Documente Academic
Documente Profesional
Documente Cultură
Fuel
journal homepage: www.elsevier.com/locate/fuel
a r t i c l e i n f o a b s t r a c t
Article history: In this study, blends of biodiesel and softwood derived fast pyrolysis oil (FPO) were produced, character-
Received 4 August 2016 ized, and evaluated for thermodynamic and emission performance in a single cylinder research engine.
Received in revised form 23 January 2017 FPO was supplied from an in-house fluidized bed reactor as well as a commercial source. Separate
Accepted 28 January 2017
FPO-biodiesel blends from both FPO sources were prepared using initial volumetric ratios of 80:20 and
Available online 10 March 2017
60:40 (biodiesel:FPO, by volume). After blending, mixing, and a 24 h settling period, two layers formed
and the top, biodiesel-rich layers containing about 5 and 10 vol% FPO were decanted and characterized
Keywords:
on the basis of a thermogravimetric analysis, viscosity, acid number, water content, elemental analysis,
Fast pyrolysis of biomass
Bio-oil blends
and heating value. Significant decreases in viscosity, acidity, and water content from the original FPO val-
Diesel engine idated blending as means of extracting compounds suitable for use as fuels from pyrolytic liquids in bio-
Combustion diesel. A single cylinder, common rail, direct injection diesel engine was used to analyze the combustion
Emissions performance of the FPO fuel blends against neat diesel and biodiesel. Fuel performance was characterized
Biodiesel on the basis of a thermodynamic combustion analysis, as well as corresponding exhaust emission mea-
surements for CO2, CO, unburned hydrocarbons, particulate matter, and NOx.
Engine testing revealed similar indicated efficiencies for biodiesel and diesel at all considered engine
operating modes, while blend fuels showed indicated efficiencies between 75 and 95% of diesel values.
In general, FPO fuels exhibited increased ignition delays and shorter combustion durations with greater
FPO blend concentrations, though this could be partially compensated for using a pilot injection strategy.
The longer ignition delays of the blend fuels resulted in overly lean regions of the cylinder, which pro-
duced largely premixed combustion events contributing to brake specific CO and uHC emissions up to
1.5 and 3.5 greater than diesel, respectively. Specific PM emissions were 41–62% lower for blend fuels
than diesel. Both blends of in-house FPO showed similar PM emission performance, however at higher
concentration than low blend commercial fuel.
Ó 2017 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.fuel.2017.01.093
0016-2361/Ó 2017 Elsevier Ltd. All rights reserved.
J. Laesecke et al. / Fuel 199 (2017) 346–357 347
Nomenclature
ASTM American society for testing and materials LBIH low blend in house fuel
aTDC after top dead center LBC low blend commercial fuel
C commercially produced FPO TAN total acid number
CAD crank angle degrees TGA thermogravimetric analysis
COV coefficient of variation PM particulate matter
DTG derivative thermogravimetric analysis P pressure [bar]
FPO fast pyrolysis oil uHC unburnt hydrocarbons
GHV gross heating value V volume [m3]
GIMEP gross indicated mean effective pressure [kPa] Wi indicated work [J]
dQ n
HB high blend dh
net heat release rate [J/CAD]
HRR heat release rate [J/CAD] gcollected fast pyrolysis oil collection efficiency
IH in house produced bio oil gth;i indicated thermal efficiency
LB low blend / equivalence ratio
circulating fluidized beds, and cyclonic, auger and ablative reac- common rail) [10,11], or studies using research fuel systems not
tors. Variations in reactor design, operation conditions, and initial suitable for production applications [12,13].
feedstock can impact the water content, heating value, acidity, In the current work, fast pyrolysis oil was produced in a flu-
and composition of pyrolysis products. Residence times of <2 s, idized bed reactor and subsequently blended with biodiesel. This
ground biomass <3 mm, immediate char removal to reduce sec- blending technique was applied for two ratios using two sources
ondary cracking, and immediate vapour quenching are characteris- of biomass pyrolysis oil, resulting in four fuel blends. The
tic of pyrolysis systems optimized to produce fast pyrolysis oils biodiesel-FPO blends resulted in a substantial transfer of biodiesel
(FPO) with higher bio-oil yields and thus greater heating values [1]. soluble compounds from the FPO into the biodiesel. The prepared
Fast pyrolysis oils have been investigated as potential fuels in blends were characterized on the basis of a thermogravimetric
various heat and power generation systems and as feedstock for analysis, heating value, acidity, water content, pH, viscosity, and
high-value chemical production including hydrocarbon genera- elemental composition. The results are compared against values
tion; however, properties of pyrolysis oil present serious chal- for commercially available diesel and biodiesel fuels.
lenges to implementation as a transportation fuel [3]. Research of To assess the suitability of the blended FPO fuels for use as a
pyrolysis oil performance in power systems has considered com- drop-in fuel, combustion testing was performed on a common rail,
bustion engines, gas turbines, and furnaces. ASTM specifications direct injection, single cylinder diesel engine at six operating
for liquid pyrolysis oil however remains limited to industrial burn- points. Thermodynamic performance and exhaust emissions of
ers due to poor material compatibility with modern internal com- the biodiesel-pyrolysis oil blends were compared with those for
bustion systems, often resulting in fuel delivery hardware failure neat biodiesel and diesel.
