Prediction of product distribution in fine biomass pyrolysis in fluidized beds based on proximate analysis

Bioresource Technology 175 (2015) 275–283

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Prediction of product distribution in fine biomass pyrolysis in fluidized beds based on proximate analysis Sung Won Kim ⇑ Global Technology, SK Innovation, 325 Exporo, Yuseong-gu, Daejeon 305-712, Republic of Korea

h i g h l i g h t s  A model was developed to predict product distribution from biomass pyrolysis.  The model is based on proximate analysis and hydrodynamics in fluidized beds.  Relationships between product yields and fluidization condition were derived.  Gas and char yields are a strong function of temperature and vapor residence time.

a r t i c l e

i n f o

Article history: Received 4 September 2014 Received in revised form 18 October 2014 Accepted 20 October 2014 Available online 28 October 2014 Keywords: Biomass pyrolysis Fluidized bed Prediction Product distribution Proximate analysis

a b s t r a c t A predictive model was satisfactorily developed to describe the general trends of product distribution in fluidized beds of lignocellulosic biomass pyrolysis. The model was made of mass balance based on proximate analysis and an empirical relationship with operating parameters including fluidization hydrodynamics. The empirical relationships between product yields and fluidization conditions in fluidized bed pyrolyzers were derived from the data of this study and literature. The gas and char yields showed strong functions of temperature and vapor residence time in the pyrolyzer. The yields showed a good correlation with fluidization variables related with hydrodynamics and bed mixing. The predicted product yields based on the model well accorded well with the experimental data. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Pyrolysis is one of the most promising technologies of biomass utilization, which converts the biomass to bio-oil, char and gasses depending on the pyrolysis conditions. The pyrolysis is a thermal degradation of materials in the absence of oxygen. The pyrolysis can be a promising option for lignocellulosic biomass conversion because bio oils derived from biomass pyrolysis could act as feedstocks for producing hydrocarbons that may be readily integrated into existing petroleum refineries or future bio-refineries (Kim et al., 2013a). In past decades, significant progress has been made in developing various pyrolysis technologies, and successfully testing them on laboratory, pilot and industrial scales. Recently, extensive research has been focused on fast pyrolysis using the fluidized bed process because of its high heat transfer and the short residence time for vapor to obtain a high yield of bio oil. Many studies ⇑ Tel.: +82 42 609 8314; fax: +82 42 609 8804. E-mail address: [email protected] http://dx.doi.org/10.1016/j.biortech.2014.10.107 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.

have been carried out to determine the parametric influence of operating conditions on the yield of pyrolysis products of various biomasses for the design and optimum operating conditions of fluidized bed pyrolyzers. They found that the product distribution is affected by operating parameters as well as biomass type (Garcia-Perez et al., 2008; Lee et al., 2008; Heo et al., 2010). The description of the pyrolysis process including product distribution is particularly challenging because it involves a great deal of physical and chemical transformations and produces a large number of product species. In spite of the technological progress and large amount of experimental data reported in the literature, there is still considerable debate over the reaction mechanism controlling the distribution of pyrolysis products (Garcia-Perez et al., 2008). As a result, existing models aiming to predict the rates or yields of the released pyrolytic volatiles are still supported by empirical data (Neves et al., 2011). Neves et al. (2011) suggested an empirical model to approximate the pyrolysis behavior of most biomass, where the product distribution is a function of temperature. However, the product distribution is affected by reactor type and geometry in addition to the chemical structure of the biomass

276

S.W. Kim / Bioresource Technology 175 (2015) 275–283

in experiments of this study and literature, which address pyrolysis of fine biomass less than 2 mm to minimize the impact of internal heat transfer inside feedstock. A comparison between model results and experimental data was done to show its predictive capability for pyrolysis of various biomasses in fluidized beds.

Bio oil product Moisture

Biomass

Organics

Condensable liquid

Gas

Gas product Char product

Ash

Char

2. Methods

Fig. 1. Biomass pyrolysis concept for the predictive model in this study.

