi-empirical model are about twice as high as the measured value for PCCI combustion. The Arrhenius pre-exponential factor and the scaling factor in the soot formation and oxidation rate equations, respectively, should be calibrated according to the temperature to obtain a unified formulation of diffusion-controlled and low-temperature combustion. Furthermore, enhanced evaporation and fuel–air mixing rate due to increased surface area of much smaller child droplets formed by the splashing of high Weber number parent droplets impinging on the cylinder liner decrease the soot-formation rate [62]. However, in this study, the splashing of the impinging droplets is not considered by the rebound/slide spray–wall interaction model in transient spray combustion simulations.

On the other hand, remarkably, in the CDC case with SOI of −10 °CA ATDC, the difference between the calculated and measured soot emissions is almost negligible because immediately before the spray tip impinges on the stepped lip of the piston bowl, the rich-premixed phase of combustion starts ahead of the spray tip at a 750 K temperature. Therefore, spray droplets do not interact with the piston bowl during the subsequent diffusion combustion phase either. Instead, STPL decreases after the rich FVAM jet splits into two streams. The lower branch spreads down in the piston bowl, while the upper branch penetrates toward the squish region after spreading over the stepped lip of the piston bowl, during which a lifted diffusion flame with a temperature of approximately 2800 K is formed in the vicinity of the stoichiometric mixture zone. A diffusion flame surrounds the main mixture and extends along the sprays’ axes up to the stepped lip of the piston bowl. Then, the diffusion flame spreads into the piston bowl and the squish region while maintaining its high temperature. There is a time lag between the spreading of the diffusion flame and earlier advection of rich mixture to these zones. A flame propagation in diesel combustion case with an SOI of −10 °CA ATDC is shown in the next section.

3.2. Effect of Injection Timing on Engine Performance and Emissions

For a DI diesel engine operating at a low load and speed without diluting the intake charge with EGR, the mode of combustion is primarily determined by injection timing. ESI timing results in premixed combustion having several AHRR peaks of short duration after the compressed FVAM auto-ignites, which is followed by the main combustion events with major heat releases. Furthermore, for very advanced injection timings, fuel spray has sufficient time to penetrate, evaporate, and mix with the air before auto-ignition occurs. Lean combustion of the well-mixed FVAM takes place in a lower-pressure and -temperature environment compared to CDC, resulting in very low NO_x emissions [63]. However, for advanced injection timings, the risk of wetting the piston bowl and/or the cylinder liner increases due to limited spray break-up and evaporation rates in a lower pressure and temperature environment. Wall wetting causes, near the walls, both low hydrocarbon oxidation and high soot-formation rates, leading to reduced CE due to incomplete combustion. On the other hand, diffusion-controlled combustion following a single late injection occurs at a significantly higher pressure and temperature, which results in a considerable increase in NO_x emissions.

In this section, AHRR and in-cylinder pressure histories, CE, NO_x, soot, and CO emissions of the single-cylinder DI diesel research engine are calculated via 3D CFD simulations for a sequence of SOI timings ranging between −70 °CA and −10 °CA ATDC.

The effects of SOI timing on in-cylinder pressure and AHRR histories, which are calculated via 3D CFD simulations, are presented in Figure 8 and Figure 9, respectively. According to CFD simulations, the highest in-cylinder pressure peak is obtained for an SOI of −30 °CA ATDC, while an SOI of −20 °CA ATDC results in the second highest peak pressure.

For ESI cases with an SOI of −30 °CA, −50 °CA, and −70 °CA ATDC, AHRR histories display a sequence of peaks with short durations due to the formation and quenching of cool flames near hot spots on valves and the cylinder head. These are followed by major reactions at higher temperatures with longer durations due to successive kinetically controlled lean combustion events. While main combustion events occur within a limited zone near the valves for an SOI of −70 °CA ATDC, they take place in a larger region inside the piston bowl with higher AHRRs and flame temperatures for SOI of −50 °CA ATDC.

In late injection cases with an SOI of −20 °CA and −10 °CA ATDC, as the FVAM auto-ignites, a sudden heat release occurs due to the kinetically controlled premixed phase of combustion, for which AHRR forms a major spike of short duration. This is followed by diffusion-controlled combustion, in which diffusion fluxes of fuel vapor and oxidizer determine the reaction rate, during which AHRR first increases gradually, forms a smaller peak, and decays sluggishly as the fuel is depleted. The highest AHRR during the premixed phase of combustion occurs for an SOI of −30 °CA ATDC.

For early injection strategy cases with an SOI of −30 °CA, −50 °CA, and −70 °CA ATDC, both AHRR peak and initial premixed combustion duration decrease as SOI is advanced, because in a cooler environment, even though ID increases, less fuel evaporates before auto-ignition.

For the late-injection strategy cases with an SOI of −20 °CA and −10 °CA ATDC, when SOI is retarded, the decrease in both the initial AHRR peak and the duration of the premixed phase of combustion can be attributed to a shorter ID. This is because less FVAM is formed downstream of the fuel sprays as the available time for spray breakup and evaporation before auto-ignition decreases, even though the temperature of the compressed air is higher. Furthermore, according to CFD simulations, premixed combustion phase durations are much shorter for late injection strategy as compared to those for early injection cases.

