Evangelos G. Giakoumis, ConstanInternal Combustion Engines Laboratory, Thermal Engin

a r t i c l e i n f o

Article history:Received 28 December 2011Accepted 26 April 2012Available online 3 June 2012

Keywords:Diesel engineBiodieselTransient operationExhaust emissionsTurbocharger lagTransient cycle

1.1. Biofuels and biodiesel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6921.2. Diesel engine transient operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6931.3. Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 693

2. Historical overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 694

* Corresponding author. Tel.: 30 210 7723529; fax: 30 210 7723531.

Contents lists available at SciVerse ScienceDirect

Progress in Energy and Combustion Science

Progress in Energy and Combustion Science 38 (2012) 691e715E-mail address: cdrakops@central.ntua.gr (C.D. Rakopoulos).eliminating, at least in part, any inefciencies caused by the use of biodiesel.Based on a large amount of published data over the last twenty years, best-t correlations are

deducted for quantication of biodiesel benets or penalties on all regulated pollutants during varioustransient/driving cycles. Also, a detailed list is provided summarizing data from all published works onthe subject during the last two decades.

2012 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6920360-1285/$ e see front matter 2012 Elsevier Ltd.doi:10.1016/j.pecs.2012.05.002tine D. Rakopoulos*, Athanasios M. Dimaratos, Dimitrios C. Rakopouloseering Department, School of Mechanical Engineering, National Technical University of Athens, Athens, Greece

a b s t r a c t

The transient operation of turbocharged diesel engines can prove quite demanding in terms of engineresponse, systems reliability and exhaust emissions. It is a daily encountered situation that drasticallydifferentiates the engine operation from the respective steady-state conditions, requiring careful anddetailed study and experimentation. On the other hand, depleting reserves and growing prices of crudeoil, as well as gradually stricter emission regulations and greenhouse gas concerns have led to an ever-increasing effort to develop alternative fuel sources, with particular emphasis on biofuels that possessthe added benet of being renewable. In this regard, and particularly for the transport sector, biodieselhas emerged as a very promising solution.The target of the present work is to review the literature regarding the effects of diesel-biodiesel

blends on the regulated exhaust emissions of diesel engines operating under transient conditions(acceleration, load increase, starting and transient cycles). The analysis focuses on all regulated pollut-ants, i.e. particulate matter, nitrogen oxides, carbon monoxide and unburned hydrocarbons; results arealso presented for combustion noise and particle size concentration/distribution. The most importantmechanisms of exhaust emissions during transients are analyzed based on the fundamental aspects oftransient operation and on the impacts the physical and chemical properties of biodiesel have relative toconventional diesel oil. Biodiesel feedstock, transient cycle and fuel injection system effects are alsodiscussed.For the majority of the reviewed transients, a decreasing trend in PM, HC and CO, and an increasing

trend in NOx emissions is established when the biodiesel ratio in the fuel blend increases. Irrespective ofdriving cycle type, the NOx emission penalty and the PM benet with biodiesel seem to increase for moreaggressive cycles/driving patterns. Moreover, biodiesels produced from unsaturated feedstocks tend toincrease the NOx emission liability, at least for older production engines; no such correlation has beenestablished for the emitted PM, HC or CO. Since the research so far stems from engines optimized fordiesel fuel, application of a revised engine calibration (e.g. EGR, injection system) can prove very useful inbiodiesel fuel blendsReview

Exhaust emissions of diesel engines operating under transient conditions with

journal homepage: www.elsevier .com/locate/pecsAll rights reserved.

. . . .

. . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . .. . . .iodi. . .

agricultural sector, which collectively form not only one of themain molecule [4e7]. The more widely used biodiesels are rapeseed

andalternative fuels emerges as a very promising, long-term, alterna-tive solution in order to achieve the desired diversication frompetroleum products. Motivated by these facts, the European Union(EU) issued a directive on the use of biofuels accounting to at least5.75% of the market for gasoline and diesel fuel sold as transportfuels by the end of 2010, and 10% up to 2020 (EU Directive 2009/28/EC). Likewise, a commitment by the US government in the early2000s to increase bio-energy three-fold in ten years enforced thesearch for viable biofuels and led to a rapid growth in the share ofbiofuels, which is expected to be boosted in the next yearsfollowing the most recent US EPA RFS2 act (Renewable FuelsStandard version 2).

In parallel during the last decades, the increasingly stringentexhaust emission regulations have dominated the automotiveindustry and forced manufacturers to new developments. Morethan in the past, a combination of both internal measures andefcient exhaust after-treatment devices will be required; an

fatty acid methyl esters (FAME).The major biodiesel advantage relative to diesel fuel is its

renewability. Relevant life cycle assessment (LCA) analyses,although highly dependent on the methodology employed andhence usually variable in their results, have shown that the source-to-wheel CO2 emissions from neat biodiesel combustion accountfor at least 50% savings with respect to petroleum diesel fuel. This isan extremely promising fact with regards to increasing globalwarming concerns from the transport sector but other issuesshould not be ignored, such as food prices and biodiversity. In viewof these concerns, it is not surprising that second-generationbiodiesels from non-edible feedstock (e.g. algae) are under devel-opment [4,5,7]. On the other hand, the major technical barriersassociated with the use of biodiesel are its higher production cost(largely owing to the cost of the feedstock, except, of course, for thecase where waste cooking oils are used), its susceptibility tooxidation as well as its poor low temperature properties. It is alsoengines/vehicles constitute an important eld, where the use ofconsumers of fossil fuels but also one of the major contributors toenvironmental pollution. Thus, automotive, truck and non-road

methyl ester (RME) in Europe and soybean methyl ester (SME) inthe US; other popular biodiesels are palm, sunower, cottonseed,waste cooking and tallow methyl esters, collectively known as3. Biodiesel physical and chemical properties . . . . . . . . . . . . . . . . . . . . . . . . . .4. Effects of biodiesel blends during transient operation . . . . . . . . . . . . . . .

4.1. Nitrogen oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.1.1. Main mechanisms and trends during transients . . . . . . .4.1.2. Parametric investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.1.3. Overall results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.1.4. Methods for compensation of the NOx increase . . . . . . . .

4.2. Particulate matter and smoke . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.2.1. Main mechanisms and trends during transients . . . . . . .4.2.2. Diesel particulate filter regeneration rate . . . . . . . . . . . . .4.2.3. Overall results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.2.4. Particle number concentration and distribution . . . . . . .

4.3. Carbon monoxide and unburned hydrocarbons . . . . . . . . . . . . . . . .4.4. Combustion noise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5. Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6. Future research directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Details of the papers dealing with transient exhaust emissions from bReferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction

1.1. Biofuels and biodiesel

Energy security is a key ingredient for the economic stability ofevery country, particularly for those that do not have adequatefossil or nuclear resources. In this regard, depleting reserves andrising prices of crude oil, having placed negative loads on the tradebalances of the non-oil producing countries, are posing a severethreat to the world economy since petroleum prices form the basisof the world industrialization. These facts, coupled with thecontinuously growing concern over global warming and environ-mental degradation in general (e.g. acid rain, smog, climatechanges), have accentuated the public and scientic awareness andled to a substantial effort to develop alternative fuel sources.Among those, biofuels have assumed a leading role since theypossess the critical benet of being renewable, showing thus an adhoc advantage in reducing the emitted carbon dioxide (CO2) [1].One of the most prominent areas where the high demand forpetroleum-based fuels manifests itself is the transportation and

E.G. Giakoumis et al. / Progress in Energy692environmentally friendly solution, even though on a supplemen-tary basis, can be the use of biofuels [2].. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 694

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 696. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 696. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 696. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 699. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 701. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 702. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 702. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 702. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 704. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 704. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 705. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 706. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 708. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 709. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 709esel-diesel blends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 710. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 713

The term biofuel refers to any fuel that is derived from biomass,such as sugars, vegetable oils, animal fats, etc. Biofuels made fromagricultural products (oxygenated by nature) reduce the coun-tries dependence on oil imports, support local agriculturalindustries and enhance farming incomes, while offering benetsin terms of sustainability and reduced particulate matter (PM)emissions. What is equally important, from an economic point ofview, is that they are more evenly distributed than fossil ornuclear resources, since they can be produced domestically.Consequently, they constitute a long-term measure to, at least inpart, increase energy security and diversity. Among the biofuelscurrently in use or under consideration, biodiesel (methyl or ethylester) is considered as a very promising fuel for the transportationsector since it possesses similar properties with diesel fuel, it ismiscible with diesel practically at any proportion, and iscompatible with the existing distribution infrastructure; more-over, biodiesel is less toxic than petrodiesel and is also bio-degradable [3,4]. Biodiesel is produced by transesterication ofvegetable oils, animal fats or recycled cooking oils, and consists oflong-chain alkyl esters, which contain two oxygen atoms per

Combustion Science 38 (2012) 691e715highly doubtful that biodiesel could ever be available in sufcientquantities to offset the use of petrodiesel.

collective form of transient cycles. The analysis that follows willprimarily focus on the two most inuential diesel engine pollut-ants, PM and NOx, but results will be also presented for the othertwo regulated pollutants CO and HC, as well as for particle sizeconcentration and combustion noise radiation.