2.1. Methodology and model structure

(Cao et al., 2004). Especially, the pyrolytic oil yield in a fluidized bed is affected by the fluidizing conditions, such as gas velocity, the physical properties of bed material and static bed height or vapor residence time (Cui and Grace, 2007). Therefore, biomass thermochemical conversion development requires methodologies for prediction of the product distribution, whereby an appropriate balance between empiricism and fundamentals related with characteristics of biomass and flow dynamics should be considered to describe the complex range of interrelated phenomena (Ioannidou et al., 2011). When considering biomass thermal conversion, proximate analysis is one of the most important characterization methods. This consists of determining the moisture, ash, volatile matter and fixed carbon contents of the raw biomass. These values are essential as certain moisture, volatile matter and fixed carbon affect both thermal behavior and plant design (Garcia et al., 2013). Recently, a thermogravimetric model for pyrolysis product distribution in a captive batch reactor was proposed based on the proximate analysis of three agricultural residues (Ioannidou et al., 2011). The model enabled the possibility to evaluate a mass balance during the thermal treatment of the residue. However, the model did not take into account the possible activity of a second reaction (Ioannidou et al., 2011) which can possibly come from a variation in operating conditions. Also, it is still far from being applied in a fluidized bed pyrolyzer due to different operating principles between reactor types. A predictive model was developed to describe the general trends of product distribution in fluidized bed for biomass pyrolysis. The model was made of mass balance based on proximate analysis and an empirical relationship with operating parameters. The empirical relationships between product yields and fluidization conditions in fluidized bed reactors were derived from the data

The biomass pyrolysis concept for this study is based on the fact that the precise mechanism is not clearly known because of extremely complex reactions that take place during the biomass pyrolysis process (Ioannidou et al., 2011). A simplified two-step concept of the pyrolysis was introduced to consider possible secondreaction which is affected by operating condition as Fig. 1. The concept assumes that biomass is composed of three components, i.e., organics (volatile and fixed carbon), moisture and ash based on proximate analysis. It is assumed that all organics and water are devolatilized from biomass and ash is left as part of solid residue as a first step in the pyrolysis process. The organics are partly converted into gas and char, depending on operating conditions in the second step of the pyrolysis process. Finally, unconverted organics, pyrolytic water and moisture are produced as bio oil or a liquid product from the fluidized bed reactor (Neves et al., 2011). The whole solid fraction of ash and pyrolytic char is produced as char, assuming that all the ashes in the feed material remain in the char. Taking into account the simplified concept and the proximate analysis results of the input material (moisture: M, volatile matter: VM, fixed carbon: FC and ash: A), it is possible to calculate mass balance as Eqs. (1)–(3).

Feedstock ¼ M þ OR þ A

ð1Þ

where OR is organics (volatiles and fixed carbon) as Eq. (2).

OR ¼ VM þ FC

ð2Þ

Product ¼ bio oil þ gas þ char

ð3Þ

Assuming that each feedstock for pyrolysis in this model has the same content of moisture as the sample used in proximate analysis, the mass of bio oil is calculated as the sum of the moisture content and mass of unconverted organics obtained from char and gas yields as Eq. (4).

Table 1 Proximate analysis of a set of biomasses used in this study. Raw material

Jatropha seedshell cake Mallee Eucalyptus loxophleba (woody fraction) Cassava stalk Cassava rhizome Jatropha seedshell waste Quercus acutissima Miscanthus sinensis Palm kernel shell Jatropha seedshell cake Miscanthus sinensis var. purpurascens Sugar cane bagasse Corn cob Corn stover Eucalyptus grandis Cassava rhizome Cassava stalk Japanese larch Radiata pine sawdust Oriental white oak a

Organics (sum of volatile and fixed carbon).

Proximate analysis [wt%]

Reference

Moisture

Volatile

Fixed carbon

Ash

2.7 0.0 0.0 0.0 2.7 8.3 10.2 5.9 2.7 8.0 6.8 4.6 8.5 6.2 1.8 2.4 8.8 7.6 10.4

79.8 81.9 79.9 77.7 79.8 73.9 71.2 71.3 79.8 74.9 76.9 79.9 76.7 80.9 81.5 81.2 91.0a 92.2a 87.6a

14.1 17.6 14.1 18.2 14.1 16.7 15.2 17.8 14.1 15.7 10.8 13.7 8.2 12.4 13.1 11.2

3.4 0.5 6.0 4.1 3.4 1.0 3.4 5.0 3.4 1.4 5.3 1.6 6.1 0.5 3.5 5.1 0.2 0.2 2.0

This study Garcia-Perez et al. (2008) Pattiya (2011) Kim et al. (2013a) Lee et al. (2008) Bok et al. (2013) Kim et al. (2013b) Heo et al. (2010) Carrier et al. (2013)