Start of Ignition (SOIgn) in °CA BTDC according to SOI is presented in Figure 10. In the PCCI cases, SOIgn is not very sensitive to SOI, but it is slightly delayed if SOI is retarded. For the ESI cases with SOI timings of −70 °CA ATDC and −50 °CA ATDC, FVAM, which auto-ignites between −26 °CA ATDC and −25 °CA ATDC, experiences a sequence of kinetically controlled lean low-temperature combustion events. These are unevenly distributed in time and have lower AHRR peaks compared to the initial spikes of the premixed phase of combustion for late injection cases. However, for late injection cases, SOIgn is strongly influenced by SOI timing, as shown in Figure 10.

For PCCI and diffusion-controlled combustion cases, ID in °CA according to SOI in °CA ATDC is presented in Figure 11. For ESI cases with an SOI of −70 and −50 °CA ATDC, ID is rather long due to the prevailing lower temperature environment. In these cases, a sequence of unevenly distributed, kinetically controlled lean combustion events with a variety in peak AHRRs is observed. On the other hand, for late single-injection cases with SOI timings of −30, −20, and −10 °CA ATDC, the ID, which decreases as the injection timing is retarded, is quite short because in-cylinder pressure and temperature increase significantly as the piston approaches TDC. For late injection cases, diffusion-controlled main combustion with a gradual increase and a subsequent decay in AHRR is preceded by the premixed phase of combustion, which manifests itself with high AHRR of short duration immediately after auto-ignition.

NO_x formation history according to °CA during compression and expansion strokes is presented in Figure 12. For early injection timings—that is, SOI timings of −70 °CA and −50 °CA ATDC—almost zero NO_x emissions are observed due to the low-temperature combustion. In late injection cases—that is, SOI timings of −30, −20, and −10 °CA ATDC—NO_x emissions are much higher due to predominantly diffusion-controlled high-temperature combustion. NO_x emissions increase gradually, as both the initial peak value of AHRR and the duration of the premixed phase of diffusion-controlled combustion increase with advancing SOI. The premixed flame-zone temperature reaches a peak at an SOI of approximately −30 °CA ATDC and then rapidly decreases due to decreasing initial AHRR peak and its duration as SOI timing is advanced further.

The effect of SOI timing on soot formation and oxidation according to °CA during the compression and expansion strokes is presented in Figure 13. As SOI is delayed, soot production rate increases, and the soot-formation period decreases. The soot-formation period is significantly longer for PCCI combustion compared to diffusion-controlled combustion due to slower PAH formation caused by the decreased reaction rate of partially mixed fuel in an environment with oxygen deficiency as the fuel film wetting the cylinder liner evaporates. However, in diffusion-controlled combustion, soot forms inside the diffusion flame from its precursors, PAHs, ahead of the premixed rich combustion zone, downstream of the leading region of the spray, at a significantly higher reaction rate. On the other hand, as SOI is delayed, soot oxidation rate increases, leading to a decreasing trend in the soot oxidation period for both PCCI and diffusion-controlled combustion. In PCCI combustion, soot production persists due to wall wetting, and soot oxidation is rather limited because of low in-cylinder temperatures [64]. However, in diffusion-controlled combustion, soot production occurs in the zone enclosed by the diffusion flame front, where higher temperatures, above 1600 K, prevail as compared to that of the rich-premixed combustion zone, 825 K, and low oxygen concentration conditions exist. This results in high soot-production rates during the injection, which leads to peaks in soot mass as soot oxidation begins. Then, soot concentration is dramatically reduced by the oxidation process, which occurs in an environment with excessive oxygen at significantly higher temperatures, e.g., 2700 K, [9,57,65].

The effect of injection timing on the evolution of CO in the combustion chamber according to °CA during the compression and expansion strokes is presented in Figure 14.

Production and oxidation rate of CO in the diffusion-controlled combustion cases are higher as compared to those for the PCCI cases. In diffusion-controlled combustion cases, a minor amount of CO left in the combustion chamber is discharged to the exhaust manifold during the exhaust stroke. Engine-out emissions of NO_x, soot, and CO on the parts per million (ppm) scale according to SOI timing are listed in

Table 5. Engine-out emissions.

SOI [°CA BTDC]	70	50	30	20	10
NOx [ppm]	4.0	54.8	3533.3	2368.6	1336.7
Soot [ppm]	147.8	206.8	20.5	13.6	17.1
CO [ppm]	134.3	935.0	22.4	12.6	16.4

.

CE and IMEP according to SOI timing are presented in Figure 15. As SOI timing is advanced, both CE and IMEP decrease due to incomplete combustion of the fuel in a lower-pressure and -temperature environment, impingement of the sprays on the piston bowl and/or cylinder liner, and displacement of the center of combustion away from TDC. For early injection cases, CE and IMEP are significantly lower than those of CDC. On the other hand, in the late injection timing cases, the sprays do not impinge on the cylinder liner but directly enter the piston crown as the droplets break up and evaporate prior to diffusion-controlled combustion in higher-pressure and -temperature gas mixture. Because of higher temperatures in the environment, fuel droplets evaporate and mix rapidly with air. FVAM and rich-premixed combustion phase products are oxidized in high-temperature lifted diffusion flames surrounding incomplete combustion products. The highest CE and IMEP are obtained for the SOI of −10 °CA ATDC case.

For various SOI timings, evolution of sprays according to °CA after SOI in top and oblique side views are presented in Figure 16.