The usual approach when analyzing alternative fuels impacts onexhaust emissions is by discussing the differing physical andchemical properties of the various blends against those of thereference fuel. Consequently, the composition and properties ofbiodiesel, together with its combustion and emission formationmechanisms, will form the basis for the interpretation of theexperimental ndings. What is equally important is that, for therst time, emphasis will be placed on the specic attributes anddiscrepancies encountered during transients too (most notably inthe form of turbocharger lag), which may enhance or alleviate thedifferences observed between the biodiesel blends and the neatdiesel operation. Through the analysis that follows it is believedthat

Light will be shed into the underlying exhaust emission trendsunder transient conditions,

Possible differentiations in the emission mechanisms withrespect to steady-state conditions will be identied,

Valuable relations will be proposed based on the large amountof experimental data surveyed regarding the quantitative effectof biodiesel blend on engine emissions during transient cycles,

Potential problems or inefciencies associated with biodieseluse will be highlighted,

Conclusions will be derived that can prove valuable to bothengine manufacturers, (inter)national administrations andinstitutions, and nally

andThe use of biodiesel during steady-state operation has beenresearched heavily during the last decades (e.g. [4,6e14]). Forexample, a search in the Elsevier database at the end of 2011 withthe term biodiesel returns more than 2000 papers that containthis word in the articles keywords. The majority of the researchersreport decrease in the amount of emitted soot (this decrease isroughly proportional to the increase in the biodiesel ratio in the fuelblend), decrease in unburned hydrocarbons (HC) and carbonmonoxide (CO - both already low enough for most turbocharged DIdiesel engines), and slight to moderate increase of nitrogen oxides(NOx). The exact percentage of emission change relative to the neatdiesel operation varies considerably; it depends on the originatingvegetable oil or fat as well as the biodiesel percentage in the fuelblend and the examined test conditions as demonstrated, forexample, by a statistical analysis published by the US EPA in 2002[15]. Another comprehensive study has been reported in [12]comparing various methyl esters with their originating vegetableoils (blends of 10 and 20% v/v with diesel fuel), and four reviews(none of which focused on transient conditions) are availableregarding performance, combustion, emissions and emissioncontrol during diesel engine operation [6,8,16,17].

1.2. Diesel engine transient operation

Although the diesel engine has for many decades dominatedthe medium and medium-large transport sector based on its reli-ability, fuel efciency and the ability to operate turbocharged,discrepancies in the form of increased exhaust smokiness andnoise have delayed its inltration (and popularity) in the highlycompetitive automotive market. Traditionally, the study of dieselengine operation has focused on the steady-state performance.However, the majority of daily driving schedules involve transientconditions. In fact, only a very small portion of a vehicles operatingpattern is true steady-state; one notable example is when cruisingon a motorway.

The fundamental aspect of turbocharged transient conditionslies in their operating discrepancies compared with the respectivesteady-state ones. Whereas during steady-state operation, enginespeed and fueling, hence all the other engine and turbochargerproperties remain essentially constant, under transient conditionsboth the engine speed and fuel supply change continuously.Consequently, the available exhaust gas energy varies, affecting theturbine enthalpy drop and, through the turbocharger shaft torquebalance, the boost pressure and the air-supply to the enginecylinders are inuenced too. However, due to various dynamic,thermal and uid delays, mainly originating in the turbochargermoment of inertia, combustion air-supply is delayed comparedwith fueling, adversely affecting torque build-up and vehicledriveability [2]. As a result of this delay in the response between air-supply and fueling, particulate and gaseous emissions peak waybeyond their acceptable, steady-state values [2,18,19], as isdemonstrated in Fig. 1. Acknowledging these nowadays wellestablished facts, various legislative directives in the EU, the US andJapan, have drawn the attention of manufacturers and researchersto the transient operation of diesel engines in the form of transientcycles certication for new vehicles [20,21]. Not surprisingly, theresearch on transient diesel engine operation has also expanded(e.g. [18,19,22,23,24,25]), emphasizing on the (experimental)investigation of exhaust emissions.

1.3. Objectives

The target of the present work is to review the literatureregarding the impacts of diesel-biodiesel blends on the (regulated

E.G. Giakoumis et al. / Progress in Energyexhaust) emissions of compression ignition engines under the verycritical transient conditions encountered in the every-day engine/vehicle operation i.e., acceleration, load increase, starting and in the

54321Time (s)

Boos

t Pre

ssur

e

Fu

el P

ump

Rac

k Po

sitio

n

Soot

Opa

city

NO

x

Fig. 1. Qualitative fueling vs. air-supply (boost pressure) response highlighting theturbocharger lag and its inter-relation with smoke opacity and NOx emission spikes.

Combustion Science 38 (2012) 691e715 693 Interesting areas for future research will be located.

2. Historical overview

Despite the plethora of published papers on steady-state dieselengine emissions when using biodiesel, the inherent difculties inmeasuring and analyzing exhaust emissions during transients hasled to a narrower investigation in this subject so far. Table A in theAppendix provides a list of the published papers in internationalJournals and well established Conferences, as well as reports fromrenowned research centers in chronological order that all deal withexhaust emissions during (truly) transient conditions, when theengine runs on various biodiesel blends [26e103]. A few commentsand conclusions derived from the list of the surveyed publicationsare summarized below:

All the investigations in Table A deal with four-strokepassenger cars and medium/heavy-duty or non-roadengines/vehicles. This is not unexpected since four-strokeengines have dominated the vehicular industry during thelast decades. To this end, earlier two-stroke studies have been

particular transient schedules that contribute mostly, althoughinstantaneous emissions were sometimes gathered too, e.g.[34,35,49,65,71], however on a second by second basis. None-theless, these investigations fall technically into the transientcategory too, and will actually comprise a very important partof the discussion that follows owing to their popularity andvast amount of published results.

As is also documented in Fig. 3, roughly 30% of the experi-mentations focused on the American FTP heavy-duty transientcycle, and another one quarter on the European passenger carNEDC. Paradoxically, the American light-duty cycles and theEuropean ETC heavy-duty one have only sporadically beendealt with.

The most investigated biodiesel blends are B20 (i.e. 20% v/vbiodiesel-blended with 80% conventional diesel) and B100(neat biodiesel) (Fig. 4). The most popular methyl ester in theresearch so far is soy-derived; it has been the subject in half ofthe studies and 40% of all investigated biodiesels during tran-sients (Fig. 3).

number decreases as the number of double bonds increases, i.e.as the methyl ester becomes more unsaturated. Highly

lica

E.G. Giakoumis et al. / Progress in Energy and Combustion Science 38 (2012) 691e715694excluded from the current analysis and the quantitative resultsthat will be presented later in the text. The same holds forindirect injection engines, minimal data from which areincluded, only for the sake of completeness. Moreover, to thebest of the authors knowledge, there are no works available inthe open literature concerning industrial, locomotive ormarine engines (load increase) transient emissions withbiodiesel.

The research was initiated in the early 90s mainly in the US onheavy-duty, highway engines, and has intensied heavily in therecent years, as is documented in Fig. 2, particularly withrespect to light-duty engines/cycles.

The impacts of biodiesel blends on engine emissions duringvarious legislative transient cycles constitute the most prolicsegment of the research (Fig. 3). Only a few publications (e.g.[50,63,73,75,84,103]) investigated discrete transient schedules,i.e., specic acceleration, load increase and starting events. Theobvious advantage of the latter category is that instantaneoustransient emissions were measured applying high responseexhaust analyzers.

All researchers of the former category (transient cycles) treatedand analyzed their measurements in a quasi steady-statemanner by providing cumulative emission values throughoutthe cycle (recall that this is actually the objective of a legislateddriving cycle), without provision of the parts of the cycle or its

Pub

0

2

4

6

8

10

12

14

16

18

Nu

mb

er o

f P

ap

ers

/R

ep

ortsFig. 2. List of published papers/reports on transient diesel engintion Year3. Biodiesel physical and chemical properties

In this section, the most important biodiesel physical andchemical properties will be briey discussed with respect to thereference diesel fuel; these properties are summarized in Table 1.The discussion will provide some valuable information for theanalysis that follows, since many of the emission mechanisms arebased on the different attributes between biodiesel and petrodiesel.From the data provided in Table 1 it can be concluded that biodieselwith respect to the reference diesel fuel contains/has:

1. 1012% by wt. oxygen (leads to proportionally lower energydensity, necessitating greater mass of fuel to be injected inorder to achieve the same engine power output).

2. ultra low natural sulfur content (considered a soot precursor),with vegetable derived methyl esters having practically zerosulfur content, and animal based ones extremely small;however, this advantage seems to gradually fade owing to thecontinuous desulfurization of the petroleum diesel fuel.

3. higher cetane number (represents the ignitability of the fuel,with higher CN leading to shorter ignition delay); cetanee emissions with biodiesel blends in a chronological order.

he t

E.G. Giakoumis et al. / Progress in Energy and Combustion Science 38 (2012) 691e715 695saturated esters such as those derived from coconut, palm,tallow and used cooking oil have the highest cetane numbers.

4. lower heating value owing to the oxygen content (greater massneeds to be injected in order to assume the same enginepower).

5. higher viscosity (leads to less accurate operation of the fuelinjectors, and to poorer atomization of the fuel spray, increasein the Sauter mean diameter of the fuel droplets and of thebreak-up time; viscosity is related to the degree of unsaturationand to the fatty acid chain length (the latter varies considerablyin the pure esters but barely in the nal biodiesels)).