Pattiya and Suttibak (2012) Park et al. (2008a) Park et al. (2008b) Park et al. (2009)

Table 2 Experimental details used in this study. References

Material

This study

418, 764

1.00

Garcia-Perez et al. (2008) Pattiya (2011) Kim et al. (2013a) Lee et al. (2008)

303a 428a 750, 1700 1113a

1.00 0.105 1.00 0.30

Bok et al. (2013)

1500

2.0, 0.20

Silicon carbide Sand Sand Sand Fused alumina Sand

Kim et al. (2013b) Heo et al. (2010) Carrier et al. (2013)

763a 700 408, 573, 868, 1129 125, 338, 406, 513

0.94 0.15 0.85

SiC Zirconia Alumina Sand

0.40

Silica sand

300, 700, 1000 300, 700, 1000

0.15 0.15, 0.18 0.15

Emery Emery Emery

c

300, 700, 1000

Particle diameter [lm]

Fluidizing gas Particle density [kg/ m3]

Gas

Gas flow rate [liter/min]

Temperature [°C]

Reactor

Symbols in figure

Diameter [m]

Height [m]

190

3210

Nitrogen

33, 41

390–512

0.102

0.97



376 428 278 40

2600b 2600b 2600b 3930

Nitrogen Nitrogen Nitrogen Nitrogen

38–53 5.9–6.6 30 2.5–8.2

350–580 437–537 500 300–550

0.102 0.0381 0.10 0.076

0.518 0.30 0.85 0.80

j N d

200

2600b

Nitrogen

30, 70

400–550

3210 3800 2500 2670

Nitrogen Nitrogen Nitrogen

33 5 40

392–511 350–550 500

0.17c 0.1 0.102 0.08 0.075

0.30c 0.3 0.97 0.30 0.68

}

b

Nitrogen

7

393–550

0.05

0.45

h

40 40

3930 3930

Nitrogen Nitrogen

2–4 2–4

400–550 400–550

0.08 0.08

0.30 0.30

4 s

40

3930

Nitrogen

2–4

400–550

0.08

0.30

⁄

190 186.5 40 320 338

2600

S.W. Kim / Bioresource Technology 175 (2015) 275–283

Feed rate [kg/hr]

Park et al. (2009) b

Fluidized bed material

Average size [lm]

Pattiya and Suttibak (2012) Park et al. (2008a) Park et al. (2008b)

a

Feedstock

Calculated from range of feedstock size. Based on Kunii and Levenspiel (1991). Rectangular type.

277

278

S.W. Kim / Bioresource Technology 175 (2015) 275–283

Table 3 Operating parameters of fluidized bed and their correlations used in this study. Name of parameter, symbol

Equation or correlation

Reference

kg/m s

Eb ¼ 0:1qp ð1  emf ÞðU g  U mf Þ

Pemberton and Davidson (1986)

Gas flux, Eg

kg/m2s

where qp is particle density, emf is voidage at minimum fluidizing condition and Ug is gas velocity Eg ¼ qg U g

Superficial gas velocity at minimum fluidizing condition, Umf

m/s

Bed entrainment flux, Eb

Unit 2

where qg is gas density i  l h 1=2 g  33:7 for dp P 100 lm ð33:72 þ 0:0408ArÞ dp q

U mf ¼

g

1:8

U mf ¼

ðqp qg Þ0:934 g 0:934 dp 1110lg0:87 qg0:066

for dp < 100 lm

Wen and Yu (1966) Baeyens and Geldart (1974)

where lg is gas viscosity, Ar is Archimedes number, g is acceleration of gravity and dp is particle size of bed material U

Bubble fraction in fluidized bed, db

–

db ¼ U gb

Bubble rising velocity, Ub

m/s

U b ¼ ðU g  U mf Þ þ U br

Single bubble’s rising velocity, Ubr

m/s

U br ¼ 0:711ðgdb Þ

Average bubble diameter, db

m

db ¼ 0:138hb ðug  umf Þ0:42 expð2:5  105 ðug  umf Þ2  103 ðug  umf ÞÞ where hb is bed height

Nusselt number for gas-particle heat transfer, Nu

–

1=2

0:8

Nu ¼

Kunii and Levenspiel (1991) Kunii and Levenspiel (1991) Kunii and Levenspiel (1991) Cai et al. (1994)

h p dp kg

where hp is heat transfer coefficient between gas and particle and kg is thermal conductivity of gas

Bio oil ¼ M þ ORð1  Y RC  Y RG Þ

ð4Þ

In Eq. (4), YRC indicates the relative char yield, which is the char yield converted from organics, and YRG indicates the relative gas yield, which is the gas yield converted from organics based on the simplified concept in Fig. 1. YRC and YRG can be expressed as Eqs. (5) and (6).

