<<

Limitations of partially premixed

Citation for published version (APA): Wang, S. (2017). Limitations of partially premixed combustion. Technische Universiteit Eindhoven.

Document status and date: Published: 17/10/2017

Document Version: Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website. • The final author version and the galley proof are versions of the publication after peer review. • The final published version features the final layout of the paper including the volume, issue and page numbers. Link to publication

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement: www.tue.nl/taverne

Take down policy If you believe that this document breaches copyright please contact us at: [email protected] providing details and we will investigate your claim.

Download date: 24. Sep. 2021 Limitations of Partially Premixed Combustion

PROEFSCHRIFT

ter verkrijging de graad van doctor aan de Technische Universiteit Eindhoven, op gezag van de rector magnificus, prof.dr.ir. F.P.T. Baaijens, voor een commissie aangewezen door het College voor Promoties, in het openbaar te verdedigen op dinsdag 17 oktober 2017 om 16.00 uur

door

Shuli Wang

geboren te Heilongjiang, China Dit proefschrift is goedgekeurd door de promotoren en de samenstelling van de promotiecommissie is als volgt: voorzitter: prof.dr.ir. M.G.D. Geers 1e promotor: prof.dr. L.P.H. de Goey copromotor: dr.ir. L.M.T. Somers leden: prof.dr. Ö. Andersson (Lund University) prof.dr. H.B. Levinsky (University of Groningen) prof.dr. D.J.E.M. Roekaerts adviseur: dr.ir. C.A.J. Leermakers (DAF NV)

Het onderzoek dat in dit proefschrift wordt beschreven is uitgevoerd in overeenstemming met de TU/e Gedragscode Wetenschapsbeoefening. Shuli Wang (2017). Limitations of Partially Premixed Combustion. Ph.D. thesis, Eindhoven University of , Eindhoven, the Netherlands.

Copyright © 2017 by Shuli Wang.

All rights reserved. No part of the material protected by this copyright notice may be produced or utilised in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without the prior written permission of the author. A catalogue record is available from the Eindhoven University of Technology Library ISBN: 978-90-386-4349-6 Cover design by Tao Yan and Shuli Wang. PPC chemiluminescence images by the courtesy of Zhenkan Wang. Printed by Ipskamp Printing, Enschede. To my family... Contents

Nomenclature v

Summary vii

1 Introduction 1 1.1 Background ...... 1 1.2 Combustion Concepts ...... 3 1.2.1 Diesel Combustion ...... 3 1.2.2 Homogeneous Charge Compression Ignition ...... 4 1.2.3 Reactivity Controlled Compression Ignition ...... 5 1.2.4 Partially Premixed Combustion ...... 6 1.3 Fuel Appetite for Partially Premixed Combustion ...... 7 1.3.1 PPC with Diesel-like Fuels ...... 7 1.3.2 PPC with -like Fuels ...... 7 1.4 Objective and Contributions ...... 9 1.5 Outline of the Thesis ...... 10

2 Experimental Apparatus 11 2.1 The Engine ...... 11 2.2 Emission Measurement Systems ...... 14 2.2.1 Horiba Measurement System (Gases) ...... 14 2.2.2 Meter () ...... 14 2.2.3 Engine Exhaust Particle Sizer (Particle Distribution) ...... 15 2.2.4 Electron Microscopy (Particle Morphology) ...... 17 2.3 Data Acquisition System ...... 19 2.4 Definition of Related Parameters ...... 20 2.5 Conclusions ...... 24

3 Partially Premixed Combustion with RON70 Fuels 25 3.1 Comparison of ...... 25 3.1.1 Experimental Details ...... 26 3.1.2 Differences between the Three Injectors ...... 26 3.2 Effect of Air-excess ...... 29

i Contents

3.2.1 Experimental details ...... 29 3.2.2 Major Combustion Characteristics ...... 33 3.2.3 Soot and NOx Emissions ...... 38 3.3 Limitations of PPC at higher loads ...... 39 3.3.1 Experimental details ...... 39 3.3.2 Combustion Characteristics ...... 40 3.3.3 NOx-Soot Trade-off ...... 41 3.4 Conclusions ...... 43

4 Low Load Tests with High-octane Fuels 45 4.1 Experimental Details ...... 45 4.1.1 Test Fuels ...... 46 4.1.2 Test Conditions — Effects of temperature and CA50 ...... 46 4.1.3 Test Conditions — Optimize NOx Emissions ...... 48 4.2 Results and Analysis — Effects of Intake temperature and CA50 ...... 49 4.2.1 Load I ...... 49 4.2.2 Load II ...... 58 4.3 Results and Analysis — Optimize NOx Emissions ...... 64 4.3.1 Load I ...... 64 4.3.2 Load II ...... 67 4.4 Conclusions ...... 70

5 Particulate Matter Emissions 71 5.1 Introduction ...... 71 5.2 Tests with Ternary Blends ...... 73 5.2.1 Test Fuels ...... 73 5.2.2 Test Conditions ...... 74 5.2.3 Comparison of Blends of RON70 and Diesel ...... 75 5.2.4 Comparison of TRF and CRF ...... 79 5.3 PRF Blends Mix with Different Volumes of ...... 83 5.3.1 Test Fuels ...... 83 5.3.2 Test Conditions ...... 84 5.3.3 Major Results and Discussion ...... 84 5.4 Tests with Binary Blends ...... 91 5.4.1 Test Fuels ...... 92 5.4.2 Test Conditions ...... 92 5.4.3 Effect of 2-EHN on Combustion Controllability ...... 94 5.4.4 Physical and Chemical Effects on PM ...... 97 5.5 Characterization of Agglomerates’ Morphology ...... 103 5.5.1 Test Conditions ...... 103 5.5.2 Major Results and Discussion ...... 104 5.6 Conclusions ...... 109

6 Conclusions and Outlook 111

Bibliography 115 ii Contents

A Engine Exhaust Particle Sizer 127 A.1 How does EEPS works? ...... 127 A.2 Secondary Dilution Factor ...... 129 A.3 Normalized Concentrations: dN/dlogDp ...... 129

Acknowledgement 131

Curriculum Vitae 133

iii

Nomenclature

AD Aerodynamic diameter AFst Stoichiometric air fuel ratio AM Accumulation mode ATDC After top dead center BD Burn duration BDC Bottom dead center BEV Battery electric BMEP mean effective BTDC Before top dead center CAD angle degree CA10 The crank angle where 10% of the heat has been released CA50 The crank angle where 50% of the heat has been released CA90 The crank angle where 90% of the heat has been released CI Compression ignition CN CO monoxide

CO2 COV Coefficient of variation DOC Diesel oxidation catalyst DPF Diesel particulate filter EC Elemental carbon EEPS Engine exhaust particle sizer EGR recirculation EOI End of injection EVC Exhaust close EVO Exhaust valve open FSN Filter smoke number HC HCCI Homogeneous charge compression ignition HDDI Heavy duty direct injection

v Nomenclature

HFR Hydraulic flow rate ID Ignition delay IDw Ignition dwell IMEP Indicated mean effective pressure IMEPg Gross indicated mean effective pressure IMEPn Net indicated mean effective pressure ISFC Indicated specific fuel consumption IVC Intake valve close IVO Intake valve open LHV Lower heating value NM Nucleation mode MON Motor octane number NOx oxides NVO Negative valve overlap ON Octane number PAH Polycyclic aromatic PF Premixed fraction PM Particulate matter PPC Partially premixed combustion PRF Primary reference fuel PRR Pressure rise rate RCCI Reactivity controlled compression ignition RoHR Rate of heat release RON Research octane number rpm SCR Selective catalytic reduction SMPS Scanning mobility particle sizer SOI Start of injection SSCI Spark assisted compression ignition TDC Top dead center TEM Transmission electron microscopy λ Air fuel ratio 2-EHN 2-ethylhexyl nitrate

vi Summary

The is widely used in many applications due to its high efficiency, high stability, and also having an extreme flexibility for a variety of operating conditions. Despite the numerous advantages of diesel , society is increasingly concerned about the pollutants emitted by diesel engines. One way to address the problem is to alter the combustion so that less pollutants are created in the first place. Low-temperature combustion is a leading strategy designed to reduce or eliminate the two most problematic pollutants emitted by diesel engines, nitrogen oxides (NOx) and particulate matter (PM). One such combustion concept is Homogeneous Charge Compression Ignition (HCCI), which typically employs long in- mixing times prior to combustion to produce relatively uniform and lean fuel/air mixtures. As the technique has matured, HCCI has evolved toward Partially Premixed Combustion (PPC) which has a more inhomogeneous fuel/air mixture and temperature field to better control the heat release rate. However, there are still many limitations in the practical application of PPC engines. The Chapter 1 and 2 introduces the research background and experimental equipment, respectively. The Chapter 3 begins with the tests using three injectors with different nozzle geometries to compare their combustion and emission characteristics. Then, the in-cylinder emission control parameter air-excess was investigated under PPC mode to get knowledge about the influences of in-cylinder density and concentration on exhaust emissions. Moreover, the limitations of PPC at higher loads in terms of the trade-off relationship between NOx and PM emissions are discussed. Diesel and gasoline are the most important and widely used fuels for internal combus- tion engines. However, because of the complex compositions of diesel and gasoline fuels, simplified surrogate fuel like primary reference fuel (PRF) is used to represent the current transportation fuels in both experimental and simulated studies. The Chapter 4 explores the low load limit of PPC using three PRFs with research octane numbers (RON) of 70, 80 and 90, respectively. The advantage of using high-octane fuels is the longer ignition delay which enhances the mixing of fuel and air before combustion. But the attainable operating region is limited at low load and idle conditions. The main objective of this study is to extend the operating region of the engine towards lower loads with the aid of an intake air heating system. In addition, different dilution strategies are used, trying to get low NOx, soot emissions, and acceptable HC and CO emissions. The Chapter 5 deals with questions related to particulate matter emissions using various high-octane fuels with different chemical structures. For combustion under PPC mode,

vii Summary particulate matter cannot be directly extrapolated from classical soot emission measurement using a smoke meter, as typically done for conventional diesel combustion. This is because the particulate matter discharged from PPC has a low fraction of carbonaceous compounds and a significant amount of soluble organic fraction, which can not be captured by the smoke meter. The particulate number concentration and size distributions were measured with an engine exhaust particle sizer (EEPS), which allows measuring the size distribution from 5.6 to 560 nanometer. Besides, transmission electron microscopy (TEM) was also used to study the morphology of the aggregates produced by different fuels at various engine conditions. Combining the experimental results obtained from this dissertation, the limitations of PPC with single injection strategy were presented: under high load, alcohol contained fuels can alleviate the NOx-soot tradeoff relationship, but NOx-soot tradeoff is unavoidable; under lower loads, PPC with high-octane fuels can achieve stable combustion, but produce much higher HC and CO emissions than diesel due to overleaning; PPC with high-octane fuels is well known to reduce PM emissions, but it also has the risk of emitting more smaller particles, which are considered to be more harmful to health. These limitations will be further addressed in the future research.

viii Chapter 1 Introduction

Compression ignition (CI) engines are widely used for transportation and power-generation applica- tions. Improving the thermal efficiency, finding suitable alternative fuels, and further reducing the pollutant emissions are needed because of energy shortages and the increasingly stringent emission regulations. Conventional diesel combustion, as a typical representation of CI combustion, shows a trade-off between nitrogen oxides (NOx) and particulate matter (PM) emissions. Considerable effort has been devoted toward breaking through the compromise between NOx and PM emissions as they both have adverse effects on the environment and human health.

1.1 Background

Internal combustion engines have been used for transportation for over a hundred years. We would not have reached the living standard of today without the transportation performed by the millions of vehicles. The Netherlands is both a very densely populated and a highly developed country, in which transportation is a key factor in the economy. One problem is the increasing price and decreasing supply of which drives the demand for more efficient engines. Another problem is the environmental impact from harmful emissions, such as (CO), unburned hydrocarbons (HC), NOx, and PM. Moreover, carbon dioxide (CO2) attracts more and more attention because of the climate changes.

Combustion is an important source of CO2 emissions. Battery electric vehicles (BEVs) are considered as of the future to solve the energy shortage and environmental problems. However, the carbon emissions of grid powered electric cars in countries with coal based generation are no different than that of cars propelled by internal combustion engines [1]. Besides, the range of BEVs is not comparable with internal combustion engine vehicles due to the limited energy density of battery [2]. The long charging time and the limited number of charging stations also bring a lot of inconvenience to drivers [2]. Even though battery density is improving day by day, batteries for heavy-duty trucks need to be even more powerful, durable, long-lasting and high volume power density. Electric trucks do exist,

1 Chapter 1. Introduction mostly medium-duty hybrid that stop and start a lot to recharge the battery. Therefore, the heavy-duty vehicles will still rely on internal combustion engines in the foreseeable future. Improving fuel efficiency and looking for alternative fuels are two major approaches to reduce carbon emissions of heavy-duty diesel engines.

The G-20 countries account for 90 percent of the global sales, and 17 out of the 20 members have chosen to follow European regulatory pathway for vehicle emissions control [3]. Emission regulations have become increasingly stringent to control harmful emissions, which are detrimental to the environment as well as human beings. Europe first introduced heavy-duty vehicle emission standards in 1988. The “Euro” track was established in 1992 and progressing through to Euro VI in 2013. The European emissions standards are shown in Table 1.1.

Table 1.1: EU Emission Standards for HD Diesel Engines. All emission limits are given in g/kWh

Tier Date Test CO HC NOx PM Euro I 1992 (>85 kW) R-49 4.5 1.1 8.0 0.36 Euro II 1996 R-49 4.0 1.1 7.0 0.25 Euro II 1998 R-49 4.0 1.1 7.0 0.15 Euro III 2000 ESC & ELR 2.1 0.66 5.0 0.10 & 0.13a Euro IV 2005 ESC & ELR 1.5 0.46 3.5 0.02 Euro V 2008 ESC & ELR 1.5 0.46 2.0 0.02 Euro VI 2013 WHSC 1.5 0.13 0.4 0.01 a For engines of less than 0.75 dm3 swept volume per cylinder and a rated power speed of more than 3000 per minute.

To meet the increasingly stringent emission regulations, significant advancements in engine-related or fuel-related , as well as exhaust aftertreatment technologies are required. To meet the Euro VI emission requirements, diesel oxidation catalyst (DOC), diesel particulate filter (DPF), and selective catalytic reduction (SCR) catalyst are commonly used for heavy-duty diesel engines. DPF filter needs regeneration which consumes additional fuel. Besides, if the DPF filter is not properly cleaned, the leftover ash can compete with soot for space, leading to overheating of DPF and potentially causing the substrate to crack. This will allow unfiltered exhaust gas to enter to the SCR catalyst. Both a cracked DPF and a poisoned catalyst are expensive to repair. In addition, the needed urea injection requires additional , piping, and refueling stations to operate. Therefore, there is a need for more advanced combustion concepts that may use renewable fuels to achieve a low and have low tailpipe emissions to reduce the dependency on aftertreatment. However, in practical use, the advanced combustion concepts also have their limitations.

2 1.2. Combustion Engine Concepts

1.2 Combustion Engine Concepts

Diesel engine designers face the challenge of complying with ever more stringent emission standards, without compromising on engine efficiency. To accomplish these goals, many advanced combustion concepts have been proposed in addition to the traditional diesel combustion mode. It is crucial to have a thorough understanding of these combustion concepts to make better use of them. In this section, the prevailing concepts are introduced.

1.2.1 Diesel Combustion Diesel combustion is a complex, turbulent, multiphase process that occurs in a high-temperature and high-pressure environment [4]. Air alone is inducted into the cylinder during the intake , and a highly reactive fuel (diesel) is injected directly into the cylinder just before the combustion process (see Figure 1.1). It is well known that diesel combustion suffers from high NOx because of regions where temperatures are high, and it also suffers from high soot emissions due to the inevitable regions of high fuel concentration. These sources are better illustrated in an equivalence ratio versus temperature map (ϕ - T). 0-D homogeneous reactor combustion simulations were performed by Kitamura et al. [5], using a detailed n-heptane at an ambient pressure of 60 bar and a residence time of 2 ms. Regions of NOx and soot formation are provided in the ϕ - T plane, as shown in Figure 1.2. The blue curve indicates the adiabatic, equilibrium flame temperature computed for each equivalent ratio at an initial reactant temperature of 1000 K, assuming the O2 concentration is 21%. The adiabatic flame temperature is the theoretical maximum. Therefore, the real flame temperature will always be lower than the blue curve. High temperature and fuel-lean mixtures lead to high NOx emissions, whereas soot production occurs at fuel-rich mixtures within a specific temperature window. This leads to the so-called NOx-soot tradeoff.

Figure 1.1: Compression Ignition engine schematic, adapted from [6].

3 Chapter 1. Introduction

Figure 1.2: NOx and soot production zones as a function of local temperature and equivalence ratio. The adiabatic gas temperature is shown by the turquoise line at 21 percent ambient oxygen. Adapted from [5] .

The NOx-soot trade-off is also known as the Diesel Dilemma. In general, a decrease in NOx emissions is achieved at the expense of producing more soot emissions and vice versa. To break this trade-off relationship, new combustion concepts are proposed emphasizing on a more uniform mixture and limiting the local in-cylinder temperatures. One of the early attempts is a concept called Homogeneous Charge Compression Ignition (HCCI), which will be introduced afterward.

1.2.2 Homogeneous Charge Compression Ignition

HCCI is a combustion process controlled by the history of temperature, pressure, and composition of the in-cylinder charge [7], [8], and [9]. HCCI is controlled by chemical kinetics, unlike conventional diesel combustion that is mostly controlled by fuel/air mixing process. Either port or very early direct injection is used to get a premixed charge prior to ignition thus producing almost zero soot. Earlier studies on HCCI with port fuel injection of diesel found that very premature ignition and knock occurred, and relatively high intake temperatures were required to minimize the fuel condensing on the intake system [10] and [11]. Given that HCCI combustion takes advantage of highly diluted charges with either lean mixtures or high level of exhaust gas recirculation (EGR), the in-cylinder temperature remains low compared to conventional diesel combustion. As a result, low level of NOx is produced, well below the Euro VI limit of 0.4 g/kWh. Diesel combustion produces high emissions of NOx and soot, which requires using expensive aftertreament systems like lean NOx traps [12], SCR catalysts ([13]) and DPF ([14]). Thus, one significant benefit of HCCI is that it does not need to use expensive aftertreatment

4 1.2. Combustion Engine Concepts systems for NOx and soot emissions. However, the potential benefits of HCCI combustion cannot be fully realized in production applications until some obstacles can be overcome. The main limitations of this combustion mode are summarized below: • The power density is low because of the high degree of dilution needed. • Combustion efficiency is low compared to conventional diesel combustion. • The combustion noise level is high. • The operating range is limited. • Combustion phasing control is poor due to lack of an immediate control actuator. To take advantage of HCCI combustion to break the NOx-soot trade-off and to maintain the primary benefits of diesel combustion for CI engines (i.e. good controllability on combustion phasing), two main branches emerged. The first one is Reactivity Controlled Compression Ignition (RCCI) and the second branch is Partially Premixed Combustion (PPC), they are explained in next sections.

1.2.3 Reactivity Controlled Compression Ignition Researchers from the University of Wisconsin-Madison developed a compression engine combustion process using in-cylinder fuel blending. RCCI (Figure 1.3), which is a variant of HCCI, is a dual-fuel engine combustion technology. Typically, a port injection of low reactive fuel establishes the global equivalence ratio, and direct single or multiple injections of high reactive fuel to trigger the combustion. By appropriately choosing the reactivities, relative amounts, and injection timings of the fuel charges, RCCI combustion can be tailored to achieve the best fuel economy. A large amount of research has been done on the effects of injection pressure [15], injection timings [16], port fuel injection fractions [17], geometry [18], intake temperature [19] and intake pressure [19] on combustion and emission characteristics. Their results demonstrated that RCCI is a promising strategy to meet current and future emission regulations without relying on NOx and soot aftertreatment. Still, there is scope for further improvements in RCCI. Current injection strategy for RCCI is injecting 70 ∼ 90% (fuel mass) of the low reactivity fuel in the intake port and the remaining of high reactivity fuel is directly injected at pressure of 150 ∼ 800 bar. Instead, it would be interesting to inject a certain portion of the low reactivity fuel first, followed by the high reactivity fuel and then inject the remaining low reactivity fuel. This could help in controlling the combustion phasing at higher loads. Low reactivity fuels other than gasoline are also of interest to investigate, such as ethanol and n-butanol. In a future study, combustion technology group at the TU/e will launch a series of studies on RCCI. The main strengths of the RCCI strategy include [16]: • Low NOx and soot emissions. • Reduced heat transfer losses. • Increased fuel efficiency and thermodynamic efficiency. • Reduced fuel system cost. • Compliant with US Environmental Protection Agency’s 2010 emissions guidelines without exhaust aftertreatment.

5 Chapter 1. Introduction

Figure 1.3: RCCI engine schematic, adapted from [6].

For real-life use, this concept requires decent controllability. The combustion phasing is controlled by the relative proportions of the two fuels, and the combustion duration is controlled by spatial stratification of the two fuels [20]. However, the port injection or very early direct injection of the low reactive fuel may result in low combustion efficiency. It was argued that deposition of unburned premixed low reactive fuel in the and crevice regions of the cylinder was the major contributor to the relatively low combustion efficiency [20]. Therefore, a second concept (PPC) is investigated which attempts to use a single fuel and achieve easier control.

1.2.4 Partially Premixed Combustion

PPC is also a combination of HCCI and traditional diesel combustion. PPC shares similar advantages as RCCI. However, it does not need an extra fuel . In this system, a single fuel is used and injected relatively late during the compression stroke. The main difference among PPC, HCCI, and conventional diesel is the time available for the premixing of the fuel and air. Compared to HCCI, PPC provides more thermal and mixture stratification to efficiently control the combustion process. While compared to conventional diesel combustion, PPC reduces local fuel-rich regions due to the longer mixing time before combustion happens. Therefore, ignition delay (ID), which is defined as the crank angle difference between the start of injection (SOI) and the start of combustion (SOC), is a key factor for PPC. A longer ignition delay enables a larger degree of premixing and results in more premixed combustion. Ignition delay can be adjusted in several ways. One is to change the in-cylinder temperature. Cooled exhaust gas recirculation (EGR), which is the most commonly used approach on diesel engines to reduce NOx emissions, slows down the kinetic reactions and results in longer ignition delays. The second way is to change the reactivity of the fuel, which is the prevailing approach currently. This will be introduced further in the next section.

6 1.3. Fuel Appetite for Partially Premixed Combustion

1.3 Fuel Appetite for Partially Premixed Combustion

The most prominent characteristic of PPC is the requirement of creating a controllable stratified mixture before ignition. The physicochemical properties of the fuel have a significant impact on the ignition and combustion process [21]. Therefore, the study of fuels has become an increasingly important method to control and burning rate in PPC.

1.3.1 PPC with Diesel-like Fuels

Lund University has conducted extensive research on diesel PPC [22]. Their results revealed that diesel PPC could defeat the NOx-soot trade-off relationship through applying excessive EGR rate (> 65%) and lowering the (e.g., 12.4). The reduction in NOx and soot emissions has to be paid with a significant decline in combustion efficiency. The wall wetting issues caused by the earlier injection timing and the high boiling point of diesel also lead to higher HC emissions. Fatty-acid methyl ester is a group of fuels originating from animal fats and vegetable oils commonly known as . Fang et al. [23] investigated biodiesel blends on an optically accessible single-cylinder diesel engine with multiple injection strategies. Biodiesel has a strong impact on reducing engine out soot emissions, probably thanks to its oxygen content. However, it was found that biodiesel leads to increased NOx emissions compared to EN590 diesel since biodiesel typically burns faster due to the slightly higher cetane number. More recently, a second generation of alternative was developed starting from the well-known Fischer-Tropsch synthesis process [24]. It is usually indicated with XTL, where X denotes the input feedstock and TL (to liquid) indicates the liquid state of the fuel. The input feedstock can be either (GTL), coal (CTL) or renewable biomass (BTL). GTL is already produced commercially and diesel fuels blended with GTL are available in several European countries [25]. The high cetane number of GTL increases the diffusive combustion phase with detrimental effects on soot formation, annulling the benefits of the poly-aromatic- free characteristic of the GTL on soot formation under certain engine conditions [24]. The NOx emissions decrease with GTL fuel compared to diesel, the decrease is considered to occur due to the lowering of the maximum heat release rate [26]. Diesel-like fuels have higher cetane numbers, which require a high EGR rate and a low compression ratio to achieve PPC, and the PPC operating load range is typically narrow. Therefore, the diesel-like fuel is not an ideal candidate for PPC operation.

1.3.2 PPC with Gasoline-like Fuels

Kitano et al. [27] investigated the effects of cetane number and fuel distillation characteristics on PPC for the purpose of reducing NOx and soot emissions from a diesel engine with EGR. They found that lowering the cetane number reduced NOx emissions effectively by suppressing the combustion temperature. The extension of the ignition delay was effective in promoting lean mixture formation, and the mild heat release characteristics indicate the suppression of NOx generation. But it was understood that excessive suppression of ignitability resulted in an increase in HC emissions and fuel consumption at lower loads.

7 Chapter 1. Introduction

Moreover, a lighter fuel with a lower distillation temperature has an advantage of reducing soot emissions, especially at higher loads. Because this kind of fuel increases the evaporation and mixing rates of injected fuel, and decreases the formation of fuel-rich regions in the . To simultaneously reduce NOx and soot emissions, an alternative of diesel PPC is to use fuels of low reactivity (gasoline-like fuels). Kalghatgi introduced the concept of gasoline PPC in 2006 [28], inspired by the fact that a less reactive fuel can provide a much longer ignition delay without using a high EGR rate or a lower compression ratio. Soot formation can only be avoided if the final injection of fuel is sufficiently completed before combustion starts. In other words, if the ignition delay is long enough there might be no fuel-rich pockets in the cylinder favoring the soot formation. Long ignition delay leads to a greater degree of premixed charge, which shifts emission characteristics of PPC closer to HCCI than conventional diesel combustion. EGR is typically used to reduce the temperature of the charge. Moreover, the lower temperature and higher specific heat of the exhaust gases also cause an increase in the ignition delay. Therefore, PPC can reduce NOx and soot emissions simultaneously. Following the path introduced by Kalghatgi, a large amount of experimental work on PPC using gasoline-diesel blended fuel was done [29], [30], and [31]. Diesel and gasoline are mostly used in conventional engines. These fuels are composed of many complex components, straight chain , branched alkanes and aromatics. The soot emission was significantly reduced with the increase in the proportion of gasoline in the mixture because the homogeneity of the fuel-air mixture was improved due to longer ignition delay. The European gasoline (EN228) has a research octane number (RON) of 95 or 98, its resistance to auto-ignition can be exploited to extend the ignition delay, but there will be some problems with the and low load if European gasoline is directly applied on a diesel engine. In general, fuels that have higher octane number (and thus lower cetane number) are favorable for PPC operating at higher loads. In addition to gasoline, alcohol fuels have also received extensive attention from engine researchers. Alcohol fuels have been used widely, particularly in Brazil, where ethanol has a long history of use as an automobile fuel. Besides ethanol, many linear and branched members of the alcohol have been studied, with a particular focus on butanol [32], [33] and [34]. Cetane number (CN) is an inverse function of a fuel’s ignition delay. In a particular diesel engine, higher cetane fuels will have shorter ignition delay periods than lower cetane fuels. Cetane number is an inverse of the similar for gasoline. Primary reference fuel (PRF) blends are mixtures of n-heptane (CH3(CH2)5CH3) and iso-octane

((CH3)3CCH2CH(CH3)2). The research octane number (RON) or motor octane number (MON) is the volume percent of iso-octane in the PRF blends. Typically n-heptane and iso-octane are used to model the combustion of diesel and gasoline [35]. The auto-ignition behavior of a test fuel mainly depends on its structure, composition and molecular size [36]. The auto-ignition chemistry of gasoline is different from that of PRF blends because of the non-paraffinic components. Using PRF blends as primary fuels combined with toluene

(C6H5CH3) as secondary fuel can have the same RON and octane sensitivity as the target gasoline [37]. Besides, ethanol and n-butanol can also be used as secondary fuels to blend with PRF blends to get a specific RON. According to Kalghatgi [38], the optimal RON for gasoline PPC is between 70 and 85 which is much lower compared to current market gasoline. The most efficient way to improve fuel adaptability is to blend two fuels with different chemical

8 1.4. Objective and Contributions properties [21]. The approach with PRF blends mixing with other simple fuels also reduces ambiguity when one wants to compare experimental results with simulation results. For the fuel blends used in this work, their mechanisms are available in literature. High octane number fuels (iso-octane, ethanol, n-butanol and toluene) which feature a long ignition delay were chosen to blend with a low octane number fuel n-heptane to get an RON in between 70 and 90, focusing on the impact of fuel structure on PM emissions. Hich-octane fuels have the advantage of longer ignition delays to improve the homogeneity of fuel/air mixture, but ignitability problems at low load and idle conditions may occur. In this dissertation, the low load limitations of PRF blends with RON of 70, 80, and 90 are studied by applying varying higher intake temperatures. Morever, previous research only focused on soot emissions in g/kWh or in FSN [39], [40] , and [41]. Recent studies correlate fine and ultra-fine particulates to adverse human health effects [42]. The porous particulate agglomerations could deposit in the deep lung and cause diseases like asthma, pneumonectasis and nasopharyngeal carcinoma [43]. Hypothetically gasoline PPC may contribute to more fine and ultra-fine particles. Therefore, it is necessary to acquire knowledge about the particulate matter in terms of number concentration and size distribution when using high-octane blends. This is also discussed in this thesis.