[4]. Methods for increasing pyrolysis oil compatibility with modern
combustion systems such as hydrotreating, catalytic cracking, 2. Materials and methods
emulsifying, filtering, and solvent addition have proven to reduce
water content, acidity, and viscosity; however, all of these come Building on existing biomass FPO research from the literature,
at a significant financial cost to the final product [1]. guidelines in fluidized beds for pyrolysis oil production and FPO
Internal combustion engines fueled with crude pyrolysis oil characterization techniques were adopted from previous studies
have been investigated under a variety of conditions. Suspended to quantify and compare the FPO and blends against traditional
particulate, low heating value, high viscosity, elevated acidity, fuels. No. 2 pump diesel and ASTM certified neat biodiesel were
and phase separation allude to poor compatibility with modern procured as baseline test fuels from commercial suppliers (Shell
high pressure fuel systems [5,6]. Testing has shown poor combus- Canada and Whole Energy Fuels in Bellingham, WA, respectively).
tion of pyrolysis oil due to extended ignition delays and prolonged Pyrolysis fuel blends were prepared with two sources of biomass
combustion durations [7]. FPO: 1) An in-house production facility was developed to produce
In previous works investigating FPO upgrading through emulsi- experimental biomass FPO; and 2) commercially available biomass
fication, it was observed that emulsifying biodiesel and pyrolysis FPO (BTG, BV Netherlands) in order to account for potential varia-
oil in varying fractions resulted in biodiesel-rich and lower tions in pyrolysis oil composition.
lignin/water-rich layers, stabilized over a 24 h settling period [8].
Emulsions were prepared in 6:4 vol ratio of biodiesel:FPO with 2.1. Fast pyrolysis oil production
stirring rates of 1200 rpm and a 4 vol% octanol surfactant addition.
The transfer of FPO compounds to biodiesel was confirmed by In-house FPO was produced using an existing fluidized bed
changes in volume, odor and colour. Heating value, water content, reactor. The reactor was previously used for gasification studies
acidity, and viscosity of the emulsions were shown to be much and was retrofitted for FPO production. The fluidized bed reactor
more favourable in the stable emulsions than the original FPO. was fed using a biomass screw feeder by means of an air inline fee-
The procedure resulted in a novel low cost method to extract com- der. Following the reactor, entrained chars were removed by two
pounds from FPO as a potential additive to transportation fuels. cyclones in series and vapours were cooled by two sequential con-
While there is an extensive research base developed for densers cooling from 400 to 120°C and from 120 to 20°C fixed with
upgraded FPO fuels, investigations often lack experimental studies glass vessels for liquid collection. Non-condensable flue gases were
of the upgraded fuel in a combustion application. Conversely, combusted in an after-burner. A detailed discussion of the fluidized
much of the experimental research relevant to application of bio- bed system is provided elsewhere [14].
mass FPO performance in internal combustion engine systems is Biomass feedstock was sourced from ground softwood pellets.
limited to crude fuels [7,9], older fuel delivery systems (non- The ground biomass was sieved with 360 mm mesh before loading
348 J. Laesecke et al. / Fuel 199 (2017) 346–357
into the biomass hopper. FPO was produced in a fluidized bed reac- viscosity, elemental composition and heating value. All evaluations
tor operated at 500°C, with a feedrate of 1 kg/hr. Fluidized bed were performed in triplicate and the mean values are presented.
material was silica sand of Sauder mean diameter of 400 mm and
bulk particle density 1250 kg/m3. Two separate streams of nitrogen 2.3.1. Water content
were continuously fed into the reactor. The primary nitrogen The water content investigations were carried out with a volu-
stream was directed through the reactor distributor plate and pro- metric Karl Fischer Titration method. A Metrohom 794 Basic
vided the bubbling fluidized bed condition at 0.48 m/s; while, a Titrino was used in conjunction with Sigma Aldrich Hydranal Com-
secondary stream entered the reactor in a lower port introducing posite 5 titrant and dry Methanol (0.01% Water) as a solvent
entrained biomass into the reactor from a pressurized screw feeder medium.
regulating the biomass flow rate.
The collected pyrolysis oil yield was defined as the mass of 2.3.2. Acid number
pyrolysis liquids collected from the condenser (mFPO collected) rela- The Total Acid Number (TAN) was measured with an automatic
tive to the mass of biomass fed into the system (mBiomass) as shown titration system HI902 from Hanna Instruments using aqueous
in Eq. (1). potassium hydroxide (0.1 M) standard titrant and acetone supplied
mFPOcollected by Sigma Aldrich (99.9%) as the titration solvent. ASTM D664 is tra-
gcollected ¼ ð1Þ ditionally implemented in determining of acid number; however a
mBiomassfed
rapid aqueous method was used in order to conserve solvents
The average pyrolysis oil yields were 31%, significantly less more often needed with petroleum analysis [18]. A sample mass
than some literature values well above 50% [15,16]. This is attrib- of 0.10–0.25 g was dissolved in 100 ml acetone before every titra-
uted to the mass of the deposited tars in the upstream regions of tion. After titrating the sample solution, the titration system
the reactor which could not be collected. Deposition of entrained reported the end-point titrant volume.
pyrolysis chars before the collection vessels was prevented
through electrically heated transfer piping; however, a mainte- 2.3.3. Dynamic viscosity
nance protocol including acetone scrubbing of piping and cyclone The dynamic viscosity was determined using an Anton Paar
clearing to remove chars and tars was required in order to preserve rheometer MCR 501 in rotational mode. 20 ml of a fuel sample
the reactor conditions and prevent pressure build up. was filled into a couette cylinder measuring geometry with a diam-
eter of 25 mm and heated to 40°C using an external circulating
2.2. Blending methodology temperature controlled bath.
Previous works have suggested that some compounds in pyrol- 2.3.4. Thermogravimetric analysis
ysis oil may be soluble in diesel and biodiesel [1,17]. In this work, a A TA Instruments SDT Q600 was used to perform a thermo-
blend of biodiesel and FPO was used to extract fuel appropriate gravimetric analysis (TGA) to provide insight into fuel evaporation
compounds from the pyrolysis oil through this relatively simple rates. Samples were heated at a constant rate of 10°C/min from
upgrading strategy. Initial mixtures of 80:20 and 60:40 biodiesel- room temperature to 500°C. The furnace is continuously purged
FPO (by volume) were prepared using neat biodiesel and pyrolysis with a nitrogen flow of 100 mL/min to prevent oxidation of the
oil produced in house (FPO-IH) and obtained commercially (FPO-C). sample. The resulting mass loss and derivative mass loss give
Each blend was stirred at room temperature in atmospheric condi- insight into fuel evaporation characteristics.
