Potential of Cellulose-Derived Biofuels for Soot Free Diesel Combustion

Transcription

1

2 Copyright 2010 SAE International Potential of Cellulose-Derived Biofuels for Soot Free Diesel Combustion Andreas Janssen; Martin Muether; Stefan Pischinger Institute for Combustion Engines, RWTH Aachen University, Aachen, Germany ABSTRACT Today s biofuels require large amounts of energy in the production process for the conversion from biomass into fuels with conventional properties. To reduce the amounts of energy needed, future fuels derived from biomass will have a molecular structure which is more similar to the respective feedstock. Butyl levulinate can be gained easily from levulinic acid which is produced by acid hydrolysis of cellulose. Thus, the Institute for Combustion Engines at RWTH Aachen University carried out a fuel investigation program to explore the potential of this biofuel compound, as a candidate for future compression ignition engines to reduce engine-out emissions while maintaining engine efficiency and an acceptable noise level. Previous investigations identified most desirable fuel properties like a reduced cetane number, an increased amount of oxygen content and a low boiling temperature for compression ignition engine conditions. Depending on the chain length, fuel compounds vary in cetane number and boiling temperature. Therefore different blends of butyl levulinate and n-tetradecane, a long-chain alkane, were investigated in this study. To gain knowledge about the combustion process with biofuels, experiments on a single cylinder diesel research engine were performed. Engine results with blends of butyl levulinate and n-tetradcanee are compared to regular diesel fuel and rapeseed oil methyl ester (B100) with respect to all regulated emissions. With the use of butyl levulinate the soot emissions can be significantly reduced up to 100 % depending on the load point. Furthermore, the experimental results indicate an additional potential for future diesel combustion concepts such as HCCI with respect to soot and NOx emissions in the part load range when using oxygenated fuels derived from biomass. Overall, the low particulate matter emissions provide justification for the consideration of butyl levulinate as a candidate for a fuel or fuel compound in future diesel engines. As a drawback malfunctions in various components of the fuel system like fuel hoses and fuel sealings can occur with the standard engine equipment. INTRODUCTION The finite nature and instability of fossil fuel supply in combination with the carbon dioxide issue has led to an increasing and enduring investigation demand in the field of alternative and regenerative fuels. Renewable biological fuels have the potential to maintain a constant level of greenhouse gas emissions by not adding any Page 1 of 21

3 additional fossil carbon to the atmosphere. These fuels can be tailored to the demands of modern engines taking into account the catalytic production processes. With the establishment of these fuels as an additional degree of freedom, a simultaneous optimisation of the engine and fuel is possible. Through the application of optimised synthesis processes, the Cluster of Excellence Tailor-Made Fuels from Biomass (TMFB) adopts an interdisciplinary approach towards research of new, biomass-based synthetic fuels. With regard to modern combustion technologies, this approach was chosen in order to verify their potential, while simultaneously reducing the dependence on fossil fuels. The long-term goal is to determine the optimal combination of fuel components and their production processes, which is based on renewable raw materials and new combustion processes. Such a tailor-made fuel is a well-defined blend of distinct molecular components with optimised physiochemical properties for future combustion systems, which can be produced by sustainable and economical production processes. In the search for new fuels to reduce engine emissions generally and particularly soot emissions several investigations have been recently published [1, 2]. All of them show a positive impact on particulate matter emissions with an increasing amount of fuel based oxygen in the combustion chamber. Soot formation can be reduced linearly with increasing oxygen content [3, 4]. The potential of oxygenated fuels has been previously explored on the example of 1-decanol, a long-chain alcohol. A reduction by up to 90% in particulate matter emissions in comparison to conventional diesel fuel could be identified [5]. Within the Cluster of Excellence TMFB, investigations have recently focused on butyl levulinate (C 9 H 16 O 3 ) as a candidate for future fuel compounds. This oxygenated hydrocarbon can be easily produced through acid catalysis of cellulose. Despite its very low cetane number, fuel properties enable an utilization as a fuel in a compression-ignited engine exclusively or as part of a blend. Therefore, different blends of butyl levulinate and n-tetradecane have been investigated (butyl levulinate vol-% varying from 60 % to 80 %). The potential of such blends to reduce soot emission will be particularly analyzed in this paper and will be compared with baseline diesel fuel and pure rapeseed oil methyl ester, a first generation biofuel. TEST SET UP The engine used in this study was optimized to achieve best combustion behavior for lowest engine-out emissions and highest fuel efficiency. The hardware enhancements include high fuel injection pressures for improved mixture formation, high EGR levels and intensified charge air and EGR cooling to lower the temperature in the combustion chamber for a cooler combustion. Details of the test engine are listed in Table 1. Further information about the research single cylinder test engine are available in different publications [6, 7]. The research engine is capable of achieving Euro 6 emissions performance by means of a lower compression ratio, a higher maximum cylinder peak pressure, a maximum rail pressure of 2000 bar and intensified EGR cooling. These engine enhancements are expected to be commonly used on advanced diesel engines in the near future. As such, the testing capabilities on the engine exceed those of today s production engines. ENGINE OPERATION - Engine tests for the fuel investigations were carried out in four part-load points, as shown in Figure 1. Page 2 of 21

