Search Results

The accurate prediction of engine performance maps can guide data-driven optimization of engine technologies to control fuel use and associated emissions. However, engine operational maps are scarcely reported in literature and often have missing data. Assessment of missing-data resilient algorithms in the context of engine data prediction could enable better processing of real-world driving cycles, where missing data is a more pervasive phenomenon. The goal of this study is, therefore, to determine the most effective technique to deal with missing data and employ it in prediction of engine performance characteristics. We assess the performance of two machine learning approaches, namely Artificial Neural Networks (ANNs) and the extreme tree boosting algorithm (XGBoost), in handling missing data.

The transportation industry has been scrutinized for its contribution towards the global greenhouse gas emissions over the years. While the automotive sector has been regulated by strict emission legislation globally, the emissions from marine transportation have been largely neglected. However, during the past decade, the international maritime organization focused on ways to lower the emission intensity of the marine sector by introducing several legislations. This sets limits on the emissions of different oxides of carbon, nitrogen and sulphur, which are emitted in large amounts from heavy fuel oil (HFO) combustion (the primary fuel for the marine sector). A 40% and 70% reduction per transport work compared to the levels of 2008 is set as target for CO2 emission for 2030 and 2050, respectively. To meet these targets, commonly, methanol, as a low-carbon fuel, and ammonia, as a zero-carbon fuel, are considered.

Investigating combustion characteristics of oxygenated gasoline and gasoline blended ethanol is a subject of recent interest. The non-linearity in the interaction of fuel components in the oxygenated gasoline can be studied by developing chemical kinetics of relevant surrogate of fewer components. This work proposes a new reduced four-component (isooctane, heptane, toluene, and ethanol) oxygenated gasoline surrogate mechanism consisting of 67 species and 325 reactions, applicable for dynamic CFD applications in engine combustion and sprays. The model introduces the addition of eight C1-C3 species into the previous model (Li et al; 2019) followed by extensive tuning of reaction rate constants of C7 - C8 chemistry. The current mechanism delivers excellent prediction capabilities in comprehensive combustion applications with an improved performance in lean conditions.

Strict regulations on air pollution motivates clean combustion research for fossil fuels. To numerically mimic real gasoline fuel reactivity, surrogates are proposed to facilitate advanced engine design and predict emissions by chemical kinetic modelling. However, chemical kinetic models could not accurately predict non-regular emissions, e.g. aldehydes, ketones and unsaturated hydrocarbons, which are important air pollutants. In this work, we propose to use machine-learning algorithms to achieve better predictions. Combustion chemistry of fuels constituting of 10 neat fuels, 6 primary reference fuels (PRF) and 6 FGX surrogates were tested in a jet stirred reactor. Experimental data were collected in the same setup to maintain data uniformity and consistency under following conditions: residence time at 1.0 second, fuel concentration at 0.25%, equivalence ratio at 1.0, and temperature range from 750 to 1100K.

The development of highly boosted and high compression spark-ignition engines with enhanced thermal efficiencies is primarily limited by knock and super-knock. Super-knock is an excessively high intensity knock which has been related to a developing detonation process. This study investigates the knocking tendency of different gasoline surrogate fuels with varying research octane numbers (RON), octane sensitivity (S) and composition. The ξ/ɛ diagram with an enclosed detonation peninsula is used to assess the knocking tendency of different fuels. The diagram plots ξ, the ratio of acoustic to auto-ignitive velocity, against ɛ, the ratio of the transit time of an acoustic wave through a hot spot, to the heat release time (τe). Constant volume simulations of auto-ignition delay times (τi) and excitation times (τe) obtained from chemical kinetic calculations, enable calculations of ξ and ɛ.