tions at 300 rpm for 180 min with an orbital shaker and allowed
24 h to settle. After settling, two layers formed in varying volumet- 2.3.5. Elemental analysis and heating value
ric fractions, confirming the transfer of compounds from FPO to the The elemental composition and heating value of the samples
biodiesel. The upper layer of the blends was decanted and stored at were determined in triplicate by Econotech, Richmond, BC. The
4°C until they were required for further characterization and use. reported values of carbon, hydrogen and nitrogen content were
These decanted upper layers are referred to as LB – Low Blend measured used ASTM D5291, and the balance was assumed to be
(80:20) and HB High Blend (60:40), depending on the initial volume oxygen. The heating value was measured using bomb calorimetry
fractions of biodiesel and FPO. The high blend ratio was selected in accordance with ASTM D4809.
due to the findings of Jiang and Ellis, 2010 reporting the specified
blend fraction to exhibit the optimal stability [8]. The lower blend 2.4. Experimental combustion methods
fraction was selected in order to consider a median value between
neat biodiesel performance and the high blend fuel. Table 1 sum- To assess the combustion characteristics of the fuels, a single
marizes the volumetric distributions of the resulting blends. There cylinder engine test was modified with an auxiliary common rail
was greater transfer of compounds from the in house fuel of both direct injection fuel system to minimize damage to pre-existing
blend fractions in comparison to the commercial fuel. fuel hardware and to minimize the fuel system volume. An exper-
imental test matrix was developed to reflect a range of engine
2.3. Analytical techniques operating conditions and fuel injection strategies. For each operat-
ing point, the combustion process was characterized based on a
The prepared blends were characterized on the basis of acid thermodynamic analysis of the in-cylinder pressure and the
number, water content, thermogravimetric analysis, dynamic exhaust stream emissions.
Table 1
Volumetric distribution of Biodiesel-FPO Blends.
FPO Source Initial Volume Initial Volume Upper Layer Lower Layer FPO transferred to
Biodiesel (%) FPO (%) Volume (%) Volume (%) Biodiesel Volume (%)
LB – In House 80 20 85 15 5.9
LB – Commercial 80 20 84 16 4.8
HB – In House 60 40 67 33 10.4
HB – Commercial 60 40 66 34 9.1
J. Laesecke et al. / Fuel 199 (2017) 346–357 349
2.4.1. Engine and fuel system defined by the crank position h, and the specific heat ratio is taken
The engine used was a naturally aspirated, single cylinder, to be constant at c = 1.37. The net heat release rate indicates the
heavy duty four-stroke diesel engine (Ricardo Proteus) attached rate at which energy is converted by combustion and lost from
to a SP800 Mustang Dynamometer. Due to potential fuel system the cylinder by heat transfer.
failure caused by pyrolysis oil combustion testing [5], extra precau-
dQ n c dV 1 dp
tion was taken to ensure material compatibility. A high pressure ¼ p þ V ð2Þ
dH c 1 dh c 1 dh
direct injection common rail fuel delivery system was assembled
using production components from Bosch. The common rail fuel
injection system was upgraded to use stainless steel compression 2.4.2. Exhaust measurements
fittings where possible and flexible fuel lines were replaced with Exhaust emissions were taken via a heated sampling line and
biodiesel resilient [19] elastomer hoses for low pressure applica- filter and directed to an AVL Emissions Bench, (CEB II) to character-
tions (e.g., fuel injector return line). No modifications were made ize the exhaust stream O2, CO2, CO, unburnt hydrocarbons (uHC),
to the fuel injector or high pressure pump. Table 2 details the and NOx concentrations. The O2 concentration was measured using
research engine specifications. The complete engine facility is a paramagnetic detector; CO2 and CO were measured through a
described elsewhere in detail [20]. non dispersive infrared sensor (NDIR); uHC were measured with
Fig. 1 provides the general process control and instrumentation a flame ionizing detector; and NOx was measured through chemi-
components of the engine test facility. Fuel temperature was mon- luminescence. Diluted particulate matter was sampled using a sep-
itored and limited to 55°C in the fuel pail in order to minimize arate heated exhaust line in conjunction with an eductor dilutor,
potential thermal degradation of the FPO blended fuels, while the driven by bone dry air from a compressor (Ingersoll-Rand
fuel mass flow rate was gravimetrically determined with a scale IRN50H-CC). The diluted particulate matter (PM) sample line was
to measure the change in fuel mass throughout the duration of directed through a low range CO2 analyzer to determine the dilu-
the measurement period. In-cylinder pressure was measured in- tion ratio, typically in ranging 15–20:1, as well as a TSI DustTrak
situ with a piezo-electric pressure transducer (Kistler 6125C). 8533 to measure PM mass concentrations.
From the in-cylinder pressure, a first law analysis provides the
net heat release rate dQ
dh
n
, as seen in Eq. (2) below [21]. P is the mea- 2.4.3. Test points
sured in-cylinder pressure in kPa, V is the cylinder volume m3 Six engine operating points were chosen to compare the com-
bustion and emission characteristics of the four FPO blended fuels
against neat diesel and biodiesel. Five operating points utilized
Table 2 pilot injections, consistent with modern fuel injection strategies.
Test Engine Specifications. A sixth test point utilized a single injection event (no pilot injec-
Displacement 1998 cc Block Ricardo Proteus tion), to contrast the differences in the autoignition characteristics
Bore 130 mm Fuel Pump Bosch CP3.3 of the fuels.
Stroke 152 mm Injector Bosch CRIN2 Table 3 summarizes the injection criteria of each test point. hPilot
Compression Ratio 14.25:1 ECU Custom based on NI
Powertrain Control Modules
and hMain are the commanded crank angle for fuel injection for the
pilot and main fuel injection, respectively. sPilot and sMain are the
1 2
T
Gravimetric 4
Fuel
Measurement 5 3
Common
Rail
Mustang
Dynamometer
DustTrak
PM Bench
Fig. 1. General Process and Instrumentation of Engine Test Facilities. P: Pressure Transducers. T: K-Type Thermocouple. 1-Low Pressure Fuel Pump. 2-Fuel Filter. 3-Bosch CP3
High Pressure Fuel Pump. 4-Fuel Return Heat Exchanger. 5-Fuel Rail and Injector Return Line.