4 Units Engine Emission Status - Euro 6 Displacement cm³ 390 Stroke mm 88.3 Bore diameter mm 75 Compression ratio - 15 Valves per cylinder - 4 Maximum peak pressure bar 220 Fuel injection system specifications: - Bosch Piezo Common Rail System Maximum bar 2000 injection pressure Hydraulic Flow cm³/30 Rate (HFR) s 310 Nozzle hole diameter µm 109 Number of spray holes - 8 Spray Cone Angle 153 Charging - Max. 3.5 bar absolut [bar] Figure 1: Reference NEDC area (1590 kg vehicle) Full Load Points Part Load Points Engine Speed [rpm] Engine Operating Points Table 1: Hardware specifications for the research engine Three of the four part-load points are located within the reference area of the New European Driving Cycle (NEDC) while the highest load point at 2400 rpm and 14.8 bar is outside this range. Based on the new European Driving Cycle (NEDC) the maximum Indicated Specific NOx (ISNOx) and Indicated Specific PM (ISPM) emissions were estimated for each part-load point and different Euro emission levels for a 1590 kg vehicle. These targets take into account the expected impact of transient operation and cold start and include an appropriate engineering margin. The estimated NOx and PM target levels are summarized in Tables 2 and 3. Investigations on a chassis dynamometer confirmed these estimated targets Page 3 of 21

5 ISNOx 1500 rpm, 4.3 bar 1500 rpm, 6.8 bar 2280 rpm, 9.4 bar 2400 rpm, 14.8 bar Euro Euro Euro % Table 2: Estimated Indicated Specific NOx (ISNOx) Targets Required to Meet Different Euro Emissions Limits at the Four part-load Operating Points ISPM 1500 rpm, 4.3 bar 1500 rpm, 6.8 bar 2280 rpm, 9.4 bar 2400 rpm, 14.8 bar Euro Euro Euro % Table 3: Estimated Indicated Specific PM (ISPM) Targets Required to Meet Different Euro Emissions Limits at the Four part-load Operating Points The experiments were carried out with pilot injection and simulated closed loop combustion control. This was done by adjusting the fuel injection timing on each fuel so that the centre of combustion (CA50) was always located at the same degree crank angle in the engine cycle, chosen to maximise the engine efficiency, regardless of the fuel properties. Fuels with lower cetane number (CN) and longer ignition delay require earlier fuel injection timing to maintain the centre of combustion at the required crank angle position. Therefore the begin of injection varies between 5 CA BTDC at 2400 rpm, 14.8 bar and 20 CA BTDC at 1500 rpm and 4.3 bar. The difference between the energizing time of the injector and the begin of injection was measured to 160 µs. The conditions at the four part-load operating points were chosen with respect to today s achievable ranges and are listed in Table 4. The intake manifold temperature was measured and controlled directly after the intercooler and before the EGR mixing. Page 4 of 21

6 1500 rpm, 4.3 bar 1500 rpm, 6.8 bar 2280 rpm, 9.4 bar 2400 rpm, 14.8 bar Centre of Combustion Pilot Offset DOI-Pilot Rail Pressure Intake Temp. Intake Pressure CA BTDC CA µs bar K bar abs 0.5 g/kwh ISNOx 0.5 g/kwh ISNOx 0.5 g/kwh ISNOx 1.0 g/kwh ISNOx Table 4: Engine Calibration of different load points FUEL PROPERTIES The potential of fuels which can be derived easily from biomass was in the focus of interest within this study. It is common knowledge that Cellulose, separated from lignin and hemicellulose, will serve as a starting molecule for the biogenous fuel production. The separation of oxygen from the cellulose requires high amounts of hydrogen and is therefore not desirable from an energetic point of view. In the case of a Fischer-Tropsch fuel (2 nd generation biofuel), where the feed stock will be broken up into smallest molecules, a separation is not necessary anymore. Within the cluster of excellence TMFB completely new production routes based on catalytic processes will be developed. These fuels will most likely contain oxygen. One of these fuel candidates is butyl levulinate which belongs to the group of levulinic esters, which can be converted from levulinic acids by dehydration/hydrogenation or esterification, while levulinic acid is directly produced by acid hydrolysis of cellulose. A molecular structure of butyl levulinate is shown in Figure 2. The low cetane number and high oxygen content of levulinic esters are ideal fuel properties for future premixed charged compression ignition combustion systems. On the other hand the commercial use of levulinic esters is limited due to the high cost of their production. Figure 2: Molecular composition of butyl levulinate However, due to the similar molecular structure of the fuel compound and the respective feedstock, the production of these esters from cellulosic biomass represents a potential low-cost route. Estimations have shown that levulinic esters can be produced on a large scale at less than 0.50 euro/l [8]. Page 5 of 21