6. higher density (volumetrically-operating injectors injectgreater mass of biodiesel than conventional diesel fuel).

7. higher bulk modulus of compressibility (at least in part, owingto the presence of oxygen in the fuel structure, which createsa permanent dipole moment in the molecule).

Fig. 3. Quantication of investigations on transient diesel engine emissions based on tchart e right) and the specic biodiesel studied (pie chart).8. higher ash point (is a measure of the temperature to whicha fuel must be heated such that the mixture of vapor and airabove the fuel can be ignited); neat biodiesel is thus much saferthan diesel in this regard.

Fig. 4. Quantication of investigations on transient diesel eMoreover, biodiesel has no aromatic or poly-aromatic hydro-carbons (considered a soot precursor), and also exhibits betterlubricity (decreases wear, particularly in rotary and distributor-typefuel injection pumps, which employ a fuel-based lubrication).

Due to its chemical structure (it usually contains a highpercentage of unsaturated fatty acids, such as oleic, linoleic andlinolenic), biodiesel is more prone to oxidation compared withconventional diesel fuel, particularly for long-term storage. Thus,additives in the form of anti-oxidants are usually required. It is themore saturated biodiesels, e.g. from palm, coconut or tallow, thatexhibit better oxidative behavior but still inferior to mineral diesel.Moreover, biodiesel exhibits worse cold-ow properties (pourpoint and cloud point) than petrodiesel requiring cold-owimprovers. Contrary to the oxidation stability, it is the more satu-rated esters that suffer from poor low temperature attributes. Onthe other hand, the best pour and cloud point are to be found in

ype of transient schedule (bar chart e left), the specic transient cycle employed (barhighly saturated FAMEs, particularly RME.Differences in the fuel properties (viscosity, density, surface

tension, boiling point and oxygen content) between biodiesel andconventional diesel fuel are expected to affect the fuel spray

ngine emissions based on the biodiesel blend applied.

andpenetration and evaporation rate although the trends reported arenot always consistent [17]. Thus, the airfuel mixing and ametemperatures can be affected too, as reported, for example, by Choiand Reitz [104]. Specically, a marked difference between biodieseland neat diesel combustion occurs in the atomization process,given that the mean droplet size is larger when biodiesel is usedowing to its a higher kinematic viscosity and surface tension. Bothfactors are well known to affect the Sauter mean diameter of thefuel droplets, which were reported by Allen and Watts to be up to40% higher for biodiesel compared to petrodiesel [105]. This factand the different distillation curves, i.e., higher distillation curve forbiodiesel than that of the neat diesel fuel, indicate that the evap-oration process will be slower for the biodiesel, and thus couldaffect the combustion process by reducing the fraction of fuelburned in the premixed combustion phase in favor for the diffusionportion. On the other hand, the fact that biodiesel contains oxygenmay, under circumstances, enhance the reaction rate during thepremixed phase, hence lead to quicker evolution of diffusioncombustion too under higher temperatures.

Not always consistent results have been reported as regardsboth the ignition delay and the duration of combustion betweenbiodiesel and petrodiesel. The inuence of the engine operating

Table 1Comparison of key physical and chemical properties between biodiesel and auto-motive diesel fuel.

Low-sulfurautomotivediesel fuel

Biodiesel

Density/15 C (kg/m3) 820e850 870e890Kinematic viscosity/40 C (cSt) 2e3.5 3.5e6.2Cetane number w50 46e65Lower heating value (kJ/kg) w43,000 36,500e39,500Oxygen content (% weight) 0 10e12Sulfur content (ppm)

ns o

E.G. Giakoumis et al. / Progress in Energy andunit injector systems, and would not appear to be that relevantto high-pressure common rail systems, where rapid transfer ofa pressurewave does not occur [106]. In any case, an interestingstudy by Cheng at el. [112] revealed that a hydrocarbon fuelwith the same ignition delay, injected at the same timing asa neat biodiesel still produced 10% lower load-averaged NOxthan the biodiesel. Similarly, Zhang and Boehman [106] usedengine controls to eliminate injection timing differencesbetween B40 and petrodiesel, and found that B40 producedslightly lower NOx emissions under low-load but higher NOxunder high loads; these ndings probably entail that other,inherent in the biodiesel combustion, issues might be moreinuential than injection timing.

b) The fact that fuel injectors operate on a volumetric rather thangravimetric basis means that if a diesel-tuned engine runs onbiodiesel blends, a larger mass of fuel will be injected, which ismore likely to promote NOx emissions as it is expected to lowerthe local airfuel equivalence ratio and increase the respectivelocal gas temperatures (hence it is temperature-related too).

c) Moreover, the reduced fuel leakage losses in the (mechanical)fuel pump owing to the higher kinematic viscosity of biodieselcompared with the neat diesel fuel, lead to higher injectionpressures [17] and, hence, mass of injected fuel, a fact that actsin parallel to (and strengthens) the previous point.

d) Lastly, the ECU calibration may dictate a different injectionstrategy (longer injection pulse-width) based on the lower

Fig. 5. Development of NO emissions during two low-load, dynamometer acceleratioheating value of biodiesel, if an engine that is tuned for dieseloperation is required to run on biodiesel [17].

Fig. 6. Effect of soybean biodiesel content on averaged cumulative PM and gaseousemissions during the (hot) EPA Transient Cycle for a 1991 calibrated, heavy-duty dieselengine (experimental data adapted from Graboski et al. [29]).Temperature-related factors (thermal NO)

a) Local conditions during biodiesel combustion are nearer tostoichiometric compared with the neat diesel operation. Thisfact can prove quite inuential during transients, particularlyduring the crucial turbocharger lag cycles, where the airefuelequivalence ratio is expected to be lower than unity. Theunderlying mechanism here is the promoted increase in theadiabatic ame temperature, which is well known to peakaround stoichiometric conditions [113]. In other words, for theduration of the turbocharger lag, whereas during dieselcombustion the airefuel equivalence ratio might be belowunity, for biodiesel combustion, the excess oxygen inherent inthe blend leads to higher values that are now closer to stoi-chiometry, and hence promote higher gas temperatures.Interestingly, similar arguments have been raised by Muelleret al. [107] during steady-state operation. They concluded thatthe combustion process generally progressed more quickly forthe biodiesel fuels and test engine conditions examined thanfor the hydrocarbon reference fuels. Specically, owing to itshigh oxygen content, biodiesel premixes more fully during theignition delay, and a larger fraction of its heat release occursduring the premixed phase of combustion. To the extent thatthis leads to higher in-cylinder temperatures and longer resi-dence times at high temperature, this would be expected toalso increase NO emissions for the biodiesel blends.

f a medium-duty turbocharged diesel engine for neat diesel and B30 operation [73].

Combustion Science 38 (2012) 691e715 697xb) A common belief about biodiesels is that they have higher adia-

batic ame temperatures (the existence of double bonds in thefuel molecule has been associatedwith high ame temperatures[113]), which is expected to increase accordingly the emitted(thermal) NOx. Steady-state [114] and transient [36,40,48,51]emission ndings from many studies lend support to this argu-ment; biodiesels produced from unsaturated feedstocks such assoybean or rapeseed have been found to emit higher NOxcompared with the more saturated esters, such as those derivedfrom palm or tallow. However, adiabatic ame temperatures donot usually vary much with biodiesel feedstock, and measuredpeak temperatures show even less variation [17,107], so theunderlying mechanism behind the unsaturation effects remainsopen for speculation.Mueller et al. [107] found thatB100 (steady-state) combustion conditions were closer to stoichiometricduring ignition and in the standing premixed auto-ignition zonenear the ame lift-off length at higher loads than was the casewith petrodiesel. This might explain the fact why highly satu-rated biodiesels produce also lower NOx, since fuels with highcetane numbers lead to mixtures farther from stoichiometric.

for

E.G. Giakoumis et al. / Progress in Energy and698c) Another contributing cause for elevated local gas temperatureswith biodiesel combustion has been revealed by Musculus[115]. In his experimental study it was shown that actual ametemperatures inside the cylinder may be inuenced by differ-ences in soot radiative heat transfer, i.e. the heat radiation inthe combustion chamber from the formation of soot particleslowers the combustion temperatures, subsequently decreasingthe emitted NOx. Since the oxygen in a biodiesel blend is wellknown to decrease soot (Section 4.2.1), thus reducing theradiative heat transfer term, higher local gas temperatures areexpected that in turn will favor greater NOx penalty, even forfuel blends for which the equilibrium calculations have not

Fig. 7. Effect of biodiesel content on cumulative emissions during the NEDCrevealed a difference in the stoichiometric adiabatic ametemperatures. Similar ndings have been reported ina comprehensive, experimental study by Mueller et al. [107].

Fig. 8. Effect of biodiesel content on EGR valve position for a heavy-duty diesel engineoperating on SME blends and various transient cycles (experimental data adapted fromSze et al. [54]).d) An essential factor for the NOx emissions behavior in modernengines is the exhaust gas recirculation (EGR) system. Althoughthis is not a combustion but rather an engine calibration effect, itnonetheless results in temperature impacts, hence it is discussedhere. Bannister et al. [76] found that the increase of the biodieselpercentage in the fuel blend led to an approximately linearreduction in the maximum tractive force at any given speedowing to the lower energy density of methyl esters. Conse-quently, when an engine that has been calibrated for neat dieseloperation runs on biodiesel, more fuel is required to achieve thedemanded torque/vehicle speed, hence a lower EGR rate isestablished (fueling is an input to the EGR control strategy),

a Euro 4 passenger car (experimental data adapted from Lujan et al. [65]).