1.4 Objective and Contributions

PPC has a great potential to achieve high engine efficiency and ultra-low emissions of NOx and soot. This has attracted the attention of researchers and as well. However, main results in literature so far are obtained with real fuels like diesel, gasoline, or mixtures of diesel with some high-octane fuels. This brings ambiguity when one wants to compare experimental results with simulation results. In addition, the impact of fuel structure on the particulate emissions can not be studied explicitly. The objective of this work is to develop a better understanding of the combustion and emission characteristics using designed fuels. The need of specific fuel properties for PPC cannot be fulfilled using a single component fuel. The designed fuels use PRF blends as primary fuels and use ethanol or n-butanol or toluene as a secondary fuel to have a specific RON but different molecular structures. Designed fuels have the advantage that it is easier to study the influence of fuel structure on particulate emissions. Besides, the mechanisms of designed fuels are available which allows a better comparison between results of experiments and simulations. The ignition delay is the key factor of PPC. Oxygen concentration and temperature are the most important parameters influencing NOx and soot emissions. Firstly, the influence of air-excess on PPC using several blends of RON70 is investigated (Chapter 3). The most important requirement of PPC is the formation of a premixed fuel-air mixture. But at higher loads, blends of RON70 are not able to provide enough ignition delay to avoid the formation of fuel-rich regions. Hence the diesel dilemma (NOx-soot tradeoff) also occurs for these blends (Chapter 3). Blends with higher RON are expected to solve this problem, but they may cause ignitability problems at lower loads. Therefore, the possibility of running blends with RON of 70, 80, and 90 at lower loads is investigated in Chapter 4. Last but not least the effect of fuel structure and engine conditions on particle emissions in terms of particulate number concentration and size distribution is studied (Chapter 5). The main contributions are in the

9 Chapter 1. Introduction following areas and will be discussed in detail in Chapters 3, 4 and 5. • The influence of different injectors on injection duration and soot emissions. • Effect of air-excess on combustion and emission characteristics. • PPC limitations at higher loads regarding NOx-soot tradeoff. • PPC limitations at lower loads. • PPC limitations regarding very small particles (less than 30 nm in diameter). • Effect of EGR on particulate number concentration and size distribution. • Effect of fuel structure on PM emissions. • The morphology of soot aggregates using different fuels.

1.5 Outline of the Thesis

This thesis is aimed at exploring the possibility of application of fuels with an RON between 70 and 90 on a heavy duty diesel engine. The test engine, emission instruments and related parameters are introduced in Chapter 2. In Chapter 3, a comparison of different injectors running with diesel under conventional diesel combustion mode, the effects of air-excess on combustion and emission characteristics under PPC mode, and the limitation regarding NOx-soot trade-off at higher loads are discussed, respectively. Initially, the effects of three different injectors on diesel combustion is presented. This study is not originally intended to compare the effects of different nozzle parameters on the combustion and emission results, but because A was broken after finishing some experiments and it had to be replaced with a different nozzle. Then, the sensitivity of combustion parameters to changes in air-excess ratio using three blends of RON70 are investigated. Finally, PPC is well known to break the NOx-soot trade-off relationship, whether this effect still exists even at high load is also studied. In Chapter 4, three PRF blends with RON of 70, 80 and 90 were investigated. Running the engine in PPC mode with high-octane fuels has the advantage of long ignition delay to reduce fuel-rich regions, thus producing low soot emissions. However, the ignitability at lower loads is limited. The objective is to study whether stable combustion regarding low cyclic variations and pressure rise rates can be achieved through adjusting the intake temperature. In Chapter 5, the particulate matter emissions regarding particle number concentration and size distribution are investigated with different fuels (all with an RON in between 70 and 90). Besides, the morphology of soot aggregates is also studied with a transmission electron microscopy (TEM). Firstly, several ternary blends, formed by mixing ethanol or n-butanol or toluene or benzaldehyde with PRF blends, are experimentally studied emphasizing on particle number and size distribution. Secondly, to better study the effect of the fuel structure on particulate matter, several binary blends (n-butanol or toluene or iso-octane blended with n-heptane) are tested. Thirdly, to rule out the effect of different ignition delays (resulting from the different RONs of n-butanol, toluene, and iso-octane) on the soot emissions, is used. Finally, not only the particle number and size distribution, but also the morphology of soot agglomerates using various EGR rates is studied. Finally, the main conclusions and future scope of work are presented in Chapter 6.

10 Chapter 2 Experimental Apparatus

This chapter gives an overview of the thermodynamic test setup, emission measurement systems, and data acquisition system. Besides, the related parameters used in the thesis are defined.

2.1 The Engine

The engine used in this thesis, referred to as CYCLOPS, is a six cylinder DAF XE355c engine. Cylinder one is isolated as a test cylinder, and cylinders 4 to 6 are operated by the stock and are responsible for maintaining the engine speed. This engine has been updated in 2014. Cylinders 2 and 3 were used as EGR cylinders [44], the purpose of which is to generate adequate EGR flow, especially for high-boost conditions. However, when the exhaust flow is small, the reciprocating motion of the in the pump cylinders can create a that sucks out the exhaust gas, thus affecting the accuracy of the exhaust emission test. Now the cylinders 2 and 3 are no longer connected to the EGR circuit of test cylinder, and the pistons of cylinders 2 and 3 were removed. Figure 2.1 shows the schematic overview of the test engine. The base engine specifications are listed in Table 2.1. The test cylinder uses a standalone fuel injection system which is composed of a Resato double-acting air-driven high-pressure , a pressure regulator and a prototype Delphi injector. The intake air is provided by an Atlas Copco air , which can provide up to 5 bar intake pressure. Exhaust pressure is always set to 0.3 bar higher than intake pressure during the experiments to ensure a sufficient EGR rate. The EGR flow can be cooled to 30 ◦C using a variable flow of process as a medium. A heater is installed to heat up the intake charge to the desired temperature. Three injectors were tested, their specifications and the comparison results will be discussed in Chapter 3. The sketches of the cylinder, the bowl, and the injections of injectors A, B, and C are presented in Figure 2.2. When testing a fuel other than diesel, a lubricity enhancer fluid is used to prevent wear, reduce friction and provide a protective cover against corrosion. The diagram of the test cylinder is shown in Figure 2.3.

11 Chapter 2. Experimental Apparatus

Table 2.1: Geometrical properties of the engine

Based engine DAF HDDI diesel engine [mm] 130 Stroke [mm] 158 [mm] 266.7 Compression ratio [-] 15.7 Bowl shape M-shaped Bowl diameter [mm] 100 Number of [-] 4 Exhaust valve close (EVC) [oCA] -346 Intake valve close (IVC) [oCA] -153 Exhaust valve open (EVO) [oCA] 128 Intake valve open (IVO) [oCA] 344

7 6 5 4 17 415S 3 8 9 14 10

11 15 16 2

12 1

18

13

1. Air compressor 2&7. Micromotion flow meter 3. Exhaust tank 4. EGR cooler 5. Cooled EGR tank 6. EGR valve 8. Mixing tank 9. Heater 10. Injector 11. Pressure transducer 12. Charge amplifier 13. Encoder 14. Backpressure valve 15. Exhaust analyzer 16. Engine exhaust particle sizer 17. Smoke meter 18. Electric control unit & data acquisition.

Figure 2.1: Schematic overview of the test engine

12 2.1. The Engine

TDC 10CAD BTDC 32.3173 CAD BTDC 30CAD BTDC 34.2388 CAD 139° BTDC 150° 40.8522 CAD 153° BTDC

60CAD BTDC

90CAD BTDC

BDC

Figure 2.2: A sketch of the cylinder, the bowl and the included spray angle (blue line for injector A, for injector B and pink line for injector C).

13 Chapter 2. Experimental Apparatus

TDC [-360, 0, 360 °CA]

IVO [ °CA] 344 EVC [-346 °CA]

Valve overlap [30 °CA]

EVO [128 °CA]

IVC [-153 °CA] BDC

Figure 2.3: Valve timing diagram for test cylinder, see Table 2.1.

2.2 Emission Measurement Systems

Exhaust emissions are the key to environmental pollution problems. In addition to focusing on HC, CO, NOx and soot emissions, people are increasingly concerned about particle number and size. The Euro VI standards also introduced particle number emission limits in 2013. The subsections below will present the emission measurement systems used in this research.

2.2.1 Horiba Measurement System (Gases) The concentration of exhaust gaseous emissions, for example, hydrocarbons, carbon monox- ide, carbon dioxide (CO2) and NOx, are measured using a Horiba Mexa 7100 DEGR exhaust analysis system, which measures HC by flame ionization method (FID), CO and CO2 by the non-dispersive infra-red method (NDIR), and NOx by the chemiluminescent method. The

Horiba Mexa is also used to determine the EGR fraction by measuring the CO2 concentration at the inlet and exhaust.

2.2.2 Smoke Meter (Soot) An AVL 415S Smoke Meter is used to measure smoke emissions. It uses the filter paper method to determine the Filter Smoke Number (FSN defined according to ISO 10054). This smoke meter has a measurement range of 0 to 10 FSN. Zero corresponds to the original

14 2.2. Emission Measurement Systems reflection of the white paper, and 10 corresponds to no light reflection. In this thesis, a sampling volume of 1000 ml is used for all measurements. The sampled engine exhaust gas passes through clean filter paper, and the filtered soot causes blackening of the filter paper. The paper blackening is measured by a photoelectric measuring head and the result is analyzed by a microprocessor [45]. The exhaust sampling schematic of the smoke meter is shown in Figure 2.4. More details about the smoke meter can be found in [46].

Dead volume

The dead volume is the volume of the gas way from the sampling point to the filter paper.

Sample Volume

Effective length

Effective length=(Sample volume- Filter area Dead volume-Leak volume) /(Filter area)

The leak volume is caused mainly by transverse currents of the filter at the suction unit.

Exhaust flow direction

Figure 2.4: Smoke meter sampling schematic, adapted from [45]

2.2.3 Engine Exhaust Particle Sizer (Particle Distribution) The engine exhaust particle sizer (TSI EEPS 3090) was used in the experiments to obtain detailed information about particle concentration and size distribution. EEPS is a fast-response, high-resolution instrument that measures very low particle number concentrations in a diluted exhaust. It was designed specifically for measuring engine exhaust emissions using mobility classification. With real-time data collection and display capabilities, the dynamic behavior of particle emissions even during transient test cycles can be visualized and studied. It measures exhaust particles in the range from 5.6 to 560 nanometers (nm), covering most of the range of interest. Besides, it operates at ambient pressure to eliminate any concerns about the evaporating volatile and semi-volatile particles. More details about how EEPS works can be

15 Chapter 2. Experimental Apparatus found in Appendix A.1. Hot engine exhaust contains both solid particles (for example, carbonaceous soot and ash) and volatile vapors (such as water, sulfate, and hydrocarbons). In this thesis, a constant volume system (CVS) tunnel is used for emission testing. So the volatile substances have a chance to condense into nanodroplets, which are detected as particles together with nonvolatile solid particles. To measure only the solid particles, thermal conditioning is required to remove the volatiles in the sample. Figure 2.5 shows how volatile compounds from nanodroplets and how they can be eliminated or suppressed through adjusting dilution and thermal conditions. It is clearly seen that the concentration and the temperature of the volatile compounds are reduced after passing through a CVS tunnel (path A to B, as shown in Figure 2.5). In the path from A to B, the volatile compound passes its dewpoint (curve N) and nucleates into nanodroplets. Subsequent secondary dilution (path B to D) will further decrease the number concentration of nanodroplets formed in the first dilution (path A to B). However, the secondary dilution is unable to evaporate these nanodroplets due to the hysteresis effect between nucleation and evaporation [47]. Therefore, the sample needs to be heated above the evaporation point (curve E) of the volatile compound (path D to C) using a thermal conditioner. Before entering the EEPS, the volatile compounds remain in the vapor phase during the cooling process (path C to D) and will not cross the nucleation curve (curve N) again to reform nanodroplets. Thus the path A to B to D to C back to D can eliminate or suppress the nanodroplets in the sample.

Figure 2.5: Principle of thermaldilution [47].

The thermo-dilution unit (379020A-30) used in this thesis contains a rotating disk thermo- diluter (model 379020A) and a thermal conditioner (model 379030). The sample is first diluted with the rotating disk thermo-ciluter (path A to B to D, as shown in Figure 2.5). Then, the sample is heated to evaporate the volatile compounds (path D to C) using the thermal conditioner. Last, with hot dilution, the volatile compounds remain in the vapor phase then enter into EEPS. More details about the thermo-dilution unit work process are given. This

16 2.2. Emission Measurement Systems thermo-dilution unit separates sampling, dilution, and conditioning of the aerosol into several steps [47]. (1) Undiluted exhaust emissions are sampled at a rate of approximately 1.0 L/min, and a portion of the raw exhaust is captured by each cavity of the rotating disk and transported to the measurement channel where it is mixed with particle-free dilution air. Figure 2.6 shows the diluted sample of 309020A in the standalone mode and Figure 2.7 shows the air path for the combined 379020A-30. The model 379030 generates the primary dilution air for the diluter with a controlled flow of 1.5 L/min. The rotating disk thermo-diluter also features a variable dilution ratio (from 15:1 to 3000:1) and the diluter can be heated up to 150 oC to avoid measurement of condensed volatile compounds. (2) Model 379030 uses an evaporation tube, which can be heated up to 400 ◦C, to eliminate nano-droplets that may have formed during the primary dilution process. No re-condensation takes place during the cooling process (path C to D, as shown in Figure 2.5), assuming the sampled aerosol is below the dew point (on curve N, as shown in Figure 2.5) after primary dilution. (3) The primary diluted measuring gas is further diluted with secondary dilution air generated in the model 379030. The secondary dilution air flow is adjustable from 0 to 15 L/min. The EEPS spectrometer always operates at 10 L/min. In this study, real-time sampling rate and a sampling period of EEPS were set at 1 Hz and 1 minute, respectively. During all the experiments primary dilution (with rotating disk thermo-diluter) temperature was set at 150 ◦C with a dilution factor of 50 and the secondary dilution temperature was set at 300 ◦C (evaporation tube) with a dilution factor of 6.7. Thus the overall dilution ratio is 335. The reason to choose a secondary dilution factor of 6.7 is given in Appendix A.2.

Figure 2.6: Diluted sample (309020A in standalone mode) [47].

2.2.4 Transmission Electron Microscopy (Particle Morphology) Transmission electron microscopy (TEM) is a microscopy technique in which a beam of electrons is transmitted through a specimen to form an image [48]. TEM is known as one of the most powerful tools to observe and understand the morphology and nanostructure of soot particles [49], [50], [51] and [52]. The TEM grid used in this study has a diameter of

17 Chapter 2. Experimental Apparatus

Figure 2.7: Sample air path (diluted) for the combined 379020A-30 system [47].

Figure 2.8: TEM grid and lacey carbon film

3 mm and 200 mesh copper grids with a lacey carbon film, as shown in Figure 2.8. Lacey carbon films offer a greater percentage of open area, provide the thinnest possible support while maintaining adequate strength, and give practically no background interference in the TEM [53]. TEM imaging was performed on a FEI Tecnai 20 (type Sphera) operated with a 200 kV equipped with a LaB6 filament and a 1k Gatan CCD camera. As presented in Figure 2.9,

18 2.3. Data Acquisition System the TEM grid was put on the smoke meter paper to gather particles in the exhaust.

Lacy grid to get TEM sampler

Exhaust flow direction

Figure 2.9: TEM grid placement method

2.3 Data Acquisition System

The engine is equipped with all common sensors typically applied in engine experiments, such as sensors which can measure intake and exhaust , and sensors which can measure temperatures of intake, exhaust, cooling water and oil. These quasi-steady-state data, together with air and fuel flows and emission data are recorded during a period of 40 seconds at a frequency of 20 Hz by an in-house data acquisition system, see [6]. The average of the recorded data is taken as the value for the experimental point under investigation. A SMETEC Combi combustion analyzer is used to record and process the so-called fast- changing parameters, such as in-cylinder pressure (which is measured with an AVL GU21C piezo-electric pressure transducer), fuel pressure and injector current. In order to understand the cyclic variations on each experimental point, all these channels are logged at 0.1 crank angle degree (CAD) for 50 consecutive cycles. EEPS spectrometer data is processed with the soot matrix which matches more closely to the data from a TSI scanning mobility particle sizer (SMPS) spectrometer. Software version 3.2.5 and firmware version 3.1.2 are used when using EEPS. A TEM sample is collected for each measurement point, and stored in a glass container which is well cleaned in an ultrasonic cleaner. All TEM samples are measured in the Materials and Interface Chemistry Department of TU/e.

19 Chapter 2. Experimental Apparatus

2.4 Definition of Related Parameters

Indicated mean effective pressure (IMEP) The brake mean effective pressure (BMEP) is the mean effective pressure calculated from the dynamometer power (). But due to the layout of the engine, BMEP cannot be used in this thesis. Therefore, the IMEP which does not consider the mechanical efficiency is used. From the P-V diagram, an example is shown in Figure 2.10, the indicated mean effective pressure is calculated. Net indicated mean effective pressure (IMEPn) is calculated from in-cylinder pressure over the complete engine cycle (720o in a four-stroke engine) using Equation (2.1).

Where Vd is the displacement volume and P is the in-cylinder pressure. To evaluate the combustion performance at different intake pressures and exhaust pressures, the gross indicated mean effective pressure (IMEPg) which excludes the gas exchange stroke (the yellow area in Figure 2.10) is used to present engine load and calculate the indicated emissions in this thesis. IMEPg is calculated from in-cylinder pressure over compression and expansion strokes (360o in a four-stroke engine) using Equation (2.2). Figure 2.11 shows the change in in-cylinder pressure with the crank angle, and helps to understand the choice of crank angle range for calculating the mean effective pressure.

360 −360 P · dV IMEPn = (2.1) R Vd 180 −180 P · dV IMEPg = (2.2) R Vd FuelMEP The supplied fuel energy content per cycle can be normalized as FuelMEP (unit is

bar), which is calculated with Equation (2.3). Where mfuel/cycle is the fuel mass per

cycle, QLHV is the lower heating value (LHV) of test fuel and Vd is the displacement volume. m · Q FuelMEP = fuel/cycle LHV (2.3) Vd Ignition delay, Ignition dwell and Burn Duration The crank angle where 10 percent, 50 per- cent, and 90 percent of the heat has been released are referred as CA10, CA50, and CA90, respectively. Ignition delay is defined as the crank angle difference between start-of-injection (SOI) and CA10, see Equation (2.4). Ignition dwell is defined as the crank angle difference between end-of-injection (EOI) and CA10, see Equation (2.5). Burn duration is defined as the crank angle difference between CA90 and CA10, see Equation (2.6). ID = CA10 − SOI (2.4) IDw = CA10 − EOI (2.5) BD = CA90 − CA10 (2.6)

Specific gas emissions The indicated specific emission of x (ISx) is calculated with Equation (2.7). Qair+Qfuel −6 · [x] · 10 · Mx · 3600 IS = Mexhaust (2.7) x Power x can be HC or CO or NOx emissions. In this equation, Qair and Qfuel are air

20 2.4. Definition of Related Parameters

flow (g/s) and fuel flow (g/s), respectively. Mexhaust (g/mol) and Mx (g/mol) are the molar masses of exhaust gas and x. Besides, [x] is the concentration of x in volume (ppm). And Power represents engine power (kW). If EGR is used for a constant intake pressure, the Qair will reduces, but the Mexhaust is almost constant.

Figure 2.10: In-cylinder pressure vs displacement volume Pressure

EVC IVC EVO IVO TDC

-360 -180 0 180 360 oCA

Figure 2.11: In-cylinder pressure vs crank angle

21 Chapter 2. Experimental Apparatus

Efficiencies Fuel chemical energy is converted into mechanical work in several stages in an internal combustion engine. Combustion efficiency is a measure of how efficiently fuel energy is converted into useful energy. When calculating the combustion efficiency in this thesis, only combustible species like HC and CO in the exhaust are considered, see Equation (2.8). Possible effects of other components are ignored. Indicated specific emissions of HC (ISHC) and CO (ISCO), and indicated specific fuel consumption (ISFC) should be measured to calculate combustion efficiency. The LHV of HC emissions is assumed to be that of the test fuel.

IS · LHV η = 1 − i=CO,HC i i (2.8) comb ISFC · LHV P fuel The next stage is the conversion of heat released in the combustion chamber into

mechanical energy, pushing the piston down. Thermodynamic efficiency (ηth) indicates the efficiency of conversion of heat to work in the compression and expansion strokes, and therefore includes the combustion efficiency. It is calculated with Equation (2.9), and more details can be found in [54].

180 −180 P · dV ηth = (2.9) ηcombR· mfuel˙ · LHVfuel

EGR ratio The EGR fraction is experimentally determined from the ratio of the CO2 volume

concentration in the intake gas to the CO2 volume concentration in the exhaust gas, see Equation (2.10). This method is commonly used in engine research [22] and [54]. The EGR gas was cooled down before blending with air, then introduced into the intake.

CO EGR = 2inlet (2.10) CO2outlet

When water is condensed and trapped in the system, this method will yield a minor

error in the calculated EGR ratio. O2 concentration and specific thermal capacity of inducted charge confirmed to be the two main parameters to control NOx emissions in diesel engines [55]. In future experiments, to have more precise control, controlling the intake oxygen concentration rather than EGR ratio is recommended.

Air excess ratio (λ) In this thesis, λ is calculated from the Brettschneider Equation (2.11) [56], which provides the λ-value based on exhaust emissions and fuel composition. In this

equation, [XX] is the gas concentration in volume, HCV is the atomic ratio of

to carbon in the fuel, OCV is the atomic ratio of oxygen to carbon in the fuel, and Cf is the number of carbon atoms in each of the hydrocarbons molecules being measured by Horiba Mexa 7100 DEGR. The λ-value calculated by Equation (2.11) is accurate within 1%, independent of the test fuel [56], when all input parameters are accurately known. This equation compares all of the oxygen in the numerator and all of the sources of carbon and hydrogen in the denominator. λ obtained as a dimensionless value to describe different mixtures: stoichiometric λ = 1, lean combustion λ > 1 and rich combustion λ < 1.

22 2.4. Definition of Related Parameters

! CO NO HCV 3.5 OCV  [CO2] + [ ] + [O2] + [ ] + × − × [CO2] + [CO] 2 2 4 3.5+ [CO] 2 CO2 λ =     (2.11) HCV OCV  1 + 4 − 2 × [CO2] + [CO] + Cf × [HC] Rate of Heat Release (RoHR) Figure 2.12 displays different combustion phases. Premixed combustion phase (bc) has the high heat-release rates characteristic. In this phase, fuel and air are better mixed during the ignition delay period and combustion occurs rapidly in a few crank angle degrees. In the mixing-controlled combustion phase (cd), the burning rate is mainly controlled by the fuel/air mixing process. The heat release rate in the mixing-controlled combustion phase may or may not reach a second peak, which is usually lower than the peak formed in the premixed phase [57]. The late combustion phase (de) has a lower heat release rate, because the kinetics of the final burnout processes become slower as the in-cylinder temperature reduces during the expansion stroke. In this thesis the RoHR is calculated with Equation (2.12), based on the first law of thermodynamics assuming the contents of the cylinder can be modeled as an ideal gas. Note that derivatives of pressure and volume (V) with respect to crank angle (α) can easily be calculated from the pressure trace and a combustion chamber volume model, respectively. In addition, the ratio of specific heats (γ) is correlated with in-cylinder temperature, which is also taken into account.

Figure 2.12: Typical heat release rate diagram of compression ignition diesel com- bustion process. Figure is adapted from [57].

dQ γ dV 1 dp = p + V (2.12) dα γ − 1 dα γ − 1 dα Premixed fraction Figure 2.13 shows an example of the heat release rates of a primary ref- erence fuel with a RON of 70 (PRF70) at three different loads without using EGR and the fitted Gaussian profiles represent the premixed combustion. The combustion

23 Chapter 2. Experimental Apparatus

of the PRF70 is dominated by premixed combustion at low load. However, it is dominated by mixing-controlled combustion at high load. And at medium load PRF70 experiences the intermediate combustion process from mainly premixed combustion to mainly mixing-controlled combustion. The premixed combustion phase is controlled by chemical kinetics, and its volumetric heat release is faster than mixing-controlled combustion phase. The premixed fraction (PF) is introduced in the previous study [58], [59] and [6], and defined as a ratio of heat release from premixed combustion to the total heat release. Premixed- and diffusive- heat release rates are typically not separated, a Gaussian profile as presented in Equation (2.13) is fitted to the heat release rate to estimate the premixed heat release. The Gaussian profile is a mathematical representation of the premixed reaction phase. Still, it provides a convincing measure of the premixing degree. 2 (x−x0) − G (x) = h · e 2·α2 (2.13)

In Equation (2.13), x0 represents the central position of the peak, and h and α stand for the height and width of the Gaussian profile, respectively.

600 D ] 600 600 HRR CA D ] CA D ] Gaussian fit 400 400 400

200 200 200

0 0 0

-200 -200 -200 Rate of Heat Release [J/ Rate of Heat Release [J/ -20 -10 0 10 20 30 Rate of Heat Release [J/CA -20 -10 0 10 20 30 -20 -10 0 10 20 30 Crank Angle [CAD ATDC] Crank Angle [CAD ATDC] Crank Angle [CAD ATDC]

Figure 2.13: Heat release rates (red line) and Gaussian profiles (blue dashed line) of PRF70 at three different loads (load increases from left to right, 5 bar, 10 bar and 15 bar IMEPg, respectively) without using EGR.

2.5 Conclusions

In this chapter, the experimental equipment was described and definitions of related pa- rameters were given. Due to the layout of the engine, BMEP is not used in this thesis. To evaluate the combustion performance at different intake and exhaust pressures, IMEPg is used to present engine load and calculate the indicated emissions. Experimental results will be presented in Chapters 3, 4, and 5.

24 Chapter 3 Partially Premixed Combustion with RON70 Fuels

This chapter starts with a comparison of the effects of different injectors on combustion and emission characteristics for conventional diesel combustion. Then the influence of air-excess (by adjusting the boost level) on PPC combustion using different fuels of RON 70 with the main focus on combustion characteristics and regulated exhaust emissions are presented. Finally, the limitations of PPC for NOx-soot trade-off at higher loads will be discussed. 1

3.1 Comparison of Injectors

This study is not originally intended to compare the effects of different nozzle parameters on the combustion and emission results. But injector A was broken after finishing some experiments (the results are shown in Section 3.2) and it had to be replaced with a different nozzle. Injector B (not a new injector) was replaced temporarily to continue the experiments. But its soot emissions are much higher than injector A, which will be discussed in detail afterward. Delphi could not offer the exactly same injector as injector A in a limited time period, so injector C which is currently in production was chosen and installed in Cyclops

1 The major content of this chapter is retrieved from the following papers: Wang, S., et al. (2015). Study on low temperature and conventional diesel combustion with fuel blends of RON70. European Combustion Meeting, Budapest. Wang, S., et al. (2014). Effect of air-excess on partially premixed combustion with blends of RON70. Proceedings SPEIC14 – Towards Sustainable Combustion, November 19-21, 2014, Lisboa, Portugal (pp. 1-8). Wang, S., et al. (2015). Effect of Air-excess on Blends of RON70 Partially Premixed Combustion. Flow Turbulence Combustion. doi:10.1007/s10494-015-9685-2.

25 Chapter 3. Partially Premixed Combustion with RON70 Fuels to do experiments (the results are shown in Section 3.3 and Chapters 4 and 5). To assess the consequences of theses changes, this study was performed.

3.1.1 Experimental Details Injector A was used mostly for HCCI experiments, so a high intake temperature was always used. The comparison of different injectors was done with diesel at an engine speed of 1250 rpm to estimate the impact of nozzle design. Fuel injection pressure and intake temperature were set to 1500 bar and 90 oC, respectively. EGR and λ were controlled at 45% (± 2%) and 1.45 (± 0.05), respectively. CA50 was kept constant at 10 CAD ATDC at different loads for the purpose of investigating the emission characteristics in the same combustion setting for the different injectors.