7 Today, levulinic esters are already known as an additive for Diesel fuel to increase the oxgen content [9]. Butyl levulinate consists of about 30% oxygen. Due to the molecular structure, particularly the small number of c-atoms and the oxygen content of an ester, butyl levulinate has a cetane number below the measurement limit, which is 18 in an ignition quality tester (IQT), the tester used for this study. Thus, cetane number improvers are needed to ignite the fuel in a compression ignition engine. Butyl levulinate can not be used as a pure chemical to substitute the fuel but needs to be blended with fast burning molecules. Therefore n-tetradecane, a long-chain paraffin with a cetane number of 95, was selected. N-tetradecane can not be easily derived from biomass. But its high cetane number provides justification for the consideration as a fuel blend. In future n-tetradecane can be substituted by other high cetane number fuels like GtL or BtL fuel. The most appropriate fuel composition was investigated in a first step. To gain valuable information about the potential of future fuels from biomass, promising fuel candidates need to be compared to today s available fuels. Therefore, additional investigations were carried out with Fatty Acid Methyl Ester (FAME). The FAME used in this study was a Rapeseed Methyl Ester (RME) that complied with EN specifications. Rapeseed Methyl Ester is a 100 % first generation biofuel and therefore called B100 in this paper. It consists of saturated as well as unsaturated fatty acids with 12 to 22 carbon atoms. The catalytic reaction is initiated by adding methanol to rapeseed oil. The outcome of this process is Rapeseed Methyl Ester as well as raw glycerine. Using the difference in density, B100 and glycerine can be separated. B100 is the most common fossil diesel-like substitute in the European Union. In the following all test results will be compared to available Diesel results which were carried out in a previous study using the same engine calibration [10]. The Diesel fuel meets the requirements of the EN 590 specifications. The corresponding properties of all investigated fuels are listed below in Table 5. Unit EN590 Diesel butyl levulinate B100 n-tetra-decan Density (15 C) kg/m³ Viscosity (40 C) mm 2 / s Table 5: Properties of investigated fuels Page 6 of 21 ~ ~ Carbon content w-% Hydrogen content w-% Oxygen content w-% Aromatics content mg/k g Molecular weight g/mol Lower Heating value kj/kg Boiling temperature C Cetane number <

8 RESULTS To evaluate the benefit of optimized fuel properties attention naturally focuses on the part-load test conditions that influence performance in the NEDC emissions cycle. For this reason, this section is devoted to an assessment of these part-load conditions. Engine measurements were taken by varying the EGR rates over a wide range to adjust NOx emissions for the corresponding Euro-level. All results presented in this paper are illustrated at the Euro 6 NOx emission limit. The analysis of butyl levulinate as a candidate for future fuel compounds or even as fuel surrogate was carried out in two steps. At first, different blends of butyl levulinate and n-tetradecane with varying n-tetradecane content from 10vol%-40vol% were investigated with regard to their combustion behavior, particularly the ignition delay time. These burning characteristic investigations allow valid conclusions about the available time for mixture preparation which is getting a more dominant role for future homogeneous or partly homogeneous combustion systems. Using the most appropriate blend, all emissions including noise emissions were measured in all four load points in a second step. Overall, this approach allows a conclusion about the potential of butyl levulinate to reduce carbon dioxide emissions and to benefit from a cleaner combustion in future combustion systems at the same time. INVESTIGATION OF COMBUSTION PERFORMANCE USING DIFFERENT FUEL BLENDS Due to the low cetane number the blending of butyl levulinate with fast igniting fuel compounds is a prerequisite to start the combustion in a compression ignition engine. Additionally, the blending offers opportunities to adjust desired fuel properties like ignition delay time or oxygen content. Within this research study three different blends were investigated and extensively analyzed. The ignitibility of a air/fuel mixture is highly depending on the intake air temperature and the boost pressure. The combustion chamber temperature needs to be high enough to start the combustion. Thus, lower part load points are more critical concerning the general ability to start the combustion. Therefore, the investigations of different fuel blends were carried out with pilot injection at a low part load point of 1500 rpm and 6.8 bar. To maintain equal testing settings and to assure valid comparability conditions all measurements were carried out at a constant ISNOx level. Furthermore the centre of combustion was kept constant. Figure 3 shows the oxygen content and the measured ignition delay time for the three investigated fuel blends compared to standard diesel fuel. The content of the ignition improver n-tetradecane was varied for these investigations from 20- vol % to 40- vol %. The enrichment of only 10- vol % of n-tetradecane was not sufficient to ignite the fuel at lower load conditions. Page 7 of 21

