Combustion of fast pyrolysis bio-oil and blends in a micro gas turbine

Biomass and Bioenergy 115 (2018) 174–185

Contents lists available at ScienceDirect

Biomass and Bioenergy journal homepage: www.elsevier.com/locate/biombioe

Research paper

Combustion of fast pyrolysis bio-oil and blends in a micro gas turbine a,b,∗

a

b

T

a

Marco Buffi , Alessandro Cappelletti , Andrea Maria Rizzo , Francesco Martelli , David Chiaramontia,b a b

CREAR, Industrial Engineering Department, University of Florence, Viale Morgagni 40, 50134 Florence, Italy Renewable Energy COnsortium for R&D, RE-CORD c/o University of Florence, Viale J. F. Kennedy, 184, 50038 Pianvallico, Florence, Italy

A R T I C LE I N FO

A B S T R A C T

Keywords: Biofuels Bioliquids Fast pyrolysis bio-oil Gas turbine Combustion Emissions

The use of Fast Pyrolysis Bio-Oil (FPBO) as fuel was studied in a small scale non-regenerated micro gas turbine, set up in a dedicated test rig. The system includes a dedicated injection line and a modified combustor to burn efficiently high volume fractions of FPBO-in-ethanol solution. The effect of the larger combustor volume improved the quality of the combustion of the reference diesel oil and pure ethanol as regards exhaust emissions, while maintaining unchanged fuel consumptions of the original configuration. Tests with 20/80 and 50/50% (volume fractions) of FPBO/ethanol blends showed successfully and stable engine operation. By increasing the FPBO volume fraction in the fuel blend, an increase in CO emissions was observed - probably due to the larger droplets derived from the more viscous fuel - as well as in NOX emissions - probably due to fuel-bound nitrogen. Considering the proposed modifications and FPBO/ethanol blend as fuel, the engine reached an overall electrical efficiency higher than that measured with benchmark diesel fuel. A final run with 100% FPBO feeding showed unstable combustion with the presence of carbon deposits in the hot parts of the system, showing that the present configuration requires further modifications to achieve this goal. Guidelines were provided for the implementation of further upgraded solutions for MGTs towards viscous, acidic and aqueous fuels feeding.

1. Introduction

and acetic acid, exhibiting low pH and high TAN (Total Acid Number). The water contained in pyrolysis oil comes from both biomass moisture and reaction stages (process water) [4], the latter including both dehydration and degradation reactions [5]. As regards fuel storage, phase separation typically occurs if the water content in the biocrude is higher than 30–45% [3]. Biomass-derived pyrolysis oil also presents high viscosity compared to fossil liquid fuels, and cannot be heated above 353 K [6–8]. This is a significant difference respect to other bioliquids such as vegetable oils, where the fuel viscosity can be reduced to diesel fuel range just heating above 373 K [9] in order to achieve similar performance compared to conventional fuels [10–12]. Moreover, due to its high oxygen content, pyrolysis oil shows very poor ignition properties and low energy density compared to fossil fuels. An extensive description of fast pyrolysis oil properties and analysis methods has been published by Oasmaa and Peacocke [13]: they provided guidelines in the design of process equipment and power generation systems. Presently, considering commercial applications, two sets of FPBO quality for burners are covered by ASTM D 7544 [14], and the corresponding European standards for boiler FPBO grades are being developed under CEN in the EU [15].

1.1. Fast Pyrolysis Bio-Oil Bioliquids can substitute part of fossil fuel demand for energy generation. Among these, viscous biocrudes derived from thermochemical processing of lignocellulosic biomass represent an opportunity to store energy and convert it in CHP units. Their use is particularly suited for stationary power generation, since in the energy sector the technology can be adapted to the properties of fuel, reducing the fuel upgrading costs. Differently, the transport sector requires the compliance with specified fuel norms, thus biocrudes and bioliquids must be further processed before their use into existing commercial vehicles, increasing their final cost [1]. In order to feed viscous bioliquids in internal combustion engines and/or gas turbines for small scale CHP units, significant modifications to injection and combustion systems are however required [2,3]. Lignocellulosic-derived bioliquids such as Fast Pyrolysis Bio-Oil (FPBO), show significant differences compared to fossil fuels in terms of chemical and physical properties. A large amount of oxygenated components is present in FPBO: it has a polar nature and cannot be mixed with hydrocarbons. The bio-oil contains water and organic acids, like formic

∗

Corresponding author. CREAR, Industrial Engineering Department, University of Florence, Viale Morgagni 40, 50134 Florence, Italy. E-mail address: marco.buffi@unifi.it (M. Buffi).

https://doi.org/10.1016/j.biombioe.2018.04.020 Received 8 October 2017; Received in revised form 20 April 2018; Accepted 24 April 2018 0961-9534/ © 2018 Elsevier Ltd. All rights reserved.

Biomass and Bioenergy 115 (2018) 174–185

M. Buffi et al.

Nomenclature AC AFR APU ASTM C CC CEN CHNOS CHP DAQ DIN EL EN

ER FPBO GPU HDPE HHV ISO LHV MGT NDIR NDUV OC ps PO/EtOH SMD T TAN Ts

Alternate current Air-Fuel Ratio Auxiliary Power Unit American Society for Testing and Materials Compressor Combustion Chamber European Committee for Standardization (in French: Comité Européen de Normalisation) Carbon Hydrogen Nitrogen Oxygen Sulphur (elemental analysis) Combined Heat and Power generation Data AcQuisition German Institution for Fuel Norms and Specifications (in German: Deutsches Institut für Normung) Electrical Load European Norm

Equivalence Ratio Fast Pyrolysis Bio-Oil Ground Power Unit High-Density PolyEthylene Higher Heating Value International Organization for Standardization Lower Heating Value Micro Gas Turbine Non Dispersive Infra-Red (sensor type) Non Dispersive Ultra-Violet (sensor type) Original Combustor Pressure sensor Fast Pyrolysis bio-oil/Ethanol blend Sauter Mean Diameter Turbine Total Acid Number Temperature sensor

1.2. Viscous, acidic fuels in gas turbines

1.3. Aim of the work

Considering FPBO as fuel for energy generation systems based on internal combustion engines, the main barriers related to its use are the high viscosity and viscosity, and the low heating value. As summarized by Van De Beld et al. [3,16] and other literature on diesel engines [17], significant modifications to the engine are necessary, from fuel storage to the nozzle. A comprehensive overview of the pyrolysis oil use in engines and turbines was given by Chiaramonti et al. [18], that highlighted how corrosion resistant materials, in-line cleaning implementations, and heated feeding lines have to be adapted to operate with the aggressive properties of the biocrude. Czernik and Bridgwater [19] reported a series of experiments on gas turbines which investigated the use of viscous fuels using different methods to accommodate some of the peculiar properties of biomass pyrolysis oils. An advantage of these systems consisted in the “silo” type combustion chamber located outside of the turbo-compressor body, a component that can be easily modified and optimized than annular combustors. This was indeed demonstrated by several researchers from the University of Twente, which studied the effect of vegetable oil [20] and pyrolysis oil [21] in a 50 kW DG4M-1 (electrical power out), a radial design micro gas turbine formerly used as APU. The use of pyrolysis oil and blends was successfully tested and the emissions characterized, showing that pyrolysis oil generates higher CO emissions that other bioliquids. Generally, high CO emissions indicates low combustion efficiency, thus the re-design of combustor becomes crucial. Recently, Beran and Axelsson [22] developed a new concept of combustor able to burn efficiently pyrolysis oil (at full load, from 70 to 100% power) for an OPRA OP16, an all radial single-shaft gas turbine rated at 1.9 MW. Main features of the system are the increased combustor volume to allow the combustion of bioliquids with low energy content, and a dedicated air blast nozzle. They found that the maximum allowed droplet size of the pyrolysis oil spray should be about 50–70% of the droplet size for diesel fuel, to achieve efficient combustion. Other studies and tests on a regenerated small turbine engine were performed by Seljak et al. [23] with the liquefied wood, a biocrude produced from the solvolysis of lignocellulosic biomass in acidified glycols, with fuel properties rather similar to pyrolysis oil. The required MGT adaptations were studied together with the testing methodology and fuel properties [24], leading to the design of an operational prototype [25], tested with different compositions of fuels [26,27]. However, the impeller was operated in absence of an electrical generator, thus requiring further studies to be operated in energy generation applications.