3.1.2 Differences between the Three Injectors The specifications of the three injectors are listed in Table 3.1. It is widely known that Diesel injector nozzles with different geometries affect the spray atomization and fuel/air mixing process, and directly influence the combustion and exhaust emissions [60], [61], [62], [63], [64], [65], [66], [67], [68] and [69]. The hydraulic flow rate (HFR) of an injector is defined as the flow volume per minute at 100 bar injection pressure, and it is measured with a pressure source and not a completed injector. A longer injection duration is required to inject the specific amount of fuel into the cylinder if the HFR of an injector is smaller than other injectors [68]. Therefore, as shown in Figure 3.1, injector A has the longest injection duration due to the lowest HFR, followed by injector C and injector B.

Table 3.1: Specifications of the three injectors

Injector A Injector B Injector C Nozzle hole number [-] 8 7 7 Injector nozzle hole size [mm] 0.151 0.207 0.180 Included spray angle [◦] 153 150 139 Hydraulic flow rate [L/min] 1.2 1.6 1.4

Injector A has more nozzle holes and smaller nozzle hole diameters compared with Injectors B and C. Figure 3.2 shows that injector A produces lower HC emissions compared to injectors B and C because of the better fuel atomization and air utilization, which is supported by previous studies [66] and [68]. Figure 3.3 shows that injector A produces the least soot emissions as well, because of the larger number of nozzle holes and improved fuel atomization due to the smaller orifice diameter. These three injectors show similar ignition dwells (CA10-EOI) but different ignition delays (CA10-SOI), as can be seen in Figure 3.4 and Figure 3.5, respectively. Although combustion starts at a similar crank angle before EOI for the three injectors (see Figure 3.4), the combustion of injectors B and C starts sooner after SOI. The longer ignition delay, and hence longer premixing, explains the lower soot emissions of injector A. Injector B injects the specific amount of fuel in less time due to its larger HFR,

26 3.1. Comparison of Injectors thus less fuel/air mixing and this promotes the soot formation. It is noted that the injector B is an old injector, so the experimental results can also be influenced by a progressive aging of the nozzle. From literature it it known that a heavier aging effect will cause a decrease in lift-off length and vapor phase penetration, even reduces the spray momentum [69].

16

14

12

10

8 Injection Duration [CAD] Duration Injection 6 Injector A Injector B Injector C 4 4 6 8 10 12 14 16 IMEPg [bar]

Figure 3.1: Injection durations of three injectors at different loads.

0.45 Injector A Injector B 0.4 Injector C

0.35

0.3

0.25

0.2 ISHC [g/kWh] ISHC

0.15

0.1

0.05 4 6 8 10 12 14 16 IMEPg [bar]

Figure 3.2: HC emissions of three injectors at different loads with diesel.

27 Chapter 3. Partially Premixed Combustion with RON70 Fuels

0.7 Injector A Injector B 0.6 Injector C

0.5

0.4

0.3 PM[g/kWh]

0.2

0.1

0 4 6 8 10 12 14 16 IMEPg [bar]

Figure 3.3: Soot emissions of three injectors at different loads with diesel.

2 Injector A Injector B 0 Injector C

-2

-4

-6 Ignition dwell [CAD] dwell Ignition

-8

-10 4 6 8 10 12 14 16 IMEPg [bar]

Figure 3.4: Ignition dwells of three injectors at different loads with diesel.

28 3.2. Effect of Air-excess

7

6.5

6

5.5

5 Ignition delay [CAD] delay Ignition

4.5 Injector A Injector B Injector C 4 4 6 8 10 12 14 16 IMEPg [bar]

Figure 3.5: Ignition delays of three injectors at different loads with diesel.

3.2 Effect of Air-excess

Diesel and gasoline are mixtures of many compounds, including alkanes, alkenes, cycloalka- nes, and aromatics [70] and [71], which very depend on refinery sources and time of year. As introduced in Chapter 1, the optimal RON for gasoline PPC is between 70 and 85. To reduce the ambiguity when one wants to compare experimental results to modeling, test fuels with available mechanisms are preferred. N-heptane and iso-octane (or their blends) are very good candidates to model the combustion of diesel and gasoline [35]. Ethanol and butanol are also very suitable candidates in this respect. At some countries, gasoline is mixing with ethanol (10 vol%) to reduce the carbon footprint. Based on these factors, three blends with RON70 were designed and tested. Intake pressure changes both the in-cylinder charge density and oxygen concentration. Therefore, it is of interest to investigate the sensitivity of combustion characteristics of the three test blends to the change of air-excess ratio.

3.2.1 Experimental details

Test Fuels

It is well known that octane number can affect engine performance range, combustion characteristics and emissions of PPC engines. The higher the octane number, the higher temperature the fuel can withstand before igniting. Alcohols, especially ethanol and n-butanol, have been recognized as suitable fuels for PPC engines due to their high octane numbers. In this study, three fuel blends with RON70 were designed and tested: n-heptane/ iso- octane (PRF70), n-heptane/ iso-octane/ ethanol (ERF70) and n-heptane/ iso-octane/ n-butanol (BRF70). The lower heating values (LHV) of the three blends are different due to the different

29 Chapter 3. Partially Premixed Combustion with RON70 Fuels

chemical compositions. However, the energy content of a stoichiometric mixture (LHVstoi) is nearly identical. Hence, the combustion temperature can be expected to be similar as well. The related properties of the test fuels are shown in Table 3.2.

Table 3.2: The properties of the test blends

Blend properties PRF70 ERF70 BRF70 N-heptane (vol%) 30 43 35 Iso-octane (vol%) 70 37 40 Additive - Ethanol N-butanol Oxygen (wt.%) 0 7.77 6.11 Density (kg/cm3) 687.0 703.2 712.7 RON/MON 70/70 72.9/61 72.9/68 RON test method DIN EN ISO 5164 MON test method DIN EN ISO 5163 Anti-knock index 70 67 70.5 Stoichiometric air-fuel ratio 15.11 13.75 13.99 LHV (MJ/kg) 44.47 41.00 41.27 LHVstoi (MJ/kg) 2.76 2.78 2.75

Test Conditions

In this study, injection pressure and injection duration were kept constant for the three blends. Injector A was used with a single injection strategy and an injection pressure of 1000 bar was applied. However, for different test fuels the fuel mass injected into the cylinder varies a little due to the different densities of the test fuels, as shown in Figure 3.6. Still, the input fuel energies of the three blends are similar. Thus the aim load 7 (±0.3) bar (IMEPg) was achieved. Experiments were conducted with different air-excess ratios, which are altered by changing air mass through varying the intake pressure. During the measurements, the engine speed was fixed at 1250 rpm with 40% EGR. The intake temperature was controlled at 40 ◦C to avoid the impact of intake temperature changes on combustion and emissions. As can be seen from Figure 3.7 and Figure 3.8, the maximum pressure rise rate (PRR) and combustion efficiency increase as combustion phasing (CA50) advances. PRR should not exceed a certain limit to prevent engine damage and excessive combustion noise. In this study, the maximum PRR is controlled not to exceed 20 bar/CAD. For this reason CA50 has been kept constant at 8 CAD to restrict the maximum PRR, and maintain a relatively good combustion efficiency. Hence, the emission characteristics of these blended fuels are investigated at the same combustion setting for each point. λ varies from 1.2 to 2.3 through adjusting the intake pressure. The operating points are plotted in Figure 3.9 (a). The combustion phasing of ERF70 becomes more and more difficult to control as λ decreases towards stoichiometric conditions compared to PRF70 and BRF70, as shown in Figure 3.9 (b). Moreover, note that PRF70 and BRF70 require a higher boost to achieve the same λ as ERF70 due to their lower oxygen content.

30 3.2. Effect of Air-excess

Figure 3.6: Fuel mass and fuel energy per cycle as a function of λ

Figure 3.7: Maximum PRR vs CA50

31 Chapter 3. Partially Premixed Combustion with RON70 Fuels

Figure 3.8: Combustion efficiency vs CA50

Figure 3.9: (a) Changing λ through changing intake pressure. (b) The corresponding CA50 when λ changes.

32 3.2. Effect of Air-excess

3.2.2 Major Combustion Characteristics

Plots shown in Figures 3.10, 3.11, and 3.12 are the in-cylinder pressures, the rate of heat release rates, and injector current signals of PRF70, ERF70, and BRF70 at various intake pressures, respectively. Apart from the fact that the mixture becomes leaner at higher intake pressure, the higher charge density is also favorable to spray atomization, evaporation, and fuel/air mixing throughout the entire combustion process [72]. In the relevant temperature regime of this work, the ignition delay becomes shorter with decreasing λ at near stoichiometric conditions [73]. However, the decreased intake pressure leads to lower pressure and density levels, consequently, results in longer ignition delays and deteriorate the fuel/air mixture formation. Hence, the observation that fuels become more difficult to ignite near stoichio- metric conditions is a pressure and density effect rather than a λ effect. Therefore, fuels need to be injected early into the combustion chamber to get a better fuel/air mixture at low boost conditions, as shown in Figure 3.13. In this study low temperature heat release (LTHR) happens only at low boost conditions where the fuel needs to be injected early into the combustion chamber. The LTHR is presented as the small peak on the RoHR curve. The premixed combustion phase is typically fast, the injected fuel combusts at once. During the mixing-controlled combustion phase, the combustion speed is determined by the rate at which mixture becomes available for burning, typically followed by an even slower late combustion phase. When the absolute intake pressure is set to 2.0 bar, PRF70 starts to show a two-stage combustion characteristic which is referred to as the combination of premixed combustion and mixing-controlled combustion, as shown in Figure 3.10. This is due to the fact that the ignition delay becomes so short at high boost conditions that injection and combustion starts to overlap. However, ERF70 and BRF70 still show almost pure premixed combustion. When the absolute intake pressure is set to 1.3 bar, ERF70 needs to be injected earlier into the combustion chamber than PRF70 and BRF70 due to its higher resistance against auto-ignition (Figure 3.13). Moreover, the heat release is faster and narrower at low boost conditions, which leads to a higher global combustion temperature.

The premixed fractions of the three blends decrease with the increase of λ, which is well illustrated in Figure 3.14. Although these three blends have a similar RON around 70, there is a clear difference in the premixed fraction. A lower λ corresponds to a lower intake pressure which results in a longer ignition delay and thus a higher premixed fraction. While larger λ requires higher intake pressure, which can improve the mixture formation but shorten the ignition delay, thus leading to lower premixed fractions. Fuel with highest chemical ignition delay (ERF70) allows earlier injection and provides more time for mixture formation, thus presenting the highest premixed fraction and lowest dependency on changing conditions. It is concluded that the observed dependence of the premixed fraction on λ is a complex interaction between chemical and physical aspects since in the experimental work not only λ but at the meantime boost conditions are also varied.

33 Chapter 3. Partially Premixed Combustion with RON70 Fuels

Figure 3.10: Cylinder pressures and RoHR data of PRF70 at various intake pressures

Figure 3.11: Cylinder pressures and RoHR data of ERF70 at various intake pressures

34 3.2. Effect of Air-excess

Figure 3.12: Cylinder pressures and RoHR data of BRF70 at various intake pressures

0.05 0 PRF70 PRF70 0.04 ERF70 -5 ERF70 BRF70 BRF70

0.03 -10

0.02 -15 PM[g/kWh]

EURO VI ATDC] [CAD SOI 0.01 -20

0 -25 1.2 1.4 1.6 1.8 2 2.2 1.2 1.4 1.6 1.8 2 2.2 [-] [-]

Figure 3.13: Start of injection as a function of λ

35 Chapter 3. Partially Premixed Combustion with RON70 Fuels

Figure 3.14: Premixed fraction and linear polynomial profile as a function of λ. The linear polynomial profile is a fit for the premixed fraction.

CA50 is near constant for all the measurements by design, but clear trends can be seen with respect to the ignition dwell (Figure 3.15) and burn duration (Figure 3.16) of the three blended fuels. In Figure 3.15 and 3.16, the dotted line a, line b, and line c connect the same ignition dwell points of the three blended fuels under various intake pressures, respectively. The ignition dwell is used to describe the separation between the fuel injection and combustion events. The target of PPC concept is to separate the end-of-injection and the start-of-combustion. Therefore a positive ignition dwell avoids local fuel-rich zones which can cause soot emission. Ignition dwells decrease when λ increases, this correlates strongly with the change of premixed fraction. ERF70 has the longest ignition dwell and the biggest premixed fraction. Consequently, its heat release rate is higher than the other test blends because of the faster combustion speed of premixed combustion. Figure 3.16 presents how the burn duration changes as a function of λ. The three test blends have a similar RON of 70, but ignition at PPC conditions is not determined by RON alone. The stratification will have a pronounced effect. Intake pressure increases the in-cylinder density and oxygen concentration at the same time. On one hand, higher boost levels will reduce the premixed time of fuel and air thus resulting in more mixing-controlled combustion and unequivocally extends the burn duration. One the other hand, a higher density will accelerate the mixing of oxygen and fuel [74] to reduce the mixing-controlled combustion phase and thus shorten the burn duration. Hence the burn duration is affected by these two factors together. When having the same ignition dwell, ERF70 has the shortest burn duration followed by BRF70 and PRF70, as shown in Figure 3.16. So even though at these conditions the same amount of fuel is injected, it is impossible to conclude that the burn duration is associated with the richness of the fuel alone. Since the combustion of the test fuels occur at different oxygen concentration hence

36 3.2. Effect of Air-excess different stratification. In sum, the difference in burn durations of the three blends is thought to result from different burning rates and thermal stratifications.

Figure 3.15: Ignition dwell as a function of λ

Figure 3.16: Burn duration as a function of λ. The dotted lines a, b, and c connect the same ignition dwell points for various intake pressures.

37 Chapter 3. Partially Premixed Combustion with RON70 Fuels

3.2.3 Soot and NOx Emissions Soot emissions of the three blends are all below EURO VI emission standard, as plotted in Figure 3.17. It can be concluded from the soot emissions that reduced λ does not cause more fuel-rich regions because of the extended mixing time. Reduced λ requires an earlier SOI because fuels become difficult to ignite due to the decreased boost level and in-cylinder charge density at the close-to-stoichiometric conditions. NOx emissions increase with the increase of λ, as shown in Figure 3.18. Previous research concluded that NOx levels were determined by the flame temperature, the oxygen concentration, and the residence time at the NOx forming period. As the intake pressure increases, the ignition delay decreases resulting in more mixing-controlled combustion, so the peak in-cylinder temperature drops. One may expect NOx emissions to decrease with the increasing dilution level due to the reduced global temperature. However apparently, the higher oxygen concentration and longer burn duration have a bigger effect under the experimental conditions. This is also found by Noehre et al. [22] that NOx emissions increase with intake pressure for a fixed EGR rate. The influence of intake pressure on NOx emissions is small compared to the influence of oxygen concentration [75]. Therefore, it is possible to significantly improve the NOx emissions at high boost conditions, such as applying more EGR to control the oxygen concentration. This study changed the intake oxygen concentration by changing the intake pressure, which inevitably change the in-cylinder density as well. In the next section, the intake oxygen concentration is changed by changing the EGR rate to investigate the influence of intake oxygen concentration on soot and NOx emissions at higher loads.

Figure 3.17: PM emissions as a function of λ

38 3.3. Limitations of PPC at higher loads

Figure 3.18: NOx emissions as a function of λ

3.3 Limitations of PPC at higher loads

3.3.1 Experimental details

The properties of the test fuels are shown in Table 3.2. Diesel was also tested as a reference. Before the tests start, the engine is warmed up until the lubrication oil and coolant fluid temperatures reach 85 oC. All the tests were conducted under steady state conditions, injector C and a single injection strategy were used. Two loads were tested, which are medium load (IMEPg of 10 bar) and a higher load (IMEPg of 15 bar), respectively. To achieve different loads and also keep injection duration short, various injection pressures were selected for the medium load (1200 bar) and high load (1800 bar). Different EGR rates were used for each load to adjust the intake oxygen concentration. Intake pressure was adjusted as well with the change of load (1.8 bar at medium load and 2.5 bar at high load) to ensure λ is not lower than 1.2 even when a high EGR rate (above 40%) was used, thus avoid excessive soot emissions. Due to the fact that the SOI timing will influence the air/fuel mixing, in this study the SOI timings for different test fuels were kept constant, while the SOI timing was changed for different loads to maintain stable combustion. As a consequence CA50 changes from 8 to 15 CAD ATDC when EGR increases. The engine speed was held constant at 1200 rpm which is a typical cruising speed (for older engines), and intake temperature was controlled at 40 oC.

39 Chapter 3. Partially Premixed Combustion with RON70 Fuels

3.3.2 Combustion Characteristics

Figure 3.19: Heat release rates at medium and higher loads with 0% EGR. The dotted lines represent the injector currents.

Figure 3.20: Heat release rates at medium and higher loads with 40% EGR. The dotted lines represent the injector currents.

Figure 3.19 and Figure 3.20 depict the heat release rates at two test loads with 0% EGR and 40% EGR, respectively. At 10 bar IMEPg, diesel shows both premixed and mixing-controlled combustion, and mixing-controlled combustion dominates the whole combustion process, thus exhibiting a longer burn duration than the blends. While the three blends provide more premixed combustion than diesel (at 10 bar IMEPg), the proportion of premixed combustion increases apparently as EGR rate grows. At the higher load (15 bar IMEPg), diesel presents a typical two-stage combustion process, a small amount of premixed combustion followed by a substantial amount of mixing-controlled combustion. Even though the blends of RON70 offer a bit more premixed combustion than diesel at the higher load, they are also controlled by mixing-controlled combustion, as is also apparent from the overlap between injection and

40 3.3. Limitations of PPC at higher loads combustion. Note that the heat release rate graphs illustrate that the combustion durations of the three blends are shorter than that of diesel leading to more expansion work. In Figure 3.21, the premixed fraction is presented as a function of the EGR percentage at two test loads. Diesel is dominated by mixing-controlled combustion at both medium and higher loads with premixed fractions smaller than 0.2 even if over 40% EGR is used. At 10 bar IMEPg, the three blends of RON70 present more premixed combustion as the EGR rate increases. They are dominated by premixed combustion with premixed fractions larger than 0.5 when over 30% EGR is used. However, at 15 bar IMEPg, the three blends all burn in a conventional diesel combustion mode with premixed fractions smaller than 0.3 (although at least twice that of diesel). This is due to the fact the ignition delay at 15 bar IMEPg is shorter than that of 10 bar IMEPg. At 15 bar IMEPg, thus higher intake pressure, the cylinder holds more mass and consequently will have better atomization and mixing. Moreover, as engine load increases, the residual gas temperature and the wall temperature increase [76]. This results in higher air temperature at injection, thus shortening the ignition delay and premixed fraction as well. EGR reduces the charge temperature which increases the ignition delay and hence the premixed fraction. This also holds for the medium load case (10 bar IMEPg).

IMEPg : 10bar IMEPg : 15bar 1 1 PRF70 PRF70 ERF70 0.8 0.8 ERF70 BRF70 BRF70 Diesel Diesel 0.6 0.6

0.4 0.4 Premixed fraction [-] fraction Premixed Premixed fraction [-] fraction Premixed 0.2 0.2

0 0 0 10 20 30 40 50 0 10 20 30 40 50 EGR [%] EGR [%]

Figure 3.21: Premixed fractions at medium and higher loads with various EGR rates.

3.3.3 NOx-Soot Trade-off Figure 3.22 illustrates the trade-off relationship between NOx and soot emissions for the test fuels at two test loads. At 10 bar IMEPg, the NOx emissions of all test fuels reduce at the increase of EGR rate, and NOx emissions are below 0.4 g/kWh (EURO VI) when more than 40% EGR is used. In addition, the soot emissions of the three blends are close to zero regardless of the EGR rate used. However, the soot emissions of diesel increase sharply when more than 25% EGR is used as is commonly known. Under such experimental conditions, the three blends of RON70 are dominated by premixed combustion, especially at high EGR

41 Chapter 3. Partially Premixed Combustion with RON70 Fuels rate. In fact the premixed fractions of the three blends exceed 0.5 when more than 30% EGR is used, while the premixed fractions of diesel are always lower than 0.2 even if more than 40% EGR is used. Nearly zero soot emissions of the three blends demonstrate that the better mixing of fuel and air outweighs the effect of reduced oxygen concentration on soot emissions as the EGR rate increases.

However, at 15 bar IMEPg, all fuels show a NOx and soot trade-off relationship, but oxygenated fuels (ERF70 and BRF70) still produce less soot. Previous experimental and numerical studies revealed that the ability of oxygenates to limit the soot production was strongly related to their molecular structure in addition to their oxygen content [77], [78], [79], and [80]. Fuel molecular structures including branching, the degree of saturation, carbon chain length, and oxygenate functional group all have effects on soot formation [77]. For a given number of carbon- in the molecule, the sooting tendency of oxygenated compounds increases in the following order: esters < aldehydes < ketones < alcohols [80]. At 15 bar IMEPg, all test fuels are dominated by mixing-controlled combustion and run under conventional diesel combustion mode with a premixed fraction smaller than 0.3. Under such experimental conditions, the reduction in oxygen concentration (when EGR rate increases) has a decisive influence on soot formation. The oxygen content of ethanol (in ERF70) or n-butanol (in BRF70) helps to increase the local λ, thus reducing the fuel-rich regions to produce less soot. In addition, their short carbon chain length also reduces soot emissions. It is also noticed that overall NOx emission goes up as the load increases, which is attributed to the higher in-cylinder pressure and thus higher temperature. The experiments show that it is almost impossible to achieve ultra-low NOx and soot emissions simultaneously. In other words, it is hard to increase the proportion of premixed combustion sufficiently. Therefore more research should be done to improve the premixed fraction and alleviate the NOx and soot trade-off at higher loads.

IMEPg : 10bar IMEPg : 15bar 14 1.4 25 0.7 PRF70 12 1.2 ERF70 0.6 20 BRF70 10 1 Diesel 0.5

8 0.8 15 0.4

6 0.6 10 0.3

PM [g/Kwh] PM [g/Kwh] PM NOx [g/kWh] NOx NOx [g/kWh] NOx 4 0.4 0.2 5 2 0.2 0.1

0 0 0 0 0 10 20 30 40 50 0 10 20 30 40 50 EGR [%] EGR [%]

Figure 3.22: NOx and PM vs EGR at two test loads

42 3.4. Conclusions

3.4 Conclusions

In the first part of this chapter, three injectors with different nozzle geometries are tested with diesel at several engine loads. Injector A produces the least soot emissions which is thought to be caused by the larger nozzle hole number and smaller nozzle hole size. Several important parameters (including nozzle hole number, nozzle hole size, included spray angle, hydraulic flow rate) are changed at the same time for injector A, B, and C. Therefore, it is difficult to conclude which nozzle geometry has the most important influence on spray formation based on our experimental data. Choosing the right nozzle is an important means of optimizing the combustion. At the same time, it is of great practical significance to study the effect of nozzle aging on the formation of the spray. This is also an important branch of future research. In the second part of this chapter, different air-excess ratios are achieved by changing the intake pressure to investigate the combustion and emission characteristics of three blends with RON 70 under PPC mode. Even though the three blends have a similar RON of 70, the ignition of PPC conditions is apparently not determined by RON alone. ERF70 needs to be injected earlier into the cylinder than PRF70 and BRF70 to achieve the same CA50, leading to the highest premixed fraction. Premixed fractions of the test blends decrease with the growth of intake pressure. At close to stoichiometric conditions, all the blends show a large premixed fraction (> 0.8) and their combustion is dominated by premixed combustion. Soot emissions under PPC mode are well below Euro VI emission regulations. In the third part of this chapter, the NOx-soot relationship is investigated at two loads (10 and 15 bar IMEPg, respectively) using three blends of RON 70. Diesel suffers a well-known NOx-soot trade-off relationship, so it is tested under the same experimental conditions as a reference. At 10 bar IMEPg, the test blends are dominated by premixed combustion, having premixed fractions larger than 0.5 when 30% EGR is used. However, diesel is dominated by mixing-controlled combustion and its premixed fraction is smaller than 0.2 even when 45% EGR is used. The test blends eliminate the trade-off relationship between NOx and soot emissions. As the EGR rate increases, the NOx emissions of all test fuels decrease dramatically, besides, the three blends maintain near zero soot emissions. But the soot emissions of diesel increase sharply when more EGR is applied. At 15 bar IMEPg, all fuels show a NOx-soot trade-off relationship, but oxygenated fuels (ERF70 and BRF70) still produce less soot emissions. To alleviate or even eliminate the NOx-soot trade-off at higher loads, methods that help to improve the proportion of premixed combustion or premixed fraction should be used, such as multiple injection strategies. Besides, the effect of increasing the proportion of ethanol or n-butanol in the blends on soot emissions also requires further study. The study indicates that more premixed combustion, or a bigger premixed fraction, helps to reduce soot emission. Whether premixed fraction can be used to predict the amount of soot emission requires further study. This chapter focuses on the combustion characteristics and the soot-NOx trade-off relationship of blends with RON70. Fuels with even higher RON are believed to increase the premixed fraction, but they may bring problems at lower loads due to their poor auto-ignition qualities. The next chapter will discuss this in more detail.

43

Chapter 4 Low Load Tests with High-octane Fuels

Higher octane number fuels strengthen the ignition delay advantage over diesel at high load, while they will have difficulties to ignite at low load. This chapter is investigating the potential of fuels with high octane number at low load and idle conditions. Several chosen PRF blends, which are blends of iso-octane and n-heptane with octane numbers of 70, 80 and 90, were tested in an adapted commercial diesel engine (Cyclops). Firstly, experiments were done with varying combustion phasings and intake temperatures without using EGR to study the combustion stability of several high-octane fuels at low load and idle conditions. Next, the campaign was extended by adjusting EGR rate and intake pressure to optimize NOx emissions. 1

4.1 Experimental Details

At Lund University, Manente [81] and [82] investigated the PPC operating region running a variety of gasoline fuels with various octane numbers in a heavy duty diesel engine with a compression ratio of 17.3. Figure 4.1 [82] shows the operating window used in their experiments. It is concluded that the operating region which can be realized decreases with fuel octane number when an intake temperature of 30 oC is used [81] and [82]. Some approaches can be tried to expand the operating region, such as using refined injection strategies, using boosted intake air, or using heated intake air. As shown in Chapter 3, at higher loads PPC is hard to establish and a higher octane number is preferred. The primary objective of this study is to investigate the low load limitations of high-octane fuels by changing the intake temperature.

1Part of this chapter is retrieved from: Wang, S., et al. Experimental Study on the Potential of Higher Octane Number Fuels for Low Load Partially Premixed Combustion, SAE Paper 2017-01-0750, 2017, doi:10.4271/2017-01-0750.

45 Chapter 4. Low Load Tests with High-octane Fuels

Figure 4.1: Stable operational load window as a function of RON [81], the color symbols represent fuels with different RONs and their properties can be found in [82].

4.1.1 Test Fuels Three PRF blends, which are mixtures of n-heptane and iso-octane, with RON of 70, 80, and 90 were tested, respectively. Diesel was also tested as a reference under the same experimental conditions. The main properties of the test fuels are listed in Table 4.1.

Table 4.1: Main properties of the four test fuels

PRF70 PRF80 PRF90 EN590 Diesel Iso-octane (vol%) 70 80 90 - N-heptane (vol%) 30 20 10 - RON (-) 70 80 90 - CN (-) 27 22.6 16.9 51.6 LHV (MJ/kg) 44.47 44.45 44.44 42.9 Stoichiometric air-ful ratio 15.107 15.102 15.096 ≈14.6 H/C 2.261 2.257 2.254 ≈1.85

4.1.2 Test Conditions — Effects of Intake temperature and CA50 Typically, idling conditions are represented by an engine speed of 600 (±5) rpm and no boost [83]. Experiments were done at two low loads, load I (5 bar IMEPg at 1200 rpm) and load II

46 4.1. Experimental Details

(2.5 bar IMEPg at 600 rpm). Injector C was used for all the experiments in this Chapter. To get a representative comparison between the selected fuels, the combustion phasing (CA50) and intake temperature were selected to vary systematically. Table 4.2 shows the respective ranges for CA50 and intake temperature. The remaining engine settings are given in Table 4.3. EGR was not used initially in order to improve the ignitability of high-octane fuels at the test conditions. The measurement was terminated when the combustion shows a tendency to misfire.