The effect of SOI timing on the history of in-cylinder liquid fuel mass content according to °CA ATDC is presented in Figure 17. For early injection timing cases with SOI timings of −70 and −50 °CA ATDC, the liquid spray completely evaporates before 20 °CA BTDC. In −70 and −50 °CA ATDC SOI timing cases, fuel droplets exist in the combustion chamber for 48 °CA and 28 °CA, respectively. On the other hand, for retarded injection timings, SOI timings of −30, −20, −10 °CA ATDC fuel spray completely evaporates within about 8 °CA after SOI.

STPL in the radial direction according to °CA after SOI for various SOI timings is presented in Figure 18. For the early injection cases, STPLs are higher than those for the late injection cases due to lower ambient pressure and density in the cylinder. For early injection timing cases, that is, SOI timings of −70 °CA and −50 °CA ATDC, fuel sprays impinge directly onto the cylinder liner at 4.4 and 4.9 °CA after SOI, respectively, and form wall films which spread on the liner. For an SOI of −70 °CA ATDC, subsequent spray droplets impinging on the liquid film that has formed on the cylinder liner partially splash back into the squish volume, as shown in Figure 16a. Then, splashing droplets of neighboring sprays interact with each other in the squish volume located between the piston top and cylinder head, while liquid films spread on both the squish area on top of the piston and cylinder head. Most of the droplets and liquid films evaporate before the start of auto-ignition at −25.8 °CA ATDC. On the other hand, for an SOI of −50 °CA ATDC, fuel films formed on the cylinder liner and piston top evaporate completely, and a minor number of droplets exist near the injector nozzles when the mixture auto-ignites at −25.3 °CA ATDC.

On the other hand, for the late SOI timings, SOI of −30, −20, −10 °CA ATDC, sprays do not interact with the cylinder liner. In late injection cases, STPL suddenly begins to drop with SOIgn. During combustion, evaporation rates of the fuel droplets increase rapidly as the gas temperature rises significantly, which decreases STPL.

For the late SOI timings of −30 °CA, −20 °CA, and −10 °CA ATDC, due to the higher- temperature and -pressure environment, the sprays entering the piston bowl completely evaporate within, respectively, 8 °CA, 7 °CA, and 6 °CA after SOI.

Fuel spray penetration into the combustion chamber and its interactions with cylinder liner, cylinder head, squish, and crevice areas, as well as the FVAM formation process, are visualized by the time histories of the RER contour plot according to °CA after SOI for various SOI timings in Figure 19. For early SOI of −70 °CA ATDC and −50 °CA ATDC, sprays directly impinge on the cylinder liner and accumulate in the crevice volume, which causes problems such as soot in oil and excessive liquid hydrocarbon build-up in the combustion chamber. Calculated time histories of RER distribution in the piston bowl, crevice, and squish region for various early SOI timings are quite like fuel mass fraction distribution results of a computational study which are validated via fuel-tracer PLIF intensity measurements [66]. On the other hand, for late injection timings, fuel vapor is capable of mixing with the air both inside the piston bowl and near the soot-in-oil rim region. Therefore, impingement of the fuel spray on the cylinder liner is avoided in the late injection timing cases. For SOI of −10 °CA ATDC, fuel vapor jet impinges on the stepped lip of the piston and splits into two branches, one of which enters the piston bowl, and the other enters the squish volume, while diffusion combustion proceeds inside the piston bowl.

Histories of the temperature contour plot according to °CA after SOIgn for various SOI timings are shown in Figure 20. For early SOI timings of −70 °CA and −50 °CA ATDC, premixed combustion starts near the valves and cylinder liner and within the crevice volume, where the RER is close to stoichiometric conditions. The premixed flame temperature in the burning zone is between 1250 K and 1750 K, which is much lower than those of the diffusion flames that form in late injection cases. Some of the splashed fuel droplets from the cylinder liner increase the RER in the squish region as they evaporate. Thus, in this zone, the flame temperature increases above 2000 K. For the late SOI timings of −30 °CA, −20 °CA, and −10 °CA ATDC, combustion occurs inside the piston bowl and above the soot-in-oil region. For SOI of −30 °CA ATDC, a significant amount of the fuel penetrates above the soot-in-oil rim region, and a high-temperature diffusion flame is formed between the soot-in-oil rim and cylinder head regions. For the late SOI timings of −20 °CA and −10 °CA ATDC, evaporated fuel uniformly penetrates both the piston bowl and the soot-in-oil region. In retarded injection timings, diffusion-controlled combustion occurs, and flame temperature varies between 2000 K and 3500 K depending on the local RER.

Time histories of the NO_x mass fraction contour plot according to °CA after SOIgn for various SOI timings are presented in Figure 21. For early SOI timings of −70 °CA and −50 °CA BTDC, very low amounts of NO_x are produced in the combustion chamber due to the low-temperature combustion of a homogeneous lean mixture. NO_x is formed in two hot spots that have sufficiently high temperatures. On the other hand, for the late SOI timings of −30 °CA, −20 °CA, and −10 °CA ATDC, significant amounts of NO_x are formed due to diffusion-controlled high-temperature combustion.

Time histories of the soot mass fraction contour plot according to °CA after SOIgn for various SOI timings are shown in Figure 22. For early SOI timings of −70 °CA and −50 °CA ATDC, even though soot formation is lower, soot emissions are significantly higher than those for the late injection cases due to much lower soot oxidation rates. Furthermore, soot is more uniformly distributed within the combustion chamber compared to those produced by diffusion flames formed by late injections. For the late SOI timings of −30 °CA, −20 °CA, and −10 °CA ATDC, significant amounts of soot are formed downstream of the rich-premixed phase of combustion, but due to much higher soot oxidation rates within the high-temperature diffusion flames, soot emissions are significantly lower compared to the PCCI cases.