9 Oxygen Content [w-%] min -1, 6.8 bar Ignition Delay [ CA] min -1, 6.8 bar % butyl levulinate + 40% n-tetradecane 70% butyl levulinate + 30% n-tetradecane 80% butyl levulinate + 20% n-tetradecane EN 590 Diesel Figure 3: Oxygen content and ignition delay times of investigated fuel blends An increase of the n-tetradecane fraction leads to a reduction of the oxygen content in the blend. Previous studies have proven the benefit of oxygen content to reduce smoke emissions. Soot formation can be reduced linearly with increasing oxygen content. Therefore only a necessary but minimum amount of n-tetradecane to guarantee sufficient ignition properties even at cold ambient conditions is desirable for the new fuel blend with regard to soot emissions at higher part-load points. The ignition delay was calculated from the start of the main fuel injection to the point where 5% of the total heat release had occurred as indicated by the cylinder pressure traces. Generally, with reduced amounts of the ignition enhancer combustion starts after an increased ignition delay which results in a longer mixture preparation and in consequence an improved mixture of fuel and air. This increase in ignition delay is not linear with the added fraction but progressive. Blends of butyl levulinate with amounts lower than 30- vol % n-tetradecane are already very sensitive to the ignition ability and require additional closed loop combustion control strategies to guarantee a secure and repeatable combustion start. Overall, all three tested blends have a longer ignition delay than the standard Diesel fuel. For the blend with 20- vol % n-tetradecane the available time between fuel injection and combustion start is as double as long as it is for Diesel. Even at lower load that already lead to an instable combustion. The amount of n-tetradecane in the new fuel blend also has a huge influence on further combustion characteristics like the maximum burning rate and the combustion duration which are the determining influencing variables on engine noise emissions and efficiency. Generally a longer ignition delay leads to an increased part of premixed combustion which results in a higher maximum burning rate and thus in a faster combustion. With regard to engine noise this leads to significantly higher values. Figure 4 shows the pressure traces and burned mass fractions of the investigated fuel blends. The blend with only 20 % n-tetradecane shows high amounts of unburned hydrocarbons. Differences in the combustion behavior with increasing amount of n-tetradecane are decreasing. Page 8 of 21

10 Cylinder Pressure [bar] rpm; 6.8 bar 80 % butyl levulinate + 20 % n-tetradecane 70 % butyl levulinate + 30 % n-tetradecane 60 % butyl levulinate + 40 % n-tetradecane Current Signal Crankangle [ CA] Burned Mass Fraction [-] Figure 4: Pressure traces and burned mass fractions of investigated fuel blends The trade-off between an optimal mixture preparation and acceptable noise in particular at higher engine loads requires an adjusted blending amount of n-tetradecane. Overall, these combustion performance investigations have proven, that butyl levulinate blended with 30- vol % n-tetradecane as an ignition improver is the most promising blend of a cellulose derived fuel surrogate for future combustions systems. Particularly in the lower part load range reduced blend amounts of n-tetradecane are not appropriate. Therefore this fuel blend (abbreviated as BLT) was chosen for deeper investigation including the study of the emission performance in different load points in the next step. COMBUSTION BEHAVIOR OF BLT As indicated in Table 5 the cetane number of butyl levulinate is below the measurement limit of 18. That means that also the BLT blend has a much lower cetane number than regular Diesel fuel. Today the ignition delay is well-known to increase for lower cetane number fuels. A comparison of burning characteristics for BLT and Diesel fuel in different load points is discussed in the following. The duration from 5-90 % fuel conversion in degree crank angle is defined as the combustion duration. In Figure 5 and 6 the combustion duration and the maximum burning rate are plotted against the ignition delay for Diesel and BLT fuel. The maximum burning rate is defined as the fraction of the total heat release. Page 9 of 21