Over the years, spark-ignition engine operation has changed significantly, driven by many factors including changes in operating conditions. The variation in operating conditions impacts the state of the end-gas, and therefore, its auto-ignition. This can be quantified in terms of K-factor, which weighs the relative contribution of Research Octane Number (RON) and Motor Octane Number (MON) to knocking tendency at any operating condition. The current study investigates the fuel requirements when operating an engine at increasing intake air pressures. A model engine was operated at varying intake air pressure in GT-Power software, from naturally aspirated intake air to heavily boosted intake air pressure of 4 bar absolute. The pressure-temperature information from the GT-Power model was used to calculate ignition delay times of the unburnt end-gas composed of a sensitive and a non-sensitive fuel in ChemKin software.

Pre-chamber ignition systems are considered as an effective technique to achieve an ultra-lean burn combustion. Hot combustion products and flames generated from pre-chamber combustion create high-speed turbulent jets, which ignite an ultra-lean mixture in the main chamber. This turbulent ignition can be classified as a jet and flame ignition process with thermal and chemical kinetic effects imposed on the main chamber by the pre-chamber. The purpose of this paper is to investigate the chemical effects of pre-chamber combustion products on main chamber ignition performance over a range of operating conditions in pre-chamber. A zero-dimensional pre-chamber combustion model was developed using CHEMKIN-PRO software. By varying the equivalence ratio of reactants, simulation results indicated that the pre-chamber generated more active radicals when burning around stoichiometric conditions but more low-carbon species when burning in rich conditions.

The utilization of H2 to catalytically treat NO emissions under lean-burn engine exhaust conditions was studied on Pt- and Pd-containing catalysts supported on CeO2 and MgO. The catalytic performance was examined using a fixed-bed reactor whose dry effluent gas stream was analyzed by an online FTIR analyzer. The catalysts NO conversion and N2 selectivity were measured in the range of 125-3000C with a feed gas composition of 0.05%NO/1%H2/10%O2/N2. The CeO2-based catalysts exhibited higher NO conversion, and the most effective catalyst was Pd/CeO2, with a conversion of 67% and selectivity of 70% near 2300C. The prepared solids were characterized using different techniques (BET, ICP-OES, CO pulse chemisorption, STEM, EELS and EDS) to correlate the structural and morphological properties of the metallic phase and the support with the catalytic activity. CeO2 is a more effective support as it yields higher metal dispersion and better facilitates the reduction of the Pt and Pd catalysts.

This research evaluates the capability of data-science models to classify the combustion events in Cooperative Fuel Research Engine (CFR) operated under Homogeneous Charge Compression Ignition (HCCI) conditions. A total of 10,395 experimental data from the CFR engine at the University of Michigan (UM), operated under different input conditions for 15 different fuel blends, were utilized for the study. The combustion events happening under HCCI conditions in the CFR engine are classified into four different modes depending on the combustion phasing and cyclic variability (COVimep). The classes are; no ignition/high COVimep, operable combustion, high MPRR, and early CA50. Two machine learning (ML) models, K-nearest neighbors (KNN) and Support Vector Machines (SVM), are compared for their classification capabilities of combustion events. Seven conditions are used as the input features for the ML models viz.

The present study experimentally examines the low-temperature autoignition area of isooctane within the in-cylinder pressure-in-cylinder temperature map. Experiments were run with the help of a Cooperative Fuel Research (CFR) engine. The boundaries of this engine were extended so that experiments could be performed outside the domain delimited by research octane number (RON) and motor octane number (MON) traces. Since homogeneous charge compression ignition (HCCI) combustion is governed by kinetics, the rotation speed for all the experiments was set at 600 rpm to allow time for low-temperature heat release (LTHR). All the other parameters (intake pressure, intake temperature, compression ratio, and equivalence ratio) were scanned, such as the occurrence of isooctane combustion. The principal results showed that LTHR for isooctane occurs effortlessly under high intake pressure (1.3 bar) and low intake temperature (25 °C).