Y RC ¼ ðChar  AÞ=OR

ð5Þ

Y RG ¼ Gas=OR

ð6Þ

The char and gas yields in pyrolysis products can be obtained from Eqs. (5) and (6), if YRC and YRG are known. 2.2. Experimental method A sample of Jatropha (Jatropha curcas L.) seedshell cake was acquired from an oil extraction plant in Indonesia. Proximate analysis was carried out with a Thermostep (ELTRA) analyzer according to the ASTM 5142 standard test method. The proximate analysis showed volatiles (mainly organic) of 79.80 wt% as shown in Table 1. The Jatropha seedshell cake sample was ground and sieved to an average size of 418 and 764 lm using an electric mixer and standard sieves. Silicon carbide (dp = 190 lm, qs = 3210 kg/m3) particles were used as bed material in this study. The pyrolysis system consisted of a mass flow controller (MFC), main column, screw feeder, cyclone, condensers, and accumulative flowmeter. The system was previously described in detail (Kim et al., 2014). The flow rate of nitrogen (99.9%) for fluidization was controlled by the MFC and the volume of the product gas was measured by the accumulative flowmeter (G6R, GTEC Industry). Before entering the stainless steel fluidized bed reactor (0.102 m id. and 0.97 m high), the fluidizing N2 gas was preheated in the air plenum to 450–600 °C. To prevent condensation of the pyrolysis vapor, the top of the reactor and the pipe connecting the reactor to the first condenser were maintained at 400 °C. The refrigerant was circulated in the shell of the condensers by a chiller system. The gas was sampled by a gas sampler to analyze its composition. Pyrolysis tests were conducted at 390–512 °C with N2 flow rates of 33 and 41 L/min at 25 °C (superficial gas velocities of 0.15–0.21 m/s depending on temperature), corresponding to ca. 7.5–10.4 Umf and a bubbling fluidization regime. A static bed height

of 0.2 m was maintained in the test. Biomass particles were fed at 1.0 kg/h. After completion of the feeding of biomass in the feeder, heating of the reactor was discontinued. Thereafter, the oil sampling pots, condensers, oil filter and the content of the reactor were weighted. The total yield of liquid products was calculated as the sum of the increase of mass in the sampling pots, condensers and filter. The mass of char was determined by weight, and the weight of the bed materials was subtracted from the total mass of solids for calculation of the char mass. The gas yield was calculated by overall mass balance. 2.3. Collection of literature data A number of studies on lignocellulosic biomass pyrolysis in fluidized beds have been reported in the literature. However, many works focused on biomass conversion in certain operating conditions in fluidized bed reactors, and essential information to evaluate fluidization quality such as the details of the bed material were not fully included in the report. Data from a set of investigations with detailed information, including 18 different biomass samples and specific information about the operating conditions of fluidized bed reactor, were collected as shown in Tables 1 and 2. The collected data are limited to biomass smaller in size than 2 mm to minimize the effect of internal heat transfer inside feed particles on the product distribution from pyrolysis (Kersten et al., 2005; Isahak et al., 2012). The experimental data on the pyrolytic product distribution were analyzed and correlated with the information concerning the experimental rigs, operating parameters and feed properties based on previous findings in the literature. 3. Results and discussion 3.1. Relations between product yields and operating parameters The major operating parameters of a pyrolysis process influencing the product distribution include pyrolysis temperature, sweeping gas flow rate, residence time of vapor, biomass heating rate, mineral matter and size of biomass particles (Akhtar and Amin, 2012). In addition, fluidization variables, affecting hydrodynamics and particle mixing, are the most important parameter from the viewpoint of a fluidized bed pyrolyzer (Cui and Grace, 2007). The

279

S.W. Kim / Bioresource Technology 175 (2015) 275–283 1

Relave gas yield, YRG [-]

Relave gas yield, YRG [-]

1

0.1

0.1

(a)

(b)

0.01

0.01

0.1

1

10

100

100

dBM/dp [-] 1

1

Relave gas yield, YRG [-]

Relave gas yield, YRG [-]

1000

Bed entrainment flux/Gas flux [-]

0.1

0.1

(c)