11 Combustion Duration [ CA] min -1, 4.3 bar -5% Combustion Duration [ CA] min -1, 6.8 bar % Combustion Duration [ CA] min -1, 9.4 bar % Ignition Delay [ CA] Combustion Duration [ CA] min -1, 14.8 bar % Ignition Delay [ CA] EN 590 Diesel BLT Figure 5: Influence of ignition delay time on combustion duration for Diesel and BLT fuel in different load points - Euro 6 ISNOx level Max. Burning Rate [1/ CA] Max. Burning Rate [1/ CA] % 1500 min -1, 4.3 bar -30% 2280 min -1, 9.4 bar Ignition Delay [ CA] Max. Burning Rate [1/ CA] Max. Burning Rate [1/ CA] min -1, 6.8 bar % min -1, 14.8 bar ] % Ignition Delay [ CA] EN 590 Diesel BLT Figure 6: Influence of ignition delay time on max burning rate for investigated fuel in different load points - Euro 6 ISNOx -level Page 10 of 21

12 Generally, BLT shows a much longer ignition delay, in particular at the smaller load points. The ignition delay decreases when load increases for both fuels but the differences between the fuels are getting smaller. Thus, the influence of different fuel properties including the cetane number on the ignition delay is more determining for lower load points with their lower combustion temperatures. The longer ignition delay of BLT leads to an increased part of premixed combustion where fuel is converted immediately. Hence, the maximum burning rate is higher for BLT except for the lowest part load point where a cool flame behavior can be expected. Usually, the ignition will take place in two steps, the so called 1 st and 2 nd stage ignition. At low pressures and temperatures the 1 st stage ignition is highly developed which results in a reduced maximum burning rate. Due to the increased part of premixed combustion also the combustion duration is shorter for the BLT fuel. This behavior can be observed for all part load points. Smoke Number [FSN] min -1 ; 4.3 bar Smoke Number [FSN] min -1 ; 6.8 bar Increased EGR Smoke Number [FSN] min -1 ; 9.4 bar NO x -Emissions Smoke Number [FSN] min -1 ; 14.8 bar NO x -Emissions EN 590 Figure 7: NOx/PM trade-off curves for investigated fuels in different load points PARTICULATE MATTER EMISSIONS - In conventional Diesel fuel operated compression-ignition engines a characteristic 'NOx/PM trade-off curve' is typically observed where PM-emissions increase as EGR is increased in order to reduce NOx emissions. These curves are illustrated in Figure 7 for all investigated load points. Page 11 of 21

13 It is noticeable that at the lowest load point (1500 rpm, 4.3 bar ) this typical behavior can not be seen. Smoke emissions remain very low regardless of the NOx-Level. At the 1500 rpm, 6.8 bar part-load point the same behavior can be seen for B100 and BLT, whereas regular diesel fuel shows the typical NOx/PM tradeoff curve at least below 0.5 g/kwh ISNOx. At 2280 rpm, 9.4 bar also B100 converted to the typical diesel NOx/PM trade-off curve. Only at the highest part load operating point (2400 rpm 14.8 bar ) a slight increase in smoke emissions can be observed for BLT. However, smoke emissions are much lower compared to Diesel and B100. The exact improvement in smoke emissions when using different biofuels can be seen in Figure 8 at a constant NOx-emission level min -1, 4.3 bar min -1, 6.8 bar ISPM ISPM % -100% min -1, 9.4 bar min -1, 14.8 bar ISPM % -100% ISPM % -95% Fuels Fuels Figure 8: Particulate matter emissions - Euro 6 ISNOx level Generally, the combustion proceeds without formation of particulate emissions for all three investigated fuels at the smallest load point. The combustion temperature is below the required temperature for soot formation which lays at about 1500 K. Additionally, the high relative air/fuel ratio prevents soot formation. As load increases particulate matter emissions tend to increase for Diesel and B100. With the use of B100 soot emissions can be reduced by about 85%. Differences between Diesel and B100 slightly decrease for higher loads but the soot reduction maintains on a high level. The BLT even offers remarkable additional potential: for all three lower load points fuel can be converted without soot production, even in a Diesel engine. These results indicate a soot free Diesel combustion over the whole NEDC operation area. Even at the highest load point soot emissions can be reduced by up to 95 % in comparison to Diesel using BLT fuel. Although B100 already shows a significant improvement in PM-emissions, a further reduction by another 80 % can be realized when running the engine with BLT fuel. All these benefits are caused by a more favorable local air/fuel ratio and therefore an improved mixture preparation as a consequence of the oxygen content within the fuel and the longer ignition delay. CO-EMISSIONS With respect to future combustion systems the role of CO-emissions (ISCO) in the spectrum of different regulated emissions gets more important for diesel engines. Advanced combustion systems like HCCI or low temperature combustion systems show disadvantages in the CO-emissions caused through lower Page 12 of 21