Combustion Science 38 (2012) 691e715escalating in-cylinder temperatures and NOx emissions. Forexample, Lujan et al. [65] measured NOx emissions 20.6%, 25.9%and 44.8% higher for B30, B50 and B100 blends respectivelycompared with the neat diesel case (Fig. 7). An even morecomprehensive approach was conducted by Sze et al. [54], whoactually quantied the EGR effect. They found that the lowercaloric value of biodiesel affected throttle position (up to 3.3%relative to the petroleum diesel operation), and subsequentlyother relevant engine parameters, such as the fuel rail pressure,VGT position and boost pressure, and ultimately decreased theEGR rate up to 10.5% (depending on the cycle) for the casesexamined. Typical results from their study are reproduced inFig. 8 that illustrates the EGR valve position change for B20 andB50 SME blends relative to mineral diesel fuel for various tran-sient cycles of a heavy-duty diesel engine. Apart from the mostlight-loaded cycle, biodiesel addition in the fuel blend resulted innoteworthy reduction in the EGR valve position.

Lastly, differences in the chemical kinetic pathways that formNOxwhen biodiesel is used have been reported, that may contribute tothe higher biodiesel NOx emissions. Arguments based on differencesin prompt NO formation seem to be the most common in this cate-gory. Such arguments typically rely on increased levels of CH beingproduced at the auto-ignition zone during biodiesel combustion (e.g.[116]), which leads to the production of N-atoms in the jet core fol-lowed by prompt NO formation, once themixture is convected to thediffusion ame, where oxygen and OH are present [107].

It should be mentioned that many of the issues discussed abovearise when diesel-tuned engines are tested/run on a fuel that hasdifferent properties; consequently, many of these issues could bedealt with applying an appropriate recalibration of the engine.Supporting evidence to this has been provided, for example, bySzybist et al. [16], who showed that once the inadvertent advance inSOI timing is corrected for, the NOx emissions were no longer fuelcomposition dependent. Thus, application of a different injectionstrategy e.g. in the form of a different timing and duration of

injection could compensate for biodiesels higher values of densityand bulkmodulus of elasticity, and address the primarymechanismof NOx emission increase (earlier start of injection), at least formechanically-controlled fuel injection systems.

4.1.2. Parametric investigation4.1.2.1. Impact of biodiesel feedstock. One of the most intriguingpoints regarding the biodiesel impact on NOx emissions is the effectof fuel properties and molecular structure. Since biodiesel can be

Fig. 9. Effects of density and cetane number on NOx emissions during transient cycles from four different heavy-duty diesel engines (data from Peterson et al. [36]; adapted fromMcCormick et al. [40]; data from McCormick et al. [48]; data from Knothe et al. [51]).

E.G. Giakoumis et al. / Progress in Energy and Combustion Science 38 (2012) 691e715 699Fig. 10. Statistically signicant effects of iodine number on NOx emissions during the UDDSet al. [36]; adapted from McCormick et al. [40,48]; data from Knothe et al. [51]; adapted fr, FTP and NEDC cycles for ve different diesel engines/vehicles (adapted from Petersonom Bakeas et al. [88]).

produced from a variety of vegetable or animal feedstocks, it seemsreasonable to assume that the different chemical composition andstructure of the originating oil or fat may inuence the engineprocesses and exhaust emissions. It is Peterson et al. [36] andMcCormick et al. [40,48], who have conducted the most illumi-

less saturated esters emitted the lowest NOx. A highly linear rela-tionship (R2 0.93) between iodine number (a usual measure ofthe fuels unsaturation) and emissions of NOx during the FTP wasalso established, as is illustrated in Fig. 10.

In general, it was found that high stearate fuels with few doublebonds produced signicantly less NOx than certication diesel.Unfortunately, these materials have poor cold-ow properties andsome are even solid at room temperature. Interestingly, a laterstudy by McCormick et al. [48], this time on a common rail, heavy-duty diesel engine, showed that B100 blend tests, in general,conrmed the above results, although the R2 was found lower thistime, of the order of 0.83 (Fig. 10); the latter implies fuel injectionsystem effects too, primarily through the weaker inuence of thebiodiesels higher bulk modulus of elasticity relative to neat dieselfuel as was discussed previously, but EGR denitely had animportant role too. On the other hand, the B20 test results were notfound statistically signicant, possibly owing to the much lowerabsolute values of NOx, which made feedstock effects more difcultto observe.

In parallel, Peterson et al. [36] studied biodiesel blends from 6different real-world feedstocks during the hot UDDS cycle fora medium-heavy-duty diesel engine, and identied a similarstatistically signicant correlation between iodine number and NOxemissions (Fig. 10). However, no association could be establishedbetween unsaturation and PM, or between NOx emissions and

Fig. 11. Effect of average cycle power on NOx emissions for a 2006, medium-duty dieselengine (adapted from Sze et al. [54]).

E.G. Giakoumis et al. / Progress in Energy and Combustion Science 38 (2012) 691e715700nating research on the subject so far, demonstrating the relationbetween nitrogen oxides and degree of unsaturation, both studyingthe performance of heavy-duty diesel engines during transientcycles.

McCormick et al. [40] studied a group of 21 different biodieselfuels (comprising of 7 real-world feedstocks and 14 pure fatty acidmethyl and ethyl esters) during the FTP cycle on a 1991 MY engine.The left sub-diagram of Fig. 9 shows the regression model reachedfor NOx emissions with density (R2 0.88); the relationshipbetween cetane number and NOx emissions is shown in the rightsub-diagram of Fig. 9. Since the impact of molecular structure isimplicit in either the density or cetane number, more saturatedesters which have higher cetane numbers and lower densities thanFig. 12. Collective NOx emission change when using various biodcetane number (right sub-diagram of Fig. 9), but a satisfactorycorrelation was found between NOx and density (left sub-diagramof Fig. 9).

Similar observations concerning the effect of unsaturationon NOx emissions have been reported by Knothe et al. [51],Fontaras et al. [74] and Bakeas et al. [81,88]. The latter additionallyidentied the glycerol content in the biodiesel blend as suspect forhigher NOx.

4.1.2.2. Impact of the transient cycle pattern. Another importantobjective is the investigation of the effect of the transient cyclescharacteristics on the exhaust (and in particular NOx) emissions. Anelucidating analysis on this subject was conducted by Sze et al. [54],iesel-diesel blends (data from all transient cycles of Table A).

Although the test samplewas very narrow, there is a possibility thata trend might apply here, a fact, however, which denitely requiresfurther testing in order to prove the validity (or not) of this argu-ment. To this aim, Yanowitz and McCormick [117] gathered datafrom various B20 investigations and fuel injection systems that arereproduced in Table 2; they concluded that common rail systemsmight be liable to higher NOx (and lower PM) relative to conven-tional diesel fuel, but the dataset was not adequate enough (and thestandard deviations observed high, suggesting variance of the data)to justify an unequivocal trend. Moreover, modern engines that are

and Combustion Science 38 (2012) 691e715 701who studied a 2006 MY engine operated on 7 different heavy-dutyengine and chassis-dynamometer cycles. The fundamentalconclusion reached from their analysis was that higher NOx penaltywas observed for biodiesel-fueled engines with increasing averagetransient cycle power. Based on several observations, the changesin NOx mass emission rate due to biodiesel was found to correlatevery well with the average cycle power (R2 0.99) and the fuelconsumption (R2 1.0) of the engine. The results obtained arereproduced in Fig. 11. The apparent reasons for the higher biodieselNOx liability with increasing cycle power lies in the primary NOxproduction mechanism. For a diesel engine, increasing powerresults in higher amount of injected fuel, lower airfuel equiva-lence ratios and nally higher (peak) gas temperatures. Likewise,a more aggressive cycle is characterized by more frequent andabrupt accelerations or load increases; the latter pave the way forhigher NOx too owing to the harsher (or more frequent) turbo-charger lag phases and the lower EGR rates they induce. Forexample, in Fig. 11 the more highly-loaded HWY cycle results ingreater EGR decreases with rising biodiesel blend than the lighter-loaded FTP or UDDS ones.

Comparable results have been reached in various investigationsa) by the research group of Professor Stournas [52,70,81,88], whoregularly measured higher NOx emissions over the more aggressive(non-legislated) Artemis or Athens driving cycles relative to thesofter NEDC, b) by Wang et al. [35], who measured higher NOx onthe more aggressive 5 mi route cycle over the lighter-loaded 5 mipeak one, and c) by Durbin et al. [53], who measured higher NOxliability on the more aggressive US06 cycle relative to the FTP75.Lopez et al. [71], who found similar results, concluded also that anSCR-based deNOx system behaved more efciently as regards theNOx liability for low biodiesel percentages. Whereas the NOxemissions increase with a DPF-EGR systemwas 19.3% and 23.9% forB20 and B100 respectively relative to diesel fuel, when an urea-based SCR was applied the respective differences were remark-ably reduced for the small percentage (6.4%) but also remarkablyincreased for the neat biodiesel (38%). On the other hand, Petersonet al. [28,36] systematically measured lower NOx (and sometimeshigher PM) emissions with various biodiesel blends relative to theirreference diesel fuel, which has been, at least in part, attributed tothe fact that the UDDS chassisedynamometer cycle studied wasless aggressive than the heavy-duty, FTP engineedynamometercycle (for which NOx increase with biodiesel blends is thecommon case). This reverse behavior of NOx emissions underlight-loaded conditions (recall that low loading results in higheramount of premixed-burn combustion) has been partiallyconrmed by Fontaras et al. [64] regarding the equally light-loadedurban part of the NEDC (for B50 blends only, another B100 fuelbehaved differently), at steady-state conditions [106], and is alsoevident in the right-hand values of Fig. 8.