350 J. Laesecke et al. / Fuel 199 (2017) 346–357
Table 4
Characterization of neat test fuels, in-house and commercial crude FPO, and low and high blend fuels.
Fig. 2. Derivative thermogravimetric analysis of test fuels and crude Fast Pyrolysis Oil from In house (IH) and Commercial (C) sources in Low Blend (LB) and High Blend (HB)
ratios.
(DTG) mass loss rate for a constant furnace heating rate of 10°C/ mfuel =ðmair Þ
/¼ ð4Þ
min under a Nitrogen purge. The diesel analysis indicates a peak ðmfuel =mox Þstoich
mass loss rate half of that of biodiesel or FPO blends; however,
A summary of the diesel normalized oxygen equivalence ratios for
across a temperature range roughly 50°C broader than compared
the operating modes with pilot injections is shown in Fig. 3. Ulti-
biodiesel and FPO fuels. LB-IH and LB-C fuels showed little devia-
mately, the equivalence ratio will have a strong impact on combus-
tion from neat biodiesel DTG curves with the exception of a slightly
tion performance and exhaust emissions [24,25]. Blended fuels
lower peak mass loss value. HB-IH and HB-C exhibit similar, albeit
generally show lower overall equivalence ratios than that of biodie-
slightly higher mass loss rates until around 220°C. At temperatures
sel. The general trend of the oxygenated fuels showing increasingly
above this, HB-IH continues to increase to a maximum mass loss at
lower global equivalence ratios relative to diesel at higher loads
a temperature below that of biodiesel but greater than all other
points is a function of greater fuel mass, and thus greater relative
fuels tested. The HB-C mass loss follows the same qualitative
oxygen present for a relatively constant air volume.
trends as that of the biodiesel, but at lower temperatures.
At low temperatures (50–100°C) the crude FPO samples indi-
3.3. Combustion performance
cated mass loss rates consistent with previous literature highlight-
ing the release of highly volatile compounds and water fractions
The efficacy with which the fuel energy was converted to work
[8]. The predominant peaks in both crude FPO samples from 100
was quantified using the gross indicated thermal efficiency. Gross
to 150°C also correspond to other findings which have been attrib-
indicated thermal efficiencies of neat biodiesel and blend fuels
uted in other works to the decomposition of organic acids, phenols,
were calculated according to Eq. (5) and normalized against the
and furans [17]. Ultimately, the DTG characterization revealed that
respective diesel indicated efficiency for a given operating mode.
the volatility of the blend fuels showed similar trends to that of
neat biodiesel. DTG analysis was performed in triplicate to assure Wi
gth;i ¼ ð5Þ
the repeatability of the analysis. mf GHV
The normalized indicated efficiencies are compared in Fig. 4. It are shown in Figs. 6 and 8. Fig. 6 indicates a significant variation in
is generally reported that the thermal efficiency of biodiesel is the pilot ignition delay of the different fuels, which can be attribu-
comparable to engines running with mineral diesel with some table to the expected lower cetane numbers of the FPO blends. Pre-
reports suggesting slightly greater efficiency [26]. A similar trend vious work has estimated the cetane number of crude pyrolysis oil
is seen here with biodiesel efficiencies no more than ±2% from to be in the range of 5–25 [27]. As such, increasing blend fractions
the indicated efficiency of diesel combustion. At higher loads of low cetane number pyrolysis oil with high cetane biodiesel
(modes 1 and 2) the thermal efficiencies of the blend fuels are gen- would result in a lower cetane number, and longer ignition delay,
erally lower than biodiesel. At lower loads, (modes 3–5) a signifi- for blended fuels (relative to diesel) with increasing FPO content.
cant decrease in the efficiency of the blend fuels is noted, relative Using the peak heat release rate during the pilot combustion as
to diesel. In particular, the indicated efficiency for HBIH fuel in an indicator of the ignition delay (and CN), an increase in the igni-
modes 3–5 is lower than the diesel efficiency by 25%. tion delay (and associated decrease in CN) is noted for increasing
The coefficient of variation (COV) of GIMEP shown in Fig. 5 is an FPO fractions (7.5, 6, and 4.5 CAD aTDC for biodiesel, low
indicator of combustion stability, such that a lower COV represents blend, and high blend, respectively).
a more stable operating condition. For operating modes with pilot After the pilot combustion, the subsequent combustion process
conditions, the COV increases with decreasing load. Across all is generally assumed to be mixing controlled. During the early
piloted engine modes, biodiesel and diesel perform very similarly phases of mixing controlled combustion (here, 0–2 CAD), biodiesel
and show roughly half the COV values as the blend fuels. The and diesel have similar heat release rate slopes, as do the blended
GIMEP COVs of both low blend fuels were consistently within 2% fuels. These similarities are present approximately until the end of
of one another. The greatest GIMEP COV is seen in the single injec- the main fuel injection. Due to the mixing controlled nature of
tion case of the high blend in-house fuel, which is likely due to the combustion, the heat release rate during the injection is expected
very low load and long ignition delays. A comparison of the single to be dependent on the fuel injection rate. Fig. 7 indicates the fuel
injection operation with lower load setting of mode 5 indicates energy injected per cycle for each operating mode and, due to the
that the pilot injection for the high blend fuel reduces GIMEP constant injection duration, indicates the average energy addition
COV by 62%. rate during injection. The higher fuel-energy injection rate for die-
The heat release rate histories of two operating points using a sel and biodiesel results in higher mixing controlled heat release
pilot injection (Modes 1 and 3, 700 kPa and 580 kPa, respectively) rates during the injection process. The distinctive heat release
trends during early mixing controlled combustion are attributable
to the differences in injected fuel energy between the neat and
blend fuels. The overall trend of all fuels show well defined pre-
mixed and mixing controlled combustion stages. While the pre-
mixed combustion varies as a result of varying fuel ignition
delays (i.e., cetane number), the main heat release profiles are sim-
ilar. The differing peak heat release rates during the mixing con-
trolled combustion are a result of varying fuel injection
quantities and heating values. The variation in fuel injection quan-
tity – despite constant injection pressure and duration for all fuels
– is due to the differing fuel viscosities of each fuel. Absolute fuel
mass injected during each operating mode is outlined in Table A.1.