14 exhaust temperatures which result in a reduced oxidation rate. Probably, the reduced combustion temperature is responsible for the higher CO-emissions in small load points, as Figure 9 illustrates. ISCO min -1, 4.3 bar +70% ISCO min -1, 6.8 bar +105% 0 0 ISCO min -1, 9.4 bar +95% ISCO min -1, 14.8 bar Fuels Fuels Figure 9: CO-emissions - Euro 6 ISNOx level BLT fuel leads to much higher CO-emissions in comparison to Diesel and B100. At an average, emissions are nearly twice as high as with regular Diesel fuel. This behavior can be seen in all load points except the highest one where CO-emissions are already on a very low level. The increased emissions with BLT are caused by the longer ignition delay which leads to locally very lean air/fuel ratios. Hence, the incomplete combustion which results in increased CO-emissions is the consequence of decreased combustion temperatures which are not sufficiently high for a complete oxidation anymore [10]. HC-EMISSIONS The HC-emission behavior is broadly similar to the CO emission trends. The exact results are illustrated in Figure 10 for Indicated Specific HC-emissions at the Euro 6 NOx-emission level target. Page 13 of 21

15 ISHC min -1, 4.3 bar +125% ISHC min -1, 6.8 bar % min -1, 9.4 bar min -1, 14.8 bar ISHC % ISHC Fuels Fuels Figure 10: HC-emissions - Euro 6 ISNOx level Overall, the blend of butyl levulinate and n-tetradecane shows a significant potential for future soot free compression ignition combustion systems. On the other hand, drawbacks concerning higher HC- and CO-emissions particularly in the NEDC operation area have to be taken into account. An improved catalyst light-off has to be ensured to compensate these disadvantages. Additionally, combustion efficiency decreases slightly due to higher incomplete combustion as a result of the higher HC- and CO-emissions. But in consequence of the soot free combustion much longer regeneration intervals of the Diesel particulate filter (DPF) is possible to prevent losses in fuel consumption. ENGINE EFFICIENCY- Finite nature of fossil fuel supply in addition with increased fuel costs put the fuel consumption into the focus of public interest. All future engine applications have to provide high efficiency performance. Page 14 of 21

16 Figure 11 displays the indicated efficiency for Diesel and BLT. Losses can be separated into the charge cycle, blow-by, wall heat, real combustion, incomplete combustion and the process losses. Figure 11: Split of losses Load Point: 2280 rpm, 9.4 bar Constant centre of combustion Figure 12: Isentropic exponents Load Point: 2280 rpm, 9.4 bar Constant centre of combustion The indicated efficiency of BLT lies about two percentage points below the efficiency of EN 590 Diesel, which is the result of three main aspects: process losses, incomplete combustion and different combustion timing. Due to the lower heating value of BLT, which results in increased injection amounts, the isentropic exponent is significantly lower for BLT in comparison. The different isentropic exponents are illustrated in Figure 12, calculated between -60 and 60 CA for EN 590 Diesel and BLT. The thermal efficiency of the process strongly depends on the isentropic exponent of the charge. The losses caused by real charge increase significantly about over one percentage point as the isentropic exponent decreases during injection. The significant decrease of the isentropic exponent during combustion is caused by higher temperatures close to the top dead centre. Additionally, the higher HC- and CO-emissions of BLT lead to a further reduction of the efficiency as mentioned before, since the unused chemical potential of the exhaust gas has to be evaluated as losses. Page 15 of 21