Time histories of the CO mass fraction contour plot according to °CA after SOIgn for various SOI timings are presented in Figure 23. CO formation is a transient step, and CO oxidizes to CO₂ in high-temperature regions to complete combustion. For early SOI timings of −70 °CA and −50 °CA ATDC, CO is produced at temperatures between 1200 K and 1500 K, and its oxidation to CO₂ is completed in regions where temperatures are above 2000 K. For late SOI timings of −30 °CA, −20 °CA, and −10 °CA ATDC, CO formed by the premixed phase of combustion of the rich mixture downstream of the sprays is oxidized to CO₂ within the high-temperature diffusion flames to complete the combustion. For late SOI timings of −30 °CA, −20 °CA, and −10 °CA ATDC, CO is produced inside the FVAM, where RER is higher than unity. Then, the oxidation step from CO to CO₂ is completed at high diffusion flame temperatures, which are between 2000 K and 3500 K depending on the SOI timing. The effects of injection timing on engine-out emissions, spray, and combustion characteristics are listed in

Table 6. Effects of SOI timing on spray, combustion characteristics, and engine-out emissions.

SOI [°CA BTDC]	70	50	30	20	10
Spray Penetration Length [mm]	>65	>65	44.8	42.6	29.8
Ignition Delay [°CA]	44.2	24.7	8.4	6.9	4.1
Start of Ignition [°CA]	25.8	25.3	21.6	13.1	5.9
Combustion Efficiency [%]	58.5	84.3	98.9	99.8	99.9
IMEP [bar]	1.99	2.34	3.23	3.64	3.89
NOx [ppm]	4.04	54.8	3533.3	2368.6	1336.7
Soot [ppm]	147.8	206.8	20.5	13.6	17.1
CO [ppm]	134.3	935.0	22.4	12.6	16.4

.

3.3. Effect of SIA on Engine Performance and Emissions

As SOI timing is advanced, due to lower in-cylinder pressure and density, both STPL and the Sauter mean diameter of droplets increase, and the spray conical angle decreases. Furthermore, evaporation rates of fuel droplets decrease due to the lower temperature of the environment. Therefore, fuel vapor-air mixing is incomplete. Furthermore, for a given engine geometry, if fuel is injected before a critical SOI which depends on SIA, injection pressure and in-cylinder charge motion, direct impingement of fuel sprays on the cylinder liner or after having been bounced by the piston squish area, indirect impingement of fuel droplets on the liner may cause wetted wall problem at auto-ignition time. According to CFD simulations, for an injector inclusion angle of 146° and an injection pressure of 800 bar, if the SOI is advanced beyond −30 °CA ATDC, due to the lack of complete evaporation in a lower-temperature environment prior to interaction, at earlier stages of the compression stroke, diesel fuel sprays impinge on and wet the cylinder liner. Therefore, the ESI strategy causes wet wall problems in PCCI combustion.

On the other hand, it is possible to eliminate the wall-wetting problems for ESI by reducing the SIA, which enables the completion of FVAM formation without fuel spray interaction with the walls for longer STPLs. Therefore, for early injection cases which increase STPL, it is possible to reduce or totally avoid the wall-wetting issues by utilizing an injector with reduced SIA [17].

In this study, for ESI strategy, the effects of SIA on spray tip penetration, the interaction of sprays with cylinder walls, and in-cylinder combustible mixture formation, as well as DI diesel engine performance and emissions, were analyzed via CFD simulations. In the case of ESI with an SOI timing of −50 °CA ATDC, the effects of wetting of cylinder walls by sprays in a DI diesel engine operating in PCCI combustion mode are analyzed for various SIAs of 146°, 140°, 130°, 120°, 100°, and 80°. This injection timing is selected because of significantly reduced NO_x emissions compared to late-injection cases due to lower combustion temperatures, and higher CE compared to that of the SOI timing of −70 °CA ATDC case due to less fuel impingement on the cylinder liner. At ESI with an SOI timing of −50 °CA ATDC, interactions of fuel spray with the cylinder liner and piston bowl for various SIAs are illustrated in Figure 24.

At ESI with an SOI of −50 °CA ATDC, in-cylinder pressure histories for various SIAs are presented in Figure 25. The highest in-cylinder peak pressure is obtained for an SIA of 100°, which is followed by 80° and 120°. On the other hand, the lowest in-cylinder peak pressures are obtained for injectors with wider SIAs of 140°, 130°, and 146°, in ascending order, with minor differences.

For an SOI at −50 °CA ATDC, the effect of SIA on AHRR history is presented in Figure 26. It is observed that for an SOI of −50 °CA ATDC, ID is insensitive to SIA, because for all SIAs considered in this study, auto-ignition starts at approximately −26 °CA ATDC. On the other hand, the AHRR history for each SIA differs considerably during the combustion process: the SIA of 80° results in the highest peak AHRR at −20 °CA ATDC. The second highest peak AHRR is observed at −16.5 °CA ATDC for the SIA of 100°.