Table 4.2: A full-factorial matrix combination of CA50 and intake temperature for the test fuels

Test fuels CA50 sweep (CAD ATDC)1 Intake temperature sweep (oC)2 PRF70 0∼10 40∼120 PRF80 0∼10 40∼120 PRF90 0∼10 60∼120 Diesel 0∼10 40∼80 1 Change step: 2.5 CAD 2 Change step: 20 oC

Table 4.3: The settings of engine parameters

Engine settings Load I Load II IMEPg (bar) 5 2.5 Injection pressure (bar) 600 600 Injection strategy Single injection Intake pressure (bar) 1.2 1 Engine speed (rpm) 1200 600

As shown in Figure 4.2, CA50 is varied between 0 and 10 CAD ATDC by adjusting SOI for different intake temperatures. As to be expected, PRF90 needs to be injected earlier into the cylinder to achieve the same CA50 as other test fuels at the same engine conditions, especially when a relatively low intake temperature is used. Figure 4.2a shows that the SOI of PRF80 and PRF90 is advanced as the intake temperature reduces to get the same CA50. However, intake temperature has an ignorable effect on the SOI of PRF70 and diesel (Figure 4.2b). Initially, the intake temperature range was planned to change from 40 to 120 oC. But the temperature ranges for PRF90 and diesel do not cover the full initial range, since PRF90 shows a tendency to misfire at 40 oC and diesel extensively above 80 oC. These results also hold for load II. Figure 4.3 displays the λ at two test loads. The λ at load I and load II are larger than 2.1 and 3.1, respectively. λ is little influenced by CA50, the small deviations in λ can be ascribed to the small adjustments of fuel injection quantity, since the fuel mass was fine-tuned for different CA50 values to maintain the set IMEPg and compensate for the changes in efficiency.

47 Chapter 4. Low Load Tests with High-octane Fuels

The global λ reduces when intake temperature is increased mainly due to the reduced air density.

0 2 Tin=40oC Tin=40oC Tin=60oC o 0 Tin=60 C Tin=80oC Tin=80oC Tin=100oC Tin=100oC o -2 o -5 Tin=120 C Tin=120 C -4

-6 -10 -8

PRF80

-10 PRF70 Start of injection [CAD ATDC] [CAD injection of Start Start of injection [CAD ATDC] [CAD injection of Start PRF90 Diesel -15 -12 -2 0 2 4 6 8 10 12 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC] CA50 [CAD ATDC]

(a) Load I with PRF80 and PRF90 (b) Load I with PRF70 and diesel

Figure 4.2: SOI vs CA50 at load I

3 4.5

2.8

2.6 4

2.4

2.2 3.5 [-]

[-] 2

6 6 1.8 3

1.6 Tin=40oC Tin=40oC o PRF70 o PRF70 1.4 Tin=60 C 2.5 Tin=60 C Tin=80oC PRF80 Tin=80oC PRF80 PRF90 PRF90 Tin=100oC Tin=100oC 1.2 Diesel Diesel Tin=120oC Tin=120oC 1 2 -2 0 2 4 6 8 10 12 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC] CA50 [CAD ATDC] (a) Load I (b) Load II

Figure 4.3: λ vs CA50 at two test loads

4.1.3 Test Conditions — Optimize NOx Emissions Based on the results of the previous studies on intake temperature and CA50, intake tempera- ture was set to 40 oC for all fuels except PRF90 and CA50 was set at 6 CAD ATDC for all fuels. As mentioned before, in order to ensure stable combustion, the intake temperature of PRF90 was set to 60 oC. EGR was used to reduce NOx emissions, and in the meantime, intake pressure was adjusted accordingly to reduce HC and CO emissions and also investigate whether the increased intake pressure can help to avoid the growth of soot emissions. Table

48 4.2. Results and Analysis — Effects of Intake temperature and CA50

4.4 shows the range of variations of EGR and intake pressure, and the engine settings are displayed in Table 4.3.

Table 4.4: EGR and intake pressure settings

Low load Idle condition Fuels EGR (%) Pin sweep (bar) Pin sweep (bar) 10 1.2∼1.6 1.0∼1.7 20 1.2∼1.6 1.3∼1.7 PRF70 30 1.2∼1.6 1.3∼1.8 40 1.2∼1.7 1.3∼2.0 10 1.2∼1.6 1.0∼1.7 20 1.2∼1.7 1.3∼1.8 PRF80 30 1.2∼1.9 1.3∼1.9 40 1.2∼2.0 1.3∼2.0 10 1.2∼1.7 1.1∼1.7 20 1.2∼1.8 1.3∼1.9 PRF90 30 1.2∼2.1 1.3∼1.9 40 1.2∼2.1 1.3∼2.0 10 1.2∼1.5 1.0∼1.4 20 1.2∼1.5 1.0∼1.4 Diesel 30 1.2∼1.6 1.0∼1.4 40 1.2∼1.6 1.0∼1.4

4.2 Results and Analysis — Effects of Intake tempera- ture and CA50

This section is separated into two main parts. In the first part, the load I case is discussed in terms of combustion stability, ignition dwell, burn duration, emissions, and efficiency. In the second part, the lower load case (2.5 bar IMEPg) with an (600 rpm) is addressed with a similar focus except for efficiency because at idle condition this is of less interest. All experiments were performed with 0% EGR.

4.2.1 Load I

Combustion Stability The coefficient of variation of the IMEPg (COVIMEP) is used to evaluate the cyclic variations. It is widely used in engine studies [57] as defined in Equation

(4.1) where xi and ¯x in Equation (4.2) represent the IMEPg of a single combustion cycle and the mean IMEPg of 50 consecutive cycles (N=50), respectively. q N 2 i=1(xi − ¯x) /N COVx = · 100% (4.1) P ¯x

49 Chapter 4. Low Load Tests with High-octane Fuels

1 N ¯x = · x (4.2) N i i=1 X Figure 4.4 shows the COVIMEP of three PRF blends and diesel. It is seen that PRF90 shows a higher COVIMEP when CA50 is around TDC. The apparent maximum of COVIMEP for PRF90 at an intake temperature of 80 °C is not understood. According to [57] the cyclic variations of IMEP should be lower than 10% to avoid drivability problems. According to a recent study [84], the upper limit value of COVIMEP on 5% is more appropriate. Hence, the test engine running with three PRF blends guarantees stable combustion since their

COVIMEP are well below 5% and even lower than that of diesel at many of the test conditions.

5 Tin=40oC o Tin=60 C PRF70 Tin=80oC 4 PRF80 Tin=100oC PRF90 Tin=120oC Diesel

3

[%]

P

E

M I

2 COV

1

0 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.4: COVIMEP at load I based on 50 consecutive tests

COVIMEP is an important indicator of stable combustion, but not the only indicator, the pressure rise rate (PRR) should also be considered. The security criterion depends much on the engine used. This engine has proven to work with pressure rise rates of 40 bar/CAD. Nevertheless, PRR should not exceed a certain limit to prevent engine damage and excessive combustion noise. The maximum PRR will be lower than 20 bar/CAD and remain on the safe side if CA50 is well controlled after 6 CAD ATDC, as can be seen in Figure 4.5.

Ignition Dwell The ignition dwell of the three PRF blends and diesel at different test conditions is shown in Figure 4.6. A positive ignition dwell can avoid local fuel-rich zones thus producing less soot. As can be observed in Figure 4.6, only PRF90 provides positive ignition dwells in the complete range of combustion phasing regardless the changes of intake temperatures. PRF80 and PRF70 have positive ignition dwells only for late combustion phasing and with low intake temperatures. This correlates well with their ignitabilities.

50 4.2. Results and Analysis — Effects of Intake temperature and CA50

Hence, PRF90 is a reasonable choice if a separation between injection and combustion event is required.

40

Tin=40oC o 35 PRF70 Tin=60 C o PRF80 Tin=80 C PRF90 Tin=100oC 30 Diesel Tin=120oC

25

20

PRRmax [bar/CAD] PRRmax 15

10

5 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.5: Maximum pressure rise rate (PRRmax) as a function of combustion phasing (CA50) and intake temperature for the three PRF blends and diesel at load I. Yellow regions indicates an acceptable PRRmax level.

5

0

Ignition dwell [CAD] dwell Ignition -5

Tin=40oC Tin=60oC PRF70 o PRF80 Tin=80 C PRF90 Tin=100oC Diesel Tin=120oC -10 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.6: Ignition dwell as function of CA50 at load I

51 Chapter 4. Low Load Tests with High-octane Fuels

Burn duration Figure 4.7 demonstrates how burn durations of the test fuels change as a function of CA50 at various intake temperatures. The burn durations of PRF70 and PRF80 increase with the increased intake temperature resulting from the reduced ignition dwell. The higher the intake temperature, the larger the spray driven combustion phase becomes and this results in a longer burn duration. The burn durations of PRF90 are the shortest, this is in agreement with the ignition dwell results, as shown in Figure 4.6. PRF90 provides sufficient premixing and is dominated by premixed combustion. Therefore, it burns faster, and intake temperature is of less importance. The RoHR of diesel at 80 oC with a CA50 around 10 CAD ATDC (condition A in Figure 4.7), together with the RoHR of PRF90 using the same intake temperature and a similar CA50 (condition B in Figure 4.7) are shown in Figure 4.8. It is clearly seen that the combustion of PRF90 separates from the injection event, while the combustion of diesel starts before the EOI and shows a two-stage combustion and thus a longer burn duration.

70

Tin=40oC PRF70 Tin=60oC PRF80 60 PRF90 Tin=80oC Diesel Tin=100oC 50 Tin=120oC

40 A IDw<0

30

Burn duration [CAD] duration Burn 20

10 B 0 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.7: Burn duration plotted against CA50 at load I. Heat release rates of operating conditions A and B are presented in Figure 4.8.

Limitations of PPC at Load I — High HC Emissions Figure 4.9 shows that PRF70 and PRF80 produce higher HC emissions at 40 oC intake temperature particularly in a retarded combustion phasing. Data for PRF90 at an intake temperature of 40 oC is not available due to the unstable combustion. When an intake temperature higher than 80 oC is used, HC emissions of PRF70 and PRF80 remain nearly constant as a function of combustion phasing. However, the HC emissions of PRF90 increase considerably (from 0.7 to 1.3 g/kWh) as CA50 is retarded nearly independent of the intake temperature. According to Sylvain Mendez [85], the main cause of HC emissions is flame quenching, including bulk quenching

52 4.2. Results and Analysis — Effects of Intake temperature and CA50

800 Diesel 700 PRF90 Injector current for Diesel Injector current for PRF90 600

500

400

300

200 Rate of heat release [J/CAD] release heat of Rate

100

0 -20 -10 0 10 20 30 40 50 60 Crank angle [CAD ATDC]

Figure 4.8: Heat release rates of diesel and PRF90 at 80 oC with a CA50 around 10 CAD ATDC.

and wall quenching. PRF90 presents mainly premixed combustion (IDw>0) regardless of intake temperature and combustion phasing. In premixed combustion, bulk quenching is characterized as incomplete oxidation in over-lean regions rather than locally over-rich regions. All tests at load I are done with a global λ higher than 2.1. Besides, the IDw increases slightly as the CA50 is retarded. Figure 4.9 also shows that the HC emissions of the three PRF blends decrease with the increase of intake temperature. This is because a higher intake air temperature facilitates complete combustion. Wall quenching is a process which delays the chemical reaction rates near the combustion chamber walls due to a temperature drop induced by the cylinder wall. PRF90 has the longest IDw and needs to be injected earlier into the cylinder to have the same CA50 as PRF70 and PRF80. Therefore PRF90 has a higher probability of reaching the cylinder wall to produce more HC emissions. However, even the earliest SOI timing of PRF90 is after -15 CAD ATDC (Figure 4.2a), so all fuels are injected in the bowl according to Figure 2.2. Figure 4.9 also displays that PRF90 produces the lowest HC emissions when CA50 is near TDC. PRF90 has the shortest burn duration, so the corresponding maximum combustion temperature is the highest which results in less HC emissions. In this study, the fact that the earlier injection timings produce lower HC emissions indicates that bulk quenching causes the main HC emission rather than wall quenching under the test conditions.

Limitations of PPC at Load I — High CO Emissions CO emissions in engines are mainly caused by incomplete combustion. They are influenced by both in-cylinder temperature and air/fuel mixture equivalence ratio. As shown in Figure 4.10, CO emissions increase as

53 Chapter 4. Low Load Tests with High-octane Fuels

3

Tin=40oC PRF70 Tin=60oC PRF80 PRF90 2.5 Tin=80oC Diesel Tin=100oC Tin=120oC 2

1.5 ISHC [g/kWh] ISHC 1

0.5

0 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.9: HC emissions plotted against CA50 at load I

the combustion phasing is retarded for all PRF blends due to the reduction in in-cylinder temperature caused by the expansion. The maximal average in-cylinder temperature is computed based on the ideal gas law without considering heat losses, to give an indication of the in-cylinder temperature at different test conditions, as shown in Figure 4.11. For conventional diesel combustion, CO emissions are typically formed in fuel-rich regions, and they are easily oxidized to CO2 afterwards due to the high combustion temperature. For PPC, the incomplete combustion of over-lean fuel/air mixture is the dominant mechanism responsible for the CO emissions at low EGR rates [86]. Experimental results demonstrate that CO emissions have a strong dependence on both CA50 and intake temperature. As shown in Figure 4.11, both CA50 and intake temperature have a large impact on the average in-cylinder temperature. When a specific intake temperature is used, such as 60 oC, the fuel with a higher octane number tends to produce more CO emissions. However, at such experimental conditions, the average maximum in-cylinder temperature of a higher octane number fuel is higher, which is expected to reduce CO emissions to a certain extent. As all PRF blends operate with almost the same global λ, the fuel with a higher octane number has the largest ignition dwell leading to more over-lean zones. Under these “over-mixed” zones, local combustion temperature may not be high enough to completely oxidize CO to CO2 on engine time scales [86]. Higher intake temperature results in higher combustion temperature and promotes complete combustion, hence the decrease in CO emissions. Moreover, higher intake temperature also reduces the dilution level, thus forming less over-lean zones. Given the fact that very little soot is formed (see Figure 4.12), apparently, no fuel-rich zones are present. Therefore, the hypothesis that over-leaning is the principal mechanism for CO emissions is strengthened.

54 4.2. Results and Analysis — Effects of Intake temperature and CA50

8

Tin=40oC o 7 Tin=60 C PRF70 o Tin=80 C PRF80 o 6 Tin=100 C PRF90 Tin=120oC Diesel

5

4

ISCO [g/kWh] ISCO 3

2

1

0 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.10: CO emissions plotted against CA50 at load I

2200 o PRF70 2100 Tin=40 C Tin=60oC PRF80 PRF90 Tin=80oC 2000 Diesel Tin=100oC 1900 Tin=120oC

1800

1700

1600

1500

1400

1300

1200 Maximum average in-cylinder temperature [K] temperature in-cylinder average Maximum 1100 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.11: Maximal average in-cylinder temperature plotted against CA50 at load I

55 Chapter 4. Low Load Tests with High-octane Fuels

Soot Emissions As shown in Figure 4.12, three PRF blends produce much less soot compared to diesel, especially for late CA50. According to [87] for PRF blends, the sooting tendency of doped flames (methane diffusion flames doped with the vapors of PRF blends) exhibits a regular and monotonic behavior and reduces with increasing n-heptane mole fraction. The level of soot volume fraction increases steadily with the increase of and branching of the main carbon chain. However, soot emissions increase with increasing amount of n-heptane under the engine test conditions because n-heptane’s low octane number shortens the ignition delay time which weakens the mixing of fuel and air. Fuel-rich zones are recognized to produce more soot emissions, so the mixing degree of air/fuel mixture has a significant influence on the formation of soot emissions. For positive IDw conditions, combustion is dominated by premixed combustion. There is sufficient time to mix fuel and air to reduce fuel-rich regions (λ<1) thus producing less soot emissions. Once a diffusion flame is established, soot emissions will increase significantly due to the inevitable presence of fuel-rich regions. It is found that the conditions where combustion starts before EOI (IDw<0) show diffusive burning and produce relatively high soot emissions. Correspondingly, conditions with a positive ignition dwell all tend to have low soot emissions, even close to zero, and nearly independent of CA50.

0.2

Tin=40oC 0.18 o Tin=60 C PRF70 o 0.16 Tin=80 C PRF80 Tin=100oC PRF90 o Diesel 0.14 Tin=120 C

0.12

0.1

0.08 ISPM [g/kWh] ISPM 0.06

0.04

0.02

0 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.12: ISPM plotted against CA50 at load I

Thermodynamic Efficiency and Fuel Consumption The thermodynamic efficiency (defined in Chapter 2) as a function of CA50 with various intake temperatures for the three PRF blends is plotted in Figure 4.13. The main trend is clear. All PRF blends have comparable thermodynamic efficiencies (from 45% to 48%), which are higher than the thermodynamic efficiency using diesel, independent of combustion phasing. Thermodynamic efficiency

56 4.2. Results and Analysis — Effects of Intake temperature and CA50

50

45

40 Tin=40oC

Tin=60oC PRF70 Thermodynamic efficiency [%] efficiency Thermodynamic Tin=80oC PRF80 PRF90 Tin=100oC Diesel Tin=120oC 35 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.13: Thermodynamic efficiency plotted against CA50 at load I

260

Tin=40oC PRF70 Tin=60oC PRF80 PRF90 250 Tin=80oC Diesel Tin=100oC Tin=120oC 240

230 ISFC [g/kWh] ISFC 220

210

200 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.14: Fuel consumption plotted against CA50 at load I

57 Chapter 4. Low Load Tests with High-octane Fuels decreases when a higher intake temperature is used due to the increased heat losses, despite the increased combustion efficiency resulting from the reduction in HC and CO emissions with higher intake temperatures. Another contribution to thermodynamic efficiency is due to the exhaust heat losses. Unfortunately, the temperature of exhaust gases is not measured in this study. PRF80 shows the best thermodynamic efficiency when an intake temperature of 40 oC is used. Figure 4.14 displays that under such experimental conditions the indicated specific fuel consumption (ISFC) of PRF80 is the lowest. Moreover, PRF80 has a COVIMEP lower than 2% and can produce near zero soot emissions.

4.2.2 Load II

Combustion Stability Figure 4.15 and Figure 4.16 display the COVIMEP and PRRmax at load II (2.5 bar IMEPg at 600 rpm), respectively. Combustion at low load is more stable compared to that at idle condition. Still, the combustion stability of the three PRF blends is guaranteed at idle condition if CA50 is controlled after 6 CAD ATDC, since the COVIMEP then remains well below the 5% limit and maximum PRR is below 20 bar/CAD.

8

Tin=40oC PRF70 7 Tin=60oC PRF80 PRF90 Tin=80oC o Diesel 6 Tin=100 C Tin=120oC

5

[%] P

E 4

M I

COV 3

2

1

0 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.15: COVIMEP at load II based on a 50-cycle measurement

Ignition Dwell At load II with an idle speed of 600 rpm, three PRF blends and even diesel all provide positive ignition dwells, as shown in Figure 4.17. As expected, PRF90 has the longest IDw compared to other fuels. The main differences in IDw among the test fuels correlate well with their ignitabilities. Fuels ignite easier at higher intake temperature. Thus the IDw reduces with the increase of intake temperature. The positive IDw results in little soot formation even for diesel, as shown in Figure 4.18. Three PRF blends produce near zero

58 4.2. Results and Analysis — Effects of Intake temperature and CA50 soot. The differences should be taken with caution since the absolute values are very small and approach the limit of detection.

40

Tin=40oC o 35 PRF70 Tin=60 C PRF80 Tin=80oC PRF90 Tin=100oC 30 Diesel Tin=120oC

25

20

15 PRRmax [bar/CAD] PRRmax 10

5

0 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.16: Maximum pressure rise rate as a function of combustion phasing and intake temperature for the three PRF blends and diesel at load II

12

Tin=40oC PRF70 Tin=60oC PRF80 PRF90 10 Tin=80oC Diesel Tin=100oC Tin=120oC 8

6

4 Ignition dwell [CAD] dwell Ignition

2

0 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.17: Ignition dwell plotted against CA50 at load II

59 Chapter 4. Low Load Tests with High-octane Fuels

0.01

Tin=40oC o PRF70 Tin=60 C o 0.008 PRF80 Tin=80 C PRF90 Tin=100oC Diesel Tin=120oC

0.006

0.004 ISPM [g/kWh] ISPM

0.002

0 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.18: Soot plotted against CA50 at load II

Rate of Heat Release Figure 4.19 and 4.20 show the rate of heat release (RoHR) of test fuels at two extreme combustion phasings (0 and 10 CAD ATDC) with an intake temperature of 60 oC. It is worth noting that the RoHR of PRF90 decreases and burn duration increases with the delay of combustion phasing. PRF90 has the highest resistance to auto-ignition, so it needs to be injected earlier into the cylinder than other test fuels. The longer mixing time allows fuel and air to premix to such an extent that mixtures are lean. The slower combustion at late combustion phasing results from the fuel/air mixture becoming gradually more dilute with increased IDw.

Limitations of PPC at Load II — High HC Emissions Figure 4.21 shows the HC emissions at load II for the test fuels at various combustion phasings and intake temperatures. Three PRF blends produce much higher HC emissions than diesel regardless of combustion phasing and intake temperature at both load I and load II. Besides, they are difficult to be oxidized by diesel oxidation catalyst (DOC) due to the low exhaust temperature. The late combustion phasing and relatively low intake temperature have a significant influence on the burning of PRF90 and PRF80 because they provide longer IDw at such conditions which cause the fuel to be over-diluted. The local λ goes toward global λ due to the long ignition delay. Therefore, the chemical reaction rate reduces and becomes prone to incomplete combustion. As a result, the HC emissions of PRF90 display a strong combustion phasing dependency especially when the intake temperature is below 100 oC. This also applies to PRF80 when the intake temperature is set at 40 oC. Even though the global λ of diesel is similar to the PRF blends, the local diesel/air mixture is richer due to the shorter ignition delay. Thus, less HC emissions are produced.

60 4.2. Results and Analysis — Effects of Intake temperature and CA50

CA50:0 CAD ATDC 600 PRF70 PRF80 500 PRF90 Diesel

400

300 Tin=60oC

200 SOI

Rate of heat release [J/CAD] release heat of Rate 100

0

-20 -15 -10 -5 0 5 10 15 20 Crank angle [CAD ATDC]

Figure 4.19: Rate of heat release and SOI timing at load II using an intake tempera- ture of 60 oC and CA50 at 0 CAD ATDC

CA50:10 CAD ATDC 600 PRF70 PRF80 500 PRF90 Diesel

400

300 Tin=60oC

200 SOI

Rate of heat release [J/CAD] release heat of Rate 100

0

-20 -15 -10 -5 0 5 10 15 20 Crank angle [CAD ATDC]

Figure 4.20: Rate of heat release and SOI timing at load II using an intake tempera- ture of 60 oC and CA50 at 10 CAD ATDC

61 Chapter 4. Low Load Tests with High-octane Fuels

6

Tin=40oC o Tin=60 C PRF70 5 o Tin=80 C PRF80 Tin=100oC PRF90 Tin=120oC Diesel 4

3 ISHC [g/kWh] ISHC 2

1

0 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.21: HC emissions plotted against CA50 at load II

Limitations of PPC at Load II — High CO Emissions The three PRF blends produce comparable CO emissions as diesel only at early combustion phasing when using a high intake temperature, as shown in Figure 4.22. All fuels provide positive ignition dwells and the experiments are done with an λ higher than 3.1. Hence, CO emissions at idle condition should be caused by over-leaning rather than the existence of fuel-rich zones and by insufficient in-cylinder temperature. PRF90 tends to create more over-mixed zones to produce CO emissions due to its longest IDw. In addition, its combustion occurs in the expansion stroke with a slower burn rate and a longer burn duration. Hence, its in-cylinder temperature is not high enough to achieve complete combustion, thus producing more CO emissions. The CO emissions are plotted against HC emissions as shown in Figure 4.23. As can be seen from Figure 4.23 that the HC and CO emissions of the PRF blends show a strong correlation. Under the test conditions, the combustion which is dominated by premixed combustion produces increasing HC and CO emissions at the same time. Diesel produces the least HC and CO emissions regardless of the changes in combustion phasing and intake temperature.

62 4.2. Results and Analysis — Effects of Intake temperature and CA50

30

Tin=40oC o Tin=60 C PRF70 25 o Tin=80 C PRF80 Tin=100oC PRF90 Tin=120oC Diesel 20

15 ISCO [g/kWh] ISCO 10

5

0 -2 0 2 4 6 8 10 12 CA50 [CAD ATDC]

Figure 4.22: CO emissions plotted against CA50 at load II

25 PRF70 PRF80 PRF90 20 Diesel

15

10 ISCO [g/kWh] ISCO

Tin=40oC o 5 Tin=60 C Tin=80oC Tin=100oC Tin=120oC 0 0 1 2 3 4 5 6 7 8 ISHC [g/kWh]

Figure 4.23: CO emissions vs HC emissions at load II

63 Chapter 4. Low Load Tests with High-octane Fuels

4.3 Results and Analysis — Optimize NOx Emissions

EGR was not used in the previous study (Section 4.2). In Section 4.2, the effects of intake temperature and combustion phasing on combustion stability and emissions are discussed. In the following section, EGR will be applied to reduce NOx emissions, and intake pressure was consequently adjusted to avoid the increased soot emissions. The aim of this study is to find a good combination between EGR and intake pressure to achieve the purpose of simultaneous reduction of NOx and soot.

4.3.1 Load I

The HC, CO, NOx, and soot emissions of PRF70, PRF80, PRF90, and diesel using 1.2 bar intake pressure (Pin) with various EGR rates at load I are shown in Figure 4.24. HC and CO emissions of the three PRF blends are higher than that of diesel regardless of EGR rate. As is expected, NOx emissions reduce when a higher EGR rate is used. This reduction in NOx emissions results from the lower combustion temperature and in-cylinder oxygen concentration. PRF90 is dominated by premixed combustion and has a shorter burn duration than the other test fuels, thus a higher peak combustion temperature leading to higher NOx emissions. EGR reduces both in-cylinder temperature and oxygen concentration when using a constant intake pressure. Consequently a higher EGR rate may lead to the formation of fuel-rich regions and also slows down the soot oxidation rate. However, the growth of the EGR rate also increases the ignition delay, allowing longer time for fuel to mix with air before combustion starts to reduce soot emissions. As can be seen in Figure 4.24, the soot emissions of diesel increase as the growth of EGR rate, but the soot emissions of the three PRF blends remain the same and even decrease when EGR rate increases. It is concluded that the CO and soot emissions of diesel are due to the presence of fuel-rich regions in the cylinder, and the low soot emissions of PRF blends are mainly due to the better fuel/air mixing resulting from the longer ignition delays. This conclusion is supported by the following experiment. When a higher intake pressure (Pin = 1.5 bar) is used, as shown in Figure 4.25, the CO and soot emissions of diesel are less than that of experiments using 1.2 bar intake pressure, especially for higher EGR rates conditions. However, the soot emissions of the PRF blends increase for a higher Pin due to the reduced ignition delay. The purpose of this study is to optimize NOx emissions, but from the above experimental results it can be seen that NOx emissions are still much higher than the Euro VI emission regulations (0.4 g/kWh). Figure 4.26 shows the HC, CO, NOx, and soot emissions of PRF70 using 40%, 60%, and 65% EGR rate at 1.5 bar intake pressure. With the further increase in EGR rate, NOx emissions are reduced and meet the Euro VI emission regulations. Soot emissions are also reduced because of the extended ignition delay and thus enhancing the fuel/air mixing. However, a large increase in HC and CO emissions has created new challenges since the traditional diesel oxidation catalyst (DOC) cannot work properly with the low-temperature exhaust gases. In previous studies, trapped hot residual gases strategies, including using negative valve overlap (NVO) together with up to 70% EGR [88] and using the rebreathing without using external EGR [89], were investigated to increase the reachable

64 4.3. Results and Analysis — Optimize NOx Emissions

Pin = 1.2 bar Pin = 1.2 bar 1 7 PRF70 6 PRF80 0.8 PRF90 PRF70 Diesel PRF80 5 0.6 PRF90 Diesel 4

0.4 3 ISHC [g/kWh] ISHC ISCO [g/kWh] ISCO 2 0.2 1

0 0 10 20 30 40 10 20 30 40 EGR [%] EGR [%]

Pin = 1.2 bar Pin = 1.2 bar 12 1 PRF70 PRF70 10 PRF80 PRF80 PRF90 0.8 PRF90 Diesel Diesel 8 0.6 6 0.4

4

ISPM [g/kWh] ISPM ISNOx [g/kWh] ISNOx 0.2 2

0 0 10 20 30 40 10 20 30 40 EGR [%] EGR [%]

Figure 4.24: HC, CO, NOx, and soot emissions of all test fuels for different EGR rates at load I; Intake pressure is constant at 1.2 bar.

65 Chapter 4. Low Load Tests with High-octane Fuels operating region of high-octane fuels and get low NOx and soot emissions as much as possible. The results suggest using NVO with NVO injection at idle condition and change to rebreathing strategy at a higher engine load. However, low NOx and soot emissions together with an acceptable combustion efficiency without using post-processors cannot be achieved even using such a method. At load I, one way to reduce HC and CO emissions is to use DOC, which requires a minimum exhaust temperature of about 200 oC to light off. Multiple injection strategies may also help to reduce HC and CO emissions by creating a concentration gradient of fuel/air mixing to reduce over-mixing.

