Effect of biomass-sulfur interaction on ash composition and agglomeration for the co-combustion of high-sulfur lignite coals and olive cake in a circulating fluidized bed combustor

Bioresource Technology 198 (2015) 325–331

Contents lists available at ScienceDirect

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

Effect of biomass-sulfur interaction on ash composition and agglomeration for the co-combustion of high-sulfur lignite coals and olive cake in a circulating fluidized bed combustor Murat Varol a,b,⇑, Aysel T. Atimtay a a b

Department of Environmental Engineering, Middle East Technical University, Ankara 06800, Turkey Department of Environmental Engineering, Akdeniz University, Antalya 07058, Turkey

h i g h l i g h t s

g r a p h i c a l a b s t r a c t


Bed agglomeration was not seen

during the tests.
Sulfur in coal prevented

agglomeration in the bed.
High-sulfur content in the coal helped

to hinder ‘‘Potassium Silicate” formation.
It is advantageous to add high-S coals to the combustor for olive cake combustion.
Limestone addition transfer K from Arcanite to Potassium Calcium Sulfate phase.

a r t i c l e

i n f o

Article history: Received 23 July 2015 Received in revised form 3 September 2015 Accepted 4 September 2015 Available online 11 September 2015 Keywords: Co-combustion Olive cake High-sulfur lignite coal Circulating fluidized bed Agglomeration

a b s t r a c t This study aimed to investigate the effect of biomass-sulfur interaction on ash composition and agglomeration for the co-combustion of high-sulfur lignite coals and olive cake in a circulating fluidized bed combustor. The tests included co-combustion of 50–50% by wt. mixtures of Bursa-Orhaneli lignite + olive cake and Denizli-Kale lignite + olive cake, with and without limestone addition. Ash samples were subjected to XRF, XRD and SEM/EDS analyses. While MgO was high in the bottom ash for Bursa-Orhaneli lignite and olive cake mixture, Al2O3 was high for Denizli-Kale lignite and olive cake mixture. Due to high Al2O3 content, Muscovite was the dominant phase in the bottom ash of Denizli Kale. CaO in the bottom ash has increased for both fuel mixtures due to limestone addition. K was in Arcanite phase in the cocombustion test of Bursa/Orhaneli lignite and olive cake, however, it mostly appeared in Potassium Calcium Sulfate phase with limestone addition. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Fluidized bed combustion technology has an advantage of burning different types of fuels in the same combustor. Due to this

⇑ Corresponding author at: Department of Environmental Engineering, Akdeniz University, Antalya 07058, Turkey. Tel.: +90 242 227 4400/4382; fax: +90 242 310 6303. E-mail address: [email protected] (M. Varol). http://dx.doi.org/10.1016/j.biortech.2015.09.016 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.

advantage, most of the fluidized bed combustors all over the world are recently designed to operate with different solid fuels. However, mixing different fuels may cause some operational problems in the system and may also lead to complete shutdown of the system. When two or more fuels are burned in a fluidized bed combustor, chemical reactions and/or physical interactions may occur between ash particles due to different ash characteristics which result in operational problems such as agglomeration of bed material, slagging, fouling and corrosion of heat exchanger tubes (Hupa, 2008). Although these operational problems are main