For an SOI of −50 °CA ATDC, the effect of SIA on the liquid-mass content history of the spray is presented in Figure 27. The ESI of a narrow inclusion angle fuel spray targeting the stepped lip of the piston improves the mixing of the fuel vapor with air by splitting the spray into two branches. Therefore, as the fuel droplets evaporate, the lowest amount of liquid mass in a fuel spray exists for a 100° SIA, followed by an 80° case. Furthermore, as the in-cylinder air motion in a planar section through the cylinder axis is considered, clockwise and counter-clockwise vortices in the piston bowl and in the squish region, respectively, help the dispersion of droplets in both branches of the split spray, thereby significantly improving the FVAM-formation process. The two highest AHRR peaks occur for SIAs of 100° and 80° because the liquid contents of these sprays are the least. On the other hand, the lowest main AHRR occurs for an SIA of 130° since the liquid content of the spray is the highest due to cylinder liner wetting.

For an SOI of −50 °CA, the effect of SIA on the history of STPL in the radial direction is presented in Figure 28. STPL decreases as the SIA becomes smaller. Sprays with an SIA bigger than or equal to 120° impinge on the cylinder liner, whereas those with an SIA of 80° and 100° interact with the curved portion and stepped lip of the piston bowl, respectively.

For various SIAs of 140°, 130°, 120°, 100°, and 80°, the evolution of spray penetration in oblique side and top views, as well as RER contour plots in a plane through the axes of the cylinder and a spray, in the instants of 5, 10, 15, and 20 °CA after SOI, are presented in Figure 29, Figure 30, Figure 31, Figure 32 and Figure 33, respectively. For an SIA of 140°, the fuel sprays impinge on the cylinder liner, forming a rich FVAM near the liner, which becomes lean before auto-ignition. On the other hand, the combustion of rich FVAM that has formed in the crevice in a low-temperature environment increases CO emissions. For an SIA of 130° case, sprays impinge on the top flat face of the piston, forming a liquid film which then spreads on the cylinder liner. Part of the liquid film rises toward the cylinder head, while the rest penetrates the crevice. Later, rich-mixture zones form below the cylinder head near the cylinder liner and in the crevice as the liquid films evaporate due to the rising combustion chamber temperature. On the other hand, a very lean mixture, which does not auto-ignite, is formed near the soot-in-oil rim region due to partial evaporation of the droplets of the main sprays before they interact with the piston. For an SIA of 120°, the fuel sprays impinge on the piston squish area and split into two streams, which spread toward the piston bowl and the cylinder liner. The upper portion partly rises toward the cylinder head along the liner and partly drains down to the crevice between the cylinder liner and piston after hitting the liner, while the lower branch enters the piston bowl, where it mixes and reacts with swirling turbulent air. In the case of an SIA of 100°, the spray impinges on the tip of the soot-in-oil rim, splitting into two branches. Part of the upper stream impinges on a valve after being deflected by the vertical face between the soot-in-oil rim and the piston squish area and partly spills over the piston squish area, while the rest enters the piston bowl, flows in a tumbling motion in a clockwise direction in the re-entrant region, and mixes with the air inside the piston bowl. Spray-targeting the edge of the piston bowl promotes mixing prior to auto-ignition and reduces CO emissions [17]. Moreover, narrow SIA results in shorter STPL, which reduces soot emissions due to enhanced FVAM formation. In the case of 80° SIA, the fuel spray impinges on the curved part of the piston bowl and splits into two branches, both of which remain inside the piston bowl due to the clockwise tumbling motion of the air, forming a large, rich FVAM zone. Therefore, spray impingement on the cylinder liner is prevented by narrow SIAs for ESI strategy.

For various SIAs of 140°, 130°, 120°, 100°, and 80°, in-cylinder temperature and NO_x and soot emission mass fraction contour plots, in a planar section through the cylinder axis and a spray axis, in the instants of 5, 10, 15, and 20 °CA after SOIgn are presented in Figure 34, Figure 35, Figure 36, Figure 37 and Figure 38, respectively.

For an SIA of 140°, a negligible amount of NO_x is formed in the squish region, which is only because the premixed combustion of the lean FVAM is confined to a small fraction of the combustion chamber near the cylinder head and valves, where the temperatures range between 1750 K and 2250 K. Therefore, the smallest NO_x emissions occur for an SIA of 140° among the various SIA cases considered in this study.

However, after the impingement of sprays on the cylinder liner, a considerable amount of soot is formed not only in the crevice but also near the cylinder head due to the combustion of rich mixtures formed by the evaporation of both the liquid film and splashed droplets, respectively. Local temperatures of the flames reach values as high as 2250 K immediately after auto-ignition. A significant amount of soot is emitted by the engine due to small oxidation rates of the soot particles transported to the zone with a mean temperature of 1370 K outside the lean-premixed flames of limited extent and duration, which form under the valves and later expand in the piston bowl.

In the case of an SIA of 130°, the lean FVAM under the valves auto-ignites first, which is followed by the formation of premixed flames in the richer mixture region under the cylinder head near the cylinder liner. Both premixed flames propagate toward the cylinder axis due to the counter-clockwise tumbling motion of the mixture in the upper part of the combustion chamber, while they simultaneously expand toward the piston bowl. NO_x emissions increase as compared to those of a case of an SIA of 140° due to both higher temperatures ranging between 1750 K and 2750 K and a bigger extent of premixed flames. Soot production and oxidation locations are like those for the case of an SIA of 140°. A portion of the soot formed in rich-mixture zones, i.e., in the crevice, near the cylinder liner, and beneath the cylinder head, is transported by in-cylinder charge motion. Some soot particles are oxidized by propagating premixed flames that have temperatures above 1750 K. On the other hand, soot particles transported to the zone outside lean-premixed flames cannot be oxidized because the temperatures are below 1500 K. The most severe soot emissions occur for an SIA of 130° due to both the biggest soot production as the highest amount of fuel accumulates in the crevice and the smallest amount of soot oxidation since lean-premixed flames remain in the upper portion of the combustion chamber.