The objective of this study is to investigate the operation of a small scale non-regenerated gas turbine when a viscous, acidic and aqueous bioliquid such as FPBO is used as fuel. A gas turbine test rig was developed with several new components that replaced the original ones. The MGT combustor was redesigned, in order achieve higher temperature in the primary combustion zone. Details of the combustor redesign followed previous studies by Cappelletti et al. [28] and more recent developments Beran and Axelsson [22], enlarging the combustion volume to maximize droplets residence times. Previous promising tests on a former version of the MGT test rig with standard fuel (diesel), biodiesel and vegetable oil [29–31], provided evidence that there was still room to improve the system to fast pyrolysis oil feeding. Moreover, other studies on viscous bioliquids in micro gas turbine were carried out on a Capstone C30 LF [32,33], focusing on heating the feeding line above 373 K to reduce oil viscosity and achieve good atomization and a stable combustion for several hours of operation. Thus, the re-designed MGT rig includes: a new re-designed combustor, two pilot flames for start up/shut down, a new control system, and a new injection line based on a tri-fuel system. A preliminary study on the spray performance was carried out in order to estimate the quality of the atomization. Then, the test campaign evaluated the MGT performance and emissions with reference diesel fuel and ethanol, comparing the original and the new configuration of the combustor (at equal spray performances). The innovative proposal of the work is based on tests with FPBO and ethanol blends at 20/80 and 50/50% (volume fractions). 2. Materials and methods 2.1. Fuel analysis Five different fuel formulations were considered for MGT tests: pure FPBO, denatured ethanol, two FPBO/ethanol blends at different volume fractions (20/80% and 50/50%), and commercial diesel fuel as benchmark fuel. FPBO was provided by BTG Biomass Technology Group BV (The Netherlands). The bio-oil was obtained by flash pyrolysis of pine wood in a rotating cone reactor at process temperature of 773 K and residence less than 1 s (according to the description of BTG fast pyrolysis reactor [34]). The FPBO is a homogenous liquid with a dark-brown uniform color and no suspended solids, subsequently filtered by BTG in order to reduce the amount of ash and solids, and thus to enable direct applications of the oil on both boilers and gas turbines. The filtered FPBO was sent to Italy and stored in a 1 m3 HDPE tank, few 175

Biomass and Bioenergy 115 (2018) 174–185

M. Buffi et al.

MGT configuration, based on the criteria given in the next section. The injection system is a triple-fuel injection line, which includes a heating section before the fuel nozzle. Diesel oil is stored into the original metallic tank, while the other bioliquids (i.e. pyrolysis oil/blends and ethanol) are stored into HDPE tanks. Piping and fittings are made of stainless steel AISI 316 and connect each device of the injection line, including two T-filters equipped with a 100 μm cartridge. The selected main fuel pump, fully compatible with pyrolysis oil viscosity and corrosion properties, is a Netzsch Nemo NM 015 SY 08 S 48 B, which can pressurize the fuel delivery line up to 2.5 MPa. Injection pressure is regulated by a manual bypass needle valve, and the fine adjusting is made varying pump rotational speed by an inverter. The system includes three Watlow Cast X heaters which can heat up the fuel up to 393 K, and two separated pilot injectors with their dedicated injection system for MGT warm up and transient load operations. The rig is equipped with several thermocouples and pressure sensors to allow MGT control and monitoring. Measurements of pressure, temperature, power output and fuel flowrate are monitored and sampled by a dedicated data acquisition system (DAQ). Data acquisition hardware is controlled by Labview software.

days after its production. The ethanol used for the research work was a commercial denatured hydrous alcohol (ethanol at 94% volume fraction). The FPBO and commercial denatured ethanol used for the present campaign were analyzed and results are reported in Table 1, including the benchmark commercial diesel fuel. FPBO has higher viscosity and density than diesel oil, and a low pH, unusual as gas turbine fuel. The lower energy content of a fuel can be partly compensated by increasing the fuel mass flow, while the corrosion and wear from an acid liquid require switching to compatible materials as stainless steel or selected polymers. The solids content of the filtered FPBO, measured according the procedure developed by Oasmaa and Peakocke [13], was below 0.5% mass fraction, thus suitable for spray applications [35]. As already mentioned, the high viscosity of pyrolysis oil represents a challenge for fuel atomization in gas turbines. In order to reduce the viscosity of pyrolysis oil, ethanol can be used as solvent. Seven samples of PO/EtOH at different ratio of pyrolysis oil-in-ethanol were characterized in terms of viscosity and LHV, using a Lauda Proline viscometer PV 15 and a LECO AC 500 isoperibol calorimeter. Blends of pyrolysis oil-in-ethanol were prepared in 30 dm3 HDPE tanks at 20% and 50% volume fractions of pyrolysis oil in denatured hydrous ethanol, and then mixed for few minutes. Fuel batches were prepared about 30 min before each test run.

2.3. Emissions measurement Two independent systems for the analysis of exhaust gases concentration of the MGT engine were adopted. The main instrument consists in a portable on-line gas analyzer (Greenline 8000), which measures CO, CO2, O2, and NOX concentrations. Emission levels during diesel fuel and ethanol tests were also double-checked by a portable onboard exhaust gas analyzer (Sensors Semtech DS) with NDIR (Non Dispersive Infra-Red) sensor for CO/CO2 measurement and NDUV (Non Dispersive Ultra-Violet) for NO/NO2 measurement. Measuring range, resolution and accuracy of the instrumentation is reported in Table 2. All emissions concentrations are reported in dry basis, corrected and normalized at 15% volume fraction of oxygen in exhaust gases.

2.2. Experimental setup The small scale gas turbine rig was designed as a test bench suitable for investigating the combustion of viscous bioliquids and biocrudes in actual power generation units and conditions. The rig was specifically built considering the fuel properties of vegetable oils [12] and fast pyrolysis oil [36], the requirements in terms of material compatibility; and the higher flow rate necessary to generate the same power. The rig is composed by multi-fuel injection line, control panel, turbine housing (sub-assembly) and AC generator. Separately, a 0–25 kW load bench dissipates the electrical power generated by the turbine. A schematic of the MGT rig layout is shown in Fig. 1. The MGT section is based on a micro gas turbine used either as auxiliary power unit (APU), or military ground unit (GPU), produced by AiResearch-Garrett Corporation, model GTP 30–67. The engine was acquired from Avon Aero Supply, Inc. (USA). The turbine shaft rotates at the fixed rotational speed of about 5550 rad s−1, with a maximum AC output of 25 kVA, 400 Hz, 120/208 V. Thanks to its simple, radial design, the selected technology is relatively easy to operate with a rather broad range of fuels and accessible in case modifications are required [12,29]. A detailed scheme of the test rig injection line is reported in Fig. 2. The fuel atomizer is a single, pressure swirl duplex nozzle with integral flow divider, placed in a reverse flow silo combustion chamber. Combustor casing and liner were completely re-designed from the original