4.1.2.3. Impact of the fuel injection system. Finally, the enginefueling system, e.g. high-pressure, electronically controlled orpump-line-nozzle might also be related to the NOx emission trend;unfortunately, owing to the inherent difculties involved, there arefew studies that have focused on different engine technologies,which could substantiate this hypothesis. A rst nding has alreadybeen discussed in Section 4.1.2.1 as regards iodine number impactson NOx emissions between electronic unit injector [40] andcommon rail engines [48]. Another interesting nding was repor-ted by Sharp et al. [37], who found that for two electronicallycontrolled engines running on the FTP cycle, the NOx emissionsincrease depended almost linearly on the fuel blends oxygencontent, but this was not the case for a third, pump-line-nozzle fuelinjected engine (operating at a much lower injection pressure),

E.G. Giakoumis et al. / Progress in Energywhich exhibited much lower sensitivity to B20 and B100 runs.systematically equipped with common rail injection systemsemploy also EGR and other antipollution control, which affect theemission trend signicantly. Hence, the effect of the injectionsystem alone cannot be easily isolated from the interaction of theother engine sub-systems.

4.1.3. Overall resultsIn 2002, the US EPA published a comprehensive analysis of

biodiesel impacts on exhaust pollutants with a large amount ofemissions data from the 1980s and early 1990s gathered andanalyzed in order to quantify the effect of biodiesel blends on allregulated pollutants [15]. Those data included many two-strokeengine results, some steady-state cycles, and were limited tonorth-American engines, primarily heavy-duty highway onestested during the FTP cycle with SME blends. For the current survey,an update of the EPA formulas has been performed for all regulatedpollutants in order to also include measurements from the lastdecade from

newer engines with modern injection systems and EGRcontrol,

passenger car and light-duty engines running on chassis-dynamometer cycles,

European and Japanese engines/vehicles, and a wider range of real-world biodiesel feedstocks.

To this aim, all transient measurements reported in Table Awerecollected and analyzed. Fig. 12 summarizes the effects of biodieselblends on transient NOx emissions by providing a best-t approx-imation to the collected measurements. From Fig. 12, a moderateincreasing trend is established with rising biodiesel blend ratio aswas also the result of the earlier EPA report demonstrated in thesame gure. It is surprising that although a considerable amount ofnew data has been included in the current statistical analysis, theearlier EPA best-t curve remains practically unchanged for blendratios up to 50% despite the diversity in the investigated engines,vehicles, feedstocks and cycles. This result can be explained asfollows. On the one hand the majority of the newer productionengines are equipped with EGR, which as was discussed previouslytends to increase the NOx emission penalty. On the other hand,these newer production engines that are included in the dataset areprimarily passenger cars tested on the relatively light-loaded NEDC.Most probably, the synergistic effect of these two contradicting

Table 2Average B20 emissions change relative to neat diesel operation for north-Americanonly, 4-stroke engines during transient cycles by fuel injector type (standard devi-ation in parentheses; from Yanowitz and McCormick [117]).

Fuel injector type Count NOx (%) PM (%) CO (%)

Common rail 4 4 (1) 23 (10) 13 (6)Electronic unit 15 1 (3) 18 (13) 13 (8)Hydraulic unit 2 1 (15) 7 (16) 12 (29)Mechanical 1 8 6 17Pump-line-nozzle 6 0 (2) 13 (6) 25 (5)

Rotary 8 1 (4) 21 (11) 19 (6)

for example, by Eckerle et al. [119] in their simulation approach. Inview of the required ECU calibration when using biofuel blends, aninteresting study was conducted by Magand et al. [120] studyingthe performance of an engine during the NEDC running ona mixture of Fischer-Tropsch, ethanol and RME. They found thatwhen the ECU is appropriately re-adjusted to cater for the differentphysical and chemical properties of their biofuel blend, the initiallyhigh NOx penalty of the order of 60% relative to the reference dieselfuel was reversed into an impressive benet of almost 22% withoutsacrice in the PM reduction. Likewise, Zhang and Boehman [106]and Ireland et al. [121] applied both EGR and injection timingrecalibration during steady-state tests in order to reduce NOx frombiodiesel combustion, while maintaining PM at levels lower thanthat of petrodiesel operation, however at the expense of fuel ef-ciency. On the other hand, Tat and van Gerpen [122] proposeda technique for the detection of the biodiesel presence in the fuel,which was to be accompanied by a retard of the static injectiontiming.

and Combustion Science 38 (2012) 691e715factors produces a cancel-out result and practically maintains theanticipated biodiesel NOx liability to the extent of the earlierinvestigations for 5e50% v/v blends. However, since the NEDC(from which many data are included) rather underestimates real-world driving conditions by incorporating soft and linear acceler-ations, it is speculated that higher biodiesel NOx penalty will beactually encountered during real-life driving.

4.1.4. Methods for compensation of the NOx increaseOwing to the (usual) NOx emission liability with biodiesel

against conventional diesel fuel, it is reasonable to suggest that aninjection retard in order to delay the ignition timing might provebenecial; this was, for example, the focal point of the work con-ducted by Starr studying a heavy-duty engine running on SMEblends during the hot FTP cycle [31].

Based on the fact that NOx emissions seem to correlate withcetane number, some researchers investigated the effect of ignitionimprovers. For example, Sharp et al. [100] found that the addition ofethyl-hexyl-nitrate (EHN) was unsuccessful for a B20 SME blend,whereas di-tert-butyl peroxide (DTBP) could lower NOx by 6.2%while maintaining the 9.1% PM reduction; similar observationswere made by Starr [31] for each timing tested. Nuszkowksi et al.[101] measured decreasing NOx with increasing both 2-EHN orDTBP, with the greater impact observed on older technology(1990s) engines. McCormick et al. [43] conrmed the above resultsof DTBP, but they too later [48] argued that the success might beactually lower for newer engines.

Graboski et al. [29] proposed that NOx neutral blends withbiodiesel could be produced by altering either the base fuelaromatic content or natural cetane number. To this aim, a low-aromatic fuel (23.9%) was treated with biodiesel (to make B20) anda small amount of cetane (n-hexadecane) to raise the cetanenumber of the test blend to the same value as certication fuel. Thelower aromatic content of the fuel compared to certication fuelwas enough to offset the NOx increase produced by the biodiesel. Atthe same time, the particulate matter was reduced relative to thecertication fuel by 24%. It was argued that fuel reformulation withoxygen and aromatic control is a route to producing NOx neutralbut PMereducing fuels. The same research group in a more recentstudy [43] found that lowering the base fuel aromatics contentfrom 31.9% to 7.5% for an SME B20 blend managed to reduce NOx by6.5%.

Based on the information gained from the studies thatconcentrated on the biodiesel feedstock impact on NOx emissions(Section 4.1.2.1), an alternative route for compensating the NOxliability is revealed, by designing the ideal triglyceride feedstock forbiodiesel. A biodiesel with a combination of one or zero doublebonds tominimize NOx production and a low pour point to improvecold weather operation would have both environmental andclimatic advantages [36]. In view of this, Chapman et al. [118]investigated a blending approach that involved the blending ofshort-chained, saturated methyl esters with biodiesel fuel, i.e.a blend of 60% capryllic acid methyl ester and 40% capric acidmethyl ester was chosen so that its short hydrocarbon chain lengthwould help offset the adverse effect of saturation on cold-owproperties. This approach was found to produce a 2.8% reductionin NOx emissions over all examined modes, and also improved thefuels cold-ow property. Unfortunately, the high cost of the satu-rated methyl esters renders it economically unfeasible.

In most modern engines, a variety of operating parameters suchas the EGR system and the injection timing are electronicallymonitored and adjusted by the engine ECU. Thus, a notablecompensation of NOx emissions could be achieved througha revised engine calibration with regards to the injection system

E.G. Giakoumis et al. / Progress in Energy702(discussed in Section 4.1.1) or the EGR control, as has been revealed,4.2. Particulate matter and smoke

4.2.1. Main mechanisms and trends during transientsAs mentioned in the Introduction, biodiesel-blended (oxygen-

ated) fuels have been found capable of (substantially) decreasingparticulate matter during steady-state operation. In general, similarresults have been reported during transient conditions too for bothpassenger cars and heavy-duty diesel engines. Fig. 13 illustratestypical sunower biodiesel effects on smoke opacity developmentduring a load increase transient event at constant engine speed ofa passenger car engine; smoke opacity measurement was accom-plished applying a high response analyzer particularly suited totransient experimentation. Clearly, smoke was found to decline thehigher the biodiesel content in the fuel blend with remarkabledecrease observed for the B70 or B100 blends. Similar results havebeen reported for various acceleration transients by Rakopouloset al. [73].