17 Heat is added in the combustion of EN 590 Diesel closer to the upper dead centre in comparison to BLT. The posterior start of combustion of BLT constitutes the third reason for its lower efficiency. The resulting pressure and burned mass fraction traces are shown in Figure 13. Figure 13: Pressure and Burned Mass Fraction trace - Load Point: 2280 rpm, 9.4 bar - Constant centre of combustion Even a constant centre of combustion does not lead to a similar combustion behavior with different fuels. Particularly for fuels with significantly different ignition delays further combustion control strategies to adjust the maximum burning rate seem to be favorable to maintain the engine efficiency. ENGINE-NOISE For diesel engine applications the engine noise is a critical limit with tightened constraints. The engine noise is characterized through the combustion sound level (CSL), which can be described as following: CSL is a method to predict engine radiated noise based on the analysis of cylinder pressure data, using an average of 50 cycles obtained from the thermodynamic test bench measurements. The CSL value combines cylinder pressure analysis with appropriate engine structure weighting functions for direct and indirect combustion noise as well as mechanical and flow noise. It divides all forces and impacts into two groups according to their characteristic curves. The first group follows the cylinder pressure curve over the crank angle and therefore represents the direct combustion noise share. A correlation to the cylinder pressure gradient exists. The second group follows the torsion force curve over the crank angle and therefore represents the indirect combustion noise share. The overall noise outside of the engine was calculated taking into account attenuation through the engine structure and mechanical noise. Through this method it is possible to separate mechanical and combustion noise sources. Engine noise did prove to be a critical parameter in the engine optimization and imposed a constraint on engine efficiency improvements [11]. Page 16 of 21

18 The comparison of the three investigated fuels with regard to noise emissions supports the following results: Generally, the noise emissions of B100 are within the Diesel range and no remarkable differences can be observed, as Figure 14 illustrates. Particularly for higher load points the combustion of BLT causes significantly increased noise emissions close to 3 db which correlate to a doubling in engine noise. The higher engine noise emissions are caused by an increased part of premixed combustion and therefore higher maximum burning rates. Future investigations should explore an advanced injection strategy, for example a second pilot injection, which probably offers additional potential to reduce noise emissions to the range of the Diesel level min -1, 4.3 bar min -1, 6.8 bar CSL [db] 85 CSL [db] min -1, 9.4 bar +2.8 db min -1, 14.8 bar +2.7 db CSL [db] 85 CSL [db] Fuels Fuels Figure 14: Combustion noise - Euro 6 ISNOx level To summarize the key results, Table 6 shows the emission behavior of BLT fuel in comparison to Diesel fuel: 1500 rpm 4.3 bar 1500 rpm 6.8 bar 2280 rpm 9.4 bar 2400 rpm 14.8 bar PM-emissions o CO-emissions o HC-emissions o Noise-emissions + o : Improvement of 80% / 2.5 db or more +: Improvement of up to 80% / 2.5 db o: No significant fuel effect visible -: Decline of up to 80 % / 2.5 db - -: Decline of 80% / 2.5 db or more Table 6: Summary of emission reduction potential of BLT in comparison to Diesel fuel Page 17 of 21

19 The significant potential of BLT fuel regarding the soot emissions is obvious. Over a wide load range the engine can be operated without any engine-out particulate matter emissions. The need for a DPF is limited to the full load operation. In the NEDC driving cycle a regeneration of the DPF is not required anymore when operating the engine with BLT with the given calibration settings including a constant centre of combustion. On the other hand several drawbacks appeared when running the engine with butyl levulinate. Due to the oxygen in the fuel and the ester-bond in particular malfunctions in various components of the fuel system like fuel hoses and fuel sealings occured. Investigations have shown that the used sealings are not appropriate for aggressive biofuels like butyl levulinate. Corrosion of fuel system components did not appear during these investigations but it seems to be an issue, especially for long-term investigations. Further investigations are required to retool the engine for future fuel requirements. Probably fuel hoses and sealings with different materials like PTFE, better known as Teflon can solve these problems. Although the drawbacks did not appear for B100 a consideration is necessary for all oxygenated future fuels. CONCLUSIONS The target of these investigations was to explore the potential of butyl levulinate as a candidate for future fuels derived from biomass in a Diesel combustion engine. The results support the following conclusions: Butyl levulinate does not ignite without blending with a fast igniting fuel component. Therefore n-tetradecane was chosen even with the knowledge that this component cannot be produced out of biomass today. Investigations have shown that a blend of 70- vol% butyl levulinate and 30- vol% n-tetradecane offers the highest potential to reduce soot emissions while maintaining an acceptable noise level for future combustion systems. Emission results indicate a soot free Diesel combustion over the whole NEDC operation area. Even at the highest load point soot emissions can be reduced by up to 95 % in comparison to Diesel using BLT fuel. As a drawback higher HC-, CO- and noise emissions have to be taken into account. Thus, an improved injection strategy like a second pilot, split injection or reduced rail pressure presents additional potential to further reduce those emissions. Even with closed loop combustion control to adjust the centre of combustion, different combustion and charge properties result in decreased engine efficiency. The utilisation of fuels derived from biomass offers completely new ways of a clean energy conversion technology out of biomass. Future combustion systems such as low temperature combustion or partly homogeneous charged compression ignition can be further improved when ideal fuel parameters are determined. The fuel can be considered as an additional design parameter which offers significant potential for future production engines. Within this study, butyl levulinate could be identified as a first promising cellulose derived fuel compound for a soot free Diesel combustion. Page 18 of 21