Though all fuels exhibited a similar heat release behaviour for
mode 1, this was not the case for the higher blend fuel at the med-
ium load condition (mode 3). As shown in Fig. 8, a distinct heat
release is not evident for the pilot combustion event for the HB-
IH fuel. Furthermore, operation with HB-IH exhibited a delayed
Fig. 4. Diesel normalized gross indicated thermal efficiency of tested fuels for pilot and higher peak heat release rate than other fuels, likely because
injected test modes 1–5. the pilot injected fuel was well premixed and converted during
the main heat release event. Due to the pilot fuel not igniting,
the HB-IH peak heat release is delayed by 2 CAD from all other
4.0% Diesel Biodiesel LBIH LBC HBIH
fuels and results in a 16% and 32% greater peak heat release rate
relative to diesel and biodiesel, respectively. While low blend fuels
3.5% still exhibit a pilot heat release, the extension of the ignition delay
Coefficient of Variance GIMEP
190
Pilot= -20 , main = -6 CAD aTDC
170 τ Pilot =0.45 ms, τmain = 1.2 ms
Injector Command
150
70 LBC
50 HBIH
30
10
Fig. 6. Mode 1 Heat Release Rate of 5 Test Fuels. High load (GIMEP diesel 700 kPa).
but with peak heat release rates retarded by 1 CAD. HB-IH had a
longer ignition delay than biodiesel and the LB fuels, as well as
lower and retarded peak heat release rate (52% that of biodiesel
and 5 CAD later). This very retarded combustion timing is likely
the reason for the very high COV noted in Fig. 5 for the single injec-
tion operating mode. The neat diesel fuel exhibited the longest
ignition delay with a peak heat release rate 3 CAD after that of
the HBIH fuel, leading to the conclusion that the cetane number
of the HBIH fuel was substantially reduced from the original neat
biodiesel and was more similar to that of the presumed lower die-
sel value. Ultimately, the extended fuel ignition delays suggest that
optimized use of the blended fuels would require modified injec-
tion timing if a single injection event strategy were to be used.
Such a strategy may be relevant for older, non-common rail injec-
tion system applications.
An overview of the ignition delay for single and piloted injec-
tion strategies (modes 1–5) is presented in Fig. 10. Here the igni-
tion delay is charcterized as the change in crank position from
the commanded start of injection to 5% integral heat release
Fig. 7. Fuel energy injected per cycle for engine test modes with pilot injection.
Error bars denote measurement standard deviation.
(CA5). At higher engine loads, the ignition delay generally
decreased for all of the fuels. Biodiesel consistently displayed the
150
90 Biodiesel
LBIH
70 HBIH
LBC
50
Diesel
30
10
Fig. 8. Mode 3 Heat Release Rate of 5 test fuels. Medium load (GIMEP diesel 450 kPa).
354 J. Laesecke et al. / Fuel 199 (2017) 346–357
220 was less variable between the fuels. To characterize this, the com-
bustion duration is quantified as the crank rotation between 10%
Injector
Command and 90% integral heat release and is shown in Fig. 11. The FPO
180
= -6 ºCAD aTDC blends have shorter combustion durations in comparison to bio-
main
diesel and diesel. A study of diesel-FPO emulsions concluded that
Heat Release Rate (J/ºCA)
τ main = 1.2 ms
140
combustion duration decreases with increasing FPO fractions
Diesel which was attributed an increased premixed combustion fraction
Biodiesel and shorter mixing controlled combustion fraction than diesel fuel
100 LBIH [31]. This trend is also seen in Fig. 8 and may be attributable to lar-
LBC
ger premixed combustion fractions of the blend fuels, as well as the
delay in mixing controlled combustion to diesel and biodiesel.
60 HBIH
Fig. 12. Specific uHC and CO emissions normalized to diesel for piloted injection operating modes. Error bars denote measurement standard deviation.
Fig. 13. Specific NOx/PM emissions normalized to diesel values for piloted injection operating modes. Error bars denote measurement standard deviation.
shows greater NOx emissions than diesel only in the highest load to the similarities of the LBIH and LBC fuels in heating value, water
operating mode, potentially as a result of the fuel bound oxygen content, viscosity, and volatility, it is proposed that differences in
and the greater in-cylinder temperatures achieved at the highest extracted compounds from the respective FPO feedstock, poten-
engine operating load. Despite heating values similar to biodiesel tially including aromatic compounds, may attribute to the vari-
and a greater fuel oxygen content, low blend fuels show lower ances in PM formation. While beyond the scope of this
NOx emission (85% of diesel). A study on biodiesel-FPO emulsions investigation, a comprehensive characterization of the FPO compo-
similarly reported 7–16% lower NOx emissions from biodiesel-FPO sition is the focus of future work. Differences in fuel composition
emulsions containing 5–15% wood pyrolysis oil [11]; while, a sep- are evident in the relative of oxygen content in the fuels. While
arate investigation with wood pyrolysis oil-biodiesel blend oxygenated species in diesel fuels result in lower PM emissions
reported a 17% reduction in specific NOx emission at high load [24,33], the LBC fuel produced the lowest PM emissions, yet had
from neat biodiesel [35]. the lowest oxygen content of the pyrolysis FPO derivative fuels.
The specific PM emissions from neat biodiesel ranged between This indicates that the fuel oxygen content alone is not a sufficient
70 and 85% of diesel emissions. It is reported that the addition of metric for characterizing the PM formation propensity of FPO-
oxygenates in diesel fuels help to oxidize carbon to CO and CO2, biodiesel blends.
thereby inhibiting PM formation and promoting oxidation [24].