Pin = 1.5 bar Pin = 1.5 bar 1 7 PRF70 6 PRF80 0.8 PRF90 Diesel PRF70 5 0.6 PRF80 PRF90 4 Diesel

0.4 3 ISHC [g/kWh] ISHC ISCO [g/kWh] ISCO 2 0.2 1

0 0 10 20 30 40 10 20 30 40 EGR [%] EGR [%]

Pin = 1.5 bar Pin = 1.5 bar 10 0.5 PRF70 PRF70 PRF80 PRF80 8 PRF90 0.4 PRF90 Diesel Diesel

6 0.3

4 0.2

ISPM [g/kWh] ISPM ISNOx [g/kWh] ISNOx 2 0.1

0 0 10 20 30 40 10 20 30 40 EGR [%] EGR [%]

Figure 4.25: HC, CO, NOx, and soot emissions of all test fuels for different EGR rates at load I; Intake pressure is constant at 1.5 bar.

66 4.3. Results and Analysis — Optimize NOx Emissions

PRF70 PRF70 12 0.1 ISHC 10 ISCO ISNOx 0.08 Pin=1.5bar

8 0.06 6 0.04

4 [g/kWh] ISPM

0.02

2 HC/CO/NOx emission [g/kWh] emission HC/CO/NOx 0 0 40 45 50 55 60 65 40 45 50 55 60 65 EGR [%] EGR [%]

Figure 4.26: HC, CO, NOx, and soot emissions of PRF70 using high EGR rates at load I, and intake pressure is constant at 1.5 bar.

4.3.2 Load II The HC, CO, NOx, and soot emissions of three PRF blends and diesel using 1.3 bar intake pressure with various EGR rates at load II are shown in Figure 4.27. There are two main causes of HC emission in diesel engines. One is the ignition delay period, where fuel is over mixing and gets leaner than the lean combustion limit. The other one is the under mixing of fuel. At idle condition, over mixing has a greater effect on the formation of HC emissions [90]. Figure 4.27 shows that HC emissions reduce slightly at the increase of EGR rate. The ignition delay will increase as the EGR rate increases, providing longer time for fuel to mix air and thus leading to more over-mixed regions. Hence, higher HC emissions are expected. However, the oxygen concentration decreases as the EGR increases, thereby reducing the formation of over-lean fuel/air mixture. As the EGR increases, the reduced oxygen concentration plays a greater role in the HC formation under the test conditions. The CO emissions only increase slightly for 40% EGR. In addition, the NOx emissions reduce as the growth of EGR rate as expected. The soot emissions of all test fuels at idle condition are low. The HC, CO, NOx, and soot emissions of three PRF blends with 40% EGR using different intake pressures at load II are shown in Figure 4.28. Only CO emissions decrease obviously with the rise of intake pressure, while HC, NOx, and soot emissions are almost not affected by the change of intake pressure. It would be possible to run the engine with higher EGR rates to further reduce NOx emissions and does not significantly deteriorate the combustion efficiency, but this requires additional experiments to verify.

67 Chapter 4. Low Load Tests with High-octane Fuels

Pin = 1.3 bar Pin = 1.3 bar 2 7 PRF70 PRF80 6 PRF90 PRF70 1.5 Diesel 5 PRF80 PRF90 4 Diesel 1

3 ISHC [g/kWh] ISHC ISCO [g/kWh] ISCO 2 0.5 1

0 0 10 20 30 40 10 20 30 40 EGR [%] EGR [%]

Pin = 1.3 bar Pin = 1.3 bar 25 0.014 PRF70 PRF70 PRF80 0.012 PRF80 20 PRF90 PRF90 Diesel 0.01 Diesel

15 0.008

10 0.006 ISPM [g/kWh] ISPM ISNOx [g/kWh] ISNOx 0.004 5 0.002

0 0 10 20 30 40 10 20 30 40 EGR [%] EGR [%]

Figure 4.27: HC, CO, NOx, and soot emissions of all test fuels with different EGR rates at load II, and intake pressure is constant at 1.3 bar.

68 4.3. Results and Analysis — Optimize NOx Emissions

40% EGR 40% EGR 2 7 PRF70 PRF70 PRF80 6 PRF80 1.5 PRF90 PRF90 5

4 1

3 ISHC [g/kWh] ISHC ISCO [g/kWh] ISCO 2 0.5 1

0 0 1.2 1.4 1.6 1.8 2 1.2 1.4 1.6 1.8 2 Pin [bar] Pin [bar]

40% EGR 40% EGR 25 0.01 PRF70 PRF70 20 PRF80 0.008 PRF80 PRF90 PRF90

15 0.006

10 0.004

ISPM [g/kWh] ISPM ISNOx [g/kWh] ISNOx 5 0.002

0 0 1.2 1.4 1.6 1.8 2 1.2 1.4 1.6 1.8 2 Pin [bar] Pin [bar]

Figure 4.28: HC, CO, NOx, and soot emissions of three PRF blends with 40% EGR at load II, and intake pressure is increased from 1.3 bar to 1.9 bar.

69 Chapter 4. Low Load Tests with High-octane Fuels

4.4 Conclusions

Experiments were carried out on a modified heavy duty diesel engine to investigate partially premixed combustion at low load conditions. Three PRF blends with research octane numbers of 70, 80, and 90 were tested. Diesel was also tested as a reference. Engine experiments were conducted at loads of 5 bar IMEPg (1200 rpm) referred to “load I” and 2.5 bar IMEPg (600 rpm) referred to “load II” using a single injection strategy. Firstly, engine operating conditions such as intake temperature and combustion phasing were adjusted to investigate their effects on combustion and emission characteristics without using EGR. EGR was applied afterward to optimize NOx emissions, and intake pressure was adjusted correspondingly to avoid the significant increase in HC, CO, and soot emissions. Stable combustion in terms of coefficient of variation of IMEPg and maximum pressure rise rate is guaranteed at both low load and idle condition using the three PRF blends. At load I, PRF90 presents positive ignition dwells at all test conditions and produces near zero soot emissions. But PRF90 requires a higher intake temperature (60 oC) than PRF80 and PRF70 to maintain a stable combustion. At load II, the three PRF blends as well as diesel provide positive ignition dwell and produce low soot emissions. But the longer ignition delays of PRF blends promote the formation of over-lean fuel/air mixtures, thus producing high HC and CO emissions. Intake pressure can efficiently reduce CO and soot emissions, but does not have an obvious effect on HC emissions. It must be noted that the mass flow of the exhaust gas is not sufficient to spin the fast and boost the intake side of the engine under realistic idle condition. Higher boost pressures can be achieved by the use of a or two-stage turbocharger system. NOx emissions reduce sharply with the growth of EGR rate. When a higher than 60% EGR is used for PRF70, its NOx emissions are below Euro VI emission legislation. But the excessive EGR rate also leads to a significant increase in HC and CO emissions. The diesel oxidation catalyst requires a minimum exhaust temperature of about 200 oC, so the HC and CO emissions from the lower loads and idle condition using high-octane fuels may bring new challenges to the traditional DOC. Reducing excessive mixing can help to reduce HC and CO emissions at lower loads, so multiple injection strategies are believed to be a useful approach by increasing the fuel stratifications.

70 Chapter 5 Particulate Matter Emissions

Recent studies correlate fine particulates and ultra-fine particulates to adverse human health effects. The porous particulate agglomerates could deposit in the deep lung and cause diseases like asthma, pneumonectasis and nasopharyngeal carcinoma [43]. High-octane fuels have shown a significant advantage in reducing soot emissions, but whether they will emit more nanoparticles is worth studying. Load, dilution, and fuel structure can be the major factors influencing that. Therefore, it is necessary to acquire knowledge about the particulate matter in terms of number concentration and size distribution. 1

5.1 Introduction

Particulate matter is a generic term for emitted particulates in the exhaust that includes elemental carbon (EC), unburned or partially burned fuel, oil, bound water, wear metals and sulfate, [91] and [92]. Particles differ in size, composition, solubility and therefore also in their toxic properties. Figure 5.1 gives an example of particles measured over a transient cycle from HD diesel engines [93]. Volatile matter accounts for a large proportion of PM. These volatiles may remain in the gas phase, condense on existing solid particles, and nucleate

1Part of this chapter has been retrieved from the following papers: Wang, S., et al. (2016). Combustion and Emission Characteristics of a Heavy Duty Engine Fueled with Two Ternary Blends of N-Heptane/Iso-Octane and Toluene or Benzaldehyde. SAE Technical Paper 2016-01-0998, doi:10.4271/2016-01-0998. Wang, S., et al. (2017). Effects of exhaust gas recirculation at various loads on diesel engine performance and exhaust particle size distribution using four blends with a research octane number of 70 and diesel. Energy Conversion and Management. https://doi.org/10.1016/j.enconman.2017.03.087 Wang, S., et al. (2017). Particulate Matter Emission from a Heavy-duty Diesel Engine with Three Binary Blends. The 12th Conference on Sustainable Development of Energy, Water and Environment Systems – SDEWES Conference, Dubrovnik.

71 Chapter 5. Particulate Matter Emissions and form new particles depending on temperature and other conditions [94]. Figure 5.2 [95] shows the soot formation process from gas phase to solid agglomerated particles. The soot formation processes are pyrolysis, nucleation, coalescence, surface growth, agglomeration.

While the sixth process, oxidation, converts hydrocarbons to CO, CO2 and H2O at any point in the process. More details about these processes can be found in [95].

Figure 5.1: The composition of particles from a heavy-duty diesel engine tested in a transient cycle on an engine test bench [94]. The composition may vary widely for different engine designs, operating conditions, and fuel compositions.

Figure 5.2: Schematic diagram of the soot formation process from gas phase to solid agglomerated particles. This figure is from [95].

Figure 5.3 shows a typical bimodal particle size distribution, which is composed of particles from the nucleation mode (NM) and from the accumulation mode (AM). Engine exhaust particles can exhibit unimodal (NM or AM) or bimodal distribution (NM+AM), depending on operating conditions. Particles smaller than 20 nm are commonly referred to as nucleation mode while those larger than 20 nm are typical of accumulation mode. Nuclei particles are mainly due to semi-volatile components originating from unburned fuel, lubricant oil, and incomplete combustion products [96]. The accumulation mode is composed primarily of carbonaceous agglomerates and adsorbed material. They are formed in locally fuel-rich regions of the flame [96]. At higher-pressure conditions and the anoxic environment when a

72 5.2. Tests with Ternary Blends high level of EGR is used, soot nucleation is promoted, which contributes to the growth of the existing soot nuclei [97]. The soot oxidation rate declines and particles in the nucleation mode with reactive surfaces tend to grow and agglomerate to generate large particles. With an increase in the number of particles, the coagulation rate increases and therefore larger particles are formed, resulting in an increase of the particulate aerodynamic diameter.

Figure 5.3: A typical bimodal particulate size distribution

5.2 Tests with Ternary Blends

In this study, several blends with a RON of about 70 have been designed. Through our own fuel design, the fuel structure and activities are easily controlled. This makes it easier to study the influence of fuel structure on particulate emissions. Also the effects of ignition delay on combustion and emission characteristics are also investigated. The limited number of components makes that the designed fuels can be handled in numerical simulations, since chemical mechanisms exist for these blends in contrast to regular gasoline or diesel. Besides, the chemical and spectral purity makes them suitable for use in optically accessible engines as well. Diesel is also tested for reference.

5.2.1 Test Fuels

Fuels with a RON of 70 are identified close to optimal for PPC [98]. N-heptane (CH3(CH2)5CH3)/ iso-octane ((CH3)3CCH2CH(CH3)2) blends are used as base fuels. Ethanol (C2H5OH) and n-butanol (C4H9OH) are selected as alcohol candidates to blend with the base fuels separately

73 Chapter 5. Particulate Matter Emissions

to get a RON around 70. Toluene (C6H5CH3) is selected as an aromatic component, which is considered to increase soot emissions, to blend with the base fuel to get a RON around

70 as well. A fuel that contains both oxygen and a phenyl group (C6H5) is an aromatic alcohol is made. Initially, benzyl alcohol (C6H5CH2OH) which is an aromatic alcohol was chosen, but a clear phase separation was found when blending it with the base fuel. Instead, benzaldehyde (C6H5CHO) is chosen in the end. All test fuels are denoted as PRF (to ensure the consistency of abbreviation format, PRF in Chapter 5 refers to PRF70), ERF, BRF, TRF, and CRF, respectively. The main properties of these blends are shown in Table 5.1. The mixing ratio of CRF is the same as that of TRF to compare the influence of benzaldehyde and toluene on particulate emissions.

Table 5.1: Properties of the test blends

Blends PRF ERF BRF TRF CRF Vol-% n-heptane 30 52 35 36 36 Vol-% iso-octane 70 23 40 39 39 Vol-% additive 0 25 25 25 25 Additive - Ethanol n-Butanol Toluene Benzaldehyde RON 70 70 72.9 70.8 - DCN 27 25.8 26.1 27.4 30.6 LHV (MJ/kg) 44.47 40.18 41.27 43.33 40.72 LHVst (MJ/kg) 2.76 2.79 2.75 2.77 2.81 AFst 15.11 13.42 13.99 14.63 13.49 H/C ratio 2.26 2.41 2.32 1.91 1.80 O/C ratio 0 0.0961 0.0582 0 0.047

5.2.2 Test Conditions

Before the tests started, the engine was warmed up until the lubricating oil and coolant fluid temperature reached 85 oC. All the tests were conducted under steady state conditions and various loads were investigated using a single injection strategy. Injector C was used for all the experiments presented in Chapter 5. The engine speed was set at 1200 rpm which is a typical highway cruising speed (for older engines) and intake temperature was controlled at 40 oC. At low load, intake pressure was set at 1.2 bar instead of ambient pressure to ensure the stable combustion of these blends. Intake pressure was adjusted according to the operating load to provide a λ higher than 1.2 even when 45% EGR was used to avoid excessive soot emissions. SOI of different test fuels is the same for a selected load, but it varies for different loads. Compared with low load (SOI -14 CAD ATDC), the SOI at medium and high loads are more towards TDC to reduce cyclic variations and achieve acceptable maximum PRR. More fuel is required when increasing load. Therefore the injection pressure is varied in order to keep injection duration short. The detailed experimental settings are summarized in Table 5.2.

74 5.2. Tests with Ternary Blends

Table 5.2: Experimental settings

IMEPg (bar) 5 (± 0.2) 10 (± 0.2) 15 (± 0.2) Engine speed (rpm) 1200 Injection pressure (bar) 600 1200 1800 Intake pressure (bar) 1.2 1.8 2.5 Exhaust pressure (bar) 1.5 2.1 2.8 EGR (%) 0∼40 0∼45 0∼45 Intake temperature (oC) 40

5.2.3 Comparison of Blends of RON70 and Diesel

Ignition Delay

Figure 5.4 illustrates the effects of EGR on ignition delay times for the test fuels at various loads. The ignition delay of CRF was also shown in Figure 5.4, but the comparison between TRF and CRF will be discussed in Section 5.2.2. At low load with the same SOI, the start of combustion for the four blends is delayed compared to diesel. Even a misfire occurred for BRF when 30% EGR is used. Therefore the maximum EGR rate used for the four blends is reduced. The increase in EGR rate leads to lower in-cylinder temperature and oxygen concentration, which correspondingly results in longer ignition delay and a shift of the main combustion to a later stage. The differences in ignition delay between the blends of RON70 and diesel become less as load increases and even converge at high load. ERF and BRF show a slightly longer ignition delay compared to PRF and TRF, which correlates with their lower cetane numbers.

Particulate Emissions of the Test Fuels at Different Loads

The particle number concentration and size distribution are measured and compared in this study. At low load with 20% EGR (the highest stable typical EGR level at low load), four blends (PRF, ERF, BRF, and TRF) show an increase in the number of particles at aerodynamic diameter (AD) below 20 nm compared to diesel as depicted in Figure 5.5. At low load, the particle size distribution of diesel is unimodal. Most particles are produced in the accumulation mode and the peak number concentration with an AD in between 78 and 82 nm. For each fuel, in general, the particle number shifts upwards, and its distribution curves move towards larger size when engine load increases. At medium load the oxygenated blends (ERF and BRF) still produce most particles in the nucleation mode, and the peak number concentration has an AD of 8 ∼ 12 nm. However, the size distribution profiles of PRF and TRF shift towards larger size and produce most particles in the accumulation mode with the peak number concentration with an AD of 50 nm. Still, the four blends produce much less particles in both nucleation mode and accumulation mode compared to diesel at medium and high loads.

75 Chapter 5. Particulate Matter Emissions

IMEPg : 5bar IMEPg : 10bar 25 25

20 20

15 15

10 10 Ignition Delay [CAD] Delay Ignition Ignition Delay [CAD] Delay Ignition 5 5

0 0 0 10 20 30 40 50 0 10 20 30 40 50 EGR [%] EGR [%]

IMEPg : 15bar 25 PRF 20 ERF BRF TRF 15 CRF Diesel

10

Ignition Delay [CAD] Delay Ignition 5

0 0 10 20 30 40 50 EGR [%]

Figure 5.4: Ignition delay vs EGR at various loads

Effects of EGR on Particulate Emissions

Only the particulate emissions of PRF, ERF, BRF, TRF and diesel at 10 bar IMEPg are discussed. A similar trend can be also found at 5 bar and 15 bar IMEPg. As shown in Figure 5.6, the size distribution curves of all test fuels shift towards larger particles as the EGR increases. There is a noticeable increase of the particles (AD > 50 nm) for PRF, TRF and diesel when more than 30% EGR is used, while ERF and BRF still produce most particles with an AD smaller than 40 nm even when 40% EGR is used. This must be associated with the local equivalence ratio since the ignition delays of all blends are comparable. With similar injection velocities (i.e. injection pressures) the air entrainment is the same for all, while the local equivalence ratio of ERF and BRF in the spray is lower because they contain hydroxyl groups, which might explain the observed phenomenon. The peak concentration occurs at a diameter range near 10 nm for all selected fuels without using EGR. When 24% EGR (± 2%) is used, the peak concentrations of particles are around 11 nm, 11 nm, 50 nm, 52 nm, and 60 nm for ERF, BRF, PRF, TRF, and diesel, respectively. The increased EGR rate reduces the number of particles in the nucleation mode but does not have a significant influence on the concentration of particles in the accumulation mode for ERF and BRF. For non-oxygenated fuels, the increased EGR rate reduces the number of particles with small AD but enhances

76 5.2. Tests with Ternary Blends

Figure 5.5: Particulate number concentration and size distribution at various loads. Right column is a zoomed plot to distinguish the results of the four blends individu- ally.

the number of particles with large AD, hence the total particle mass increases. However, the increase in the total particle mass is more significant for diesel with an increased EGR rate, which is in agreement with the steeper rise of soot emissions as presented in Figure 5.7. Diesel produces most particles with an AD larger than 100 nm when more than 35% EGR is used, and these particles dominate the mass size distribution due to the cubic dependence of particle diameter. The cooled EGR is a well-known technique to reduce NOx emissions. The increase of EGR rate results in a lower oxygen concentration which favors the formation of soot, and besides, the increased EGR reduces peak in-cylinder temperatures and leads to incomplete soot oxidation. The use of high EGR rates leads to the higher amounts of soot emitted in a given exhaust volume where coagulation, accumulation, condensation of volatile fractions on the particles and surface growth are likely to happen [99]. All these factors result in an increasing concentration of larger diameter particles in the accumulation mode.

77 Chapter 5. Particulate Matter Emissions

Figure 5.6: Particle number concentration and size distribution with various EGR rates at medium load for ERF, BRF, PRF, TRF, and Diesel.

IMEPg : 10bar 2 PRF ERF 1.5 BRF TRF Diesel

1 PM [g/kWh] PM 0.5

0 0 10 20 30 40 50 EGR [%]

Figure 5.7: Soot emissions on mass base vs EGR rate at medium load

78 5.2. Tests with Ternary Blends

5.2.4 Comparison of TRF and CRF The compositions of diesel fuel is quite complex, including straight-chain alkanes, branched- chain alkanes, naphthenes, alkanes, and aromatics [100]. Polycyclic aromatic hydrocarbons (PAHs) generated in flames are considered as precursors of soot [101]. For conventional diesel aromatics account for a significant portion. Toluene is a representative of many of these aromatic compounds [102]. Toluene (C6H5CH3) is a mono-substituted benzene derivative, consisting of a CH3 group attached to a phenyl group. Benzyl alcohol (C6H5CH2OH) was targeted first for comparison. But a clear phase separation after 24h was found when blending benzyl alcohol with the base fuel. Instead, benzaldehyde (C6H5CHO) was chosen. It is an organic compound consisting of a benzene ring with a formyl substituent. Benzaldehyde is also the simplest aromatic aldehyde and an oxygenated fuel as well. Therefore, it is of interest to compare the PM emissions of TRF and CRF.

Heat Release Rate Plots shown in Figure 5.8 are RoHR data and injector current signal for TRF and CRF at low load (5 bar IMEPg), medium load (10 bar IMEPg) and a higher load (15 bar IMEPg), respectively. TRF has longer ignition delays than CRF at low and medium loads and correspondingly exhibits more premixed combustion. While at 15 bar IMEPg, TRF and CRF have almost the same ignition delay. The sufficiently high in-cylinder temperature causes both fuels to burn shortly after being injected into the cylinder and resulting in diesel-like heat-release rates. At low and medium loads when EGR rate is increased from 0% to 30%, test fuels become harder to combust due to the reduced in-cylinder temperature and show longer ignition delay and more premixed combustion. However, at 15 bar IMEPg 30% EGR is not enough to increase the premixed combustion considerably because of the high reactivity of test fuels under such experimental conditions.

Soot Emissions As shown in Figure 5.9, at low and medium loads TRF produce near zero soot emissions, slightly lower than CRF. However, at 15 bar IMEPg, their soot emissions increase sharply with the growth of EGR rate. TRF and CRF have longer ignition delays at low and medium loads compared to high load. Therefore fuel and air are well premixed thus less fuel-rich regions are formed where it is easy to generate soot. Glassman [103] points out that fuel structure has a substantial effect on the sooting tendency of diffusion flames, while little influence in premixed flames. At low load, both TRF and CRF are dominated by premixed combustion, but TRF has a higher premixed fraction due to its longer ignition delay than CRF. Hence TRF produces less soot. It is seen from Figure 5.9 that at 10 bar IMEPg the soot emissions of CRF apparently increase compared to TRF when more than 40% EGR is used. Figure 5.10 shows that the λ of TRF and CRF approaches 1.3 when 40% EGR is used. At conditions with low oxygen concentration, the local λ of TRF is closer to the global λ than for CRF due to its longer ignition delay. Hence the higher soot emissions formed for CRF. At 15 bar IMEPg, TRF and CRF have a similar ignition delay. They are both dominated by a large amount of mixing-controlled combustion, where fuel structure largely influences the soot emissions. When the global λ reduces to 1.3 when 37% EGR is used (Figure 5.10), the soot emissions of

79 Chapter 5. Particulate Matter Emissions

TRF and CRF begin to increase rapidly. However, CRF produces slightly less soot emissions than TRF at 15 bar IMEPg. This is because the oxygenated benzaldehyde in CRF will raise the local λ at the flame lift-off length, thus increasing the oxygen content participating in the oxidation of soot in a diffusion flame [104].

Figure 5.8: Injector current (low range) and RoHR (high range) at different loads with 0% and 30% EGR, respectively.

Particle Number Concentration and Size Distribution

Figure 5.11 presents the particulate number concentration and size distribution of TRF and CRF for different EGR rates at low load, medium load, and high load, respectively. The particle number size distribution of TRF exhibits unimodal distributions for 0% EGR rate at all test loads, and the particles with an AD of 10 nm have the highest concentration. At low load, when 30% EGR is used for TRF, its particles in the NM reduce but do not form more particles in the AM. However, for CRF the small-size particles (AD < 20 nm) reduce and number of large particles increases significantly as EGR rate increases. At medium load, both TRF and CRF produce more particles in the AM when EGR rate increases. When 42% EGR is used, CRF produces much more particles in the AM than TRF, and this is also consistent with the results of mass based soot emissions (Figure 5.9). The λ decreases with the increase of EGR

80 5.2. Tests with Ternary Blends

IMEPg : 5bar IMEPg : 10bar 0.2 0.2 TRF TRF CRF CRF 0.15 0.15

0.1 0.1 PM [g/kWh] PM PM [g/kWh] PM 0.05 0.05

0 0 0 10 20 30 40 50 0 10 20 30 40 50 EGR [%] EGR [%] IMEPg : 15bar 2 TRF CRF 1.5

1

PM [g/kWh] PM 0.5

0 0 10 20 30 40 50 EGR [%]

Figure 5.9: Soot emissions of TRF and CRF for various EGR rates at different loads.

IMEPg : 5bar IMEPg : 10bar 3 2.5

TRF TRF 2.5 CRF CRF

2 [-]

[-] 2

6 6 1.5 1.5

1 1 0 10 20 30 40 50 0 10 20 30 40 50 EGR [%] EGR [%] IMEPg : 15bar 2.5

TRF CRF

2 [-] 6 1.5

1 0 10 20 30 40 50 EGR [%]

Figure 5.10: λ of TRF and CRF for various EGR rates at different loads.

81 Chapter 5. Particulate Matter Emissions rate. Under such experimental condition, CRF is more likely to have fuel-rich regions because of its shorter ignition delay, hence enhancing particle agglomeration and resulting in higher soot emissions. At 15 bar IMEPg with 37% EGR, CRF produces more particles with an AD in between 20 nm and 100 nm than TRF, but CRF produces slightly less particles with an AD larger than 100 nm. Since particles with larger AD contribute more to the total soot mass, so CRF produces less soot emissions than TRF as shown in Figure 5.9.

Figure 5.11: Effects of EGR and operating load on particulate number concentration and size distribution for TRF and CRF.

82 5.3. PRF Blends Mix with Different Volumes of Ethanol

5.3 PRF Blends Mix with Different Volumes of Ethanol

Studies on ethanol PPC already indicated that zero soot emission is achievable [105] and [106]. In this study four blends, all composed of n-heptane, iso-octane and ethanol but with different volume fractions, were tested under PPC mode with various dilution strategies. Diesel was also tested for comparison. To focus on the role of ethanol, all blends were designed to have the same RON.

5.3.1 Test Fuels

To make fuel blends as close to RON 70 as possible, a regression model (Equation (5.1)) from Lund University [107] is used to calculate the required volume percentages of each fuel component:

RONpredicted = β + αNXN + αT XT + αEXE + γNT XNXT + γNEXNXE 2 2 2 (5.1) +γTEXT XE + δN2 XN + δT 2 XT + δE2 XE

where β is a constant and α, γ and δ are coefficients for linear, interaction, and quadratic terms, respectively. The specific values of these parameters can be found in [107]. X represents liquid volume fractions of each compound (N for n-heptane, T for toluene, I for iso-octane and E for ethanol). Required volume fractions to make the blends and estimated RON values are shown in Table 5.3, and some other fuel properties are shown in Table 5.4.

Table 5.3: Design of the test blends

Fuel Ethanol vol% Iso-octane vol% N-heptane vol% RON E05 5 60 35 69.95 E15 15 41 44 69.69 E25 25 23 52 69.98 E38 38 0 62 70.04

Table 5.4: Fuel properties of the blends containing ethanol

Fuel H/C ratio O/C ratio LHV (MJ/kg) AFst E05 2.3562 0.0625 43.5153 14.3120 E15 2.5035 0.1601 41.8059 13.1230 E25 2.6129 0.2328 40.1396 12.2358 E38 2.7196 0.3037 38.0379 11.3711

83 Chapter 5. Particulate Matter Emissions

5.3.2 Test Conditions

A particular dilution condition is determined by both intake pressure and EGR rate. Therefore, both of these factors are varied in this study to investigate their influence on PM. The intake pressure is varied from 1.4 bar to 2.2 bar (±0.02 bar) with a step of 0.2 bar with a constant EGR rate of 40%. EGR rate is varied from 30% to 50% with a step of 10% at two reference points, with a constant intake pressure set at 1.8 bar and 2.2 bar, respectively. Intake temperature is set to 52oC, the same as used for the test condition of RON values [108]. To get a FuelMEP (introduced in Chapter 2) of 15 bar (± 0.5 bar) and a combustion phasing at 8 CAD ATDC (± 0.3 CAD ATDC), injection duration and SOI timing are varied. The corresponding IMEPg is 6.9 bar (± 0.3 bar). A single injection is used and engine speed is kept constant at 1200 rpm. The experimental settings are summarized in Table 5.5.