(d) 0.01

0.01 0.1

1

1

10

Ug/Umf [-]

Relave Temperature [-]

Relave gas yield, YRG [-]

1

0.1

(e)

0.01 0.1

1

10

Relave residence me of vapor [-] Fig. 2. Effect of operating parameters on relative gas yield: (a) size ratio of biomass to bed material, (b) flux ratio of bed entrainment to gas, (c) relative temperature, (d) fluidization number and (e) relative residence time of vapor (for symbol identification, see Table 2).

mixing of binary mixtures such as biomass and bed materials is mainly affected by the particle size ratio and gas velocity (Wu and Baeyens, 1998). Also, the heat transfer and hydrodynamics in the fluidized bed are governed by bubble behavior (Kim et al., 2003, 2013b). To assess the relationships between the relative gas and char yields and operating parameters for Eqs. (4)–(6) in the model, the collected data from this study and literature (Table 2) were plotted against various operating parameters as summarized in

Table 3. The dimensional parameters, temperature and vapor residence time were modified into dimensionless terms considering reactor design and scale-up because the yields are dimensionless. A concept of relative temperature (TR) was introduced as in Eq. (7) to consider that the organics are devolatilized over 100 °C (373.15 K) at atmospheric conditions.

TR ¼

T py  373:15 373:15

ð7Þ

280

S.W. Kim / Bioresource Technology 175 (2015) 275–283

1

Relave char yield, YRC [-]

Relave char yield, YRC [-]

1

0.1

0.1

(b)

(a) 0.01 0.1

1.0

0.01 0.3

Bubble fracon, δ [-]

3

Relave Temperature [-]

Relave char yield, YRC [-]

Relave char yield, YRC [-]

1

0.1

(c) 0.01

0.1

(d) 0.01

0.1

1

10

0.01

Relave residence me of vapor [-]

0.1

Ash content [-]

Fig. 3. Effect of operating parameters on relative char yield: (a) bubble fraction in bed, (b) relative temperature, (c) relative residence time of vapor and (d) ash content (for symbol identification, see Table 2).

where Tpy is the pyrolysis temperature in Kelvin (K). Also, the vapor residence time (tvapor) was modified into a relative residence time of vapor (trv), where tvapor is divided by 2 s as in Eq. (8).

t rv ¼

t vapor 2

ð8Þ

Two seconds was chosen as a normalization factor because it is an optimal vapor residence time to prevent the secondary cracking of volatiles in pyrolysis (Carlson et al., 2008; Kim et al., 2013b), indicating the gas or char yield changes possibly before and after the 2 s of vapor residence time. Finally, several parameters to predict the relative gas and char yields were determined by comparison of correlation coefficients from the correlation diagram on a trial-and-error strategy based on previous findings about the effect of operating parameters in the literature (Neves et al., 2011). The results with comparatively good correlation were plotted against each other in the correlation diagrams as shown in Fig. 2 for the relative gas yield and in Fig. 3 for the relative char yield. The relative temperature and the relative residence time of vapor commonly showed a strong correlation with the gas and char yields. The pyrolysis temperature is a major parameter in determining the yields of char, gas and liquid. At the lower temperature (<300 °C) char is the main product. At the higher temperature, the production of light gases is favored (Neves et al., 2011). The vapor residence time is also an important factor.

Longer residence time allows for secondary reactions to occur which form either additional gases or char, both of which reduce the bio oil yield (Carrier et al., 2013). Additionally the relative gas yields were well correlated with the size ratio of biomass feedstock to bed material (dBM/dp), the flux ratio of bed entrainment to gas (Eb/Eg) and fluidizing number (Ug/Umf). The bed mixing and heat transfer are affected by the particle size ratio in the bed and gas velocity. A high size ratio of biomass feedstock to bed material could cause a segregation of binary mixture in the bed. However, smaller bed particles attribute better mixing of the bed and higher heat transfer to biomass or primary pyrolysis products, which possibly induce further reaction of primary volatile into gas product. Higher gas velocity could be positive for bed mixing, but the excessive increase of gas velocity leads to the entrainment of bed materials to dilute phase above the bed and further cracking through the contact of hot particles and volatiles, contributing to an increase in the gas yield (Kim et al., 2013b). The relative char yields showed a strong correlation with the bubble fraction in the fluidized bed. Poor mixing in the fluidized bed leads to high char yields due to incomplete pyrolysis (Kim et al., 2013b). The poor mixing could be overcome by vigorous bubble behavior from increasing the gas velocity. Additionally, an increase in bubble formation increases the pyrolysis reactivity of biomass owing to an increase in the local heat transfer of bed material to biomass (Kim et al., 2003). The relative char yields