326

M. Varol, A.T. Atimtay / Bioresource Technology 198 (2015) 325–331

concerns for the combustion of all types of solid fuels, they are particularly important for the combustion of biomass containing high alkali and alkaline earth metals (Szemmelveisz et al., 2009). Although the agglomeration problem was firstly studied for coal combustion (Reddy and Mahapatra, 1999; Vuthaluru and Zhang, 1999), it became more important when several biomass fuels started to be co-combusted with coals (Arvelakis et al., 2001; Lin et al., 2003; Zheng et al., 2007; Vamvuka et al., 2008; Toscano and Corinaldesi, 2010; Yang et al., 2011; Silvennoinen and Hedman, 2013; Duan et al., 2015). Alkali metals, K and Na, in biomass play an important role in the formation of agglomeration during co-combustion of biomass in fluidized beds (Hupa, 2012; Silvennoinen and Hedman, 2013). These alkali metals can react with silica sand (bed material) at temperatures in the range of 700–900 °C forming low-melting point eutectics like alkali silicates (Scala and Chirone, 2006, 2008; Lokare, 2008). Alkali silicates can form a molten layer on particle surfaces making the particles sticky during combustion (Hupa, 2012). When the system operates for a long time, the sticky layer on the surface of the particles causes formation of permanent bonds between sand particles (Scala and Chirone, 2008). The particles in the bed hit and adhere onto each other due to this sticky layer, causing the particles increase both in size and number. If this is allowed to continue, agglomeration of the whole bed material takes place and total shutdown of the combustor cannot be avoidable. Although the formation of silicates is not well defined, a solid–gas reaction between silica and alkali chloride vapors is given as a possible way which forms liquid alkali silicate (K2SiO3) (Hupa, 2012). If there is no alkali chloride available, silica can react with other alkali vapors such as alkali hydroxides forming alkali silicate (Hupa, 2012). For the combustion of biomass with ash containing high amounts of K or Na, reaction of SiO2 with alkali oxides or salts are possible as reported by Hupa (2012). Quartz sand, which is mainly composed of SiO2, has a melting temperature of around 1450 °C (Lin et al., 1997). SiO2 can react with alkali oxides or salts in the ash, forming eutectic mixtures with melting temperatures of 874 °C for alkali oxides and 764 °C for alkali salts (Grubor et al., 1995). These temperatures are lower than the melting temperature of SiO2 as well as the melting temperatures of individual components (891 °C for K2CO3 and 851 °C for Na2CO3). The main precaution to prevent total agglomeration is lowering the alkali content in the combustor by continuously removing the ash formed and replacing the bed material with fresh silica sand (Hupa, 2012). It is also important to use alternative bed materials in order to minimize the silica content which eventually reacts with alkalis (Hiltunen et al., 2008). Different types of materials are used as an alternative to silica sand. Dolomite, alumina and limestone (Ninduangdee and Kuprianov, 2015), porous alumina (Shimizu et al., 2006), blast-furnace slag and olivine sand (Davidsson et al., 2008) can be used as alternative bed materials. Using these alternative bed materials may eliminate the formation of molten silicates. However, silica sand is commonly preferred bed material even with the high-alkali fuels due to high cost of alternative bed materials (Hupa, 2012). Moreover, some problems such as high attrition and entrainment rates, chemical instability, and plugging of air nozzles and windbox were reported during the usage of these alternative bed materials (Khan, 2007). Another measure to take against agglomeration is to add some materials such as kaolin, dolomite, lime, and alumina into the bed. However, these materials have limited usage due to low efficiency and high operational cost. Co-combustion of high-alkali biomass with highsulfur coals is another solution to solve the bed agglomeration problem during combustion of high-alkali biomass. Sulfur in the

coal may react with alkali oxides forming alkali sulfates. Consequently, Si cannot find enough alkali metals to form viscous alkali silicates. Before using any fuel in real applications, it is a good practice to estimate the tendency of the fuel ash to agglomerate. The agglomeration tendency of a fuel ash is estimated by using some indices. Bed Agglomeration Index (BAI) is the one that is commonly used. It is the ratio of iron oxides to the sum of potassium and sodium oxides in the fuel ash. The ratio is given in Eq. (1). If the ratio is less than 0.15, it is possible that the agglomeration will be seen (Bapat et al., 1997).

BAI ¼

%ðFe2 O3 Þ %ðK2 O þ Na2 OÞ

ð1Þ

The BAI is directly related to melting temperature of the fuel ash. If K and Na contents are high in fuel ash, this lowers the BAI and increases the tendency of ash to agglomerate. On the other hand, Ca and Mg contents in the ash were reported to have an increasing effect on BAI due to increase in the melting temperature of the fuel ash (Loo and Koppejan, 2008). In the literature, while burning biomass itself or co-combusting of several biomasses with coal, some operational problems regarding bed agglomeration or defluidization of the bed have often been reported. Control methods such as use of alternative bed materials (sillimanite, bauxite, calcite, magnesite, silica, and alumina), mineral additives (clay, kaosil, bauxite, and carbonate), pre-treatment (water washing, Al and Ca pre-treatment) of fuels for mitigating particle agglomeration were extensively studied (Vuthaluru and Zhang, 1999; Liu et al., 2007; Sun et al., 2008; Vamvuka et al., 2008). Agglomeration tendency of fluidized bed combustor ashes were also investigated in several studies (Anthony and Jia, 2000; Brus et al., 2005; Zevenhoven-Onderwater et al., 2006; Liu et al., 2009). Although there are several studies about the ash-related problems for fluidized bed combustion of biomass in the literature, there is no study conducted with olive cake and high-sulfur Turkish lignites to investigate ash and agglomeration characteristics of these fuels. This study aims to fill this gap. This study also aims to investigate the effect of limestone addition on agglomeration characteristics as well as the effect of sulfur-biomass interaction on the ash composition by burning olive cake with Bursa-Orhaneli and Denizli-Kale lignites in a CFBC.