Table 5.5: Experimental settings and variations XX XXX Fuel XXX E05, E15, E25, E38, Diesel Parameter XX Intake pressure (bar) 1.4 1.6 1.8 2.0 2.2 EGR rate (%) 40 40 30, 40, 50 40 30, 40, 50 Exhaust pressure (bar) 1.7 1.9 2.1 2.3 2.5 Engine speed (rpm) 1200 Intake temperature (oC) 52

5.3.3 Major Results and Discussion

Effects of Intake Pressure on Heat Release Rate

Figure 5.12 depicts the heat release rates using 40% EGR for five test fuels at different intake pressures. Premixed combustion dominates the combustion of ethanol containing fuels. However, the premixed combustion reduces slightly as the intake pressure increases (Figure 5.13). This phenomenon is most noticeable for E05 at an intake pressure of 2.2 bar. Increased intake pressure shortens the ignition delay since the increasing in-cylinder oxygen concentration and temperature accelerates the reaction. Smaller ignition delay leads to a more mixing-controlled condition. Diesel always has less premixed combustion compared to ethanol containing fuels, and the second peak of heat release appears when the intake pressure increases above 1.4 bar.

Soot Emissions

Ethanol containing fuels have the advantage of reducing soot emissions, and this is more pronounced for the blends with more than 5% ethanol content. As shown in Figure 5.14a, the mass based soot emissions of ethanol containing blends are much lower than the Euro VI emission legislation (0.01 g/kWh), regardless of intake pressure. On the contrary, although soot emissions of diesel can be suppressed substantially by increasing intake pressure, they

84 5.3. PRF Blends Mix with Different Volumes of Ethanol

Figure 5.12: Injector current (low range) and RoHR (high range) using different intake pressures for the test fuels with a 40% EGR

85 Chapter 5. Particulate Matter Emissions

EGR 40% 1

0.9

0.8

0.7

0.6

0.5

0.4

Premixed fraction [-] fraction Premixed 0.3 E05 0.2 E15 E25 0.1 E38 Diesel 0 1.4 1.6 1.8 2 2.2 Intake Pressure [bar]

Figure 5.13: Premixed fractions vs intake pressure for 40% EGR are still much higher than ethanol containing fuels (Figure 5.14b). Figure 5.15a shows that the soot emissions of E05 increase substantially as the EGR rate increases and they are higher than that of the other blends. This is due to the shortest ignition delay and minimal oxygen content of E05. However, the soot emissions of E05 are still much lower (factor 100) compared to diesel. Figure 5.15b shows that the soot emissions of diesel increase dramatically with the growth of EGR rate due to the reduced in-cylinder oxygen concentration and in-cylinder temperature.

EGR 40% EGR 40% 0.01 1.2 E05 Diesel E15 E25 1 0.008 E38

0.8 0.006 0.6

0.004 ISPM [g/kWh] ISPM ISPM [g/kWh] ISPM 0.4

0.002 0.2

0 0 1.4 1.6 1.8 2 2.2 1.4 1.6 1.8 2 2.2 Intake Pressure [bar] Intake Pressure [bar] (a) Ethanol containing blends (b) Diesel

Figure 5.14: Soot emissions vs intkae pressure with 40% EGR

86 5.3. PRF Blends Mix with Different Volumes of Ethanol

Pin=1.8 bar Pin=1.8 bar 0.01 1.2 E05 Diesel E15 E25 1 0.008 E38

0.8 0.006 0.6

0.004 ISPM [g/kWh] ISPM ISPM [g/kWh] ISPM 0.4

0.002 0.2

0 0 30 40 50 30 40 50 EGR [%] EGR [%] (a) Ethanol containing blends (b) Diesel

Figure 5.15: Soot emissions vs EGR using 1.8 bar intake pressure

Effects of Intake Pressure on Particulate Emissions Particle size distributions were measured for each fuel at different intake pressures and are represented in Figure 5.16. Diesel has a more complex composition including longer chain alkanes and aromatics. During combustion, diesel is more inclined to generate more aromatic rings and larger poly-aromatic hydrocarbons which are easily aggregated at low temperature and low-oxygen environment to form large particles. Diesel has longest ignition delay at the intake pressure of 1.4 bar, but it produces the highest soot emissions (Figure 5.14b) and the most particles in the AM at that condition. This could be due to the lower in-cylinder charge density and hence poorer diesel break-up during the premixing process, thus creating richer conditions locally [109]. Increased intake pressure reduces the particle concentration of diesel significantly in the whole diameter range, which indicates that higher oxygen concentration and in-cylinder density have a greater impact on the soot reduction for diesel. Figure 5.16 shows that E05 produces particles from both NM and AM, while E15, E25, and E38 all emit particles mainly from NM. The differences in particle size distribution also explain why E05 produces a bit higher mass based soot emissions than E15, E25, and E38. It is worth to note that intake pressure has a smaller effect on the PM emissions of ethanol containing blends than on that of diesel, because of their higher volatility, larger premixed fraction, and lower

AFst (lower AFst is more likely to result in leaner mixtures for the same entrainment speed).

87 Chapter 5. Particulate Matter Emissions

#106 E05 #106 E15 ]

] 8 15

3 3

6 10 4 5

2 dN/dlogDp [#/cm dN/dlogDp dN/dlogDp [#/cm dN/dlogDp 0 0 5 10 20 50 100 300 600 5 10 20 50 100 300 600 Particle Diameter [nm] Particle Diameter [nm]

#106 E25 #106 E38 ]

] 15 15

3 3

10 10

5 5 dN/dlogDp [#/cm dN/dlogDp dN/dlogDp [#/cm dN/dlogDp 0 0 5 10 20 50 100 300 600 5 10 20 50 100 300 600 Particle Diameter [nm] Particle Diameter [nm] #108 Diesel ] 2 3 Pin=1.4bar Pin=1.6bar 1.5 Pin=1.8bar Pin=2.0bar 1 Pin=2.2bar 0.5

dN/dlogDp [#/cm dN/dlogDp 0 5 10 20 50 100 300 600 Particle Diameter [nm]

Figure 5.16: The influence of intake pressure on particulate number concentration and size distribution for each test fuel with 40% EGR.

Comparison between ethanol containing fuels and diesel regarding PM emissions is shown in Figure 5.17. Diesel produces much more particles in the AM which contribute significantly more to the particulate mass than particles in the NM. This is consistent with Figure 5.14 (mass based soot emissions). However, although almost all ethanol containing fuels produce close to zero soot emissions, they produce more particles in the NM than diesel which may bring greater damage to human health. It is concluded from Figure 5.16 and 5.17 that ethanol containing fuels do not require high intake pressure to reduce soot emissions, and their particles in the NM are nearly independent of the intake pressure especially for blends with a high volume percent of ethanol.

88 5.3. PRF Blends Mix with Different Volumes of Ethanol

#108 Pin=1.4bar #107 Pin=1.6bar ]

] 2 15

3 3 1.5 10 1 5

0.5 dN/dlogDp [#/cm dN/dlogDp dN/dlogDp [#/cm dN/dlogDp 0 0 5 10 20 50 100 300 600 5 10 20 50 100 300 600 Particle Diameter [nm] Particle Diameter [nm]

#107 Pin=1.8bar #107 Pin=2.0bar ]

] 8 3

3 3 6 2 4 1

2 dN/dlogDp [#/cm dN/dlogDp dN/dlogDp [#/cm dN/dlogDp 0 0 5 10 20 50 100 300 600 5 10 20 50 100 300 600 Particle Diameter [nm] Particle Diameter [nm] #106 Pin=2.2bar

] 15 3 E05 E15 10 E25 E38 Diesel 5

dN/dlogDp [#/cm dN/dlogDp 0 5 10 20 50 100 300 600 Particle Diameter [nm]

Figure 5.17: Comparison of the particulate emissions of test fuels under different intake pressures with 40% EGR.

Effects of EGR on Particulate Emissions

Figure 5.18 finally shows the particulate number concentration and size distribution of all test fuels with different EGR rates using an intake pressure of 1.8 bar. An increase of EGR rate increases the number concentration of diesel particles over the whole range of aerodynamic diameters measured. There is a dramatic increase in the particle number concentration with an AD larger than 100 nm when EGR rate increases to 50%. E05 and E10 produce more particles in the AM and less particles in the NM as EGR increases. However, the particulate number and size distribution of E25 and E38 are little influenced by the change of EGR. They produce most particles in the NM. This is in agreement with mass based soot emissions presented in Figure 5.15. Because the particles in the NM have an AD smaller than 20 nm and contribute little to the soot mass, therefore, the particles in the AM dominate the mass. Figure 5.19 shows the heat release rates of all test fuels at 1.8 bar intake pressure with 30%, 40%, and

89 Chapter 5. Particulate Matter Emissions

7 7 ]

] #10 E05 #10 E15 3 3 3 3

2 2

1 1

0 0 dN/dlogDp [#/cm dN/dlogDp dN/dlogDp [#/cm dN/dlogDp 5 10 20 50 100 300 600 5 10 20 50 100 300 600 Particle Diameter [nm] Particle Diameter [nm]

7 7 ]

] #10 E25 #10 E38 3 3 3 3

2 2

1 1

0 0 dN/dlogDp [#/cm dN/dlogDp dN/dlogDp [#/cm dN/dlogDp 5 10 20 50 100 300 600 5 10 20 50 100 300 600

Particle Diameter [nm] Particle Diameter [nm] ] 3 #108 Diesel 6 EGR 30% 4 EGR 40% 2 EGR 50%

0

dN/dlogDp [#/cm dN/dlogDp 5 10 20 50 100 300 600 Particle Diameter [nm]

Figure 5.18: The influence of EGR rate on particulate number concentration and size distribution for each test fuel with an intake pressure of 1.8 bar.

50% EGR, respectively. Ethanol containing blends are dominated by premixed combustion, and their ignition delays increase slightly with increase of EGR rate. This effect is most obvious for diesel. Diesel presents more premixed combustion as EGR rate increases, but it is still dominated by mixing-controlled combustion. The increase of particles in the AM when EGR increases for E05, E15, and diesel is not because of the increase in mixing-controlled combustion, nor because the ignition delay is shortened. It is known that the particles in the NM are mostly composed of volatile material [110], and they could be purely consisting of volatile material so there is not enough particle surface area available for the volatile material to preferentially condense on [111]. Therefore, the increase of particles in the AM for E05, E15, and diesel are not due to the adsorption or agglomeration of particles in the NM, but formed in locally fuel-rich regions of the flame. EGR technique is widely used for the reduction of NOx emissions. However, it promotes the soot formation as the engine operates at richer mixtures where the availability of oxygen during the combustion process is limited. Increased soot

90 5.4. Tests with Binary Blends

EGR 30% EGR 40%

800 800

600 600

400 400

200 200

RoHR [J/CAD] RoHR [J/CAD] RoHR

0 0

-10 0 10 20 -10 0 10 20 Crank Angle [CAD ATDC] Crank Angle [CAD ATDC] EGR 50%

800 E05 E15 E25 E38 600 Diesel

400

200 RoHR [J/CAD] RoHR

0

-10 0 10 20 Crank Angle [CAD ATDC]

Figure 5.19: Injector current (low range) and RoHR (high range) using different EGR rates for all test fuels at 1.8 bar intake pressure.

emissions due to the use of high EGR rates creates a possible environment for coagulation, accumulation, and condensation of volatile fractions on the particles thus generate larger size particles [112] and [99]. The oxygen content of E25 and E38 is high, thus increasing the local λ to have lower soot emissions.

5.4 Tests with Binary Blends

In the studies so far, test fuels are designed by adding oxygenates or aromatics to PRF blends to get a similar RON. However, even if the RON is almost the same, ignition delay is still different because it is not influenced by RON alone (MON or RON-MON is also a parameter, and CN is the most appropriate factor in determining the ignition delay of test fuels in diesel engines). In addition, the proportion of fuel used in each blend varies. Therefore, both the physical and chemical properties are changed. This research focuses on the investigation of the effect of toluene and n-butanol on the PM in terms of number concentration and size distribution by decoupling the physical and chemical properties as much as possible.

91 Chapter 5. Particulate Matter Emissions

5.4.1 Test Fuels The same volume percentage (70%) of iso-octane, toluene and n-butanol were blended with n-heptane separately. These three binary blends are called PRF, TH and BH, respectively. Firstly, these blends were tested without using a cetane number improver 2-ethylhexyl nitrate

(C8H17NO3). Here 2-EHN is short for 2-ethylhexyl nitrate and acts as an improver of CN when small quantities are added. Then, TH and BH were tested with using 2-EHN to achieve the same ignition delay as PRF. Unfortunately, the measurements of BH with 2-EHN were not possible due to the fact that a 2-micron filter in the fuel supply system was quickly and consistently contaminated, and consequently obstructed the fuel supply to the injector. Apparently, a certain chemical reaction occurs between BH and 2-EHN, but no information is found in the literature. The properties of the neat fuels are enumerated in Table 5.6 and the test fuels for different experiments are listed in Table 5.7.

Table 5.6: The main properties of a single fuel

Cetane LHV Density Auto-ignition Boiling Fuel number (MJ/kg) (kg/m3) temperature (K) point (K) N-heptane 53 [113] 44.57 [114] 684.1 [115] 493 [115] 371.4 [115] Iso-octane 15 [113] 44.43 [114] 691.9 [116] 691 [116] 372.2 [116] Toluene 9 [113] 40.59 [114] 866.8 [117] 753 [117] 383.6 [117] n-Butanol 17 [113] 33.08 [114] 809.8 [118] 616 [118] 390.25 [118] 2-EHN - 29.85 [119] 960 [119] 488 [119] 483 [120]

Table 5.7: Test fuels for different experiments

Experiments Test fuels AFst PRF 15.11 Different ignition delays TH 13.89 BH 12.22 PRF 15.11 Equal ignition delays TH+2-EHN (3.1 vol %) ≈ 13.73

5.4.2 Test Conditions This study was also done with three different loads, 5 bar IMEPg, 10 bar IMEPg and 15 bar IMEPg, respectively. According to the results of the previous study, as shown in Figure 5.4, BRF with a RON of 70 cannot achieve a stable combustion when 30% EGR is used at low load. Although the RON of TH and BH are not tested, TH should have a RON higher than 70 and BH should have a RON lower than 70. Because toluene has a RON of 106 and n-butanol has a RON of 96. Therefore, the experimental conditions are modified and summarized in Table 5.8. Combustion phasing was varied to investigate how test fuels without/with 2-EHN affect the combustion controllability and particulate number and size distribution.

92 5.4. Tests with Binary Blends

Table 5.8: Experimental settings

IMEPg (bar) 5(± 0.2) 10(± 0.2) 15(± 0.2) Engine speed (rpm) 1200 CA50 (CAD ATDC) change from -4 to 12 if possible Intake pressure (bar) 1.5 2.0 2.5 Fuel pressure (bar) 600 1200 1800 Exhaust pressure (bar) 1.8 2.3 2.8 Intake temperature (oC) 40 EGR (%) 30

Figure 5.20: ID vs CA50 at different loads with various injection timings.

93 Chapter 5. Particulate Matter Emissions

5.4.3 Effect of 2-EHN on Combustion Controllability

Figure 5.20 shows the ignition delays of PRF, TH and BH at different loads without using 2- EHN, respectively. TH has the longest ignition delay, followed by BH, and then PRF regardless of the operating load. At low load, in-cylinder temperature and pressure are relatively low compared to at higher loads. It is noteworthy that for TH controlling combustion phasing through adjusting the SOI timing at low load is hard especially when CA50 approaches TDC, as shown in Figure 5.21. Given that TH provides long ignition delay for fuel to premix with air before ignition when SOI is before -20 CAD ATDC, the fuel chemistry plays an important role under this HCCI-like combustion mode. As shown in Figure 5.22, TH displays two-stage ignition behavior. Moreover, when TH and air have a sufficient mixing time, advancing SOI from -21 to -23 CAD ATDC causes overleaning and leads to lower combustion speed and a retarded combustion phasing. This is consistent with the finding of [121] that burn duration increases with ignition delay for mainly premixed combustion. If mixing time is long, the mixture becomes leaner and will gradually approach the global λ. As the most reactive scalar is on the rich side of stoichiometry, ignition delay will become longer. Figure 5.23 (left side) shows that the ignition delay of TH is shortened to that of PRF by using 3.1% 2-EHN by volume for CA50 sweep measurements. It is clearly seen from Figure 5.23 (right side) that TH with cetane improver 2-EHN (TH+C) has a good controllability similar to PRF and its operational window regarding combustion phasing is enlarged compared to TH alone.

IMEPg ≈ 5bar 15 PRF TH BH 10

5 CA50 [CAD ATDC] 0

-5 -25 -20 -15 -10 -5 0 SOI [CAD ATDC]

Figure 5.21: SOI vs CA50 at low load

To gain a more thorough understanding of the influence of 2-EHN on combustion characteristics, the RoHRs of PRF, TH and TH+C with a constant combustion phasing (CA50 at 5 CAD ATDC) at all test loads are plotted in Figure 5.24. TH is dominated by premixed

94 5.4. Tests with Binary Blends

TH at IMEPg ≈ 5bar 600 SOI & CA50 [CAD ATDC] -13 & 8.5 -15 & 4.4 500 -21 &3.6 -23 & 5.1 SOI advances from -21 to -23 CAD ATDC 400 SOI advances from -13 to -15 CAD ATDC

300

200 RoHR [J/CAD]

100

0

-100 -15 -10 -5 0 5 10 15 20 25 30 Crank Angle [CAD ATDC]

Figure 5.22: Heat release rates of TH with different injection timings at low load.

30 15 PRF PRF TH TH 25 BH BH TH+C 10 TH+C 20

15 5

ID [CAD] ID 10 0 5 ATDC] [CAD CA50

0 -5 -5 0 5 10 15 -25 -20 -15 -10 -5 0 CA50 [CAD ATDC] SOI [CAD ATDC]

Figure 5.23: Combustion controllability in terms of ID and CA50 of TH when 2-EHN is added at 5bar IMEPg. TH+C indicates added cetane improver.

95 Chapter 5. Particulate Matter Emissions combustion from low to high load, while the proportion of premixed combustion decreases as the engine load increases, as shown in Figure 5.25. PRF and TH+C present almost the same heat release rates, especially for 10 bar and 15 bar IMEPg. At low load, the heat release rate of TH+C differs slightly from PRF due to its slightly lower ignition delay. It is concluded that an equal ignition delay can lead to the same heat release rate and thus the same combustion mode. Hence the impact of fuel structure on soot emissions can be better studied.

IMEPg : 5bar IMEPg : 10bar 600 1000 PRF PRF TH 800 TH 400 TH+C TH+C 600

200 400

200 RoHR [J/CAD] RoHR RoHR [J/CAD] RoHR 0 0

-200 -200 -10 0 10 20 30 -10 0 10 20 30 Crank Angle [CAD ATDC] Crank Angle [CAD ATDC]

IMEPg : 15bar 800 PRF 600 TH TH+C 400

200 RoHR [J/CAD] RoHR 0

-200 -10 0 10 20 30 Crank Angle [CAD ATDC]

Figure 5.24: Heat release rates of PRF, TH and TH+C with a constant combustion phasing at 5 CAD ATDC for different loads.

96 5.4. Tests with Binary Blends

1 PRF TH 0.8 TH+C

0.6 PF [-] PF 0.4

0.2

0 5 10 15 IMEP [bar]

Figure 5.25: Premixed fractions of PRF, TH and TH+C for different loads at a constant combustion phasing of 5 CAD ATDC.

5.4.4 Physical and Chemical Effects on PM

PM emissions are influenced by both physical effects and chemical properties. Physical effects include temperature, pressure, λ, ignition delay and so on [122]. In this study, only two physical factors are investigated, λ and ignition delay, to find out their effects on PM emissions. Changing a fuel will inevitably change the stoichiometric air/fuel ratio. Ignition delay differs if the cetane number of fuel and engine operating condition are not held the same. When ignition delay is not constant during a change of test fuel, the physical effects on the soot processes must be taken into account in addition to the chemical effects caused by the fuel. Even if the combustion is mainly controlled by mixing-controlled combustion, the ignition quality of the fuel influences the soot formation as it will change the lift-off length, see [122] and [123]. Cetane improver 2-EHN is added to TH to lower the ignition delay to approximately the same ignition delay as PRF. Thus the chemical effects caused by different fuels on PM emissions can be investigated separately. For a diffusion flame, fuel structure and composition have a significant influence on the sooting tendency [103] and [124]. A ranking of sooting tendencies related to the fuel structure is proposed based on a large number of previous studies, from highest to lowest ranking are: polyaromatics > aromatics > alkynes > alkenes > alkanes > alcohols [125]. As shown in Figure 5.26, TH produces less soot emissions compared to PRF at low load, while it produces more soot emissions at higher loads. The soot emissions of BH are the lowest and increase with increasing load. At low load, the combustion of TH is dominated by premixed combustion

97 Chapter 5. Particulate Matter Emissions where the very long ignition delay has a dominant effect on soot formation and produces most particles in the NM, as shown in Figure 5.27. TH has the longest ignition delay independent of operating loads and combustion phasings. However, more mixing-controlled combustion appears with the increase in load and PM emissions shift to the AM. Hence, it is concluded that the fuel structure has a bigger impact on PM emissions than ignition delay when the mixing-controlled combustion starts to influence the whole combustion process. Even though BH provides shorter ignition delays compared with TH, it contains oxygen and hence has a lower AFst. BH also has a much lower sooting tendency, hence both factors result in lower soot emissions and less particles in the AM.

0.014 PRF TH 0.012 BH

0.01

0.008

0.006 ISPM [g/kWh]

0.004

0.002

0 5 10 15 IMEPg [bar]

Figure 5.26: Soot emissions of PRF, TH, and BH with a constant CA50 (5 CAD ATDC) at different loads. The soot emission of BH at low load is zero.

Figure 5.28 shows the premixed fractions of PRF, TH, and BH with a constant CA50 (5 CAD ATDC) at all test loads. Apparently there is not a critical PF value to predict the existence of the AM, this value is closely related to the fuel used. As shown in Figure 5.28, PRF mainly produces particles in the AM with a PF of 0.6 at low load. TH produces most particles in the AM when PF reduces to 0.77 at medium load. However, BH still shows a bimodal particle size distribution with a PF of 0.54 producing most particles in the NM and the particles in the AM also increases at medium load. Given that toluene itself is A1, it is highly prone to produce PAHs [126]. TH can certainly accelerate PAH formation. Therefore, the lower soot emissions of TH compared to PRF at low and medium loads is mainly because of its longer ignition delay. Figure 5.29 displays the particle size distributions of PRF, TH, and TH+C with a constant CA50 (5 CAD ATDC) at 10 bar IMEPg. Figure 5.29 shows that TH+C produces less particles with an AD smaller than 60 nm, but more large particles (AD > 60 nm) are produced.

98 5.4. Tests with Binary Blends

Toluene is mainly consumed by the H-abstraction reactions by OH and H radicals forming a benzyl radical, and the ipso-substitution and decomposition reactions forming a phenyl radical [126]. Benzyl radical and phenyl radical are two of the most important precursors of PAH growth [127]. When TH+C no longer has the advantage of a longer ignition delay compared to PRF, substantial amounts of big particles in the AM are formed.

#107 IMEPg : 5bar #107 IMEPg : 10bar 4 2 PRF PRF

TH TH

] ] 3 3 3 BH 1.5 BH

2 1

1 0.5

dN/dlogDp [#/cm dN/dlogDp [#/cm dN/dlogDp

0 0 10 20 40 80 150 300 600 10 20 40 80 150 300 600 Dp [nm] Dp [nm]

#107 IMEPg : 15bar 2 PRF

] TH 3 1.5 BH

1

0.5 dN/dlogDp [#/cm dN/dlogDp

0 10 20 40 80 150 300 600 Dp [nm]

Figure 5.27: Particulate number concentrations and size distributions of PRF, TH and BH with a constant CA50 (5 CAD ATDC) at different loads.

99 Chapter 5. Particulate Matter Emissions

1 PRF TH 0.8 BH

0.6 PF [-] 0.4

0.2

0 5 10 15 IMEPg [bar]

Figure 5.28: Premixed fractions of PRF, TH, and BH with a constant CA50 (5 CAD ATDC) at different loads.

#106 IMEPg : 10bar 6 PRF TH TH+C 5

4

] 3

3

dN/dlogDp [#/cm dN/dlogDp 2

1

0 10 20 40 80 150 300 600 Dp [nm]

Figure 5.29: Particulate number concentrations and size distributions of PRF, TH and TH+C with a constant CA50 (5 CAD ATDC) at 10 bar IMEPg.

100 5.4. Tests with Binary Blends

#106 CA50: 3 CAD ATDC #106 CA50: 5 CAD ATDC 5 5 PRF PRF

TH TH ]

] 4 4 3 3 TH+C TH+C

3 3

2 2 dN/dlogDp [#/cm dN/dlogDp dN/dlogDp [#/cm dN/dlogDp 1 1

0 0 10 20 40 80 150 300 600 10 20 40 80 150 300 600 Dp [nm] Dp [nm]

#107CA50: 11.5 CAD ATDC 3 PRF

2.5 TH ]

3 TH+C 2

1.5

1 dN/dlogDp [#/cm dN/dlogDp 0.5

0 10 20 40 80 150 300 600 Dp [nm]

Figure 5.30: Particulate number concentrations and size distributions of PRF, TH and TH+C for different combusiton phasings at 10 bar IMEPg.

Next, the effect of changing combustion phasing on PM emissions is studied. Figure 5.30 shows the particle size distributions of PRF, TH, and TH+C for different combustion phasings at 10 bar IMEPg. When 2-EHN is added to TH, TH+C shows the same ignition delay and RoHR as PRF, as shown in Figure 5.31. The accumulation mode is composed primarily of carbonaceous agglomerates and adsorbed material. They are formed in locally fuel-rich regions of the flame, see [93] and [96]. TH+C produces more particles (AD > 60 nm) in the AM compared to PRF, this must be due to its higher sooting tendency which is a chemical/structure related property. It is seen from Figure 5.30 that the particulate number concentrations of PRF and TH+C increase with the retarded CA50 due to the decreasing combustion temperature. However, the particulate number and size of TH are almost independent of the change of CA50. Figure 5.31 shows the heat release rates of PRF, TH, and TH+C with different CA50 at 10 bar IMEPg. TH is dominated by premixed combustion and its premixed fraction increases as CA50 is retarded. Although the retarded CA50 results in a lower combustion temperature which can reduce the soot oxidation rate and increase

101 Chapter 5. Particulate Matter Emissions the soot emissions, it also increases the mixing time before combustion starts reducing fuel-rich regions leading to lower soot emissions. However, PRF and TH+C are dominated by mixing-controlled combustion at medium load. Besides, the change of CA50 does not have a noticeable impact on their ignition delays. So the decrease in combustion temperature, which mainly determines oxidation, plays a decisive role in the increase of PM emissions.

Figure 5.31: Injector current (low range) and RoHR (high range) for different CA50 at 10 bar IMEPg.

102 5.5. Characterization of Agglomerates’ Morphology

5.5 Characterization of Agglomerates’ Morphology

Previous studies show that high-octane number fuels can dramatically reduce soot emissions, but they may still cause an environmental or health problem since they generate a larger number of smaller particles compared to diesel. In this work, a FEI Tecnai 20 (type Sphera) transmission electron microscope was used to visualize the nanostructures of aggregated particles, and get more knowledge about the morphological characteristics of BH (a mixture of n-heptane and n-butanol and it is introduced in Section 5.3) and diesel [128].

5.5.1 Test Conditions In this study, the TEM was operating with a 200 kV LaB6 filament. The TEM grid having a diameter of 3 mm and 200 mesh copper grids with a lacey film was used to collect particle samples. Lacey carbon films are similar to the holey carbon films but they offer a greater percentage of open area, provide the thinnest possible support while maintaining adequate strength, and give practically no background interference in the TEM [53]. Firstly, diesel was tested, and the engine conditions are summarized in Table 5.9 (Nr. 1∼3). Secondly, BH was tested using the same engine conditions as shown in Table 5.9 (Nr 4∼6). However, TEM grids which were used to collect particle samples of BH remain clean. Thirdly, a significant amount of EGR was used and the injection pressure was lowered for BH to increase the injection duration and reduce the premixing of the BH/air mixture to try to increase soot emissions, and the test condition is shown in Table 5.10 (Nr. 7). However, the maximum PRR under such experimental conditions are extremely high, and engine knock occurs. In order to successfully collect the particles when the engine runs with BH, engine operating conditions were adjusted even more as shown in Table 5.10 (Nr. 8∼10). CA50 was kept constant at 5 CAD ATDC for all the tests except for Nr. 7 (CA50 is 10 CAD ATDC) for the reason to reduce its PRR.