281

S.W. Kim / Bioresource Technology 175 (2015) 275–283 1

1

+ 25%

0.9

0.8

Measured char yield [wt %]

0.8

Measured gas yield [wt %]

+ 25%

0.9

0.7 0.6

- 25%

0.5 0.4 0.3

0.7 0.6

- 25%

0.5 0.4 0.3 0.2

0.2

(a)

0.1

(b)

0.1 0

0 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0

1

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Calculated char yield [wt %]

Calculated gas yield [wt %]

1

+ 20%

0.9

Measured bio oil yield [wt %]

0.8 0.7

- 20%

0.6 0.5 0.4 0.3 0.2

(c)

0.1 0 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Calculated bio oil yield [wt %] Fig. 4. Comparison between calculated and measured (a) gas, (b) char and (c) bio oil yields (for symbol identification, see Table 2).

showed comparatively weak correlation with ash content even though the ash content was reported as an important parameter in secondary reaction to char formation (Akhtar and Amin, 2012). Biomass contains trace amounts of inorganic compounds such as metals and extraneous organic materials that appear in pyrolysis ash. High amounts of ash in biomass lead to large char formation since the inorganic elements in the ash are known to catalyze char forming reaction during both primary and secondary pyrolysis (Akhtar and Amin, 2012; Kim et al., 2014). However, a clear interrelationship between the char yields and total ash contents of biomass was not observed in this study potentially due to the influence of the varying concentration of the related inorganic elements in ashes depending on biomass sources. Interestingly, the Nusselt number for the heat transfer rate in the bed showed weak correlations with the relative gas and char yields as shown in Fig. S1. The heat transfer rate is an important factor in the product distribution: rapid heating to a reaction temperature that is too low or too high will adversely affect the product yield. However, heat transfer behavior in the fluidized bed of small particles differs from that of larger ones in a qualitative sense due to the different predominant mechanism (Kim et al., 2003). The different mechanism with bed particle size attributes

the weak correlations between product yields and the Nusselt number. 3.2. Prediction of product yields From the parameter analysis, it was concluded that the relative gas yield is a strong function of five operating parameters: TR, trv, dBM/dp, Eb/Eg and Ug/Umf. The relative gas yield values (YRG) in the present and previous studies (Table 2) have been correlated with the five dimensionless parameters to apply to the model as, 0:05 Y RG ¼ 0:00029T 0:99 R t rv

 0:11  1:14  0:09 dBM Eb Ug dp Eg U mf

ð9Þ

with a correlation coefficient of 0.87 and a standard error estimate of 0.105. The relative char yield values (YRC) have been also correlated with the three parameters: TR, trv and db. Additionally the ash content was included in developing the correlation to reflect an influence of ash for the char formation as Eq. (10), even though it has a comparatively lower interrelationship than others. 0:05 0:05 Y RC ¼ 0:23T 1:14 t 0:07 A r v db R

ð10Þ

282

S.W. Kim / Bioresource Technology 175 (2015) 275–283

calculated values as shown in Fig. 4c. The calculated bio oil yields predict the experimental data within ±20%. A comparison of model predictions for temperature effect on the product distribution in this and previous studies (GarciaPerez et al., 2008; Pattiya and Suttibak, 2012) is summarized in Fig. 5. The model provides good predictions over the range of data that accounts for temperature effect and different biomass sources. In particular, the model prediction of char yields is better than others. The prediction results are encouraging in that the model can predict the optimum temperature range to maximize the bio oil yield at given operating conditions in a fluidized bed reactor as shown in Fig. 5a.

80 70

(a)

Bio-oil yield [wt %]

60 50 40 30

This study

20

Garcia-Perez et al. (2008)

10

Paya and Subak (2012)

3.3. Discussion on the developed model

0 300

400

500

600

700

Temperature [oC] 60

This study

Gas yield [wt %]

50

(b)

Garcia-Perez et al. (2008) Paya and Subak (2012)

40 30 20 10 0 300

400

500

600

700

Temperature [oC] 60

4. Conclusions This study

(c)

50

A predictive model was satisfactorily developed to describe the general trends of the product distribution in a fluidized bed for lignocellulosic biomass pyrolysis. The model was made of mass balance based on proximate analysis and an empirical relationship with operating parameters including fluidization hydrodynamics. The gas and char yields showed a strong function of temperature and vapor residence time in the pyrolyzer. The yields showed good correlation with fluidization variables related with hydrodynamics and bed mixing. The predicted product yields based on the model well accorded to the experimental data.