Returning to Fig. 6, we can expand on the latter remarks byobserving cumulative PM emissions from a heavy-duty dieselengine during the FTP cycle with respect to the biodieselpercentage fuel blend. Reduction up to 66% (hot start) or 63% (coldstart) was measured when running the engine with 100% SMEcompared with petroleum diesel fuel, without sacrice in engineefciency. Interestingly, PM emissions were reduced proportionallyto the oxygen content of the fuel blend (oxygen content being

15 20 25 30 35Time (s)

0

2

4

6

8

10

12

14

16

Sm

oke

Op

ac

ity

(%

)

Neat diesel30% biodiesel70% biodiesel100% biodiesel

Fig. 13. Smoke opacity development during a 26e90 Nm load increase transient event

at 1661 rpm for various dieselesunower biodiesel blends (adapted from Armas et al.[50]).

0.21%, 2.37%, 4%, 7.24% and 11.03% respectively for the 0, 20, 35, 65and 100% by volume biodiesel-diesel blends). Similar results for theemitted PM were documented in Fig. 7, this time regarding theNEDC (cold starting effects are incorporated here).

Based on the research conducted so far, the benecial effect ofbiodiesel blend on smoke opacity values/PM emissions duringtransients, as has been documented in the representative Figs. 6, 7and 13 can be, primarily, attributed to the following factors:

diesel fuel) compounds that are generally considered to act assoot precursors.

Further, and in order to account for the lower heating value ofbiodiesel, fuel consumption must increase for the samedemanded engine torque. Hence, the ECU strategy may dictatean earlier start of injection and a decrease in the exhaust gasrecirculation (EGR) rate, both of which result in elevatedtemperatures inside the cylinder that promote soot oxidation(again at the expense of NOx).

iese

Fue

0.00.00.00.00.0

E.G. Giakoumis et al. / Progress in Energy and Combustion Science 38 (2012) 691e715 703 Increased oxygen concentration in the biodiesel blend, whichaids the soot oxidationprocess, irrespective of fuel composition.Soot formation, caused by high temperature decomposition,mainly takes place in the fuel-rich zone at high temperaturesand pressures, specically within the core region of each fuelspray, effectively preventing carbon atoms fromparticipating insoot precursor reactions. If the fuel is partially oxygenated, as isthe case with biodiesel, it possesses the ability to reduce locallyfuel-rich regions and limit sootnucleationearly in the formationprocess, thus reducing PM emissions and smoke opacity.Further, the formation of soot is strongly dependent on engineload, with higher loads (e.g. cruising portions of the drivingcycle) promoting higher temperature, longer duration of diffu-sion combustion (where particles aremostly formed) and loweroverall oxygen availability (airefuel equivalence ratio). Thelocally very low values of airefuel ratio experienced duringturbocharger lag at the onset of each acceleration and loadincrease, enhance the above mechanism, which is morepronounced thehigher the engine rating, i.e., thehigher the full-fueling to no-fueling difference. The excess oxygen of biodieselaids in maintaining these airefuel equivalence ratio discrep-ancies during turbocharger lag (where soot is primarilyproduced [2]) milder relative to the neat diesel transient oper-ation, however, at the expense of NOx as was argued in Section4.1.1. Despite the signicant positive effect of biodiesel on PMemissions, it has been reported in the literature that thebiodiesel-bound oxygen may be under-utilized. This is due tothe fact that methyl esters undergo decarboxylation, whichyields a CO2 molecule directly from the ester [16,123]. Thus, theoxygen in the biodiesel blend is used less effectively to removecarbon fromthepool of soot precursors (compared, for example,with ethanol- or ether-diesel blends);

Combustionwithbiodiesel shifts tomorecontrolledmode,whichmeans that there is more time available after diffusioncombustion for soot oxidation at a high temperature environ-ment. In other words, the previously mentioned injection andcombustion advance acts in favorof the soot destructionprocess;

Biodiesel is characterized by lower airfuel equivalence ratio(Table 1), which reduces the possibility of fuel-rich regions inthe non-uniform fuelair mixture [8];

Different structure of soot particles (see Section 4.2.2); Longer carbon chains of the biodiesel compared with the dieselfuel;

Absence of aromatic (primarily) and sulfur (secondarily, owingto the continuously decreasing sulfur content in conventional

Table 3Summary of composite transient particulate composition from a 1997, heavy-duty d

Fuel DOC Volatile organic fraction (g/kW h)

Total Oil

Diesel No 0.0039 0.023B20 No 0.0044 0.024B100 No 0.0043 0.024Diesel Yes 0.0027 0.009B20 Yes 0.0024 0.008

B100 Yes 0.0020 0.009 0.0In view of these strong PM-decreasing mechanisms duringbiodiesel blends combustion, it is not surprising that 90% of alltransient measurements correspond to PM reduction when addingbiodiesel in the fuel blend. However, there have been a fewexceptions too. For example, Fontaras et al. [64] reported animpressive 177% increase of PM emissions over the NEDC (it waspostulated that this was due to the increase in the SOF), Petersonand Reece [28] reported a 19.2% increase of PM emissions whena rapeseed-derived neat biodiesel was tested compared with theneat diesel fuel during the UDDS chassis-dynamometer cycle, andsimilar trends were reached by Durbin et al. [38].

Most of the studies so far have also concluded that, despite thesometimes substantial PM decrease, biodiesel-fueled enginesproduce a higher fraction of SOF in their exhausted PM than whenpetroleum-baseddiesel fuel is used; a typical summaryof compositetransient particulate composition is provided in Table 3 for a heavy-duty diesel engine running on the FTP transient cycle for two bio-diesel blends and the nominal diesel fuel. As is made obvious inTable 3, SOF emissions generally increasedwith increasing biodieselcontent, although it has been claimed that the exact percentageactually varies fromengine to engine.Moreover, the lube-oil portionof the SOF was found to be unaffected by the biodiesel content.Although the exact mechanism is not absolutely clear, this higherSOF percentage of biodiesel has been attributed to its lower vola-tility (higher boiling point) as well as to heavy fuel-related organiccompounds that remain intact through combustion [37,51,52].

There are two important issues that need to be addressed too:

a) Contrary to NOx emissions (Section 4.1.2.1), no correlationcould be established between biodiesel feedstock and transientPM emissions; the latter suggests that the positive impacts ofbiodiesel on PM are most probably not associated with specicinternal physical or chemical properties or attributes of thefuel, such as, for example, the number of double bonds, butrather depend on the oxygen content. On the other hand,a dependence on density and cetane number has been claimedin one study, namely with cetane number lower than 45 anddensity higher than 895 kg/m3, PM emissions were substan-tially elevated [40]; nonetheless no explicit trend has beenestablished, and this nding has not been conrmed insubsequent studies.

b) The transient cycles characteristics do not seem to correlatethatwellwith PMaswas the casewithNOx (in Section 4.1.2.2, R2

forNOx had an impressive value of 0.99); from the same study of

l engine during the FTP transient cycle (adapted from Sharp et al. [37]).

Non-volatile (g/kW h)

l & other PM SO3 H2O Soot16 0.137 0.013 0.08420 0.118 0.009 0.06419 0.070 0.001 0.02417 0.101 0.001 0.07116 0.079 0.004 0.051

11 0.040 0.003 0.017

Sze at el. [54], a lower but still healthy value of 0.87 wasestablished for the coefcient of determination of PMemissionswith respect to the average power cycle. It seems then that thehigher the average cycle power (or the more aggressive thecycle), the higher the benet of biodiesel-blended fuels on PMemissions relative to neat diesel oil. Obviously, the higher theloading or the aggressiveness of a cycle, the lower the airfuelequivalence ratios or the harsher and the more frequent theturbocharger lag phases induced respectively; both lead tomore intense soot production, where the benecial effects ofbiodiesel (most notably its increased oxygen content) prevailover the neat diesel operation. Moreover, high loadings andabrupt accelerations both dictate lower EGR rates through theusually applied ECU calibration, which again act in favor oflower PM emission rates at the expense of NOx.

Despite the well established positive effects of biodiesel on PMemissions during fully-warmed-up transients discussed so far, its useduring cold starting has been found to exhibit an adverse behavior.This hasbeenprimarilyattributed to thehigher initial boilingpointofbiodiesel with respect to conventional diesel fuel, which leads to

E.G. Giakoumis et al. / Progress in Energy and704more difcult fuel evaporation at lowambient temperatures, and thehigher viscosity of biodiesel, which reduces the rate of spray atom-ization. Both phenomena, which are enhanced with very high bio-diesel blends, lead to worse fueleair mixing and, thus, more intensesoot formation at low temperatures; the latter increase wasmeasured up to 80% for an automotive, HSDI diesel engine, whencomparing B100 with neat diesel operation [50]. Later in the warm-up phase, when the engine assumes its normal operating tempera-ture, the advantages of biodiesel combustion prevail, and sootemissions decrease compared with the neat diesel fuel. These coldstarting results have been conrmed by Rakopoulos et al. [84] duringvarious hot starting tests performed on an engine dynamometer fora medium-duty turbocharged bus/truck engine, as is demonstratedin Fig. 14, where it is made clear that starting not only increases thepeak in smoke opacity but also the duration of unacceptable blacksmoke emissions. Interestingly, extended pressure variability wasalso noticed for the diesel-biodiesel blend during the starting eventof Fig. 14 compared with the neat diesel operation. Likewise duringtransient cycles, for the previously mentioned example of Fontaraset al. [64], the increase actually rose to 278% when only the cold(urban) ECE15 was isolated. Similar observations were made byFig. 14. Development of smoke opacity during hot starting of a medium-duty turbo-charged diesel engine for neat diesel and B30 operation [84].Macor et al. [86] again for the cold-started ECE15. On the other hand,Graboski et al. [29] measured higher PM emissions during the cold-started runs of the heavy-duty FTP compared with the hot ones, butin either case a signicant benet relative to neat diesel operationwas established for each biodiesel blend tested. Bearing in mind thearguments raised in the last paragraphs, it seems that it is the fully-warmed-up, urban engine operation that mostly benets as regardsPM emissions reduction from the use of biodiesel.