Table 5.9: Engine test conditions for diesel

Nr. 1 2 3 Test fuel Diesel IMEPg (bar) 6 Injection pressure (bar) 1200 EGR (%) 20 30 35 Intake pressure (bar) 1.2 Intake temperature (oC) 40 Soot (FSN) 0.34 1.32 2.26 Engine speed (rpm) 1200

103 Chapter 5. Particulate Matter Emissions

Table 5.10: Engine test conditions for BH

Nr. 7 8 9 10 Test fuel BH IMEPg (bar) 6 7.5 Injection pressure (bar) 800 800 700 700 EGR (%) 50 39 39 41 Intake pressure (bar) 1.2 1.5 Intake temperature (oC) 40 80 Soot (FSN) 0.02 0.07 0.22 0.28 Engine speed (rpm) 1200

5.5.2 Major Results and Discussion

Ignition Dwell

BH (Nr. 4∼6) provides longer ignition dwells compared to diesel (Nr. 1 ∼ 3) under the same experimental conditions due to the lower cetane number, as shown in Figure 5.32. Zero soot emissions were produced due to the better fuel/air mixing and higher local λ. Besides, TEM grid failed to collect any particles. To increase the soot emissions to reach the detectable range, the injection pressure was reduced and the EGR rate was raised (Nr. 7). However, at such experimental condition, the soot emission is still low and ignition dwell increases dramatically so that a slight knock occurred. To be able to collect aggregated particles when the engine runs with BH, the engine operating conditions are further adjusted to allow BH to burn more easily. Therefore, a greater amount of BH is injected into the cylinder, and the intake temperature is also increased. Compared to the tests labeled 4 to 7, the ignition dwell of BH is reduced, finally resulting in detectable soot emissions (0.07 ∼ 0.28 FSN) and the TEM grid successfully collects the particles from the exhaust.

EEPS Results

Figure 5.33 shows that test fuels all exhibit a unimodal particle number size distribution under the corresponding test conditions. Diesel produces particles mainly in the AM and its particle number increases with the EGR rate. Moreover, the soot oxidation rate declines due to the lowered combustion temperature and lower oxygen concentration caused by the increased EGR rate. Consequently, smaller particles with reactive surfaces tend to grow and agglomerate to generate large particles. However, BH mainly produces particles in the NM, and its particle number is little influenced by the change of EGR rate when the same operating conditions as for diesel were used. This is obviously due to the long ignition dwell. Fuel is mostly premixed to significantly reduce fuel-rich regions, thus little carbonaceous agglomerates are formed. As a result, only particles in the NM are generated. Tests with higher intake temperature, longer burn duration, and lower injection pressure (Nr. 8 ∼ 10) produce most particles in the AM similar to the diesel experiments. As shown in Figure 5.34, due to the shorter ignition dwell,

104 5.5. Characterization of Agglomerates’ Morphology

14

12

10

8

6

4

Ignition dwell [CAD] dwell Ignition 2 Nr. 1~3 0 Nr. 4~7 Nr. 8~10 -2 1 2 3 4 5 6 7 8 9 10 Test Number [-]

Figure 5.32: Ignition dwell vs test number

BH presents more mixing-controlled combustion and consequently shows a similar particle size distribution as diesel. But most of the particles of BH still have a smaller AD compared to that of diesel. This may be caused by its higher premixed fraction.

TEM Images As mentioned earlier, when BH was tested with the same experimental conditions of diesel, it always produces zero soot emissions (0 FSN). Hence the TEM grids failed to collect any particles. When a higher EGR rate and a lower injection pressure are used (Nr.7), BH produces detectable soot emissions (0.02 FSN) but a slight knock occurred. Figure 5.35 shows the corresponding TEM image. The particles are not evenly distributed on the lacy grid, and most of the collecting area is clean. At such test condition, the engine experiences a high PRR and a high cyclic variation. So the reliability of the TEM image as shown in Figure 5.35 is not high. Hence, further analysis of test Nr. 7 is not useful. Figures 5.36 and 5.37 show TEM images of soot aggregates for diesel and BH, where the scale bars from top to bottom indicate 1 µm, 200 nm and 100 nm, respectively. A previous study [129] has shown that PPC mainly produces chain-like particles. The conventional diesel combustion shows this particle morphology of soot particles as well [130], [131], [132], [133], [134] and [135]. As observed in the figures, the vast majority of soot aggregates appears to be of a chain-like structure where primary particles are attached closely to each other. It is easy to find separate chain-like particles in the TEM image of test Nr.1, as shown in Figure 5.36. However, these chain-like particles re-agglomerate, either tightly or loosely, with the increase of EGR rate to generate even larger particles. The size of primary particles produced

105 Chapter 5. Particulate Matter Emissions

#107 Diesel #107 BH 10 10 Nr. 1

Nr. 2 Nr. 4 ]

] 8 8 Nr. 5 3 3 Nr. 3 Nr. 6 Nr. 7 6 6

4 4 dN/dlogDp [#/cm dN/dlogDp dN/dlogDp [#/cm dN/dlogDp 2 2

0 0 10 20 50 100 300 500 10 20 50 100 300 500 Dp [nm] Dp [nm]

#107 BH 10

Nr. 8

] 8

3 Nr. 9 Nr. 10 6

4

dN/dlogDp [#/cm dN/dlogDp 2

0 10 20 50 100 300 500 Dp [nm]

Figure 5.33: Particle number concentrations and size distributions for diesel and BH at various engine test conditions.

when burning diesel is not the same, the average diameter is around 20 nm. As shown in Figure 5.37, it is easy to observe that some primary particles of BH have an AD around 10 nm. The results of EEPS for tests Nr.4 ∼ Nr.7, as shown in Figure 5.33, also indicate that BH can produce a significant number of particles with an AD around 10 nm. But rarely any primary particles smaller than 10 nm are observed in Figure 5.37. Apparently, the particles measured by the EEPS with an AD less than 10 nm should not be solid carbon soot particles as also mentioned in [136] and [137]. Moreover, compared to diesel, the produced primary particles when burning BH are smaller probably due to the fuel structure and short carbon chain length. Earlier studies showed that EEPS with the original instrument matrix (IM-2004) underes- timates the sizes of particles with an AD larger than 50 nm [138]. PM emissions specifically

106 5.5. Characterization of Agglomerates’ Morphology

800 Nr. 1 700 Nr. 4 Nr. 8 600

500

400

300

200

Rate of heat release [J/CAD] release heat of Rate 100

0

-20 -10 0 10 20 30 Crank angle [CAD ATDC]

Figure 5.34: Injector current (low range) and RoHR (high range) for tests Nr.1, Nr.4, and Nr. 8.

Figure 5.35: TEM image of the collected particles emitted by BH with engine knock (Nr.7).

107 Chapter 5. Particulate Matter Emissions

Figure 5.36: TEM images of the individual and aggregated particles emitted by diesel (Nr. 1∼3). carbonaceous aggregates have different charging characteristics compared to nearly spherical particles. Aggregates get more charges at a given mobility diameter in a unipolar charging environment, correspondingly causing the aggregate mobility diameters to be underestimated [138] and [139]. A new instrument matrix (IM-Soot) was used for all the EEPS measurements in this chapter. But after comparison of the TEM results with EEPS results, it appears that the particle size of aggregates is still underestimated. Correspondingly, the particle number concentrations in the 20-120 nm size range are overestimated by EEPS. This is also confirmed by previous studies [140] and [141].

108 5.6. Conclusions

Figure 5.37: TEM images of the individual and aggregated particles emitted by BH (Nr. 8∼10).

5.6 Conclusions

High-octane fuels have shown a great advantage in reducing soot emissions. In this chapter the effects of fuel structure and selected engine parameters on particulate emissions regarding number concentration and size distribution using designed fuels were studied. The TEM was also used to visualize the nanostructure of aggregated particles. Firstly, several ternary blends with a RON 70 were tested at different loads. In general, the particle size distribution curves shift upwards and towards bigger sizes as the engine load increases for all test fuels.

109 Chapter 5. Particulate Matter Emissions

Compared to other fuels, alcohol containing fuels yield less soot emissions and produce considerably more particles in the nucleation mode. When a higher EGR rate is applied, in general the particulate number concentration in nucleation mode decreases, while more particles in accumulation mode are generated. Secondly, several binary blends were designed to further investigate the effect of different fuel structures on PM emissions. If ignition delay is not kept constant, the physical effects on the soot formation must be considered in addition to the fuel-induced chemical effects. 2-EHN was used to equalize the ignition delay of test fuels. Then the physical effects (mixing) are decoupled from the fuel-induced chemical effects. Fuel structure has less effect on the PM emissions for purely premixed combustion but has a significant influence on the sooting tendency for mixing-controlled combustion. The fuel structure plays an increasingly important role in PM emissions with the reduction of premixed fraction. Finally, the TEM was used to get more knowledge about the morphological characteristics of BH (70 vol% n-butanol blended with 30 vol% n-heptane) and diesel. The aerodynamic diameters of primary particles of diesel are generally larger than that of BH. The aggregated particles increase as the EGR rate increases. Comparing the results from TEM and EEPS, it is concluded that the particle size of aggregates may be underestimated and the number concentrations may be overestimated by the EEPS.

110 Chapter 6 Conclusions and Outlook

The heavy-duty vehicles will still rely on internal combustion engines in the foreseeable future. The major challenge facing society is to reduce the impact of the growth of energy use on local and global environment. Therefore, there is a need for more advanced combustion concepts that may use renewable fuels to achieve a low carbon footprint and low tailpipe emissions. Low temperature combustion (LTC) has a great potential to achieve high engine efficiency and ultra-low emissions of NOx and PM. A representative of such an LTC concept is PPC. It uses direct injection and provides more moderate mixing times which is key to closely link the ignition of PPC with fuel injection event, hence providing more control than HCCI. The charge distribution of PPC however is more heterogeneous at ignition compared to HCCI, though chemical kinetics still play a significant role. Previous research concluded that diesel fuel is not an ideal candidate for PPC operation. In this study, several ternary and binary blends, all having a RON in the range from 70 to 90, were designed and tested with a single injection strategy. Firstly, the effect of air-excess on PPC combustion and emission characteristics is studied. At a higher boost pressure, the cylinder charge becomes leaner as more fresh air mass is forced into the cylinder for the same amount of injected fuel. Moreover, a higher intake pressure provides more oxygen, less wall-wetting, and thus improved fuel-air mixture quality. Consequently, at a higher boost condition, fuels need to be injected later into the cylinder as they ignite more easily. In the relevant temperature regime of this work, the chemical ignition delay times become shorter with decreasing λ at near stoichiometric conditions. However, decreased boost pressure and therefore overall lower pressure and density levels respectively lead to longer chemical ignition delay times and deteriorations in mixture formation. Consequently, the observation that fuels become more difficult to ignite at conditions near global stoichiometry is a pressure and density effect rather than a λ effect. Although PPC has many merits, it also presents some limitations, as shown in Figure 6.1.

• At medium load (10 bar IMEPg), the test blends of RON 70 are dominated by premixed combustion, and produce ultra-low NOx and soot emissions simultaneously when more

111 Chapter 6. Conclusions and Outlook

Figure 6.1: Limitations of PPC presented in this study.

than 40% EGR is used. When the engine load rises to a higher load (15 bar IMEPg), these blends of RON 70 are controlled by mixing-controlled combustion. They show a NOx-soot tradeoff relationship similar to diesel. The increased soot emissions at higher loads are because of the presence of fuel-rich regions which are formed due to the lack of fuel and air mixing. This study indicates that more premixed combustion helps to reduce soot emissions. Therefore, even higher-octane fuels can be used to promote the mixing of fuel and air. Moreover, the test results show that oxygen containing fuels (in this study, ethanol and n-butanol are tested) can help to increase local λ for a given ignition delay, thus lowering soot emissions.

• Higher-octane fuels can help to increase the premixing of fuel and air, but they bring problems at lower loads due to their poor auto-ignition qualities. Three primary reference fuels having a RON of 70, 80, and 90 were tested at lower loads. Stable

combustion regarding the COVIMEP and maximum PRR is guaranteed by increasing the intake temperature and controlling the CA50. Results show that PRF90 requires a higher intake temperature (60 oC) to maintain a stable combustion than PRF80 and PRF70 (40 oC). Soot emissions are not a problem at lower loads since the test PRF blends provide sufficiently long ignition delay compared to their short injection duration to avoid the formation of fuel-rich regions. NOx emissions can also be ultra- low when more than 60% EGR is applied. However, the ignition delay becomes so long that fuel and air are over-mixed. As a consequence of the over-mixing, their HC

112 and CO emissions are very high which are the main obstacles faced by PPC technology commercially. • High-octane fuels have shown a significant advantage in reducing mass based soot emissions. In general, as the engine load increases, the particle size distribution curves shift upwards and towards bigger sizes regardless of the test fuels. Compared to other fuels, alcohol containing fuels yield less mass based soot emissions but produce consid- erably more particles in the nucleation mode. The results also indicate that when more EGR is applied, the particulate number concentration in nucleation mode decreases, while more particles in accumulation mode are generated. It is found that the particle size distribution is primarily determined by fuel/air stratification, hence ignition delay, even for fuels containing aromatic components (toluene or benzaldehyde). Only if the stratification is identical and rich conditions are available a definite impact of fuel structure can be found. TEM images show that the aerodynamic diameters of diesel primary particles are larger than that of BH (70 vol% n-butanol blended with 30 vol% n-heptane). Furthermore, the number concentration of aggregated particles increase as the EGR rate increases. It should be noted that the highly premixed combustion at low load produces particles mainly with a geometric diameter smaller than 20 nm irrespective of fuel structure, which is considered to be more threatening to human health than large particles. Therefore, more research is needed to overcome these limitations of PPC.

Future Work on PPC at Higher Loads In the thesis, all the experiments were done with a single injection strategy. Experimental results show that the premixing of fuel and air becomes worse at higher loads and cannot produce simultaneously low NOx and soot emissions even with high-octane fuels. There are two recommended strategies to improve the fuel-air mixture at higher loads: • Multiple injections could help to increase the fuel/air mixing and decrease fuel-rich regions to reduce soot emissions at higher loads. • Apply alcohol-containing fuels to provide a higher local λ for a given ignition delay. Alcohol fuels are possible fuels with a low carbon-footprint (bio-fuels, e-fuel).

Future Work on PPC at Lower Loads The HC and CO emissions of high-octane fuels at lower loads are high therefore the combus- tion efficiency also deteriorates. Unfortunately, the traditional diesel oxidation catalyst (DOC) has relatively lower effectiveness due to the lower exhaust gas temperatures. Therefore, low temperature oxidation catalysts need to be developed further so that HC and CO emissions can comply with the future emission standards. Apart from the study on low temperature oxidation catalysts, there are some ways may help to reduce HC and CO emissions at lower loads:

• Multiple injection strategies may also help to reduce HC and CO emissions by creating a fuel/air concentration gradient to reduce over-leaning.

113 Chapter 6. Conclusions and Outlook

the engine and run with a lower λ also helps to increase combustion efficiency at lower loads, but this would also lead to increased throttling losses. • Mode-shift strategies are also promising. One such example could be a combined spark assisted compression ignition (SSCI) - PPC, where SSCI is used for starting and low load operation while PPC is used for medium to high load operation. • RCCI is also a way to solve the low combustion efficiency of PPC at lower loads. The engine is started with diesel-like fuels (run under conventional diesel combustion mode) at cold start and then switched to LTC modes after adequate warm-up at higher loads.

Future Work with TEM It proved difficult to measure the particle morphology in PPC mode. The engine needed to be run at far from optimal settings, in conditions at which only particles in AM mode were generated. Obviously, the comparison of particle morphology characteristics of diesel and BH under the same test conditions have more practical value. In the future experiments, the collection of particles in NM may have some more focus, a series of approaches as shown below is suggested to achieve the study purpose. • Do experiments with different TEM grids regarding hole shape and hole size. • Adjust the location where the TEM grid is placed or use multiple TEM grids at the same time and place them at different locations. • Extend the collection time (30s, 60s, 90s and 120s) to see whether it is possible to collect particles in the NM directly (the results given in the thesis are based on a collection time of 6s). • Investigate the TEM samples which are collected from the exhaust where a bimodal particle number size distribution is presented.

114 Bibliography

[1] Shades of Green: Electric Cars’ Carbon Emissions Around the Globe, http:// shrinkthatfootprint.com/electric-car-emissions.

[2] J. Wang, Battery energy consumption modelling, testing and prediction : a practical case study, Ph.D. thesis, Technische Universiteit Eindhoven (2016).

[3] M. Williams, R. Minjares, A technical summary of Euro 6/VI vehicle emission stan- dards, Tech. rep., The international councial of clean transportation (icct) (2016).

[4] J. E. Dec, A conceptual model of di diesel combustion based on laser-sheet imaging*, in: International Congress & Exposition, SAE International, 970873, 1997.

[5] T. Kitamura, T. Ito, J. Senda, H. Fujimoto, Mechanism of smokeless diesel combustion with oxygenated fuels based on the dependence of the equivalence ration and tempera- ture on soot particle formation, International Journal of Engine Research 3 (4) (2002) 223–248.

[6] C. A. J. Leermakers, Efficient fuels for future engines, Ph.D. thesis, Technische Univer- siteit Eindhoven (2014).

[7] M. Yao, Z. Zheng, H. Liu, Progress and recent trends in homogeneous charge com- pression ignition (HCCI) engines, Progress in Energy and Combustion 35 (5) (2009) 398–437.

[8] S. Onishi, S. H. Jo, K. Shoda, P. D. Jo, S. Kato, Active thermo-atmosphere combustion (atac) - a new combustion process for internal combustion engines, in: 1979 Congress and Exposition, SAE International, 790501, 1979.

[9] M. Noguchi, Y. Tanaka, T. Tanaka, Y. Takeuchi, A study on gasoline engine combustion by observation of intermediate reactive products during combustion, in: 1979 SAE International Off-Highway & Powerplant Congress & Exposition, SAE International, 790840, 1979.

[10] T. W. Ryan, T. J. Callahan, Homogeneous charge compression ignition of diesel fuel, in: International Fuels & Meeting & Exposition, SAE International, 961160, 1996.

115 Bibliography

[11] A. W. (Bill) Gray, T. W. Ryan, Homogeneous charge compression ignition (hcci) of diesel fuel, in: International Spring Fuels & Lubricants Meeting & Exposition, SAE International, 971676, 1997.

[12] L. Xu, R. McCabe, LNT+in situ SCR catalyst system for diesel emissions control, Catalysis Today 184 (1) (2012) 83–94.

[13] C. Lambert, R. Hammerle, R. McGill, M. Khair, C. Sharp, Technical advantages of urea scr for light-duty and heavy-duty diesel vehicle applications, in: SAE 2004 World Congress & Exhibition, SAE International, 2004-01-1292, 2004.

[14] N. S. I. Alozie, G. Fern, D. Peirce, L. Ganippa, Influence of biodiesel blending on particulate matter (pm) oxidation characteristics, in: WCX™ 17: SAE World Congress Experience, SAE International, 2017-01-0932, 2017.

[15] N. R. Walker, A. B. Dempsey, M. J. Andrie, R. D. Reitz, Use of low-pressure direct- injection for reactivity controlled compression ignition (rcci) light-duty engine operation, SAE International Journal of Engines, 2013-01-1605 6 (2) (2013) 1222–1237.

[16] S. L. Kokjohn, R. M. Hanson, D. A. Splitter, R. D. Reitz, Fuel reactivity controlled compression ignition (RCCI): a pathway to controlled high-efficiency clean combustion, International Journal of Engine Research 12 (3) (2011) 209–226.

[17] R. M. Hanson, S. L. Kokjohn, D. A. Splitter, R. D. Reitz, An experimental investigation of fuel reactivity controlled pcci combustion in a heavy-duty engine, SAE International Journal of Engines, 2010-01-0864 3 (1) (2010) 700–716.

[18] D. Splitter, M. Wissink, S. Kokjohn, R. D. Reitz, Effect of compression ratio and piston geometry on rcci load limits and efficiency, in: SAE 2012 World Congress & Exhibition, SAE International, 2012-01-0383, 2012.

[19] D. Splitter, M. Wissink, D. DelVescovo, R. D. Reitz, Improving the understanding of intake and charge effects for increasing rcci engine efficiency, SAE International Journal of Engines, 2014-01-1325, 7 (2) (2014) 913–927.

[20] R. D. Reitz, G. Duraisamy, Review of high efficiency and clean reactivity controlled compression ignition (RCCI) combustion in internal combustion engines, Progress in Energy and Combustion Science 46 (2015) 12–71.

[21] X. Lu, D. Han, Z. Huang, Fuel design and management for the control of advanced compression-ignition combustion modes, Progress in Energy and Combustion Science 37 (6) (2011) 741–783.

[22] C. Noehre, M. Andersson, B. Johansson, A. Hultqvist, Characterization of partially premixed combustion, in: Powertrain & Fluid Systems Conference and Exhibition, SAE International, 2006-01-3412, 2006.

[23] T. Fang, C.-f. F. Lee, Bio-diesel effects on combustion processes in an HSDI diesel engine using advanced injection strategies, Proceedings of the Combustion Institute 32 (2) (2009) 2785–2792.

116 Bibliography

[24] C. Guido, C. Beatrice, S. Di Iorio, V. Fraioli, G. Di Blasio, A. Vassallo, C. Ciaravino, Alternative diesel fuels effects on combustion and emissions of an euro5 automotive diesel engine, SAE International Journal of Fuels and Lubricants, 2010-01-0472 3 (1) (2010) 107–132.

[25] A. Abu-Jrai, J. Rodríguez-Fernández, A. Tsolakis, A. Megaritis, K. Theinnoi, R. Cracknell, R. Clark, Performance, combustion and emissions of a diesel engine operated with reformed egr. comparison of diesel and gtl fuelling, Fuel 88 (6) (2009) 1031 – 1041.

[26] S. Gill, A. Tsolakis, K. Dearn, J. Rodríguez-Fernández, Combustion characteristics and emissions of fischer–tropsch diesel fuels in ic engines, Progress in Energy and Combustion Science 37 (4) (2011) 503 – 523.

[27] K. Kitano, R. Nishiumi, Y. Tsukasaki, T. Tanaka, M. Morinaga, Effects of fuel properties on premixed charge compression ignition combustion in a direct injection diesel engine, in: 2003 JSAE/SAE International Spring Fuels and Lubricants Meeting, SAE International, 2003-01-1815, 2003.

[28] G. T. Kalghatgi, P. Risberg, H.-E. Ångström, Advantages of fuels with high resistance to auto-ignition in late-injection, low-temperature, compression ignition combustion, in: Powertrain & Fluid Systems Conference and Exhibition, SAE International, 2006- 01-3385, 2006.

[29] M. Wei, S. Li, H. Xiao, G. Guo, Combustion performance and pollutant emissions analysis using diesel/gasoline/iso-butanol blends in a diesel engine, Energy Conversion and Management 149 (2017) 381 – 391.

[30] A. Torregrosa, A. Broatch, R. Novella, J. Gomez-Soriano, L. Mónico, Impact of gasoline and diesel blends on combustion noise and pollutant emissions in premixed charge compression ignition engines, Energy 137 (2017) 58 – 68.

[31] S. Lee, J. Jeon, S. Park, Optimization of combustion chamber geometry and operating conditions for compression ignition engine fueled with pre-blended gasoline-diesel fuel, Energy Conversion and Management 126 (2016) 638 – 648.

[32] A. Nayyar, D. Sharma, S. L. Soni, A. Mathur, Characterization of n-butanol diesel blends on a small size variable compression ratio diesel engine: Modeling and experimental investigation, Energy Conversion and Management 150 (2017) 242 – 258.

[33] S. Saravanan, B. R. Kumar, A. Varadharajan, D. Rana, B. Sethuramasamyraja, G. L. N. rao, Optimization of di diesel engine parameters fueled with iso-butanol/diesel blends – response surface methodology approach, Fuel 203 (2017) 658 – 670.

[34] S. M. Sarathy, P. Oßwald, N. Hansen, K. Kohse-Höinghaus, Alcohol combustion chemistry, Progress in Energy and Combustion Science 44 (2014) 40 – 102.

[35] X. Liu, H. Wang, Z. Zheng, J. Liu, R. D. Reitz, M. Yao, Development of a combined reduced primary reference fuel-alcohols (methanol/ethanol/propanols/butanols/n- pentanol) mechanism for engine applications, Energy 114 (2016) 542 – 558.

117 Bibliography

[36] A. K. Agarwal, A. P. Singh, R. K. Maurya, Evolution, challenges and path forward for low temperature combustion engines, Progress in Energy and Combustion Science 61 (2017) 1 – 56.

[37] H. W. Won, N. Peters, H. Pitsch, N. Tait, G. Kalghatgi, Partially premixed combustion of gasoline type fuels using larger size nozzle and higher compression ratio in a diesel engine, in: SAE/KSAE 2013 International Powertrains, Fuels & Lubricants Meeting, SAE International, 2013-01-2539, 2013.

[38] G. T. Kalghatgi, The outlook for fuels for internal combustion engines, International Journal of Engine Research 15 (4) (2014) 383–398.

[39] S. Wang, P. C. Bakker, L. M. T. Somers, L. P. H. de Goey, Effect of Air-excess on Blends of RON70 Partially Premixed Combustion, Flow, Turbulence and Combustion.

[40] H. Solaka Aronsson, I. Truedsson, M. Tuner, B. Johansson, W. Cannella, Comparison of fuel effects on low temperature reactions in ppc and hcci combustion, in: SAE 2014 International Powertrain, Fuels & Lubricants Meeting, SAE International, 2014-01- 2679, 2014.

[41] V. Manente, Gasoline Partially Premixed Combustion - An Advanced Internal Combus- tion Engine Concept Aimed to High Efficiency, Low Emissions and Low Acoustic Noise in the Whole Load Range, Ph.D. thesis, Lund University (2010).

[42] R. W. Baldauf, R. B. Devlin, P. Gehr, R. Giannelli, B. Hassett-Sipple, H. Jung, G. Martini, J. McDonald, J. D. Sacks, K. Walker, Ultrafine Particle Metrics and Research Consid- erations: Review of the 2015 UFP Workshop., International journal of environmental research and public health 13 (11).

[43] A. Ibald-Mulli, H. Wichmann, W. Kreyling, A. Peters, Epidemiological Evidence on Health Effects of Ultrafine Particles, Journal of Aerosol Medicine 15 (2) (2002) 189–201.

[44] C.A.J. Leermakers, C.C.M. Luijten, L.M.T. Somers, G.T. Kalghatgi, B.A. Albrecht, Experimental study of fuel composition impact on pcci combustion in a heavy-duty diesel engine, in: SAE 2011 World Congress & Exhibition, SAE International, 2011-01- 1351, 2011.

[45] AVL Smoke Meter - Emission Measurement, https://www.avl.com/-/avl-smoke-meter.

[46] Smoke Value Measurement With the Filter- Paper- Method, https://www.avl.com/ documents/10138/885893/Application+Notes (2005).

[47] TSI Incorporated, Operation Manual: Rotating disk thermodiluter model 379020a thermal conditioner air supply model 379030, Tech. rep., TSI (2011).

[48] J. C. Meyer, C. O. Girit, M. F. Crommie, A. Zettl, Imaging and Dynamics of Light Atoms and Molecules on Graphene, Nature LettersarXiv:0805.3857.

118 Bibliography

[49] K. Miyashita, Y. Fukuda, Y. Shiozaki, K. Kondo, T. Aizawa, D. Yoshikawa, D. Tanaka, A. Teraji, Tem analysis of soot particles sampled from gasoline direction injection engine exhaust at different fuel injection timings, in: JSAE/SAE 2015 International Powertrains, Fuels & Lubricants Meeting, SAE International, 2015-01-1872, 2015.

[50] T. Aizawa, N. Takahata, K. Okabe, Y. Mizutani, Effect of fuel aromatics on in-flame diesel soot nanostructure via hrtem, in: JSAE/SAE 2015 International Powertrains, Fuels & Lubricants Meeting, SAE International, 2015-01-1829, 2015.

[51] M. Sakai, H. Iguma, K. Kondo, T. Aizawa, Nanostructure analysis of primary soot particles directly sampled in diesel spray flame via hrtem, in: SAE 2012 International Powertrains, Fuels & Lubricants Meeting, SAE International, 2012-01-1722, 2012.

[52] K. O. Lee, R. Cole, R. Sekar, M. Y. Choi, J. S. Kang, C. S. Bae, H. D. Shin, Morphological investigation of the microstructure, dimensions, and fractal geometry of diesel particu- lates, Proceedings of the Combustion Institute 29 (1) (2002) 647 – 653, proceedings of the Combustion Institute.

[53] Lacey carbon films, http://www.agarscientific.com/lacey-carbon-films.html.

[54] M. Shen, Particulate Matter Emissions from Partially Premixed Combustion with Diesel, Gasoline and Ethanol, Ph.D. thesis, Lund University (2016).

[55] S. d’Ambrosio, R. Finesso, E. Spessa, Calculation of mass emissions, oxygen mass fraction and thermal capacity of the inducted charge in si and diesel engines from exhaust and intake gas analysis, Fuel 90 (1) (2011) 152 – 166.

[56] J. Brettschneider, Bosch Technische Berichte technisch-wissenschaftliche Mitteilungen aus Forschung, Entwicklung und Fertigung der Bosch-Gruppe, Vol. 6, Bosch, 1964.

[57] J. B. Heywood, Internal Combustion Engine Fundamentals, McGraw-Hill Book Com- pany, New York, 1988.