Studies reviewing similar FPO-biodiesel blend performance also
report significant decreases in PM relative to diesel and biodiesel 4. Conclusions
[11,35]. The PM emissions from the commercially derived FPO
show 30–80% emission levels of diesel. In comparison, blend fuels To investigate the application of pyrolysis compounds from
derived from in-house pyrolysis oil performed similarly to one waste biomass as drop in fuels, FPO was produced in a continuous
another producing 50–98% of the specific PM emissions as diesel fluidized bed reactor system and compared to commercially pro-
(i.e., 33–50% than the commercially sourced FPO blend fuel). Due duced FPO. The in-house production system developed here
356 J. Laesecke et al. / Fuel 199 (2017) 346–357
resulted in an FPO collection efficiency of 31%, though this was lim- combustion events. The large premixed combustion fractions
ited by tar deposition in the reactor. The in-house and commercial resulted in elevated CO and uHC emissions, and associated low fuel
FPO was blended with biodiesel to produce fuel blends with 5 conversion and indicated gross efficiencies. Similar trends of ele-
and 10% (by volume) of FPO derived compounds. vated CO and uHC emissions from pyrolysis derived fuels are in
The FPO-Biodiesel blends had 200–350% greater water content agreement with previous work [11,13,31]. The PM emissions were
than neat biodiesel, a large increase in acid number, and a 2%– consistently lower for blend fuels with the exception of the lowest
10% decrease in dynamic viscosity. The gross heating value for load operating mode, where absolute PM emissions were minimal
the low blends was not impacted, while it was reduced by 25 for all fuels. Both blends of the in-house FPO showed similar PM
and 35% for HB-IH and HB-C, respectively. The volatility of the emission performance, however these emissions were higher
blend fuels remained relatively unchanged to the neat biodiesel. (33–55%) than the blends of commercial obtained FPO. It is sus-
A characterization of the molecular composition of the blends pected that the variability in pyrolysis-derived compounds, evi-
was not performed, however future studies will include this anal- dent in the distinctive elemental compositions of the FPO
ysis for further insight into the blending processes. derivative fuels, can be attributed to the variable PM emissions.
Neat diesel, biodiesel, and the prepared fuel blends were tested FPO-biodiesel fuel blends offer a low cost solution to improve
in a single cylinder research engine to observe their combustion crude FPO fuel quality by selectively extracting biodiesel soluble
performance across five operating modes using a piloted injection compounds. Promising reductions in the biodiesel NOx effect and
strategy and one operating mode with a single injection strategy. reductions in PM emissions may contribute to future solutions to
All fuels exhibited well defined premixed and mixing controlled further reduce tailpipe emissions in the advancement of renewable
combustion profiles at higher loads (mode 1 and 2). The relative fuels. In spite of this, challenges with modern fuel hardware com-
phasing of the premixed combustion peaks indicated that in patibility and long ignition delays, which impair low load perfor-
increase in FPO resulted in longer ignition delays, which agrees mance, must be addressed to optimize FPO blends for
with an expected reduction in cetane number, though cetane num- consideration as appropriate drop-in fuel alternatives. To further
bers were not explicitly determined in the current work. Lower understand the performance of FPO-biodiesel blends, an investiga-
heat release rates were observed during the initial phases of the tion into the presence and concentrations of species should be per-
mixing controlled combustion for the blended fuels during high formed to provide better insight into the effects of FPO-derived
load operation. This was attributed to a lower average fuel- compounds on fuel characteristics, combustion processes, and
energy addition rate during injection, due to the lower heating val- emission trends in compression ignition engines.
ues of the blended fuels. For medium to low loads (modes 3–5),
only weak pilot combustion event was noted for the high blend
Acknowledgements
fuel, which resulted in large single event heat release rate profiles.
This was likely the result of the combustion being quenched by
The authors would like to acknowledge the support from the
main injection fuel. Single injection testing showed a slight retar-
Korea Institute of Science and Technology (KIST), through the
dation of the heat release rate profiles for the low blend fuels from
KIST-UBC Bioenergy on-site laboratory program, as well as the
biodiesel, while the high blend fuel had a similar heat release pro-
financial support from the Natural Science and Engineering
file to diesel. In general, the FPO fuels showed increasing ignition
Research Council (NSERC) of Canada through the Discovery Grant
delay and shorter burn duration with greater FPO blend concentra-
(RGPIN 418700-13) and the CREATE Clean Combustion Engine
tions; however, the use of a pilot injection resulted in a relatively
(CREATE 432004-13) programs. We are also grateful for the sup-
similar heat release behaviour for all test fuels at all but the lowest
port and collaboration with our colleagues at UBC, most notably
load (mode 5).
Dr. Bob Evans, Dr. Steven Rogak, Dr. Chang Soo Kim, Jeremy
The indicated efficiency of biodiesel was within 5% of diesel val-
Rochussen, Jeff Yeo, and Katharina Kirsch.
ues, while the blended fuels resulted in a reduced indicated effi-
ciency relative to diesel at lower engine loads (80% the efficiency
of diesel in modes 3–5). The extended ignition delays of the blend Appendix A
fuels, particularly at low loads, resulted in overly lean regions of
the cylinder, which in turn resulted in large, delayed, premixed See Table A.1.
Table A.1
Summary of Experimental Data Presented.