4.2.2. Diesel particulate lter regeneration rateApplication of a modern, high-efciency exhaust after-

treatment system (e.g. diesel particulate lter e DPF) beingcapable of reducing soot emissions to negligible levels, can practi-cally downgrade to a large extent the biodiesel benecial effects onPM [46,65]. Moreover, since Euro 5 specications are even stricterregarding PM, it seems that the use of biodiesel alone is not enoughto conform to these standards (as could have been the case withprevious specications), requiring a DPF in any case.

An important factor affecting the DPF regeneration rate is theoxidative reactivity of particulate matter. Boehman et al. [124]found that the use of a B20 blend during steady-state operationcan signicantly lower the balance point temperature (BPT; is theDPF inlet temperature at which the rate of particle oxidationapproximately equals the rate of particle collection). They pre-sented results showing that it is not the increased availability ofNO2 that is responsible for the decrease in BPT but rather theinherent differences in soot reactivity for different fuels, andspecically for B20, the more highly disordered soot nanostructure,such that the soot is more reactive or it is reactive at lowertemperatures. The same research group [16,125] further concludedthat the more reactive toward oxidation behavior of biodiesel sootderives from enhanced incorporation in the soot of surface oxygenfunctionality. These results were conrmed byWilliams et al. [126],again during steady-state operation, who showed that on average,the BPT is 45 C and 112 C lower, respectively, for B20 blends andneat biodiesel, than for 2007 certication diesel fuel. Filter regen-eration rate measurements indicated that biodiesel causesa signicant increase in regeneration rate, evenwhen B5 blends areemployed. Overall, their results suggested signicant benets fromthe use of biodiesel blends in engines equipped with DPFs. It wasargued that this decrease of BPT and the subsequent increase inregeneration rate might allow passive DPFs to be used in lowertemperature (i.e. urban) engine cycles, avoiding the need foractively regenerated lters and their associated fuel economypenalty. In parallel, actively regenerated systems might require lessfrequent regeneration, perhaps resulting in a lower fuel economypenalty.

There are few studies available that investigated the effects ofbiodiesel on DPF regeneration during transients in order to conrmthe above signicant (steady-state) ndings. Tatur et al. [102] foundthat under normal operating conditions, an SME B20 blend hadmarginal impact on both DPF regeneration rate and NOx adsorptioncatalyst leanerich cycle development for a passenger car runningon various American, light-duty cycles. On the other hand,Muncriefet al. [60], concluded that a diesel vehicle operating under low-loadconditions (as the refuse truck studied) with neat cottonseed orsoybean derived biodiesel had relatively cool exhaust to adequatelyoxidize the accumulated soot on the DPF with NO2. It was specu-lated, however, that this might be counter-balanced by the loweremitted PM from biodiesel combustion.

4.2.3. Overall resultsThe statistical analysis of all transient measurements from the

studies of Table A yields the quadratic best-t curve illustration

Combustion Science 38 (2012) 691e715presented in Fig. 15, where the effects of biodiesel blends on PM

iesel

E.G. Giakoumis et al. / Progress in Energy andemissions are demonstrated. Again, although rather high degree ofvariationwas observed, and documented in the R2 value (due to thefact that data from all kinds of cycles, engines, and originating oilshave been included), a compelling decreasing trend of PMemissionswith rising biodiesel blend ratios can be established, conrming in

Fig. 15. Collective PM reduction when using various biodthe most explicit way the results during steady-state operation.Finally, Fig. 16 documents the, well established during steady-

state operation, trade-off between PM and NOx. The latter,although not statistical signicant, seems to apply reasonably wellto transient cycles emissions too for the two most investigatedbiodiesel blend ratios B20 and B100, particularly for the neat blend(R2 0.46). It seems reasonable then to modify the engine cali-bration by applying a slightly higher level of EGR during biodiesel

Fig. 16. PM-NOx trade-off for B20 and B100 blends (data from all transient cycles ofTable A).combustion, in order to trade-off some of the signicant PM benetfor lower NOx emissions.

4.2.4. Particle number concentration and distributionCurrently, diesel particulate matter legislation is primarily based

-diesel blends (data from all transient cycles of Table A).

Combustion Science 38 (2012) 691e715 705on the emitted particle mass. However, particle size distribution isgaining increasing attention in terms of air quality, as it is believedthat the toxicity increases as the particle size decreases; in fact,a correlation between elevated ambient particulate matter concen-tration andhospital admissions has been suggested [127,128]. To thisaim, the EUhas legislated an intermediate Euro 5b specication levelthat also includes a particle number limit of the order of 6 1011/kmover the NEDC for both passenger car (category M) and light-dutyvehicles (categories N1 and N2), applicable from September 2011.

There is an increasing amount of recent research, based onsteady-state experimentation, that inter-relates the decrease in PMemissions from biodiesel use with an increase in the number of themore toxic nanoparticles (e.g. [16,128]), although the trend is notunambiguous [8]; similar contradicting observations hold duringtransients. Fontaras et al. [64] found that the use of neat SMEincreased total particle number up to 200% compared with the neatdiesel operation during the NEDC (measured with a CPC); a B50blend behaved similarly. This increase in the particle number arosefrom a general shift towards smaller-diameter (nano)particles andwas evident for all transient cycles examined (the transient resultswere conrmed for three steady-state runs at 50, 90 and 120 km/h).The increased viscosity of biodiesel, as well as the increase of SOF,and of the injection pressure and the advance in injection timingwith rising biodiesel percentage in the fuel blend have been iden-tied as possible causes for this behavior. Likewise, Tinsdale et al.[78] reported a 25% increase in the nucleation mode particlenumber over the NEDC for a B30 blend compared with the neatdiesel case, and Chien et al. [68], using MOUDI and nano-MOUDIdevices, concluded that as the (waste cooking derived) biodieselpercentage increased, the ultra-ne and nanoparticles numberincreased too; in fact, when neat biodiesel was used, nanoparticles

dominated the size distribution. The calculated median massdiameters for their examined blends were: 0.146, 0.144, 0.134 and0.124 mm, for neat diesel, B20, B60 and B100 respectively.

On the other hand, Macor et al. [86] found that the total numberof emitted particles decreased for both tested vehicles between 10and 20% over the examined driving cycles (NEDC, Artemis) relativeto the neat diesel case. A possible explanation that has been arguedby some researchers in such cases is the absence of sulfur in thebiodiesel. The respective size distribution proles from the samestudy revealed that the larger-diameter particles (for all fuelblends) were produced during those sections that include frequentaccelerations or high power. This behavior can be explained by thefact that generally the trend is towards larger particles withincreasing load, whereas nanoparticles are favored mainly at idlingconditions [2]. As the load increases, more fuel is injected; thisfavors the formation of larger particles primarily owing to thelonger diffusion combustion duration, to the higher combustiontemperatures, and to the reduced oxidation rate of the soot in theexpansion stroke, since there is less time available after the end ofthe diffusion combustion and also lower oxygen availability. Fon-taras et al. [74] expanded on the previous studies by including also

contribute towards lower (engine-out) emitted CO. These argu-ments (primarily the increased oxygen content) explain the clearbehavior illustrated in Fig. 6 during the (hot) FTP heavy-dutytransient cycle; likewise, Bakeas et al. [81] reported a 10% reduc-tion in (the already low emissions of) CO from a B20 blend withrespect to the neat diesel operation for a passenger car testedduring the NEDC. Similar results have been reported, among manyothers, in [27,28,29,34,36,48,52,60,66,71,81,82,85].

In Fig. 7 in contrast, a slightly increasing trend of CO emissions isobserved with increasing biodiesel percentage in the fuel blend.This was also the result reached by Rose et al. [79], Macor et al. [86],Bermudez et al. [90] and Fontaras et al. [64]; the latter reporteda 54% increase in CO emissions during the NEDC for a B50 fuelblend, which was even higher, of the order of 95%, when neatbiodiesel was tested compared with diesel fuel. Durbin et al. [53]reported similar results (increase of CO (and HC) emissions withrising biodiesel blend ratio) when an engine tted with oxidationcatalyst was tested, whereas for another, older production, enginewithout oxidation catalyst the trend was reversed. In all theseexperimentations, the CO increase (which was even higher duringthe cold-started runs or segments of the cycle) was attributed to the

E.G. Giakoumis et al. / Progress in Energy and Combustion Science 38 (2012) 691e715706biodiesel feedstock effects on the particle number concentration. Itwas found that the effect of biodiesel on total particle number,including volatile and semi-volatile particles, was variable.Although reductions were observed over the low-power (sectionsof the) cycles, and for sunower and unused frying methyl esters,the PME and the very popular in Europe RME blends were associ-ated with up to 3 times higher particle number emissions than thebase diesel fuel over the Artemis Motorway test.