2.4. Combustor design Previous works [37,38] discussed the use of pyrolysis oil in unmodified MGT combustion chambers. The design of a new combustor for unconventional fuels required the consideration of several possible solutions, including geometry, air distribution and internal air-fuel mixing [39–43]. The system must be able to ensure flexible operation and work with multiple fuels: in our work a partial re-design of the original silo-combustor (as already investigated by Cappelletti et al. [44–46] for other gas turbine solutions) was carried out. Based on the original geometry of the Garrett GTP 30–67 combustor liner, the work applied the design criteria of Beran and Axelsson [22] to increase the combustor volume. A key requirement for an adequate FPBO combustion is to increase the residence time of the oil droplets [47], which in

Table 1 Chemical characterization of fuels. Fuel properties

Unit

C H N S O pH Ash content Water content Kinematic viscosity at 313 K Density at 298 K Solids content HHV LHV

mass fraction mass fraction mass fraction mass fraction mass fraction – mass fraction mass fraction mm2 s−1 Mg m−3 mass fraction MJ kg−1 MJ kg−1

(%) (%) (%) (%) (%) (%) (%) as received

(%)

Method

Diesel fuel (EN 590 98/70/EC)

Denatured ethanol

Filtered FPBO

ASTM D5291 ASTM D5291 ASTM D5291 Internal method Calculated ASTM E70 EN ISO 6245 EN ISO 8534 EN ISO 3104 EN ISO 3675 ASTM D7579 DIN 51900-2 Calculated

86.5 ̴ 13.5 ̴ 0 < 0.0001 < 2.7 – – – 2–4.5 0.82–0.85 0 45.60 42.50

44.93 12.30 0.02 0 42.7 7 0 7.2 1.1 0.817 0 25.37 22.76

54.97 6.43 0 0.013 38.56 2.7 0.02 22.5 37.01 1.192 0.0197 18.91 17.32

176

Biomass and Bioenergy 115 (2018) 174–185

M. Buffi et al.

Fig. 1. Schematic of the MGT test rig and interconnections, including load bench and measurements [29].

Consequently, the estimated droplet residence time also increased 5 folds with respect to the original configuration. The combustor casing was also re-designed and appropriate room for the pilots injectors was added. In the original combustor, a simple swirler on primary air worked as a cooling slot for the dome. In the new geometry, the two functions were separated: an eight-blade swirler was realized in the centerline with the role of stabilizing the flame as suggested for low swirl combustors [52], and ten cooling slots were added for the external wall cooling. The air distribution in the primary zone of the combustion chamber is linked to the stoichiometric ratio and the energy content of the fuel. The stoichiometric AFR (Air-Fuel Ratio) of fast pyrolysis oil is lower than that of diesel fuel, as well as the LHV, so the air distribution has been modified in order to fit FPBO requirements. In order to achieve complete combustion of FPBO, the design aimed 0.9–0.95 ER (Equivalence Ratio) when remaining at 20 kW of electrical power output. Based on the fuel properties given in Table 1, the AFRs of selected fuels were calculated. The AFRdiesel/AFRPO is equal to 2.06 while the LHVdiesel/LHVPO is 2.44: the original air distribution system provides 1̴ 6% less air than that required for the complete combustion of pyrolysis oil. Thus, the new combustor geometry implemented a higher

the primary zone is defined as:

τres =

Vpz ρgas 1 m˙ PZ

(1)

where ṁPZ is the mass flow in the primary zone, ρgas is the combustion gas density in the primary zone, and VPZ is the volume of the primary zone. The main design parameter is the combustor load, ϑ, defined as the ratio between the reaction and residence times [48–50]:

θ=

T P31.75 VPZ exp ⎛ B 3 ⎞ ⎝ (φ) ⎠ m˙ PZ

(2)

P3 is the combustor inlet pressure, VPZ is the volume of the primary zone, B(φ) is a function of the equivalence ratio in primary zone, and ṁPZ is the mass flow in the primary zone. Generally, higher combustor efficiency is obtained by increasing the combustion load [39,48,51]. The application of this method to our case, keeping the suggested ratio between the original combustor volume and the redesigned model at about 4 (considering not-regenerated gas turbines), resulted in a significantly larger combustor. The new liner has been also increased almost two times in length to suit the turbine inlet section.

Fig. 2. Scheme of the multi-fuel injection line and the MGT sub-assembly. Abbreviations: Ts, temperature sensor; ps, pressure sensor; T, turbine; C, compressor; CC, combustion chamber; EL, electrical load. 177

Biomass and Bioenergy 115 (2018) 174–185

M. Buffi et al.

Table 2 Specifications of emissions measuring systems. Instrument

Gas sensor

Measuring range

Resolution

Accuracy

Greenline 8000

O2 CO2 CO NOx O2 CO2 CO NOx

0–25% of span calculated 0 - 4000 mg kg−1 0–1000 mg kg−1 0–25% of span 0–20% of span 0 - 8000 mg kg−1 0–1000 mg kg−1

0.1% 0.1% 1 mg kg−1 1 mg kg−1 0.1% 0.01% 1 mg kg−1 0.1 mg kg−1

± 0.1% – ± 10 mg kg−1 < 300 mg kg−1, ± 4% measured vale up to 4000 mg kg−1 ± 5 mg kg−1 < 100 mg kg± 0.1% ± 3% of reading or ± 0.1%, whichever is greater ± 3% measured vale up to 4000 mg kg−1 ± 2% of reading or ± 3% full scale, whichever is greater

Sensors Semtech DS

Hardesty [60] measured 72 mN m−1 at 298 K; Bakhshi and Adjaye [61] 29.2 mN m−1; Shihadeh and Hochgreb [17] 32.2–40 mN m−1 for NREL and ENSYN pyrolysis oil, depending on water content, and 29.3 mN m−1 for diesel No. 2. Garcìa-Pérez et al. [62] studied the multiphase structure of pyrolysis oil and measured the surface tension of bio-oil at different temperatures (from 298 to 353 K): at 318 K a significant reduction of the surface tension occurs and at around 353 K it drops below 25 mN m−1. Based on these literature data, the surface tension of pyrolysis oil was assumed equal to 34 mN m−1 at 298 K. The surface tension of the other fuels was also taken from literature [63,64]. The surface tension for pyrolysis oil-ethanol blends was estimated by averaging as a function of the mass fractions.

air mass flow in the swirler section so to compensate this difference. As regards the holes position along the liner, no significant modifications was necessary and approximately the same design was maintained. The new combustion chamber includes also two additional holes along the swirler surface for pilot nozzles. A comparison between the new and the old liner geometry is given in Fig. 3. 2.5. SMD estimation for FPBO spray This part of the work aimed at verifying if the nozzle is able to produce droplets of the right size for proper combustion. Despite recent studies selected air-assisted atomizers for pyrolysis oil [22], or liquefied wood [24,25], this particular solution could cause excessive penetration of the spray inside the combustion chamber at high fuel injection pressure [53]. In order to avoid this potential effect, considerably accentuated in the case of small scale MGTs, the existing pressure swirl nozzle has been maintained. A preliminary investigation of the expected SMD (Sauter Mean Diameter) was carried out for the selected fuels according to Jasuja’ equation [54] for kerosene, gas oil and heavy residual fuel, also suggested by Lefebvre [55] and Rashad et al. [56] for viscous fuels. SMD is the diameter of a liquid droplet whose volume-to-surface area ratio equals that of the complete spray, and is defined by the following equation:

SMD = 4,4σ 0,6vL0,16 mL0,22 ΔPl−0,43,

2.6. Design of experiment Test runs started initially using pure denatured ethanol and diesel oil as fuels in the unmodified (original) MGT combustor. Then, blends of filtered FPBO/ethanol (in abbreviation PO/EtOH) were investigated in the new combustor. The experimental plan is reported in Table 3. The warm-up procedure consists in running the engine at idle conditions (no electrical load) for 5 min since all metal parts are heated up and the engine operates at stable conditions. Afterwards, the engine is ready for tests with pure diesel fuel or pure ethanol using the main nozzle only (as shown in Fig. 2): pilot injectors are instead operated in case of FPBO (pure or blended) is fed. Electrical load is set between 5 and 20 kW, each measuring point stabilized for at least 5 min before parameter acquisition. For FPBO/blend or pure FPBO testing, the combustion is initially supported by the ethanol pilot flames and then the main injection line is activated. In order to maintain a constant power output, the ethanol mass flow in the pilot line is gradually reduced until the thermal power is completely provided by the main injection line. Shutdown reverses the procedure, finally switching to diesel fuel to maintain the system always protected by a diesel oil film.