Fuel Mode COV GIMEP Indicated Power Fuel Injected Equivalence CA5-SOI CA90-CA10 PM uHC NOx CO
(%) (kW) (grams/cycle) Ratio (CAD) (CAD) (g/kW h) (g/kW h) (g/kW h) (g/kW h)
Diesel 1 0.50% 11.67 0.14 0.50 15.7 26.6 0.55 0.31 7.92 1.32
Diesel 2 0.62% 9.23 0.13 0.40 14.9 25.5 0.71 0.39 6.73 1.35
Diesel 3 0.89% 7.17 0.09 0.33 14.2 20.6 0.33 0.72 5.97 3.98
Diesel 4 0.80% 6.75 0.10 0.32 16.7 20.7 0.41 0.78 5.90 3.54
Diesel 5 1.29% 3.86 0.06 0.27 17.4 16.4 0.09 2.18 4.24 25.80
Diesel S 1.93% 4.34 – – 15.8 12.7 – – – –
Biodiesel 1 0.53% 10.49 0.15 0.45 12.3 27.4 0.32 0.16 8.57 0.91
Biodiesel 2 0.69% 8.40 0.12 0.37 13.7 24.7 0.44 0.43 6.54 1.76
Biodiesel 3 0.75% 6.22 0.09 0.31 14.2 21.5 0.26 0.71 5.66 5.30
Biodiesel 4 0.84% 5.96 0.12 0.31 15.9 19.8 0.24 0.62 5.44 4.65
Biodiesel 5 1.39% 3.32 0.07 0.26 16.1 16.7 0.07 2.10 3.99 30.21
Biodiesel S 1.42% 3.89 – – 12.9 9.4 – – – –
LBIH 1 1.48% 9.33 0.13 0.41 14.4 24.5 0.49 0.28 6.52 1.13
LBIH 2 1.40% 8.03 0.10 0.36 15.3 21.4 0.34 0.47 6.25 2.34
LBIH 3 1.80% 5.26 0.07 0.28 15.4 19.4 0.21 0.91 4.83 9.45
LBIH 4 1.95% 5.17 0.08 0.29 16.6 18.7 0.19 0.84 4.84 6.61
LBIH 5 2.81% 3.01 0.04 0.26 15.8 17.4 0.09 2.58 3.54 35.83
J. Laesecke et al. / Fuel 199 (2017) 346–357 357
Fuel Mode COV GIMEP Indicated Power Fuel Injected Equivalence CA5-SOI CA90-CA10 PM uHC NOx CO
(%) (kW) (grams/cycle) Ratio (CAD) (CAD) (g/kW h) (g/kW h) (g/kW h) (g/kW h)
LBIH S 1.73% 3.38 – – 12.8 9.3 – – – –
LBC 1 0.56% 9.35 0.13 0.41 14.5 24.6 0.30 0.26 6.58 1.26
LBC 2 0.54% 8.15 0.11 0.37 14.0 23.9 0.21 0.39 6.61 1.87
LBC 3 0.95% 5.94 0.09 0.29 15.2 20.0 0.13 0.77 5.34 7.70
LBC 4 1.84% 5.64 0.07 0.30 16.5 19.0 0.12 0.77 5.18 5.99
LBC 5 1.65% 3.39 0.05 0.26 16.0 16.9 0.08 2.33 3.82 32.39
LBC S 1.73% 3.59 – – 12.9 9.2 – – –
HBIH 1 1.63% 9.31 0.13 0.41 14.9 24.1 0.42 0.25 6.75 1.39
HBIH 2 0.53% 8.02 0.12 0.37 14.6 23.2 0.31 0.35 6.72 2.24
HBIH 3 0.98% 5.38 0.05 0.29 18.4 16.1 0.21 0.97 4.54 13.26
HBIH 4 0.89% 5.41 0.08 0.30 18.7 16.3 0.21 0.75 5.32 9.59
HBIH 5 1.40% 3.10 0.05 0.26 18.1 15.2 0.13 2.95 3.96 41.49
HBIH S 3.75% 2.92 – – 15.5 10.8 – – – –
References [18] Shao J, Agblevor F. New rapid method for the determination of total acid
number (Tan) of bio-oils. Am J Biomass Bioenergy 2015;4:1–9. http://dx.doi.
org/10.7726/ajbb.2015.1001.
[1] Bridgwater A. Review of fast pyrolysis of biomass and product upgrading.
[19] Kass MD, Janke C, Connatser R, Lewis S, Keiser J, Theiss T. Compatibility
Biomass Bioenergy 2012;38:68–94. http://dx.doi.org/10.1016/j.
assessment of elastomeric infrastructure materials with neat diesel and a
biombioe.2011.01.048.
diesel blend containing 20 percent fast pyrolysis bio-oil. SAE Int J Fuels Lubr
[2] Bridgwater A, Meier D, Radlein D. An overview of fast pyrolysis of biomass. Org
2015;8. http://dx.doi.org/10.4271/2015-01-0888. 2015-01-0888.
Geochem 1999;30:1479–93. http://dx.doi.org/10.1016/S0146-6380(99)
[20] Rochussen J, Yeo J, Kirchen P. Effect of fuelling control parameters on
00120-5.
combustion and emissions characteristics of diesel-ignited natural gas dual-
[3] Czernik S, Bridgwater A. Overview of applications of biomass fast pyrolysis oil.
fuel combustion. SAE World Congr 2016. http://dx.doi.org/10.4271/2016-01-
Energy Fuels 2004;18:590–8. Doi 10.1021/Ef034067u.
0792.Copyright.
[4] Tester CC, Density R, Products P, Electrode G. Standard specification for
[21] Heywood JBCN-J or ABRRT. H. Internal combustion engine fundamentals.
pyrolysis liquid biofuel 2013;1:1–5. http://dx.doi.org/10.1520/D7544-12.2.
McGraw-Hill Ser Mech Eng 1988:xxix, 930 2 of plates; 1988.
[5] Chiaramonti D, Oasmaa A, Solantausta Y. Power generation using fast pyrolysis
[22] The role of catalysis for the sustainable production of bio-fuels and bio-
liquids from biomass. Renew Sustain Energy Rev 2007;11:1056–86. http://dx.
chemicals. Amsterdam, NL: Elsevier Science; 2013.
doi.org/10.1016/j.rser.2005.07.008.
[23] Murphy MJ, Taylor JD, Mccormick RL. Compendium of experimental cetane
[6] Mueller CJ. The feasibility of using raw liquids from fast pyrolysis of woody
number data. Natl Renew Energy Lab 2004:1–48. http://dx.doi.org/10.2172/
biomass as fuels for compression-ignition engines: a literature review 2013.
1086353.
http://dx.doi.org/10.4271/2013-01-1691.