Numerical simulations were conducted to systematically assess the effects of different spray models on the ignition delay predictions and compared with experimental measurements obtained at the KAUST ignition quality tester (IQT) facility. The influence of physical properties and chemical kinetics over the ignition delay time is also investigated. The IQT experiments provided the pressure traces as the main observables, which are not sufficient to obtain a detailed understanding of physical (breakup, evaporation) and chemical (reactivity) processes associated with auto-ignition. A three-dimensional computational fluid dynamics (CFD) code, CONVERGE™, was used to capture the detailed fluid/spray dynamics and chemical characteristics within the IQT configuration. The Reynolds-averaged Navier-Stokes (RANS) turbulence with multi-zone chemistry sub-models was adopted with a reduced chemical kinetic mechanism for n-heptane and iso-octane.

Alcohols are potential blending agents for diesel that can be effectively used in compression ignition engines. This work investigates the use of n-butanol as a blending component for diesel fuel using experiments and simulations. Dodecane was selected as a surrogate for diesel fuel and various concentrations of n-butanol were added to study ignition characteristics. Ignition delay times for different n-butanol/dodecane blends were measured using the ignition quality tester at KAUST (KR-IQT). The experiments were conducted at pressure of 21 and 18 bar, temperature ranging from 703-843 K and global equivalence ratio of 0.85. A skeletal mechanism for n-dodecane and n-butanol blends with 203 species was developed for numerical simulations. The mechanism was developed by combining n-dodecane skeletal mechanism containing 106 species and a detailed mechanism for all the butanol isomers.

Toluene primary reference fuel (TPRF) (mixture of toluene, iso-octane and heptane) is a suitable surrogate to represent a wide spectrum of real fuels with varying octane sensitivity. Investigating different surrogates in engine simulations is a prerequisite to identify the best matching mixture. However, running 3D engine simulations using detailed models is currently impossible and reduction of detailed models is essential. This work presents an AramcoMech reduced kinetic model developed at King Abdullah University of Science and Technology (KAUST) for simulating complex TPRF surrogate blends. A semi-decoupling approach was used together with species and reaction lumping to obtain a reduced kinetic model. The model was widely validated against experimental data including shock tube ignition delay times and premixed laminar flame speeds. Finally, the model was utilized to simulate the combustion of a low reactivity gasoline fuel under partially premixed combustion conditions.

Gasoline compression ignition (GCI) engines are becoming more popular alternative for conventional spark engines to harvest the advantage of high volatility. Recent experimental study demonstrated that high reactivity gasoline fuel can be operated in a conventional mixing controlled combustion mode producing lower soot emissions than that of diesel fuel under similar efficiency and NOx level [1]. Therefore, there is much interest in using gasoline-like fuels in compression ignition engines. In order to improve the fidelity of simulation-based GCI combustion system development, it is mandatory to enhance the prediction of spray combustion of gasoline-like fuels. The purpose of this study is to model the spray characteristics of high reactivity gasoline fuels and validate the models with experimental results obtained through an optically accessible constant volume vessel under vaporizing [2] and reactive conditions [3].

Future internal combustion engines demand higher efficiency but progression towards this is limited by the phenomenon called knock. A possible solution for reaching high efficiency is Octane-on-Demand (OoD), which allows to customize the antiknock quality of a fuel through blending of high-octane fuel with a low octane fuel. Previous studies on Octane-on-Demand highlighted efficiency benefits depending on the combination of low octane fuel with high octane booster. The author recently published works with ethanol and methanol as high-octane fuels. The results of this work showed that the composition and octane number of the low octane fuel is significant for the blending octane number of both ethanol and methanol. This work focuses on toluene as the high octane fuel (RON 120). Aromatics offers anti-knock quality and with high octane number than alcohols, this work will address if toluene can provide higher octane enhancement.