(3)

where σ is the surface tension, vL is the kinematic viscosity, mL is the fuel flow rate and ΔP is the pressure drop over the nozzle. The pyrolysis oil surface tension is higher than that of traditional fuels due to a higher water content [57,58] and a number of chemical compounds potentially contributing to a possible unbalance between non-equilibrium and equilibrium surface tensions [59]. Several authors reported surface tension figures for lignocellulosic fast pyrolysis oils: Shaddix and

Fig. 3. 3D model of the new combustor (top) versus the original Garrett GTP 30–67 model (bottom). 178

Biomass and Bioenergy 115 (2018) 174–185

M. Buffi et al.

experimentally investigated for each selected fuel batch. The fuel mass flow, measured before MGT test runs by collecting a mass of fuel, was sprayed at constant pressure (pressure drop, Table 4) in a sealed vessel at atmospheric pressure. The actual mass fuel flow was finally assessed by measuring the fuel consumption reported in Fig. 7. The expected SMDs are shown in Table 4, according to eq. (3). The increased pressure drop at the nozzle generates a 4 ̴ 0% larger SMD for pyrolysis oil than for the reference diesel fuel, at equal thermal power. As suggested by Beran et al. [22] and Shaddix et al. [66], the SMD of pyrolysis oil should be reduced up ̴50% than that of diesel oil to ensure complete combustion. This target can be partially achieved only in the case of pure ethanol, while the pyrolysis oil-in-ethanol blend increases the estimated SMD above the one of the diesel fuel. For this reason, during transient phases, the presence of pilots flames allows to increase the overall heating rate in the combustion chamber so to maintain a stable combustion (as proposed by other authors [22,67]). Moreover, fuel pre-heating before atomization has a beneficial effect on the viscosity of FPBO, which is further reduced in favor of a further lowering of the SMD. The experimental campaign aimed at proving the stable combustion of PO/EtOH blends in the re-designed combustor.

Table 3 Experimental plan of the testing campaign. Run

Fuel

Fuel pre-heat temperature

Combustor configuration

Electric load

n 1 2 3 4 5

– Diesel Ethanol Diesel Ethanol PO/EtOH blend 20/80% (volume fraction) PO/EtOH blend 50/50% (volume fraction) Pyrolysis oil

(K) 293 293 293 293 353

– original original re-designed re-designed re-designed

(kW) 0–20 0–20 0–20 0–20 0–20

353

re-designed

0–20

353

re-designed

20

6

7

Fuel preheating during FPBO tests was controlled so to maintain fuel temperatures at approximately 353 K. Test conditions are summarized in the following Table 3. Main monitored operating parameters were: fuel consumption, fuel injection pressure, exhaust/combustion temperature, and exhaust emissions concentration. Each parameter is sampled when the main fuel line is working with a single fuel type.

3.3. Gas turbine test campaign

3. Results and discussions

3.3.1. Test of the new combustor The analysis of the new combustor performance was carried out during the test runs nr. 1–4 (as shown in Table 3). The effect of a different geometry was evaluated in terms of gas turbine performances and emissions. The impact on the combustion temperature was not investigated due to the lack of possible temperature measurements in the original turbine configuration. Despite the changed geometry, no significant effects on fuel consumption were observed in the range of ± 0.5 kg h−1, thus the study of the emissions can be seen as an indicator of the combustion quality. When compared to the original combustor (OC), the redesigned combustor reduced approximately by half the CO emissions of diesel fuel over the whole load range, as shown in Fig. 5. The CO reduction was mainly due to the increased residence time in the new combustion chamber. By distributing the air along a longer liner, CO species have longer time to complete oxidation to CO2. Moreover, the new geometry promotes mixing in the primary zone and leads the entrance of a larger amount of air, where the original condition of a rich mixture of air-fuel ratio is moved to stoichiometric conditions, leading to further reduction of CO [48]. This effect, relevant

3.1. Fuel analysis The measured viscosity and LHV for various ethanol-in-pyrolysis oil solution are reported in Fig. 4. The presence of ethanol-in-pyrolysis oil reduced the kinematic viscosity almost one order of magnitude at a volume fraction of 50/50%, and below 2.5 mm s−1 at 80/20%. The effect of increasing temperature can further reduce the viscosity, as studied in detail by other authors [65]. Moreover, ethanol also increases the heating value of the blends to 21 MJ kg−1 for the 50/50% PO/EtOH and about 22.5 MJ kg−1 for the 20/80% PO/EtOH (volume fractions). 3.2. Spray calculations In order to estimate the SMD of the selected nozzle, the thermal load in the combustion chamber at nominal condition (20 kW power) was assumed for calculations. The relationship between the fuel mass flow of the selected nozzle and the fuel pressure drop over the nozzle was

Fig. 4. Effects of ethanol-in-filtered fast pyrolysis bio-oil in terms of fuel viscosity and LHV, according to UNI EN ISO 3104 (measured at 313 K) and DIN 51900-2. 179

Biomass and Bioenergy 115 (2018) 174–185

M. Buffi et al.

Table 4 Properties of fuels and results from calculations of SMD for different fuels. Nozzle orifice diameter is 0.75 mm. Fuels

Surface tension (g s

Diesel Fuel Denatured Ethanol Fast Pyrolysis Bio-Oil PO/EtOH blend at 20/80% (volume fractions) PO/EtOH blend at 50/50% (volume fractions)

−2

)

27.4 22.5 34 25.6 29.3

Density

Viscosity

−3

(kg m 840 820 1192 920 1004

)

2

(mm s 3 1.1 37 1.8 4.1

−1

)

Pressure drop

Expected SMD

(MPa)

(μm)

0.8 1 1.15 1.3 1.35

63.8 48.3 109 66.4 77.2

efficiency of 26% with natural gas as fuel. Running the MGT with ethanol, it generates higher power than diesel. This fact can be mostly due to the larger production of water vapor during combustion, which has a positive impact on the MGT performance since this fuel contains the larger hydrogen fraction compared to other fuels (as shown in Table 1). Thus, a significantly increase in the exhaust gases volumetric flow passing through the centripetal turbine (at equal thermal load) is observed, and consequently higher electrical power is generated. Despite the test run with diesel fuel showed the most stable operating conditions, it resulted lower in terms of efficiency. This effect relates to the lower fuel flowrate at the same load, which corresponds to a decreased penetration of the spray due to the reduced nozzle pressure drop. The consequence is a more compact flame and higher irradiation heat losses, as confirmed also by the visual inspection of Martin and Boateng [65] in the combustion of residential-scale oil-fired boiler. The analysis of the equivalence ratio in the primary zone of the combustion chamber also confirmed this result. Fig. 9 shows the equivalence ratio at different load conditions, considering approximately the same air mass flow at the combustor inlet given the fixed rotational speed. Despite the engine can achieve a maximum electrical power out at 25 kW when fueled with diesel oil, the point at 20 kW load was assumed as reference condition for the new combustor design. Despite higher loads would favor the combustion of pyrolysis oil due to higher heat rates, fuel flows for PO/EtOH blends have been measured beyond the operational range of the selected nozzle. The original design of gas turbine combustor is based on a rich airfuel mixture in the primary zone, a criteria which has been revisited in the design of the new combustor. As shown in Fig. 9, at 20 kW power,