20 ACKNOWLEDGMENTS This work was performed as part of the Cluster of Excellence "Tailor-Made Fuels from Biomass (TMFB)", which is funded by the Excellence Initiative by the German federal and state governments to promote science and research at German universities. REFERENCES 1. Weiskirch, C. et al. (2008): Alternative Fuels for Alternative and Conventional Diesel Combustions Systems, SAE Paper , Society of Automotive Engineers 2. Maly, R. (2007): Optimum Diesel Fuel for Future Clean Diesel Engines, SAE Paper , Society of Automotive Engineers 3. Curran, H.J et al. (2001): Detailed Chemical Kinetic Modeling of Diesel Combustion with Oxygenated Fuels, SAE Paper , Society of Automotive Engineers 4. Donahue, R.J. et al. (2000): Effects of Oxygen Enhancement on the Emissions from a DI Diesel via Manipulation of Fuels and Combustion Chamber Gas Composition, SAE Paper , Society of Automotive Engineers 5. Janssen, A. et al. (2009): Tailor-Made Fuels: The Potential of Oxygen Content in Fuels for Advanced Diesel Combustion Systems, SAE Paper , Society of Automotive Engineers 6. Lamping, M et al. (2008): Modernes Dieselbrennverfahren zur Darstellung günstiger Motorrohemissionen bei verbesserten Verbrauchsverhalten, Page 42, MTZ 01/ Busch, H. et al. (2007): Downsizing of HSDI Diesel engines potential and challenges, 6th Symposium Towards Clean Diesel Engines (TCDE), Ischia, Corma, A. et al. (2007): Chemical Routes for the Transformation of Biomass into Chemicals, Chem. Rev. 2007, 107, Huber, G.W. et al. (2006): Synthesis of Transportation Fuels from Biomass: Chemistry, Catalysts, and Engineering; Chem. Rev. 2006, 106, Janssen, A. et al. (2009): Tailor-Made Fuels for Future Advanced Diesel Combustion Engines; SAE Paper , Society of Automotive Engineers 11. Alt, N. et al. (2004): Combustion Sound Prediction within Combustion System Development. SIA Vehicle Comfort 2004 CONTACT Point of contact for questions: Andreas Janssen mailto: janssen@vka.rwth-aachen.de ( ) Page 19 of 21

21 DEFINITIONS, ACRONYMS, ABBREVIATIONS ABDC: After Bottom Dead Center BLT: Blend of 70- vol % butyl levulinate and 30- vol % n- tetradecane BtL: Biomass to Liquid BOI: Beginning of (Fuel) Injection BTDC: Before Top Dead Centre CA50: Point in the combustion process where 50% of the injected fuel mass has been converted, also called the centre of combustion CA: CN: Degrees Crank Angle Cetane Number CSL: Combustion Sound Level DOI: Duration of (Fuel) Injection DPF: Diesel Particulate Filter EGR: Exhaust Gas Recirculation FSN: Filter Smoke Number GtL: Gas to Liquid HCCI: Homogeneous Charge Compression Ignition HFR: Hydraulic Flow Rate : Indicated Mean Effective Pressure ISCO: Indicated Specific Carbon Monoxide Emissions ISHC: Indicated Specific Hydrocarbon Emissions ISNOx: Indicated Specific NOx Emissions ISPM: Indicated Specific Particulate Matter Emissions IQT: Ignition Quality Tester NEDC: New European Driving Cycle η i : Indicated efficiency Page 20 of 21

22 Δη CC : Charge Cycle losses taking into account the losses of expansion work and exhaust and intake force Δη BB : Blow By losses between piston ring and cylinder liner Δη WH : Wall Heat losses taken into account the heat transfer at the combustion chamber walls Δη rc : Losses through real Combustion, when the heat release curve differs from ideal process Δη ic : Losses through incomplete Combustion, when exhaust gas contains unburned hydrocarbons Δη Th : Losses through Thermal efficiency due to real thermodynamic properties and constant volume cycle PM: Particulate Matter PTFE: Polytetrafluoroethylene TMFB: Cluster of Excellence Tailor-Made Fuels from Biomass Page 21 of 21