4.3. Carbon monoxide and unburned hydrocarbons

Carbon monoxide emissions depend mainly on the airefuelequivalence ratio, being usually of low signicance during steady-state diesel engine operation. For fuel-rich mixtures, such asthose experienced during the turbocharger lag phase at the onset ofan acceleration or load increase event, CO concentration in theexhaust increases steadily with decreasing airfuel ratio. The use ofbiodiesel in engines is capable of reducing CO emissions primarilyowing to the higher oxygen content of the fuel blend that favorsmore complete combustion. As was the case with PM, the higherbiodiesel cetane number and the advanced injection may alsoFig. 17. CO and HC diesel oxidation catalyst efciency for two ambient temperatures (left), aet al. [76]).oxidation catalysts limited efciency during the early minutesbefore reaching its light-off temperature, backed up by EGR andinjection timing strategy effects.

An instructive comparative study that supports this argumentwas conducted by Bannister et al. [76]; they reported lower engine-out CO (and HC) emissions, whereas their tailpipe counterpartswere higher relative to neat diesel, and this was too attributed toa reduction in the catalyst conversion efciency with increasingblend ratio, as is demonstrated in Fig. 17. The latter effect was mostprobably owing to the lower biodiesel exhaust gas temperature(illustrated in the right sub-diagram of the same gure). Theaverage exhaust gas temperature was measured 2.3% lowerbetween B50 and neat diesel fuel during the NEDC. Similar resultswere reached by Muncrief et al. [60] and by Sze et al. [54], whofound the mean exhaust gas temperature from a B50 SME blend tobe 2.5% or 16 C lower than that from reference diesel fuel duringa highway cycle, and around 1% for lighter-loaded cycles such as theFTP and the UDDS. The lower exhaust gas temperature is respon-sible for lower available exhaust gas energy that in turn can explainthe extended catalyst light-off time but also its reduced efciencythroughout the whole cycle. Bannister et al. [76] also speculatednd average exhaust gas temperature (right) during the NEDC (adapted from Bannister

that this decrease in exhaust gas temperature might be responsiblefor the lower boost pressures measured, since their variablegeometry turbocharger operated on lower available exhaust gasenergy (this was conrmed by Sze et al. [54]). This fact furtherhighlights the unexpected impacts and interactions which resultfrom changes in the fuel properties, demonstrating the need forrevised calibration strategies (as is notably the case with EGRcontrol), when a diesel-tuned engine is required to run on biodieselblends.

The effect of biodiesel feedstock on CO emissions during tran-sients is not clear. As was the case with PM, the US EPA study [15]concluded that soy-derived (i.e. highly unsaturated) biodiesel led tohigher CO than RME or TME (both with much fewer data available),implying a correlation between unsaturation (or higher density orlower cetane number) and elevated CO, which has not, however,been conrmed by subsequent studies, at least on a statisticallysignicant level. Further, from Table 2 it seems that the older rotaryand pump-line-nozzle injection systems promote a higher CObenet with biodiesel blends relative to neat diesel fuel than theirmodern electronic and common rail counterparts, but no unam-biguous trend has been conrmed.

Hydrocarbons form during the ignition delay period as the resultof either very low airefuel ratios or, mainly, under-mixing of fuelthat cannot ignite or support a ame. Contrary to soot, NOx or COthat peak only during load increase or acceleration, turbocharged

by almost all researchers regarding hot tested transient cycles.Nonetheless, as was also the case with CO, the rst minutes of thecycle, where the engine is still cold, may inuence the HC emissionsremarkably, reducing the exhaust gas temperatures and the DOCefciency, and resulting in higher emissions the higher the biodieselpercentage in the fuel blend (see Fig. 17) [46,61,67,74,79]. Interest-ingly, possible sampling line issues have also been identiedby some researchers as causes for the lower HC level in theexhaust compared with the neat diesel operation. For example thelower volatility of biodiesel may lead to hydrocarbons of highmolecular weights being condensed in the exhaust pipe [37,129].Lastly, as was the case with PM and CO (and unlike NOx), it has notbeen feasible to statistically correlate the biodiesel feedstock withthe HC emissions.

Figs. 18 and 19 summarize the quantitative effects of biodieselblends on CO and HC emissions during all transient cycle studiesfrom Table A, where, despite the scattered and sometimes contro-versial data measured, a clear decreasing tendency is establishedwith increasing biodiesel blend ratio. Both the CO and HC best-tcurves are comparable in their trend and development with theolder data, the biodiesel benet however has somewhat fadedduring the last years. It seems then that inclusion of many datafrom

a) newer production, European engines with after-treatment

E.G. Giakoumis et al. / Progress in Energy and Combustion Science 38 (2012) 691e715 707diesel engine HC emissions are noticed at the onset of decelerationor load decrease too, since in the latter cases, turbocharger lageffects lead to instantaneously very high airefuel equivalence ratios[2].

The absence of aromatic content as well as the advanced injec-tion and combustion timing and the higher oxygen concentration ofbiodiesel, all promote more complete combustion reducingaccordingly the level of unburned HC. Moreover, the higher cetanenumber of biodiesel contributes to the reduction of the ignitiondelay, where an important amount of HC is formed. For example,Karavalakis et al. [52] reported a 20% reduction in HC from a B20blend compared with the neat diesel operation for a passenger cartested during theNEDC, and similar reductions have beenmeasuredFig. 18. Collective CO emission change when using various biodiecontrol in the form of diesel oxidation catalysts, which seemto operate less efciently with biodiesel,

b) many light-loaded passenger car cycles, and, predominantly,c) cold-started runs

has gradually shifted the overall trends to lower CO and HC emis-sion benets. In fact, many of the researchers from light-dutyengines/cycles actually report increases in both CO and HC withthe use of biodiesel.

Finally, with regards to CO2 emissions, the majority of the studiesconcluded that no statistically signicant changes were measuredduring transients between biodiesel blends and petroleum dieselfuel, with both moderate increases and decreases reportedsel-diesel blends (data from all transient cycles of Table A).

The only comprehensive analysis regarding biodiesel impacts oncombustion noise development during transients has been repor-

iodiesel-diesel blends (data from all transient cycles of Table A).

and Combustion Science 38 (2012) 691e715depending on the specic engine, blend and cycle examined. Thelatter nding practically conrms the results during steady-stateoperation that engine efciency is, effectively, not altered withthe addition of biodiesel in the fuel blend.

4.4. Combustion noise

The three primary sources of noise generation in a diesel engineare gas-ow, mechanical processes and combustion [130,131]. The

Fig. 19. Collective HC emission change when using various b

E.G. Giakoumis et al. / Progress in Energy708origin of combustion noise radiation lies in the (high) rate ofcylinder pressure rise, mainly after the ignition delay period, whichcauses discontinuity in the cylinder pressure frequency spectrumand increase in the level of the high-frequency region, resulting invibration of the engine block and, ultimately, in combustion noiseradiation (the characteristic diesel combustion knock). Thiscombustion noise radiation manifests itself in the domain froma few hundred up to a few thousand Hz.

There is only a handful of works available regarding biodieseleffects on noise radiation during steady-state engine operation, andwith contradicting results. Nabi et al. [132] reported overall noisereduction for a naturally aspirated engine operating at high loads,ranging from 1 dBA for a B10 blend up to 2.5 dBA for neat karanjabiodiesel (B100). Although the reduction was, in a rather simplisticmanner, attributed to the oxygen content of the biofuel, it was mostprobably the higher cetane number of biodiesel that was respon-sible for a shortening of the ignition delay period that ultimatelyreduced the cylinder pressure rise rate and the combustion noisecontribution to the total emitted noise. On the other hand, Haiket al. [133] found that the algae oil methyl ester was responsible forhigher cylinder pressure rise rates dp/d4 (a typical surrogateproperty of combustion noise) compared with the neat diesel fuelor the raw algae oil for their indirect injection diesel engine.Although Anand et al. [134] did not specically measure combus-tion noise, they too reported on the cylinder pressure rise rate forneat karanja biodiesel fuel blends, and concluded that the trend vs.neat diesel fuel was not clear; in fact the contributing factors actedin favor or against each fuel blend depending on the speed or loadconditions.ted by Giakoumis et al. [103], and is depicted in Fig. 20 that illus-trates noise development during three different accelerations ofa medium-duty engine. Although biodiesel is characterized byslightly higher cetane number, its acceleration response comparedwith the neat diesel operation was proven marginally noisier, andFig. 20. Difference in the combustion noise between a B30 blend and diesel fuelthroughout three acceleration tests [103].

andthis was attributed to its higher density that increased the totalinjected mass and the fuel mass burnt during the inuencing pre-mixed combustion phase. However, the trend was not clearthroughout each acceleration, and sometimes the differences werevery small, within the measuring units accuracy. It is thus sus-pected that higher blends might be required in order to esta