In the case of an SIA of 120°, evaporated fuel jets directly impinge on the soot-in-oil rim region, then split into two wall jets. The upper stream flows on top of the piston under the influence of the counter-clockwise tumble in the squish and climb on the liner up to the cylinder head, while the lower branches enter the piston bowl and tumble clockwise due to the air vortex in the piston bowl, and both streams simultaneously become leaner as they spread in the tangential direction and mix with the air. As the combustion chamber temperature rises further due to compression, first, an extremely lean mixture below the valves auto-ignites, and the premixed flame spreads toward the higher RER zone inside the piston bowl, eventually filling a significant portion of it. As compared to the case of an SIA of 140°, lean combustion temperatures ranging between 1750 K and 3000 K significantly increase NO_x emissions.

On the other hand, for the 120° SIA, the combustion of small, rich-mixture zones formed around the valve edges and in the crevice, respectively, by the evaporation of splashed fuel droplets and drained fuel films from the cylinder liner increases soot production. Furthermore, oxidation rates of the soot particles transported to the zone around the cylinder axis and the squish region are rather low because local temperatures range between 1200 K and 1500 K. Soot emissions are almost halved as compared to the case of an SIA of 140° due to lower soot formation in the crevice and higher soot oxidation in the much larger lean-premixed combustion zone within the piston bowl.

In the case of an SIA of 100°, auto-ignition starts in a larger zone under the valves, and premixed flame spreads rapidly, filling the upper portion of the combustion chamber and most of the piston bowl. The most evenly distributed flame temperatures are obtained due to enhanced FVAM homogeneity as the impinging sprays split and the upper streams interact with the valves and the soot-in-oil rim while the lower streams tumble. The premixed flame temperature changes between 2000 K and 3000 K, and NO_x emissions are slightly lower compared to the SIA of 120° case due to a lower average flame temperature.

In the case of an SIA of 100°, soot is formed in locally rich zones that have RERs of approximately 2.0 under the valves and inside the piston bowl. Because of the improved mixture homogeneity due to the absence of local rich zones in the crevice region and around the valve edges, soot formation decreases, while premixed flame temperatures range between 1750 K and 3000 K, and its extent increases. As compared to the bigger SIA cases, engine-out soot emissions significantly decrease, as shown in Figure 37, since soot oxidation considerably increases with both the temperature and size of the premixed flame.

In the case of an SIA of 80°, a rich FVAM zone occupying a significantly bigger volume compared to the previous cases is formed in the piston bowl after the spray impinges on the curved portion of it. However, because the splashed droplets are unable to penetrate toward the top of the combustion chamber due to the clockwise tumbling motion of the air inside the piston bowl, the mixture outside the piston bowl is extremely lean. Therefore, at first, rich FVAM near the curved portion of the piston bowl auto-ignites, and then, the rich-premixed flame propagates along the inclined straight portion of the piston bowl toward the combustion chamber axis. NO_x emissions are significantly elevated due to the much bigger size of the rich-premixed flame, with a high temperature range of 1750 K and 3000 K, than those of the lean-premixed combustion cases obtained for bigger SIAs. On the other hand, soot formed at the beginning of rich combustion near the curved portion of the piston bowl is effectively oxidized by the high-temperature flame as it propagates toward the cylinder axis, which results in the lowest soot emissions among the various SIA cases considered.

The tradeoff between engine-out NO_x and soot emissions according to SIA is presented in Figure 39. In ESI with wide SIAs, very low engine-out NO_x emissions are obtained, while soot emissions are significant. For wide SIAs, combustion takes place near the cylinder liner due to spray impingement on the liner, and the flame tends to be quenched due to heat transfer to the water jacket. Therefore, the combustion of the rich mixture at relatively lower temperatures causes excessive soot production. For narrower SIAs, more uniform FVAM can be obtained in the combustion chamber by spray-targeting the stepped lip of the piston bowl, which leads to complete high-temperature combustion of the premixed lean mixture. Therefore, soot emissions decrease significantly, while NO_x emissions increase. For the narrowest SIA considered in this study, due to high soot oxidation and high thermal NO_x formation rates, respectively, at elevated temperatures of rich combustion in the piston bowl, engine-out soot emissions are the lowest, but NO_x emissions are excessive.

Engine-out CO emissions, according to SIA, are presented in Figure 40. For the wide SIAs, incomplete combustion near wetted cylinder walls causes higher CO emissions [67]. For ESI, if spray is targeted at the stepped lip of the piston bowl by reducing SIA, the combustion of more uniform lean FVAM filling the piston bowl and the squish region results in the lowest CO emissions. Finally, in the narrowest SIA cases, high-temperature combustion in the piston bowl promotes complete combustion. However, minor CO emissions occur due to local oxygen deficiency.

CE and IMEP for various SIAs are presented in Figure 41. For wide SIA cases of 146°, 140°, and 130°, CEs are 84.3%, 78.9%, and 81.5%, respectively; while for narrow SIA cases of 120°, 100°, and 80°, CEs are 91.3%, 97.8%, and 96.3%, respectively. CE severely deteriorates due to incomplete combustion of the inadequately mixed fuel vapor with air at low temperatures for wide SIA cases. The lowest CE of 78.9% occurs for the SIA of 140° due to incomplete rich combustion of the poorly mixed fuel vapor with air in both the squish region and crevice at flame temperatures ranging between 1500 K and 2000 K after spray impingement on the cylinder liner.