Light naphtha is the light distillate from crude oil and can be used in compression ignition (CI) engines; its low boiling point and octane rating (RON = 64.5) enable adequate premixing. This study investigates the combustion characteristics of light naphtha (LN) and its multicomponent surrogate under various start of injection (SOI) conditions. LN and a five-component surrogate for LN, comprised of 43% n-pentane, 12% n-heptane, 10% 2-methylhexane, 25% iso-pentane and 10% cyclo-pentane, has been tested in a single cylinder optical diesel engine. The transition in combustion homogeneity from CI combustion to homogenized charge compression ignition (HCCI) combustion was then compared between LN and its surrogate. The engine experimental results showed good agreement in combustion phasing, ignition delay, start of combustion, in-cylinder pressure and rate of heat release between LN and its surrogate.

The blending of ethanol with PRF (Primary reference fuel) 84 was investigated and compared with FACE (Fuels for Advanced Combustion Engines) A gasoline surrogate which has a RON of 83.9. Previously, experiments were performed at four HCCI conditions but the chemical effect responsible for the non-linear blending behavior of ethanol with PRF 84 and FACE A was not understood. Hence, in this study the experimental measurements were simulated using zero-dimensional HCCI engine model with detailed chemistry in CHEMKIN PRO. Ethanol was used as an octane booster for the above two base fuels in volume concentration of 0%, 2%, 5% and 10%. The geometrical data and the intake valve closure conditions were used to match the simulated combustion phasing with the experiments. Low temperature heat release (LTHR) was detected by performing heat release analysis.

Increasing the anti-knock quality of gasoline fuels can enable higher efficiency in spark ignition engines. In this study, the blending anti-knock quality of dicyclopentadiene (DCPD), a by-product of ethylene production from naphtha cracking, with various gasoline fuels is explored. The blends were tested in an ignition quality tester (IQT) and a modified cooperative fuel research (CFR) engine operating under homogenous charge compression ignition (HCCI) and knock limited spark advance (KLSA) conditions. Due to current fuel regulations, ethanol is widely used as a gasoline blending component in many markets. In addition, ethanol is widely used as a fuel and literature verifying its performance. Moreover, because ethanol exhibits synergistic effects, the test results of DCPD-gasoline blends were compared to those of ethanol-gasoline blends. The experiments conducted in this work enabled the screening of DCPD auto-ignition characteristics across a range of combustion modes.

The blending of ethanol with primary reference fuel (PRF) mixtures comprising n-heptane and iso-octane is known to exhibit a non-linear octane response; however, the underlying chemistry and intermolecular interactions are poorly understood. Well-designed experiments and numerical simulations are required to understand these blending effects and the chemical kinetic phenomenon responsible for them. To this end, HCCI engine experiments were previously performed at four different conditions of intake temperature and engine speed for various PRF/ethanol mixtures. Transfer functions were developed in the HCCI engine to relate PRF mixture composition to autoignition tendency at various compression ratios. The HCCI blending octane number (BON) was determined for mixtures of 2-20 vol % ethanol with PRF70. In the present work, the experimental conditions were considered to perform zero-dimensional HCCI engine simulations with detailed chemical kinetics for ethanol/PRF blends.

Gasoline-ethanol-methanol (GEM) blends, with constant stoichiometric air-to-fuel ratio (iso-stoichiometric blending rule) and equivalent to binary gasoline-ethanol blends (E2, E5, E10 and E15 in % vol.), were defined to investigate the effect of methanol and combined mixtures of ethanol and methanol when blended with three FACE (Fuels for Advanced Combustion Engines) Gasolines, I, J and A corresponding to RON 70.2, 73.8 and 83.9, respectively, and their corresponding Primary Reference Fuels (PRFs). A Cooperative Fuel Research (CFR) engine was used under Spark Ignition and Homogeneous Charge Compression Ignited modes. An ignition quality tester was utilized in the Compression Ignition mode. One of the promising properties of GEM blends, which are derived using the iso-stoichiometric blending rule, is that they maintain a constant octane number, which has led to the introduction of methanol as a drop-in fuel to supplement bio-derived ethanol.