[24] Cheng A, Dibble R, Buchholz B. The effect of oxygenates on diesel engine
[7] Shihadeh A, Hochgreb S. Diesel engine combustion of biomass pyrolysis oils.
particulate matter. SAE-Paper; 2002. http://dx.doi.org/10.4271/2002-01-1705.
Energy Fuels 2000;14:260–74. http://dx.doi.org/10.1021/ef990044x.
[25] Zhou L, Boot MD, Johansson BH, Reijnders JJE. Performance of lignin derived
[8] Jiang X, Ellis N. Upgrading bio-oil through emulsification with biodiesel:
aromatic oxygenates in a heavy-duty diesel engine. Fuel 2014;115:469–78.
thermal stability. Energy Fuels 2010;24:2699–706. http://dx.doi.org/10.1021/
http://dx.doi.org/10.1016/j.fuel.2013.07.047.
ef901517k.
[26] Tesfa B, Mishra R, Zhang C, Gu F, Ball AD. Combustion and performance
[9] Moloodi S, Tzanetakis T, Nguyen B, Zarghami-Tehran M, Khan U, Thomson MJ.
characteristics of CI (compression ignition) engine running with biodiesel.
Fuel property effects on the combustion performance and emissions of
Energy 2013;51:101–15. http://dx.doi.org/10.1016/j.energy.2013.01.010.
hardwood-derived fast pyrolysis liquid-ethanol blends in a swirl burner.
[27] Suppes G, Tui Y, Regehr E. Hydrophilic diesel fuels – ignition delay times of
Energy Fuels 2012;26:5452–61. http://dx.doi.org/10.1021/ef300657d.
several different blends. State Altern Fuel Technol 1997. http://dx.doi.org/
[10] Van De Beld B, Holle E, Florijn J. The use of pyrolysis oil and pyrolysis oil
10.4271/971686.
derived fuels in diesel engines for CHP applications. Appl Energy
[28] Menkiel B, Donkerbroek A, Uitz R, Cracknell R, Ganippa L. Combustion and soot
2013;102:190–7. http://dx.doi.org/10.1016/j.apenergy.2012.05.047.
processes of diesel and rapeseed methyl ester in an optical diesel engine. Fuel
[11] Prakash R, Singh RK. Comparison of performance and emission parameters of a
2014;118:406–15. http://dx.doi.org/10.1016/j.fuel.2013.10.074.
diesel engine fueled with biodiesel and wood pyrolysis oil emulsions. SAE Tech
[29] Hribernik A, Kegl B. Influence of biodiesel fuel on the combustion and emission
Pap 2012;13:123–32.
formation in a direct injection (DI) diesel engine. Energy Fuels
[12] Chiaramonti D, Bonini M, Fratini E, Tondi G, Gartner K, Bridgwater A, et al.
2007;21:1760–7. http://dx.doi.org/10.1021/ef0700432.
Development of emulsions from biomass pyrolysis liquid and diesel and their
[30] Bittle JA, Knight BM, Jacobs TJ. Interesting behavior of biodiesel ignition delay
use in engines – Part 1: emulsion production. Biomass Bioenergy
and combustion duration. Energy Fuels 2010;24:4166–77. http://dx.doi.org/
2003;25:85–99. http://dx.doi.org/10.1016/S0961-9534(02)00183-6.
10.1021/ef1004539.
[13] Lee S, Jang Y, Kim TY, Kang K-Y, Kim H, Lim J. Performance and emission
[31] Huang Y, Han X, Shang S, Wang L. Performance and emissions of a direct-
characteristics of a diesel engine fueled with pyrolysis oil-ethanol blend with
injection diesel engine operating on emulsions of corn stalk bio-oil in diesel.
diesel and biodiesel pilot injection. SAE Int J Fuels Lubr 2013;6:785–93. http://
Proc Inst Mech Eng Part D J Automob Eng 2012;226:1119–29. http://dx.doi.
dx.doi.org/10.4271/2013-01-2671.
org/10.1177/0954407012438295.
[14] Masnadi M, Grace J, Bi X, Lim J, Ellis N, Li YH, et al. From coal towards
[32] Shihadeh A, Hochgreb S. Impact of biomass pyrolysis oil process conditions on
renewables: catalytic/synergistic effects during steam co-gasification of
ignition delay in compression ignition engines. Energy Fuels 2002;16:552–61.
switchgrass and coal in a pilot-scale bubbling fluidized bed. Renew Energy
http://dx.doi.org/10.1021/ef010094d.
2015;83:918–30. http://dx.doi.org/10.1016/j.renene.2015.05.044.
[33] Sison K, Ladommatos N, Song H, Zhao H. Soot generation of diesel fuels with
[15] Salehi E, Abedi J, Harding T. Bio-oil from sawdust: effect of operating
substantial amounts of oxygen-bearing compounds added. Fuel
parameters on the yield and quality of pyrolysis products. Energy Fuels
2007;86:345–52. http://dx.doi.org/10.1016/j.fuel.2006.07.019.
2011;25:4145–54. http://dx.doi.org/10.1021/ef200688y.
[34] Hoekman SK, Robbins C. Review of the effects of biodiesel on NOx emissions.
[16] DeSisto WJ, Hill N, Beis SH, Mukkamala S, Joseph J, Baker C, et al. Fast pyrolysis
Fuel Process Technol 2012;96:237–49. http://dx.doi.org/10.1016/
of pine sawdust in a fluidized-bed reactor. Energy Fuels 2010;24:2642–51.
j.fuproc.2011.12.036.
http://dx.doi.org/10.1021/ef901120h.
[35] Lee S, Kim T, Kang K. Performance and emission characteristics of a diesel
[17] Garcia-Perez M, Adams TT, Goodrum JW, Geller D, Das KC. Production and fuel
engine operated with wood pyrolysis oil. Proc Inst Mech Eng Part D J Automob
properties of pine chip bio-oil/biodiesel blends. Energy Fuels
Eng 2013;228:180–9. http://dx.doi.org/10.1177/0954407013502951.
2007;21:2363–72. http://dx.doi.org/10.1021/ef060533e.