Fig. 5. CO emissions (normalized at 15% O2 in normal condition at 273.15 K) at different electrical load for the two different combustor layouts (Test runs nr. 1–4). OC: original combustor configuration.

in terms of emissions, does not affect the electrical efficiency of the unit, as confirmed by the analysis of the measured fuel consumptions. CO emissions of ethanol remained approximately in the same range, probably due to the high volatility and short oxidation time of the fuel, less affected by the combustor geometry. By increasing the combustor volume, NOX emissions were also reduced for both fuels with the load, as shown in Fig. 6. The distribution of air along the redesigned combustor is more uniform and consequently the peaks of combustion temperature should be reduced, favoring NOX emissions reduction. Summarizing, the larger combustor improved the quality of the combustion as regards exhaust emissions while maintaining the fuel consumptions at same levels of the original unit. 3.3.2. Gas turbine performances Test runs nr. 3–6 studied the effect of different fuel compositions on MGT performances. Fuel mass flow consumptions are reported c. As expected, fuel consumptions were significantly influenced by the energy content of each fuel: fuel mass flow increases with the pyrolysis oil content, despite the volumetric flow slightly decreases. Fuel consumption and heating value can be used to calculate the thermal power generated in combustion chamber. Considering the data measured in Fig. 7 and the LHVs reported in Table 1, the efficiency of the MGT was calculated. Fig. 8 shows the electrical efficiency of the system at different loads. As known, the Garrett GTP 30–67 shows lower electrical efficiency than stationary regenerated micro gas turbines such as a Capstone C30 LF, which has a nominal electrical

Fig. 6. NOX emissions (normalized at 15% O2 in normal condition at 273.15 K) at different electrical load for the two different combustors layout (Test runs nr. 1–4). OC: original combustor configuration. 180

Biomass and Bioenergy 115 (2018) 174–185

M. Buffi et al.

Fig. 7. Fuel consumption during selected tests (runs nr. 3–6) at different electrical load, using the new combustor.

Fig. 9. Equivalence ratio (in primary zone) of the selected fuels at different electrical load, new combustor (test runs nr. 3–6).

power), the combustion remained stable as long as ethanol pilot injectors were used, providing approximately 50% of the overall thermal power. Once the ethanol flow in the pilots was reduced, pressure and temperature instabilities were observed, and after few seconds the engine shuts down. Pure pyrolysis oil exhibits a higher viscosity than PO/ EtOH blends, thus the quality of the atomization was not sufficient and the generated thermal power from PO combustion was not enough to support the overall combustion process. The present achievement underlines the importance of having a large primary zone and a fine atomization for the combustion of pyrolysis oil, as confirmed by the numerical studies of Sallevelt et al. [68], but also how the re-design of the injectors is necessary of the aim is 100% FPBO. A method to improve the atomization could in fact be the use of air assisted nozzles [20]. Twin-fluid atomizers are less sensitive to changes in fuel viscosity compared to pressure-swirl nozzles. The suitability of

Fig. 8. Electrical efficiency of the selected fuels (test runs nr. 3–6) at different electrical load, using the new combustor.

diesel fuel reached stoichiometric conditions causing the highest measured temperature in the combustion chamber (Fig. 10). PO/EtOH blends require more time to complete burning since the oxidation process is affected by the higher average molecular weight of pyrolysis oil, and the presence of water [47]. As confirmed by the measurements shown in Fig. 10, combustion temperatures are relatively lower for pyrolysis oil blends, and the fuel evaporation requires twice as much evaporation heat compared to ethanol. This results in a decreased flame stability (the total lifetime of a pyrolysis oil droplet is approximately 60% longer than that of an ethanol droplet of the same size [68]) and consequently, lower temperatures are reached after the primary zone section, together with more distributed temperature profile along the combustor. A final test run (nr. 7, Table 3) was performed using pure pyrolysis oil as fuel. When the rig was switched to pure PO (at 20 kW MGT

Fig. 10. Measure of the combustion temperature at different electrical load. The measure was taken in the liner, in proximity of the secondary air inlet (test runs nr. 3–6). 181

Biomass and Bioenergy 115 (2018) 174–185

M. Buffi et al.

darker than the in other parts. This is due to higher temperature reached in that zone, which in the internal part (as shown in Fig. 13d) promoted the formation of deposits originated by the partial pyrolysis of the droplets. Moreover, the presence of deposits on the right side of the combustor can be observed, a phenomena probably caused by a not fully symmetrical air-fuel mixing. The main deposits are located in the inlet channel of the turbine volute (Fig. 14), where a bend is present and promotes the impact of the char particles on the wall. This inspection indicates the incomplete combustion of pure pyrolysis oil droplets, which have been entrained into the turbine inlet due to their dimensions (not sufficient to ensure a complete combustion). These results indicated that further components, such as fuel nozzle and probably a different air distribution along the combustor, are required to efficiently burn pure pyrolysis oil.

an air-blast nozzles in the combustion of viscous biofuels, i.e. woodderived biofuel, has already been demonstrated by Seljak et al. [69]. In their research, preheated liquefied wood with a viscosity of around 100 mm2 s−1 was burned in a jet engine combustor using air-assisted atomization. However, air-assisted atomizers can provoke higher spray penetration than pressure swirl atomizer at high fuel injection pressures [53], thus the shape of the combustion chamber has to be adapted as well. 3.3.3. Emissions The exhaust emissions of CO and NOX of pyrolysis oil/ethanol blends were evaluated, as shown in Fig. 11 and Fig. 12, and compared to the diesel and ethanol cases (already reported in previous Figs. 5 and 6). As shown in Fig. 11, by increasing the pyrolysis oil content in blend with ethanol, CO emissions increase. While at idle conditions (0 kW load) high CO emissions indicating poor combustion conditions and very different values for the different fuels are observed, by increasing the load they tend to converge to more similar concentrations. At 20 kW power, the CO concentration for PO/EtOH 50/50% is 1̴ 000 mg m−3 (850 ppm), compared to the 500 ppm for diesel. This results is in line with the measurements of Beran and Axelsson [22], that found that CO emissions for pure pyrolysis oil were about two fold higher than diesel fuel. The longer residence time of the droplets in the new combustor allows a complete oxidation, keeping the CO emissions low. However, larger droplet diameters are generated in the case of pyrolysis oil combustion (as seen in Table 4), and consequently, longer time is required to burn the whole droplet. The impact of the latter on emission provokes higher CO, which is also favored by the presence of water in pyrolysis oil that further reduce the overall combustion temperature (as observed as shown in Fig. 10). Despite the lower combustion temperature of pyrolysis oil blends, fuel-NOX formation is promoted by the higher content of pyrolysis oil in the fuel blend than diesel fuel, and their concentration increases at increasing load. For the same reason, under 5 kW power the combustion temperature is not sufficiently high to promote NOX formation, as shown in Fig. 12. However, PO blends and diesel show similar trend and are almost identical, as confirmed also by the experiments of Beran and Axelsson [22] in the case of a larger volume combustor in GT. The emissions measured during the final test run (nr. 7, Table 3: short test run with 100% FPBO) under unstable combustion conditions showed CO concentrations above 5000 mg Nm−3 (4000 ppm) and even higher NOX than previous tests, confirming the previous analysis of poor atomization.