Garcia-Perez et al. (2008) Char yield [wt %]

Comparison between the developed model results and measurement data on the product distribution from pyrolysis of various biomasses showed good agreement. However, it is clear that the correlation model in this study is an empirical model and it has limited application scope because the correlations used in the model were derived from a particular set of experimental data and a restricted range of validity. Nevertheless, the empirical model is useful because it can be readily used in a comprehensive reactor model. The model can be used as a complementary tool to further analyze and check the consistency of the collected data (Neves et al., 2011). Further research over a wide range of operating parameters in units on different scales is needed to provide a general understanding and prediction of the product distribution from the pyrolysis of biomass in the fluidized bed from the viewpoint of the design and optimal operation of commercial pyrolysis reactor. In addition, elemental composition of biomass materials and products in the model should be considered for improvement in prediction capability of the model such as product properties.

40

Paya and Subak (2012) 30 20 10 0 300

400

500

600

700

Temperature [oC] Fig. 5. Effect of temperature on product distributions and their prediction based on this study: (a) bio oil, (b) gas and (c) char.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2014.10. 107. References

with a correlation coefficient of 0.78 and a standard error estimate of 0.071. The range of variables in Eqs. (9) and (10) covers 0.54 6 TR 6 1.29, 0.26 6 trv 6 11.86, 0.37 6 dBM/dp 6 32.50, 159.53 6 Eb/Eg 6 719.16, 1.78 6 Ug/Umf 6 38.63, 0.09 6 db 6 0.53 and 0.002 6 A 6 0.061. The gas and char yields calculated by Eqs. (5) and (6) predict the experimental data of the present and previous studies within ±25% as shown in Fig. 4(a) and (b). Finally, Eq. (4) of the model is applied to the calculation of the bio oil yields from the pyrolysis of various lignocellulosic biomasses by Eqs. (9) and (10) in conjunction with Eq. (3). The bio oil yields measured in the present and previous studies (Table 2) are compared with the

Akhtar, J., Amin, N.S., 2012. A review on operating parameters for optimum liquid oil yield in biomass pyrolysis. Renew. Sust. Energy Rev. 16, 5101–5109. Baeyens, J., Geldart, D., 1974. An investigation into slugging fluidized beds. Chem. Eng. Sci. 29, 255–265. Bok, J.P., Choi, H.S., Choi, J.W., Choi, Y.S., 2013. Fast pyrolysis of Miscanthus sinensis in fluidized bed reactors: characteristics of product yields and biocrude oil quality. Energy 60, 44–52. Cai, P., Schiavetti, M., Michele, G.D., Grazzini, G.C., 1994. Quantitative estimation of bubble size in PFBC. Powder Technol. 80, 99–109. Cao, Q., Xie, K.C., Bao, W.R., Shen, S.G., 2004. Pyrolytic behavior of waste corn cob. Bioresour. Technol. 94, 83–89. Carlson, T.R., Vispute, T.P., Huber, G.W., 2008. Green gasoline by catalytic fast pyrolysis of solid biomass derived compounds. ChemSusChem 1, 397–400.