[58] H. Solaka, U. Aronsson, M. Tuner, B. Johansson, Investigation of partially premixed combustion characteristics in low load range with regards to fuel octane number in a light-duty diesel engine, in: SAE Technical Paper, SAE International, 2012, pp. SAE 2012–01–0684.

[59] H. Solaka, Impact of Fuel Properties on Partially Premixed Combustion, Ph.D. thesis, Lund University (2014).

[60] P. Tu, H. Xu, D. K. Srivastava, K. Dean, D. Jing, L. Cao, A. Weall, J. K. Venus, Numerical investigation of gdi injector nozzle geometry on spray characteristics, in: JSAE/SAE 2015 International Powertrains, Fuels & Lubricants Meeting, SAE International, 2015- 01-1906, 2015.

[61] C. Yao, P. Geng, Z. Yin, J. Hu, D. Chen, Y. Ju, Impacts of nozzle geometry on spray combustion of high pressure common rail injectors in a constant volume combustion chamber, Fuel 179 (2016) 235–245.

119 Bibliography

[62] W. Ning, R. D. Reitz, R. Diwakar, A. M. Lippert, A numerical investigation of nozzle geometry and injection condition effects on diesel fuel injector flow physics, in: SAE World Congress & Exhibition, SAE International, 2008-01-0936, 2008.

[63] C. Du, M. Andersson, S. Andersson, Effects of nozzle geometry on the characteristics of an evaporating diesel spray, SAE International Journal of Fuels and Lubricants, 2016-01-2197 9 (3) (2016) 493–513.

[64] J. S. Han, P. H. Lu, X. Xie, M. C. Lai, N. A. Henein, Investigation of diesel spray primary break-up and development for different nozzle geometries, in: SAE Powertrain & Fluid Systems Conference & Exhibition, SAE International, 2002-01-2775, 2002.

[65] D. T. Hountalas, T. C. Zannis, G. C. Mavropoulos, V. Schwarz, J. Benajes, C. A. Gonzalez, Use of a multi-zone combustion model to interpret the effect of injector nozzle hole geometry on hd di diesel engine performance and pollutant emissions, in: SAE 2005 World Congress & Exhibition, SAE International, 2005-01-0367, 2005.

[66] Y. Jung, C. Bae, J. Jang, D. Kim, Improvement of premixed compression ignition combustion using various injector configurations, in: SAE 2011 World Congress & Exhibition, SAE International, 2011-01-1357, 2011.

[67] C. Bae, J. Yu, J. Kang, J. Kong, K. O. Lee, Effect of nozzle geometry on the common-rail diesel spray, in: Spring Fuels & Lubricants Meeting & Exhibition, SAE International, 2002-01-1625, 2002.

[68] S. Kang, W. Cho, C. Bae, Y. Kim, Influence of the injector geometry at 250 mpa injection in a light-duty diesel engine, in: WCX™ 17: SAE World Congress Experience, SAE International, 2017-01-0693, 2017.

[69] M. Bardi, G. Bruneaux, L.-M. Malbec, Study of ecn injectors’ behavior repeatability with focus on aging effect and soot fluctuations, in: SAE 2016 World Congress and Exhibition, SAE International, 2016-01-0845, 2016.

[70] W. J. Pitz, C. J. Mueller, Recent progress in the development of diesel surrogate fuels, Progress in Energy and Combustion Science 37 (3) (2011) 330 – 350.

[71] Y. Wang, Y. Qin, X. Deng, X. Huang, The investigation on non-linear characteristic of yawing moment of twin-tailed configuration, Procedia Engineering 67 (2013) 347 – 356, 7th Asian-Pacific Conference on Aerospace Technology and Science, 7th APCATS 2013.

[72] W. Su, Y. Lu, Effect of charge density and oxygen concentration on emissions in a high density-ltc diesel engine by retarding intake valve timing and raising boost pressure, in: SAE 2010 World Congress & Exhibition, SAE International, 2010-01-1261, 2010.

[73] K. Fieweger, R. Blumenthal, G. Adomeit, Self-ignition of s.i. engine model fuels: A shock tube investigation at high pressure, Combustion and Flame 109 (4) (1997) 599 – 619.

120 Bibliography

[74] A. Maiboom, X. Tauzia, J.-F. Hétet, Experimental study of various effects of exhaust gas recirculation (EGR) on combustion and emissions of an automotive direct injection diesel engine, Energy 33 (1) (2008) 22–34.

[75] W. F. Colban, P. C. Miles, S. Oh, Effect of intake pressure on performance and emissions in an automotive diesel engine operating in low temperature combustion regimes, in: Powertrain & Fluid Systems Conference and Exhibition, SAE International, 2007-01- 4063, 2007.

[76] S. Saravanan, G. Nagarajan, S. Sampath, A correlation for the ignition delay of a ci engine fuelled with diesel and biodiesel, International Journal of Green Energy 11 (5) (2014) 542–557.

[77] E. J. Barrientos, J. E. Anderson, M. M. Maricq, A. L. Boehman, Particulate matter indices using fuel smoke point for vehicle emissions with gasoline, ethanol blends, and butanol blends, Combustion and Flame 167 (2016) 308 – 319.

[78] P. Singh, X. Hui, C.-J. Sung, Soot formation in non-premixed counterflow flames of butane and butanol isomers, Combustion and Flame 164 (2016) 167 – 182.

[79] P. G. Aleiferis, M. K. Behringer, Flame front analysis of ethanol, butanol, iso-octane and gasoline in a spark-ignition engine using laser tomography and integral length scale measurements, Combustion and Flame 162 (12) (2015) 4533 – 4552.

[80] R. Lemaire, D. Lapalme, P. Seers, Analysis of the sooting propensity of c-4 and c-5 oxygenates: Comparison of sooting indexes issued from laser-based experiments and group additivity approaches, Combustion and Flame 162 (9) (2015) 3140 – 3155.

[81] V. Manente, B. Johansson, P. Tunestal, W. Cannella, Effects of different type of gasoline fuels on heavy duty partially premixed combustion, SAE International Journal of Engines, 2009-01-2668 2 (2) (2009) 71–88.

[82] V. Manente, C.-G. Zander, B. Johansson, P. Tunestal, W. Cannella, An advanced internal combustion engine concept for low emissions and high efficiency from idle to max load using gasoline partially premixed combustion, in: SAE 2010 Powertrains Fuels & Lubricants Meeting, SAE International, 2010-01-2198, 2010.

[83] P. C. Bakker, J. De Abreu Goes, L. M. T. Somers, B. H. Johansson, Characterization of low load ppc operation using ron70 fuels, in: SAE 2014 World Congress & Exhibition, SAE International, 2014-01-1304, 2014.

[84] G. Przybyla, S. Postrzednik, Z. Zmudka, The impact of air-fuel mixture composition on SI engine performance during natural gas and producer gas combustion, IOP Conference Series: Materials Science and Engineering 148 (1) (2016) 012082.

[85] S. Mendez, J. T. Kashdan, G. Bruneaux, B. Thirouard, F. Vangraefschepe, Formation of unburned hydrocarbons in low temperature diesel combustion, SAE International Journal of Engines, 2009-01-2729 2 (2) (2009) 205–225.

121 Bibliography

[86] S. Kook, C. Bae, P. C. Miles, D. Choi, L. M. Pickett, The influence of charge dilution and injection timing on low-temperature diesel combustion and emissions, in: Powertrain & Fluid Systems Conference & Exhibition, SAE International, 2005-01-3837, 2005.

[87] M. Kashif, P. Guibert, J. Bonnety, G. Legros, Sooting tendencies of primary reference fuels in atmospheric laminar diffusion flames burning into vitiated air, Combustion and Flame 161 (6) (2014) 1575–1586.

[88] P. Borgqvist, P. Tunestal, B. Johansson, Gasoline partially premixed combustion in a light duty engine at low load and idle operating conditions, in: SAE 2012 World Congress & Exhibition, SAE International, 2012-01-0687, 2012.

[89] P. Borgqvist, P. Tunestal, B. Johansson, Comparison of negative valve overlap (nvo) and rebreathing valve strategies on a gasoline ppc engine at low load and idle operating conditions, SAE International Journal of Engines, 2013-01-0902 6 (1) (2013) 366–378.

[90] S. M. A. Rahman, H. H. Masjuki, M. A. Kalam, M. J. Abedin, A. Sanjid, H. Sajjad, Im- pact of idling on fuel consumption and exhaust emissions and available idle-reduction technologies for diesel vehicles - A review, Energy Conversion and Management (2013) 171–182.

[91] R. Lee, J. Pedley, C. Hobbs, Fuel quality impact on heavy duty diesel emissions:- a literature review, in: International Fall Fuels and Lubricants Meeting and Exposition, SAE International, 982649, 1998.

[92] T. L. Ullman, Investigation of the effects of fuel composition on heavy-duty diesel engine emissions, in: 1989 SAE International Fall Fuels and Lubricants Meeting and Exhibition, SAE International, 892072, 1989.

[93] D. B. Kittelson, Engines and nanoparticles, Journal of Aerosol Science 29 (5-6) (1998) 575–588.

[94] H. Burtscher, Physical characterization of particulate emissions from diesel engines: a review, Journal of Aerosol Science 36 (7) (2005) 896–932.

[95] D. R. Tree, K. I. Svensson, Soot processes in compression ignition engines, Progress in Energy and Combustion Science 33 (3) (2007) 272 – 309.

[96] D. Kittelson, W. Watts, J. Johnson, On- and laboratory evaluation of combustion aerosols?Part1: Summary of diesel engine results, Journal of Aerosol Science 37 (8) (2006) 913–930.

[97] M. Lapuerta, F. J. Martos, J. M. Herreros, Effect of engine operating conditions on the size of primary particles composing diesel soot agglomerates, Journal of Aerosol Science 38 (4) (2007) 455–466.

[98] G. T. Kalghatgi, P. Risberg, H.-E. Ångström, Partially pre-mixed auto-ignition of gaso- line to attain low smoke and low nox at high load in a compression ignition engine and comparison with a diesel fuel, in: 2007 Fuels and Emissions Conference, SAE International, 2007-01-0006, 2007.

122 Bibliography

[99] L. Labecki, A. Lindner, W. Winklmayr, R. Uitz, R. Cracknell, L. Ganippa, Effects of injection parameters and egr on exhaust soot particle number-size distribution for diesel and rme fuels in hsdi engines, Fuel 112 (2013) 224 – 235.

[100] Diesel Fuels, http://monographs.iarc.fr/ENG/Monographs/vol45/mono45-10.pdf.

[101] Y. Kobayashi, T. Furuhata, K. Amagai, M. Arai, Soot precursor measurements in benzene and hexane diffusion flames, Combustion and Flame 154 (3) (2008) 346–355.

[102] P. Hellier, N. Ladommatos, R. Allan, J. Rogerson, Combustion and emissions character- istics of toluene/n-heptane and 1-octene/n-octane binary mixtures in a direct injection compression ignition engine, Combustion and Flame 160 (10) (2013) 2141 – 2158.

[103] I. Glassman, Soot formation in combustion processes, Symposium (International) on Combustion 22 (1) (1989) 295 – 311.

[104] E. M. Kurtz, D. Kuhel, J. E. Anderson, S. A. Mueller, A comparison of combustion and emissions of diesel fuels and oxygenated fuels in a modern di diesel engine, SAE International Journal of Fuels and Lubricants, 2012-01-1695 5 (3) (2012) 1199–1215.

[105] M. Shen, M. Tuner, B. Johansson, W. Cannella, Effects of egr and intake pressure on ppc of conventional diesel, gasoline and ethanol in a heavy duty diesel engine, in: SAE/KSAE 2013 International Powertrains, Fuels & Lubricants Meeting, SAE International, 2013-01-2702, 2013. doi:10.4271/2013-01-2702. [106] C. Woo, S. Kook, E. R. Hawkes, Effect of intake air temperature and common-rail pressure on ethanol combustion in a single-cylinder light-duty diesel engine, Fuel 180 (2016) 9 – 19.

[107] H. Solaka Aronsson, M. Tuner, B. Johansson, Using oxygenated gasoline surrogate compositions to map ron and mon, in: SAE 2014 World Congress & Exhibition, SAE International, 2014-01-1303, 2014.

[108] ASTM D2699 - 16e1 Standard Test Method for Research Octane Number of Spark- Ignition Engine Fuel, https://www.astm.org/Standards/D2699.htm.

[109] F. Payri, J. Benajes, R. Novella, C. Kolodziej, Effect of intake oxygen concentration on particle size distribution measurements from diesel low temperature combustion, SAE International Journal of Engines, 2011-01-1355 4 (1) (2011) 1888–1902.

[110] H. Sakurai, K. Park, P. H. McMurry, D. D. Zarling, D. B. Kittelson, P. J. Ziemann, Size-dependent mixing characteristics of volatile and nonvolatile components in aerosols., Environmental science & technology 37 (24) (2003) 5487–95.

[111] J. M. Desantes, J. Benajes, J. M. Garcia-Oliver, C. P. Kolodziej, Effects of intake pres- sure on particle size and number emissions from premixed diesel low-temperature combustion, International Journal of Engine Research 15 (2) (2014) 222–235.

[112] A. Tsolakis, Effects on Particle Size Distribution from the Diesel Engine Operating on RME-Biodiesel with EGR, Energy and Fuels.

123 Bibliography

[113] M. J. Murphy, J. D. Taylor, R. L. McCormick, Compendium of Experimental Cetane Number Data, Tech. rep., National Renewable Energy Laboratory (NREL), Golden, CO () (sep 2004).

[114] , http://www.thermalfluidscentral.org/encyclopedia/index.php/ Heat_of_Combustion.

[115] N-heptane data taken from PubChem Compound Database, https://pubchem.ncbi.nlm. nih.gov/compound/8900.

[116] Iso-octane data taken from PubChem Compound Database, https://pubchem.ncbi.nlm. nih.gov/compound/10907.

[117] Toluene data taken from PubChem Compound Database, https://pubchem.ncbi.nlm. nih.gov/compound/1140.

[118] N-butanol data taken from PubChem Compound Database, https://pubchem.ncbi.nlm. nih.gov/compound/263.

[119] 2-Ethylhexyl nitrate 97% | Sigma-Aldrich, http://www.sigmaaldrich.com/catalog/ product/aldrich/293784?lang=en{&}region=NL.

[120] J. C. Bradley, A. Williams, A. Lang, Jean-Claude Bradley Open Melting Point Dataset, http://www.chemspider.com/Chemical-Structure.56557.html (jan 2014).

[121] S. Rabl, T. Davies, A. McDougall, R. Cracknell, Understanding the relationship between ignition delay and burn duration in a constant volume vessel at diesel engine conditions, Proceedings of the Combustion Institute 35 (3) (2015) 2967–2974.

[122] L. M. Pickett, D. L. Siebers, Fuel effects on soot processes of fuel jets at di diesel condi- tions, in: SAE Powertrain & Fluid Systems Conference & Exhibition, SAE International, 2003-01-3080, 2003.

[123] L. M. Pickett, D. L. Siebers, C. A. Idicheria, Relationship between ignition processes and the lift-off length of diesel fuel jets, in: Powertrain & Fluid Systems Conference & Exhibition, SAE International, 2005-01-3843, 2005.

[124] N. Ladommatos, P. Rubenstein, P. Bennett, Some effects of molecular structure of single hydrocarbons on sooting tendency, Fuel 75 (2) (1996) 114–124.

[125] I. K. Puri, Environmental implications of combustion processes, CRC Press, 1993.

[126] B. Chen, X. Liu, H. Liu, H. Wang, D. C. Kyritsis, M. Yao, Soot reduction effects of the addition of four butanol isomers on partially premixed flames of diesel surrogates, Combustion and Flame 177 (2017) 123 – 136.

[127] N. A. Slavinskaya, U. Riedel, S. B. Dworkin, M. J. Thomson, Detailed numerical modeling of pah formation and growth in non-premixed ethylene and ethane flames, Combustion and Flame 159 (3) (2012) 979 – 995.

124 Bibliography

[128] S. Kook, R. Zhang, Q. N. Chan, T. Aizawa, K. Kondo, L. M. Pickett, E. Cenker, G. Bruneaux, Ö. Andersson, J. Pagels, E. Z. Nordin, Automated detection of primary particles from transmission electron microscope (tem) images of soot aggregates in diesel engine environments, SAE International Journal of Engines, 2015-01-1991 9 (1) (2015) 279–296.

[129] H. Seong, K. O. Lee, S. Choi, C. Adams, D. E. Foster, Characterization of particulate morphology, nanostructures, and sizes in low-temperature combustion with , in: SAE 2012 World Congress & Exhibition, SAE International, 2012-01-0441, 2012.

[130] C. T. Vuk, M. A. Jones, J. H. Johnson, The measurement and analysis of the physical character of diesel particulate emissions, in: 1976 Automotive Engineering Congress and Exposition, SAE International, 760131, 1976.

[131] K. Carpenter, J. H. Johnson, Analysis of the physical characteristics of diesel particulate matter using transmission electron microscope techniques, in: 1979 SAE International Off-Highway and Powerplant Congress and Exposition, SAE International, 790815, 1979.

[132] K. O. Lee, R. Cole, R. Sekar, M. Y. Choi, J. Zhu, J. Kang, C. Bae, Detailed characterization of morphology and dimensions of diesel particulates via thermophoretic sampling, in: Spring Fuels & Lubricants Meeting & Exhibition, SAE International, 2001-01-3572, 2001.

[133] K. O. Lee, J. Zhu, Evolution in size and morphology of diesel particulates along the , in: 2004 SAE Fuels & Lubricants Meeting & Exhibition, SAE International, 2004-01-1981, 2004.

[134] J. Zhu, K. Lee, A. Panov, J. Akers, C. Habeger, An investigation of particulate morphol- ogy, microstructures, and fractal geometry for ael diesel engine-simulating , in: 2004 Powertrain & Fluid Systems Conference & Exhibition, SAE International, 2004-01-3044, 2004.

[135] J. Song, K. O. Lee, Fuel property impacts on diesel particulate morphology, nanostruc- tures, and nox emissions, in: SAE World Congress & Exhibition, SAE International, 2007-01-0129, 2007.

[136] U. Kirchner, V. Scheer, R. Vogt, R. Ka¨gi, TEM study on volatility and potential presence of solid cores in nucleation mode particles from diesel powered passenger cars, Journal of Aerosol Science 40 (1) (2009) 55–64.

[137] I. S. Abdul-Khalek, D. B. Kittelson, Real time measurement of volatile and solid exhaust particles using a catalytic stripper, in: International Congress & Exposition, SAE International, 950236, 1995.

[138] X. Wang, M. A. Grose, A. Avenido, M. R. Stolzenburg, R. Caldow, B. L. Osmondson, J. C. Chow, J. G. Watson, Improvement of engine exhaust particle sizer (eeps) size distribution measurement – i. algorithm and applications to compact-shape particles, Journal of Aerosol Science 92 95 – 108.

125 Bibliography

[139] X. Wang, M. A. Grose, R. Caldow, B. L. Osmondson, J. J. Swanson, J. C. Chow, J. G. Watson, D. B. Kittelson, Y. Li, J. Xue, H. Jung, S. Hu, Improvement of engine exhaust particle sizer (eeps) size distribution measurement – ii. engine exhaust particles, Journal of Aerosol Science 92 (2016) 83 – 94.

[140] N. Zimmerman, K. J. G. Pollitt, C.-H. Jeong, J. M. Wang, T. Jung, J. M. Cooper, J. S. Wallace, G. J. Evans, Comparison of three nanoparticle sizing instruments: The influence of particle morphology, Atmospheric Environment 86 (2014) 140 – 147.

[141] C. H. Jeong, G. J. Evans, Inter-comparison of a fast mobility particle sizer and a scanning mobility particle sizer incorporating an ultrafine water-based condensation particle counter, Aerosol Science and Technology 43 (4) (2009) 364–373.

[142] D. Kittelson, J. Savstrom, Model 3090 Engine Exhaust Particle Sizer Spectrometer, http://www.tsi.com/uploadedFiles/Product{_}Information/Literature/Spec{_}Sheets/ 3090{_}2980244A.pdf (2005).

[143] TSI, Aerosol Statistics lognormal Distributions and dN/dlogDp, http://www. tsi.com/uploadedFiles/{_}Site{_}Root/Products/Literature/Application{_}Notes/ PR-001-RevA{_}Aerosol-Statistics-AppNote.pdf (2012).

126 Appendix A Engine Exhaust Particle Sizer

A.1 How does EEPS works?

EEPS is an effective tool for measuring engine exhaust particle emissions due to its fast- response and high-resolution. Figure A.1 shows the flow schematic of EEPS. EEPS measures the size distribution and number concentration of particle emissions in the range from 5.6 to 560 nanometers, so larger particles are removed by cyclone at the inlet of EEPS. EEPS uses a corona charger to create ions. Particles mix with the ions to a predictable charge level versus particle size. Charged particles are then introduced to the measurement region and flow down the column surrounded by sheath air. A high-positive voltage is applied to the central rod and creates an electric field which repels the positively charged particles outward according to their electrical mobility. Charged aerosol particles are detected on a column of electrometers.

127 Appendix A. Engine Exhaust Particle Sizer

Figure A.1: EEPS flow schematic [142]

128 A.2. Secondary Dilution Factor

A.2 Secondary Dilution Factor

Figure A.2 shows that aerosol from the thermal conditioner with flow QMD (standard 1.5 L/min) enters the secondary dilution mixing chamber. Secondary diluted measuring gas with flow QMG is drawn by EEPS (standard 10 L/min). QAS is the flow of the secondary dilution air from the air supply part of mode 379030. The dilution factor (DF) in secondary dilution is calculated with Equation (A.1). Excess gas glow (QEX) is not less than zero, so

QAS should be larger than 8.5 L/min according to Equation (A.2). Therefore, the minimal secondary dilution ratio is 10/1.5 (6.667). In this thesis, all the measurements were done with a secondary dilution ratio of 6.7.

Q + Q DF = AS MD (A.1) QMD

QAS = QMG + QEX − QMD (A.2)

Figure A.2: Secondary dilution in 379030

A.3 Normalized Concentrations: dN/dlogDp

Aerosol distributions are predominantly lognormal in character, so data is typically plotted on a lognormal X-axis. Particle data is plotted as a function of the concentration (dN) for each particle size bin in the simplest technique. The number concentration (dN) measured by EEPS is the concentration of particles in a given channel, which is also held for SMPS. Simply plot particle concentration against particle diameter works well for one type of instrument. However, it can be confusing to compare aerosol size distribution using the same aerosol but with instruments of different resolutions. As shown in Figure A.3, both plots of data are taken with SMPS with the same aerosol. The distribution on the left is plotted using 64 channel resolution, and the the right is the exact same data plotted with 32 channel resolution. The data on the right looks like twice the data on the left. However, this is just an illusion from different resolutions. The concentration on the right is double because the width of the bin is twice that

129 Appendix A. Engine Exhaust Particle Sizer of the left channel. The normalized concentration (dN/dlogDp), which is normalized to the bin width, allows easy comparison among different instruments. The SMPS is typically used with 64 channels per decade resolution. Therefore, to calculate the normalized concentration, the concentration value of each bin is multiplied by 64 (dN/dlogDp=1/64). The EEPS used in this thesis has 16 channels per decade, so the normalized concentration is calculated by concentration of each bin multiplied by 16. Particle size is presented on a log scale with 32 equally sized bins in the size range of 5.6 to 560 nanometers. Figure A.4 shows the comparison of normalized concentration taken with 64 channel resolution and 32 channel resolution using the same aerosol tested by SMPSTM. It is seen that the peak concentration values are very similar for different channel resolutions. It is worth noted that normalized concentrations are several orders of magnitude larger than the concentrations not normalized. This is due to the high resolution and small bin width (much less than one) of particle instruments.

Figure A.3: Particle concentration measured by SMPSTM plotted against particle diameter taken with 64 channel resolution (left) and 32 channel resolution (right). The figure is from [143].

Figure A.4: Normalized particle concentration measured by SMPS plotted against particle diameter taken with 64 channel resolution (left) and 32 channel resolution (right). Tha small difference in mode concentration is because of the slight particle bin changes in the regions near the bin boundaries. The figure is from [143].

130 Acknowledgement

This acknowledgement brings an end to my PhD thesis. I really have enjoyed and gained a lot during the past four years. The completion of this work would not have been possible without the support, help, and guidance from some individuals who contributed their invaluable assistance.

First and foremost, I would like to express my heartfelt thanks to my supervisor, Bart Somers. Bart, it would be impossible to complete this thesis without your support. Thanks for spending a lot of effort on correcting my papers and thesis. Thanks to your nice attitude, continuous encouragements, strong support and confidence in me. You praised me for my achievement and provided me with freedom to work on this project, which allows me to grow as an independent researcher. I really enjoyed working with you.

My special thank goes to my promoter, Philip de Goey. Philip, your precision in profes- sional scientific work and admirable research experience made a very strong impact on me. Your insightful suggestions have pushed me to think wise and work hard throughout my research work. I am also very grateful that you spent your holiday time on correcting my thesis instead of enjoying the sunshine on the beach. :)

I would like to thank Anne Spoelstra and Heiner Friedrich, who played important roles in TEM tests. They provided scientific advice and many insightful discussions. Apart from this, their wealth of knowledge in the field of materials and interface chemistry is inspiring. I really learned a lot from them and enjoyed the cooperation.

I would like to thank all the committee members for reviewing, commenting and approv- ing this work.

I express my sincere gratitude to my officemates, friends, and colleagues for contributing immensely to my personal and professional time at Eindhoven. Thanks for all of you making

131 Acknowledgement my PhD life filled with lots of joy and fun.

My special thanks are dedicated to my family, who have been loving, encouraging and supporting me throughout the years. Last but not least, I would like to thank my husband Wang Miao, for brightening all my days with his unconditional love. There are so many things my heart wants to say to you, all of which can be summed up in just three words — thanks for everything. I love you.

132 Curriculum Vitae

Education

• 2013—2017 PhD studies at Mechanical Engineering Department, Eindhoven University of Technol- ogy (TU/e), Eindhoven, the Netherlands.

• 2010—2013 MSc studies at the State Key Laboratory of Engines, Mechanical Engineering Depart- ment, Tianjin University, Tianjin, China.

• 2006—2010 BSc studies at Mechanical Engineering Department, Tianjin University, Tianjin, China.

Publications

• S. Wang, X. Zhu, L. M. T. Somers, and L. P. H. de Goey, “Effects of exhaust gas recirculation at various loads on diesel engine performance and exhaust particle size distribution using four blends with a research octane number of 70 and diesel", Energy Conversion and Management, 2017.

• S. Wang, K. V. D. Waart, L. M. T. Somers, and L. P. H. de Goey, “Experimental study on the potential of higher octane number fuels for lowload partially premixed combustion", SAE Technical Paper 2017-01-0750, 2017.

• S. Wang, X. Zhu, L. M. T. Somers, and L. P. H. de Goey, “Combustion and Emis- sion Characteristics of a Heavy Duty Engine Fueled with Two Ternary Blends of N- Heptane/Iso-Octane and Toluene or Benzaldehyde", SAE Technical Paper 2016-01-0998, 2016.

• S.Wang, P. C. Bakker, L. M. T. Somers, and L. P. H. de Goey, “ Effect of Air-excess on Blends of RON70 Partially Premixed Combustion", Flow, Turbulence and Combustion, 2015.

133 Curriculum Vitae

• M. Fathi, O. Jahanian, DD. Ganji, S. Wang, and L. M. T. Somers, “Stand-alone single- and multi-zone modeling or direct injection homogeneous charge compression ignition (DI-HCCI) combustion engines", Applied Thermal Engineering, 125 , 1181-1190, 2017.

• X. Luo, S. Wang, B. de Jager, and F. Willems, “Cylinder Pressure-based Combustion Control with Multi-pulse Fuel Injection", IFAC-PapersOnLine, vol. 48, no. 15, pp. 181–186, 2015.

Conference Proceedings

• S. Wang, S. Sprengers, L. M. T. Somers, and L. P. H. de Goey, “Particulate Matter Emission from a Heavy-duty Diesel Engine with Three Binary Blends", 12th conference on sustainable development of energy, water and environment systems, Dubrovnik, 2017.

• S. Wang, X. Zhu and L. M. T. Somers, “Effects of EGR at various loads on diesel engine performance and exhaust particle size distribution using four blends of RON70 and diesel”, 11th conference on sustainable development of energy, water and environment systems, Lisbon, 2016.

• S. Wang, A. J. M. De Visser, and L. M. T. Somers, “Experimental Study on Low Temperature and Diesel Conventional Combustion by fueling blends of RON70", Proceedings of the 7th European Combustion Meeting (ECM 2015), Budapest, 2015.

• S. Wang, P.C. Bakker, A.J.M. De Visser, and L. M. T. Somers, “Effect of air-excess on blends of RON70 partially premixed combustion", Proceedings SPEIC14 – Towards Sustainable Combustion, Lisbon, 2014.

134