4. Conclusions The combustion of high share of Fast Pyrolysis Bio-Oil in ethanol (1:1 vol ratio) as fuel was investigated in a test rig based on a modified APU-derived small scale gas turbine. Adaptations included a re-designed silo combustor and a modified injection line, which replaced the original components. The modified configuration of the test rig allowed stable MGT operation at different loads, with blends of pyrolysis oil/ ethanol at 20/80 and 50/50% (volume fractions). First tests with diesel oil and ethanol validated the design of the new combustor, showing lower CO emissions and slightly higher NOx emissions at full load compared to the original configuration. The effect of the larger volume of the combustor improved the quality of the combustion while maintaining similar performances. This fact can be attributed to the longer residence time (higher combustor volume) and a better distributed air inlet along the liner. The presence of pyrolysis oil in the blends significantly affects CO and NOX emissions compared to pure ethanol and diesel oil. By increasing the FPBO content in the blend, CO emissions increases as larger droplets are formed by the more viscous biofuel, while NOx emissions increase due to fuel-bound nitrogen. Regarding the electrical efficiency, tests with pure ethanol and PO/EtOH blends were higher than diesel. This fact was linked to the larger production of water vapor in combustion, and a well distributed air-fuel mixing in primary zone. In order to achieve a stable combustion with pure pyrolysis oil without the support of pilot injectors, further investigation on

3.3.4. Deposits The inspection of deposits of unburnt products on the hot path components of the MGTs/GTs is one of the most significant analysis in the turbomachinery applications. Pyrolyzed char, tars and other solids, generated by the FPBO combustion can form deposits over the turbine rotor and through the exhaust. A visual inspection after testing is of great importance to analyze the presence of deposits on the hot surfaces [24]. After each test run, combustor liner, fuel injectors and spark plug were disassembled and visually inspected to evaluate the formation of deposits. While tests nr. 1–6 (Table 3) showed only very slight visual differences, test nr. 7 with pure pyrolysis oil showed a significant presence of deposits. Fig. 13 shows the status of the combustor and its main internal components after the pure pyrolysis oil run (test nr. 7). The deposit on the spark plug is solid char formed during pyrolysis oil combustion (Fig. 13a). The nozzle is instead less covered by deposits, probably thanks to ethanol feeding at the end of the test, that is used to clean fuel piping alter pyrolysis oil fueling (Fig. 13b). The liner remained clean in the external surface, but it changed color along its length (Fig. 13c). Immediately after the primary zone, before the dilution air holes of the secondary zone, the color of the metal surface is

Fig. 11. CO emissions (normalized at 15% O2 in normal conditions at 273.15 K) of the selected fuels (test runs nr. 3–6) at different electrical load, new combustor. 182

Biomass and Bioenergy 115 (2018) 174–185

M. Buffi et al.

Fig. 14. Inlet section of turbine volute after test nr.7 with pure pyrolysis oil.

present study revealed that the use of 50% (volume fraction) of fast pyrolysis bio-oil (blended with ethanol) allowed a stable combustion at 20 kW power output in the revised configuration of the micro gas turbine.

Fig. 12. NOx emissions (normalized at 15% O2) in normal condition at 273.15 K) of the selected fuels (test runs nr. 3–6) at different electrical load, new combustor.

Acknowledgements

local heat transfer on the injection nozzle are required. A possible solution could be the modification of the atomization section, as well as the increase of temperature of the primary combustion air. Test towards 100% FPBO feeding in this MGT configuration showed unstable operation, and the analysis of carbon deposits on the hot parts of the combustor confirmed this statement. The most significant result of the

This research received funding from the CHPyro project (E!8096), Eurostar Joint and Innovation programme (2008 - 2013) with cofounding from European Union, in support of the experimental activities. Authors wish to thank BTG Bioliquids BV and BTG BTL, in particular Dr. Bert Van De Beld and Dr. Gerhard Muggen, for the supply of

Fig. 13. Status of combustor after test nr.7 with 100% FPBO: a) spark plug; b) main fuel nozzle; c) external view of the liner; d) internal view of the liner. 183

Biomass and Bioenergy 115 (2018) 174–185

M. Buffi et al.

the fast pyrolysis bio-oil used in this work.