For ESI with an SOI of −50 °CA ATDC, the highest CE of 97.8% and the highest peak in-cylinder pressure of 107 bar are obtained in the 100° SIA case due to high-temperature lean-premixed combustion of a high-quality FVAM. Flame temperatures ranging between 2250 K and 3000 K oxidize hydrocarbons and soot particles very effectively in the piston bowl and squish region. This is accomplished by spray-targeting the tip of the stepped lip of the piston bowl so that a sufficiently uniform FVAM is formed in the combustion chamber, and penetration of liquid fuel into the crevice is avoided.

Variations of IMEP and CE according to SIA have a similar trend. Therefore, the peak IMEP value of 3.37 bar and the best CE of 97.8% are obtained in the case of an SIA of 100°. This is due to spray-targeting at the tip of the stepped lip of the piston bowl so that spray interaction with the cylinder liner is avoided, which results in the best air to fuel mixing quality. On the other hand, an SIA of 140° results in the lowest IMEP of 2.26 bar due to spray impingement on the cylinder liner causing both high engine-out soot emissions and incomplete combustion because of insufficient air–fuel mixing. The effect of the spray inclusion angle on spray penetration length, combustion efficiency, performance, and engine-out emissions characteristics are listed in

Table 7. Effect of spray inclusion angle on spray, combustion characteristics, and engine-out emissions.

Spray Inclusion Angle [°]	146	140	130	120	100	80
Spray Penetration Length [mm]	>65	>65	>65	63.3	42.4	36.2
Combustion Efficiency [%]	84.3	78.8	81.5	91.3	97.8	96.3
IMEP [bar]	2.34	2.26	2.42	2.95	3.37	2.95
NOx [ppm]	54	34	345	575	498	2227
Soot [ppm]	206	403	447	210	33	33
CO [ppm]	935.0	2266	2620	770	2	8

.

3.4. Discussion

According to the performance and engine-output emissions calculated via CFD simulations in this study of a single-cylinder HDDI diesel research engine with a stepped-lip piston bowl operating at a partial load, the following conclusions are obtained:

In a single-injection strategy, for a given spray-inclusion angle, engine operation is optimal with near-zero NO_x emissions at a certain SOI timing.

In the PCCI operation, optimum net-indicated efficiency, indicated specific fuel consumption, and NO_x, soot, and CO emissions are achieved by preventing or minimizing the interaction of the fuel spray with the cylinder walls at a spray inclusion angle that depends on the SOI timing.

Ideally, the design of an injection system that can adaptively control the injected fuel amount, the SOI timing, and the spray inclusion angle according to the engine load can provide the best engine performance and engine-out emissions in both the CDC and PCCI operating ranges.

IMEP and CE can be improved, and soot and CO emissions can be significantly reduced by selecting appropriate injection timing for a given SIA and piston bowl geometry.

By considering ISFC and Maximum Pressure Rise Rate (MPRR), as well as soot, NO_x, and CO emissions in part-load CDC operation, the HDDI diesel engine’s stepped lip piston bowl geometry, injector nozzle configuration, spray-targeting, and swirl ratio have already been optimized before, based on a computational study on a light-duty diesel engine [68]. Better CE and fuel economy can be obtained by a stepped lip piston bowl, which not only enables higher AHRR during the main part of the diffusion combustion by having better access to the air in the squish region but also increases thermal efficiency because of reduced heat losses [69,70]. Even though more soot is formed during diffusion combustion, splitting the fuel spray between the piston bowl and the region above the piston crown at the ratio of the two volumes by making it impinge on the stepped lip of the piston bowl improves lean mixture uniformity and, therefore, reduces soot formation and speeds up soot oxidation. Therefore, soot emissions decrease significantly at the expense of slightly increased NO_x emissions, which can be controlled by EGR, as compared to those of the baseline re-entrant piston bowl. SOI, STPL, piston bowl shape, and the location and contour of the stepped lip can be determined according to the above criterion by CFD simulations for a late single-injection having a chosen wide SIA. Then, to minimize ISFC and emissions, the optimum SIA can be determined via CFD simulations of diffusion combustion at part load [71]. As an alternative to piston bowl geometry, a dimpled stepped lip piston bowl could have been selected. On the other hand, the piston bowl shape can be optimized to improve CE and reduce emissions for early injection with a narrow SIA at low load operation. However, for such a piston bowl geometry, CE and emissions deteriorate considerably if a late single-injection with narrow SIA is applied in high-load operation. Therefore, for a chosen piston bowl geometry, the design of a fuel-injection system capable of adaptively controlling SIA in accordance with the injection timing(s) is necessary to obtain optimal CE and engine-out emissions at all engine speeds and loads. To keep high CE and low engine-out emissions for all operating conditions, as the engine load increases, both SIA should be increased and injection timing should be delayed, to maintain spray-targeting at the stepped lip of the piston bowl, so that splashed droplets may disperse as evenly as possible both in the piston bowl and in the squish volume. However, to the best knowledge of the authors of this study, so far, an injection system capable of continuously decreasing both SIA and STPL as SOI is advanced without compromising atomization quality to avoid increased soot, and CO emissions in early injections have not been developed.