S.W. Kim / Bioresource Technology 175 (2015) 275–283 Carrier, M., Joubert, J., Danje, S., Hugo, T., Gorgens, J., Knoetze, J., 2013. Impact of the lignocellulosic material on fast pyrolysis yields and product quality. Bioresour. Technol. 150, 129–138. Cui, H., Grace, J.R., 2007. Fluidization of biomass particles: a review of experimental multiphase flow aspects. Chem. Eng. Sci. 62, 45–55. Garcia, R., Pizarro, C., Lavin, A.G., Bueno, J.L., 2013. Biomass proximate analysis using thermogravimetry. Bioresour. Technol. 139, 1–4. Garcia-Perez, M., Wang, X.S., Shen, J., Rhodes, M.J., Tian, F., Lee, W., Wu, H., Li, C., 2008. Fast pyrolysis of oil mallee woody biomass: effect of temperature on the yield and quality of pyrolysis products. Ind. Eng. Chem. Res. 47, 1846– 1854. Heo, H.S., Park, H.J., Yim, J., Sohn, J.M., Park, J., Kim, S., Ryu, C., Jeon, J., Park, Y., 2010. Influence of operation variables on fast pyrolysis of Miscanthus sinensis var. purpurascens. Bioresour. Technol. 101, 3672–3677. Ioannidou, O., Jung, C.G., Zabaniotou, A., 2011. A thermogravimetric model to predict yield product distribution in pyrolysis of agricultural biomass. Catal. Today 167, 129–134. Isahak, W.N.R.W., Hisham, M.W.M., Yarmo, M.A., Hin, T.Y., 2012. A review on bio-oil production from biomass by using pyrolysis method. Renew. Sust. Energy Rev. 16, 5910–5923. Kersten, S.R.A., Wang, X., Prins, W., van Swaaij, P.M., 2005. Biomass pyrolysis in a fluidized bed reactor. Part 1: literature review and model simulations. Ind. Eng. Chem. Res. 44, 8773–8785. Kim, S.W., Ahn, J.Y., Kim, S.D., Lee, D.H., 2003. Heat transfer and bubble characteristics in a fluidized bed with immersed horizontal tube bundle. Int. J. Heat Mass Transfer 46, 399–409. Kim, S.W., Koo, B.S., Lee, D.H., 2014. A comparative study of bio-oils from pyrolysis of microalgae and oil seed waste in a fluidized bed. Bioresour. Technol. 162, 96– 102. Kim, S.W., Koo, B.S., Ryu, J.W., Lee, J.S., Kim, C.J., Lee, D.H., Kim, G.R., Choi, S., 2013a. Bio-oil from the pyrolysis of palm and Jatropha wastes in a fluidized bed. Fuel Process. Technol. 108, 118–124.

283

Kim, S.W., Park, D.K., Kim, S.D., 2013b. Pyrolytic characteristics of Jatropha seedshell cake in thermobalance and fluidized bed reactors. Korean J. Chem. Eng. 30, 1162–1170. Kunii, D., Levenspiel, O., 1991. Fluidization Engineering. Butterworth-Heinemann, US. Lee, S., Eom, M., Yoo, K., Kim, N., Jeon, J., Park, Y., Song, B., Lee, S., 2008. The yields and composition of bio-oil produced from Quercus Acutissima in a bubbling fluidized bed pyrolyzer. J. Anal. Appl. Pyrolysis 83, 110–114. Neves, D., Thunman, H., Matos, A., Tarelho, L., Gomez-Barea, A., 2011. Characterization and prediction of biomass pyrolysis products. Prog. Energy Combust. Sci. 37, 611–630. Park, H.J., Dong, J.I., Jeon, J.K., Park, Y.K., Yoo, K.S., Kim, S.S., Kim, J., Kim, S., 2008a. Effect of the operating parameters on the production of bio-oil in the fast pyrolysis of Japanese larch. Chem. Eng. J. 143, 124–132. Park, H.J., Park, Y.K., Kim, J.S., 2008b. Influence of reaction conditions and the char separation system on the production of bio-oil from radiata pine sawdust by fast pyrolysis. Fuel Process. Technol. 89, 797–802. Park, H.J., Park, Y.K., Dong, J.I., Kim, J.S., Jeon, J.K., Kim, S.S., Kim, J., Song, B., Park, J., Lee, K.J., 2009. Pyrolysis characteristics of oriental white oak: kinetic study and fast pyrolysis in a fluidized bed with an improved reaction system. Fuel Process. Technol. 90, 186–195. Pattiya, A., 2011. Bio-oil production via fast pyrolysis of biomass residues from cassava plants in a fluidized-bed reactor. Bioresour. Technol. 102, 1959–1967. Pattiya, A., Suttibak, S., 2012. Production of bio-oil via fast pyrolysis of agricultural residues from cassava plantations in a fluidised-bed reactor with a hot vapour filtration unit. J. Anal. Appl. Pyrolysis 95, 227–235. Pemberton, S.T., Davidson, J.F., 1986. Elutriation from fluidized beds-I. Particle ejection from the dense phase into the freeboard. Powder Technol. 41, 243–251. Wen, C.Y., Yu, Y.H., 1966. A generalized method for predicting the minimum fluidization velocity. AIChE J. 12, 610–612. Wu, S.Y., Baeyens, J., 1998. Segregation by size difference in gas fluidized beds. Powder Technol. 98, 139–150.