03.056. [25] T. Seljak, T. Katrašnik, Designing the microturbine engine for waste-derived fuels, Waste Manag. 47 (2016), http://dx.doi.org/10.1016/j.wasman.2015.06.004. [26] T. Seljak, S. Rodman Oprešnik, T. Katrašnik, S.R. Opre, S.R. Oprešnik, S.R. Opre, Microturbine combustion and emission characterisation of waste polymer-derived fuels, Inside Energy 77 (2014) 226–234, http://dx.doi.org/10.1016/j.energy.2014. 07.020. [27] T. Seljak, M. Kunaver, T. Katrašnik, Emission evaluation of different types of liquefied wood, Strojniški Vestn. – J. Mech. Eng 60 (2014) 221–231, http://dx.doi. org/10.5545/sv-jme.2013.1242. [28] A. Cappelletti, A.M. Rizzo, D. Chiaramonti, F. Martelli, CFD redesign of micro gas turbine combustor for bio-fuels fueling, XXI Int. Symp. Air Breath. Engines, Busan, Korea, 2013, pp. 1199–1206, , http://dx.doi.org/10.13140/2.1.4096.1601. [29] A.M. Rizzo, Microgasturbine Fed with Liquid Biofuels: Conversion and Testing, Industrial Engineering Dept., University of Florence, Italy, 2011, http://dx.doi.org/ 10.13140/RG.2.2.33488.61445. [30] D. Chiaramonti, A.M. Rizzo, G. Riccio, A. Cappelletti, M. Prussi, F. Martelli, Adaptation of a micro gas turbine to biofuels and preliminary tests with diesel fuel, Third Int. Conf. Appl. Energy, 2011, pp. 2057–2066. [31] D. Chiaramonti, A.M. Rizzo, A. Spadi, M. Prussi, G. Riccio, F. Martelli, Exhaust emissions from liquid fuel micro gas turbines fed with vegetable oil and biodiesel, Proc. 19th Eur. Biomass Conf. Exhib, 2011, pp. 1672–1680, , http://dx.doi.org/10. 5071/19thEUBCE2011-VP2.5.17. [32] M. Prussi, D. Chiaramonti, G. Riccio, F. Martelli, L. Pari, Straight vegetable oil use in Micro-Gas Turbines: system adaptation and testing, Appl. Energy 89 (2012) 287–295, http://dx.doi.org/10.1016/j.apenergy.2011.07.031. [33] M. Prussi, D. Chiaramonti, L. Recchia, F. Martelli, F. Guidotti, L. Pari, Alternative feedstock for the biodiesel and energy production: the OVEST project, Inside Energy 58 (2013) 2–8, http://dx.doi.org/10.1016/j.energy.2013.02.058. [34] D. Meier, B. Van De Beld, A.V. Bridgwater, D.C. Elliott, A. Oasmaa, F. Preto, Stateof-the-art of fast pyrolysis in IEA bioenergy member countries, Renew. Sustain. Energy Rev. 20 (2013) 619–641, http://dx.doi.org/10.1016/j.rser.2012.11.061. [35] D. Wissmiller, Pyrolysis Oil Combustion Characteristics and Exhaust Emissions in a Swirl-stabilized Flame, Iowa State University, 2009, https://lib.dr.iastate.edu/etd/ 10889/. [36] M. Buffi, A. Cappelletti, A.M. Rizzo, F. Martelli, D. Chiaramonti, Modifies and experimental tests an a liquid fuel micro gas turbine fueled with pyrolysis oils and its blends, 25th Eur. Biomass Conf. Exhib., Stockholm, Sweden, 2017 http:// programme.eubce.com/abstract.php?idabs=14023&idses=546&idtopic=8. [37] G. López Juste, J.J. Salvá Monfort, Preliminary test on combustion of wood derived fast pyrolysis oils in a gas turbine combustor, Biomass Bioenergy 19 (2000) 119–128, http://dx.doi.org/10.1016/S0961-9534(00)00023-4. [38] Q. Lu, W.-Z.Z. Li, X.-F.F. Zhu, Overview of fuel properties of biomass fast pyrolysis oils, Energy Convers. Manag. 50 (2009) 1376–1383, http://dx.doi.org/10.1016/j. enconman.2009.01.001. [39] V.W. Greenhough, A.H. Lefebvre, Some applications of combustion theory to gas turbine development, Symp. Combust 6 (1957) 858–869, http://dx.doi.org/10. 1016/S0082-0784(57)80122-2. [40] H. Mongia, Gas turbine combustion design, technology and research - current status and future direction, 33rd Jt. Propuls. Conf. Exhib, American Institute of Aeronautics and Astronautics, Reston, Virgina (USA), 1997, , http://dx.doi.org/10. 2514/6.1997-3369. [41] D.B. Kulshreshtha, S.A. Channiwala, S.B. Dikshit, Numerical simulation as design optimization tool for gas turbine combustion chambers, Combust. Fuels Emiss. Parts A B, vol. 2, ASME, 2010, pp. 737–746, , http://dx.doi.org/10.1115/GT2010-22889. [42] D.B. Kulshreshtha, S.A. Channiwala, Kerosene fuelled tubular type combustion chamber gas turbine engine: design methodology and numerical investigations, Int. J. Turbo Jet Engines 27 (2010) 63–78, http://dx.doi.org/10.1515/TJJ.2010.27. 1.63. [43] P. Arunachalam, Selection of a suitable combustion system for a small gas turbine engine, Defence Sci. J. 38 (1988) 381–396 https://doi.org/10.14429/dsj.38.5871. [44] A. Cappelletti, F. Martelli, E. Bianchi, E. Trifoni, Numerical redesign of 100kw MGT combustor for 100% H2 fueling, Energy Procedia 45 (2014) 1412–1421, http://dx. doi.org/10.1016/j.egypro.2014.01.148. [45] A. Cappelletti, A.M.A.M. Rizzo, D. Chiaramonti, F. Martelli, CFD redesign of micro gas turbine combustor for bio-fuels fueling, XXI Int. Symp. Air Breath. Engines (2013) 1199–1206, http://dx.doi.org/10.13140/2.1.4096.1601. [46] A. Cappelletti, F. Martelli, Investigation of a pure hydrogen fueled gas turbine burner, Int. J. Hydrogen Energy (2017), http://dx.doi.org/10.1016/j.ijhydene. 2017.02.104. [47] R. Calabria, F. Chiariello, P. Massoli, Combustion fundamentals of pyrolysis oil based fuels, Exp. Therm. Fluid Sci. 31 (2007) 413–420, http://dx.doi.org/10.1016/ j.expthermflusci.2006.04.010. [48] A.H. Lefebvre, A.H. Lefebvre, Theoretical Aspects of Gas Turbine Combustion Performance, (1966). [49] A.H. Lefebvre, D.R. Ballal, D.W. Bahr, Gas turbine combustion, Angew. Chem. Int. Ed. 40 (2001) 9823, http://dx.doi.org/10.1002/1521-3773(20010316) 40:6<9823::AID-ANIE9823>3.3.CO;2-C. [50] C. Badarinath, Development of Aero Gas Turbine Annular Combustor: an Overview, Bangalore, India, (2008) http://www.combustioninstitute-indiansection.com/pdf/ DEVELOPMENT_OF_AERO_GAS_TURBINE_ANNULAR_COMBUSTOR.pdf. [51] T. Mongia, H. Reynolds, R. Coleman, E. Bruce, Combustor Design Criteria Validation. Volume II. Development Testing of Two Full-scale Annular Gas Turbine Combustors, Phoenix, Arizona (USA) (1979) http://www.dtic.mil/dtic/tr/fulltext/ u2/a067689.pdf. [52] P. Gobbato, M. Masi, A. Cappelletti, M. Antonello, Effect of the Reynolds number

Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx. doi.org/10.1016/j.biombioe.2018.04.020. References [1] P. Roy, G. Dias, Prospects for pyrolysis technologies in the bioenergy sector: a review, Renew. Sustain. Energy Rev. 77 (2017) 59–69, http://dx.doi.org/10.1016/j. rser.2017.03.136. [2] A.K. Hossain, P.A. Davies, Pyrolysis liquids and gases as alternative fuels in internal combustion engines - a review, Renew. Sustain. Energy Rev. 21 (2013) 165–189, http://dx.doi.org/10.1016/j.rser.2012.12.031. [3] B. Van De Beld, E. Holle, J. Florijn, The use of pyrolysis oil and pyrolysis oil derived fuels in diesel engines for CHP applications, Appl. Energy 102 (2013) 190–197, http://dx.doi.org/10.1016/j.apenergy.2012.05.047. [4] D. Meier, A. Oasmaa, G.V.C. Peacocke, Properties of fast pyrolysis liquids: status of test methods, in: A.V. Bridgwater, D.G.B. Boocock (Eds.), Dev. Thermochem. Biomass Convers. SE - 31, Springer Netherlands, 1997, pp. 391–408, , http://dx.doi. org/10.1007/978-94-009-1559-6_31. [5] A. Demirbas, Effect of initial moisture content on the yields of oily products from pyrolysis of biomass, J. Anal. Appl. Pyrolysis 71 (2004) 803–815, http://dx.doi.org/ 10.1016/j.jaap.2003.10.008. [6] J.P. Diebold, A review of the chemical and physical mechanisms of the storage stability of fast pyrolysis bio-oils, Natl. Renew. Energy Lab (2000) 59. [7] X. Hu, Y. Wang, D. Mourant, R. Gunawan, C. Lievens, W. Chaiwat, M. Gholizadeh, L. Wu, X. Li, C.Z. Li, Polymerization on heating up of bio-oil: a model compound study, AIChE J. 59 (2013) 888–900, http://dx.doi.org/10.1002/aic.13857. [8] D. Elliott, S.-J. Lee, T. Hart, Stabilization of Fast Pyrolysis Oil: Post Processing, Richand, Washington (USA) (2012) https://www.pnnl.gov/main/publications/ external/technical_reports/PNNL-21549.pdf. [9] B. Esteban, J.R. Riba, G. Baquero, A. Rius, R. Puig, Temperature dependence of density and viscosity of vegetable oils, Biomass Bioenergy 42 (2012) 164–171, http://dx.doi.org/10.1016/j.biombioe.2012.03.007. [10] A. Cavarzere, M. Morini, M. Pinelli, P.R. Spina, A. Vaccari, M. Venturini, Experimental analysis of a micro gas turbine fuelled with vegetable oils from energy crops, Energy Procedia 45 (2014) 91–100, http://dx.doi.org/10.1016/j.egypro. 2014.01.011. [11] F. Chiariello, C. Allouis, F. Reale, P. Massoli, Gaseous and particulate emissions of a micro gas turbine fuelled by straight vegetable oil-kerosene blends, Exp. Therm. Fluid Sci. 56 (2014) 16–22, http://dx.doi.org/10.1016/j.expthermflusci.2013.11. 013. [12] D. Chiaramonti, A.M. Rizzo, A. Spadi, M. Prussi, G. Riccio, F. Martelli, Exhaust emissions from liquid fuel micro gas turbine fed with diesel oil, biodiesel and vegetable oil, Appl. Energy 101 (2013) 349–356, http://dx.doi.org/10.1016/j. apenergy.2012.01.066. [13] A. Oasmaa, C. Peacocke, Properties and Fuel Use of Biomass-derived Fast Pyrolysis Liquids, VTT Technical Research Centre of Finland, Espoo, Finland, 2011 VTT Publication 731. [14] J. Lehto, A. Oasmaa, Y. Solantausta, M. Kytö, D. Chiaramonti, Review of fuel oil quality and combustion of fast pyrolysis bio-oils from lignocellulosic biomass, Appl. Energy 116 (2014) 178–190, http://dx.doi.org/10.1016/j.apenergy.2013.11.040. [15] A. Oasmaa, B. Van De Beld, P. Saari, D.C. Elliott, Y. Solantausta, Norms, standards, and legislation for fast pyrolysis bio-oils from lignocellulosic biomass, Energy Fuels 29 (2015) 2471–2484, http://dx.doi.org/10.1021/acs.energyfuels.5b00026. [16] B. van de Beld, E. Holle, J. Florijn, The use of a fast pyrolysis oil – ethanol blend in diesel engines for chp applications, Biomass Bioenergy 110 (2018) 114–122, http:// dx.doi.org/10.1016/j.biombioe.2018.01.023. [17] A. Shihadeh, S. Hochgreb, Impact of biomass pyrolysis oil process conditions on ignition delay in compression ignition engines, Energy Fuels 16 (2002) 552–561, http://dx.doi.org/10.1021/Ef010094d. [18] D. Chiaramonti, A. Oasmaa, Y. Solantausta, Power generation using fast pyrolysis liquids from biomass, Renew. Sustain. Energy Rev. 11 (2007) 1056–1086, http:// dx.doi.org/10.1016/j.rser.2005.07.008. [19] S. Czernik, A.V. Bridgwater, Overview of applications of biomass fast pyrolysis oil, Energy Fuels 18 (2004) 590–598, http://dx.doi.org/10.1021/ef034067u. [20] J.L.H.P. Sallevelt, J.E.P. Gudde, A.K. Pozarlik, G. Brem, The impact of spray quality on the combustion of a viscous biofuel in a micro gas turbine, Appl. Energy 132 (2014) 575–585, http://dx.doi.org/10.1016/j.apenergy.2014.07.030. [21] A. Pozarlik, A. Bijl, N. Van Alst, E. Bramer, G. Brem, Pyrolysis Oil Utilization in 50 KWe Gas Turbine, 18th IFRF Members’ Conf. – Flex. Clean Fuel Convers. to Ind (2015), pp. 1–10. [22] M. Beran, L.-U. Axelsson, Development and experimental investigation of a tubular combustor for pyrolysis oil burning, J. Eng. Gas Turbines Power 137 (2014) 31508, http://dx.doi.org/10.1115/1.4028450. [23] S.A. Rezzoug, R. Capart, Liquefaction of wood in two successive steps: solvolysis in ethylene-glycol and catalytic hydrotreatment, Appl. Energy 72 (2002) 631–644, http://dx.doi.org/10.1016/S0306-2619(02)00054-5. [24] T. Seljak, B. Sirok, T. Katrasnik, B. Širok, T. Katrašnik, Advanced fuels for gas turbines: fuel system corrosion, hot path deposit formation and emissions, Energy Convers. Manag. 125 (2016) 40–50, http://dx.doi.org/10.1016/j.enconman.2016.