Furthermore, the effect of multiple injection timings/locations/fuels (pilot-main, pilot-main-post, etc./port-direct, direct with double injectors/fuels with various reactivities), variable valve timing/CR/EGR rate/air intake temperature/coolant temperature, water injection, exhaust gas trapping, or any combination of them on engine combustion performance and emissions for optimization over a wide range of load-speed pairs and for extension of the upper load limit of PCCI combustion can be studied by experimentally validated CFD simulations in the future.

4. Conclusions

In this study, the effects of SOI timing and SIA on the combustion and emission characteristics of a single-cylinder HDDI diesel research engine for PCCI mode of combustion were investigated via CFD simulations. The simulations were performed on a full-cycle 3D CFD model with a reduced chemical kinetic mechanism of surrogate fuel n-heptane at a 1100 rpm engine speed and 3-bar IMEP operating conditions. The results of this study can be summarized as follows:

• In single injection cases with 146° SIA at both −50 and −10 °CA ATDC SOI timings, the NO_x and CO emissions predicted via CFD are consistent with the measurements. However, predicted soot emissions differ significantly for both injection timings, which is amplified in the case of PCCI combustion following the early injection.. In the PCCI combustion case, soot emissions predicted by the Hiroyasu–NSC semi-empirical model are about twice as high as the measured value. Arrhenius pre-exponential factor for soot-formation kinetics (A_sf) and scaling factor of NSC soot oxidation model (A_so) in the soot-formation and oxidation rate equations, respectively, should be calibrated according to temperature so that a unified formulation of both diffusion-controlled and PCCI combustion can be obtained.
• For an SIA of 146°, spray tip penetration rate and in-cylinder peak liquid fuel mass content decrease as SOI is delayed. If SOI is varied between −70 and −10 °CA ATDC, ID decreases from 44.2 to 4.1 °CA, while CE and IMEP increase from 58.5% to 99.9% and 1.99 to 3.89 bar, respectively. The highest and lowest in-cylinder pressure peaks of 10.1 MPa and 8.8 MPa are obtained for SOI timings of −30 and −70 °CA ATDC, respectively.
• For an SIA of 146°, in delayed injection cases where the SOI is after −30 °CA ATDC, diffusion-controlled combustion occurs. For an SOI of −10 °CA ATDC, due to complete diffusion-controlled combustion, the highest CE and IMEP of 99.9% and 3.89 bars, and the lowest engine-out soot and CO emissions of 17 ppm and 16 ppm, all respectively, are obtained. On the other hand, for an SOI of −10 °CA ATDC, excessive NO_x emissions of 1136.7 ppm occur due to higher flame temperatures compared to those of the PCCI combustion cases.
• For an SIA of 146°, in the early injection cases with SOI timings of −50 °CA ATDC and −70 °CA ATDC, PCCI combustion IDs of 24.7 °CA and 44.2 °CA, respectively, are significantly higher than those of CDC, which range between 4.1 °CA and 8.4 °CA depending on the SOI. In diffusion-controlled combustion, NO_x emissions are 1336.7 ppm, 2368.6 ppm, and 3533.3 ppm for SOI timings of −10, −20, and −30 °CA ATDC, respectively. However, in PCCI cases with SOI timings of −50 °CA ATDC and −70 °CA ATDC, low-temperature combustion causes significantly decreased NO_x emissions of 54.8 ppm and 4.04 ppm, respectively. In PCCI combustion, soot emissions of 147.8 ppm and 206.8 ppm for SOI timings of −70 and −50 °CA ATDC, respectively, are significantly higher than those of diffusion-controlled combustion due to persistent soot production caused by wall-wetting and insignificant soot oxidation rates at lower in-cylinder temperatures. In PCCI combustion, despite lower CO production rates, due to their much lower oxidation rates at lower combustion temperatures, engine-out CO emission is 648.9 ppm for SOI timings of −70 °CA ATDC and 935.0 ppm for SOI timings of −50 °CA ATDC, which are significantly higher as compared to those for the diffusion-controlled combustion.
• For an ESI with an SOI of −50 °CA ATDC, the SIA sweep reveals that, due to high-temperature lean-premixed combustion of well-mixed FVAM, an injector with a narrow SIA of 100° results in the highest CE of 97.8%, and the lowest soot and CO emissions of 33.5 ppm and 2.2 ppm, respectively, and IMEP of 3.37 bars, with acceptable NO_x emissions of 6.5 × 10⁻⁶ kg/stroke. This is accomplished by spray-targeting the tip of the stepped lip of the piston bowl so that a lean mixture is formed in most of the combustion chamber, while spreading a slightly rich mixture within the piston bowl, in the squish region and beneath the valves.
• For an SOI of −50 °CA ATDC, sprays with wide SIAs of 146°, 140°, and 130° directly impinge on the cylinder liner, while the spray with an SIA of 120° interacts with the cylinder liner after impinging on the soot-in-oil rim. The lowest CE of 78.8% and the smallest IMEP of 2.26 bar are obtained for the case of an SIA of 140° due to poor mixture formation. As the influence of SIA on engine-out emissions are analyzed, the highest soot and CO emissions occur for an SIA of 130°, which are followed by those for an SIA of 140°. However, near-zero NO_x emissions, that is, 34.7 ppm, are observed for a 140° SIA due to low-temperature combustion of poor-quality fuel–air mixture.

Optimal CE and engine-out emissions for all operating conditions can be achieved via a fuel-injection system which adaptively controls SIA in accordance with the injection timing(s) required by engine load.