184

Biomass and Bioenergy 115 (2018) 174–185

M. Buffi et al.

[53]

[54] [55] [56]

[57]

[58]

[59]

[60] [61]

and the basic design parameters on the isothermal flow field of low-swirl combustors, Exp. Therm. Fluid Sci. 84 (2017) 242–250, http://dx.doi.org/10.1016/j. expthermflusci.2017.02.001. S. Kulshreshtha, D.B. Dikshit, S.A. Channiwala, Experimental investigations of air assisted pressure swirl atomizer, Indian J. Sci. Technol 4 (2011) 126–130, http:// dx.doi.org/10.17485/ijst/2011/v4i2/29947. A.K. Jasuja, Atomization of crude and residual fuel oils, J. Eng. Power 101 (1979) 250–258. A.H. Lefebvre, Gas Turbine Combustion, second ed., Taylor and Francis Ltd, Ann Arbor, MI, 1998. M. Rashad, Y. Huang, Analysis and comparison of correlations established for prediction of SMD for pressure swirl atomizers, Proc. 2015 12th Int. Bhurban Conf. Appl. Sci. Technol. IBCAST 2015 (2015) 460–466, http://dx.doi.org/10.1109/ IBCAST.2015.7058543. J. Lehto, A. Oasmaa, Y. Solantausta, M. Kytö, D. Chiaramonti, Fuel Oil Quality and Combustion of Fast Pyrolysis Bio-oils, (2013) https://doi.org/10.1016/j.apenergy. 2013.11.040. Q. Lu, W.-Z. Li, X.-F. Zhu, Overview of fuel properties of biomass fast pyrolysis oils, Energy Convers. Manag. 50 (2009) 1376–1383, http://dx.doi.org/10.1016/j. enconman.2009.01.001. T. Tzanetakis, N. Ashgriz, D.F. James, M.J. Thomson, Liquid fuel properties of a hardwood-derived bio-oil fraction, Energy Fuels 22 (2008) 2725–2733, http://dx. doi.org/10.1021/ef7007425. C.R. Shaddix, D.R. Hardesty, Combustion Properties of Biomass Flash Pyrolysis Oils: Final Project Report, (1999), http://dx.doi.org/10.2172/5983. N.N. Bakhshi, J.D. Adjaye, Characteristics of a fast pyrolysis bio-fuel and its

[62]

[63]

[64]

[65]

[66] [67]

[68]

[69]

185

miscibility with oxygenated and conventional fuels, in: Second Biomass Conf. Am. Energy, Environ. Agric. Ind., Aug. 21–24, NREL, Golden CO 80401, CP-200-8098, pp., Portland, Oregon (USA), n.d. M. Garcìa-Pérez, A. Chaala, H. Pakdel, D. Kretschmer, D. Rodrigue, C. Roy, Multiphase structure of bio-oils, Energy Fuels 20 (2006) 364–375, http://dx.doi. org/10.1021/ef050248f. F. Wang, J. Wu, Z. Liu, Surface tensions of mixtures of diesel oil or gasoline and dimethoxymethane, dimethyl carbonate, or ethanol, Energy Fuels 20 (2006) 2471–2474, http://dx.doi.org/10.1021/ef060231c. G. Vazquez, E. Alvarez, J.M. Navaza, Surface tension of alcohol + water from 20 to 50 °C, J. Chem. Eng. Data 40 (1995) 611–614, http://dx.doi.org/10.1021/ je00019a016. J.A. Martin, A.A. Boateng, Combustion performance of pyrolysis oil/ethanol blends in a residential-scale oil-fired boiler, Fuel 133 (2014) 34–44, http://dx.doi.org/10. 1016/j.fuel.2014.05.005. C.R. Shaddix, D.R. Hardesty, Combustion Properties of Biomass Flash Pyrolysis Oils: Final Project Report - Sandia, (1999), http://dx.doi.org/10.2172/5983. C. Branca, C. Di Blasi, R. Elefante, Devolatilization and heterogeneous combustion of wood fast pyrolysis oils, Ind. Eng. Chem. Res. 44 (2005) 799–810, http://dx.doi. org/10.1021/ie049419e. J.L.H.P. Sallevelt, A.K. Pozarlik, G. Brem, Numerical study of pyrolysis oil combustion in an industrial gas turbine, Energy Convers. Manag. 127 (2016) 504–514, http://dx.doi.org/10.1016/j.enconman.2016.09.029. T. Seljak, S. Rodman Oprešnik, T. Katrašnik, Microturbine combustion and emission characterisation of waste polymer-derived fuels, Inside Energy 77 (2014) 226–234, http://dx.doi.org/10.1016/j.energy.2014.07.020.