2. Methods 2.1. Characteristics of fuels Proximate and ultimate analyses of Bursa-Orhaneli lignite, Denizli-Kale lignite and olive cake are given in Table 1. In all tests, the fuel particle size was 1–2 mm. Olive cake was selected as biomass because of its high alkali content in ash (K2O content of olive cake ash was about 50% by wt. on d.b. as given in Table 3). 2.2. Laboratory-scale Circulating Fluidized Bed Combustor (LAB-CFBC) and experimental conditions The experimental setup consists of a circulating fluidized bed combustor, a fuel feeding system, electrical heaters, and two cyclones, flue gas cooling unit and a bag filter. The experimental setup was described in detail in the previous studies (Varol et al., 2014a,b). Four sets of co-combustion tests of Bursa-Orhaneli lignite (B) + olive cake and Denizli-Kale lignite (D) + olive cake with and without limestone (L) addition were conducted in order to see the effect

327

M. Varol, A.T. Atimtay / Bioresource Technology 198 (2015) 325–331 Table 1 Proximate, ultimate analyses and calorific values of lignites and olive cake used in the study. Proximate analysis

FC

VM

Ash

Moisture

% by wt. (as fired)

HHV

LHV

kJ/kg (as fired)

Bursa-Orhaneli (B) Denizli-Kale (D) Olive cake (OC)

32.18 28.58 16.31

34.50 37.58 67.25

9.69 17.16 7.00

23.63 16.68 9.44

4469 4131 4592

Ultimate analysis

C

H

N

O

Scombustible

4161 3861 4269 Stotal

Ash

2.52 3.74 0.78

12.69 20.60 7.73

% by wt. (dry basis) Bursa-Orhaneli (B) Denizli-Kale (D) Olive Cake (OC)

68.48 55.74 54.23

4.56 3.42 5.46

1.51 1.46 1.38

11.53 17.08 30.59

1.23 1.70 0.61

FC: Fixed Carbon, VM: Volatile Matter, HHV: Higher Heating Value, LHV: Lower Heating Value.

Table 2 Operational parameters for the co-combustion tests. Test code

B-OC-L

B-OC

D-OC-L

D-OC

Coal Biomass (50% by wt.) Limestone Molar Ca/Stotal ratio Duration of fuel feeding, hh:mm Coal feeding rate, kg/h Biomass feeding rate, kg/h Thermal power, kW Air flowrate, m3/h (@20 °C, 1 atm) uo, Superficial velocity, m/s Excess air ratio (k), – Temperature of dense phase (Tdp), °C Temperature of freeboard (Tfb), °C Temperature of return leg (Trl), °C

Bursa-Orhaneli Olive cake Çan limestone 2 40:24 2.5–3.0 2.5–3.1 25–31 27–30 3.4–3.6 1.12–1.48 798 ± 6 800 ± 7 710 ± 20

Bursa-Orhaneli Olive cake – – 25:03 2.3–3.0 2.3–3.1 23–31 28–34 3.6–4.1 1.16–1.49 824 ± 11 831 ± 14 726 ± 39

Denizli-Kale Olive cake Çan limestone 2 51:25 3.2–3.5 3.1–3.4 24–27 24–27 3.4–3.6 1.09–1.42 806 ± 11 787 ± 8 698 ± 29

Denizli-Kale Olive cake – – 49:50 3.2–3.5 3.2–3.4 29–31 27–28 3.4–3.6 1.11–1.35 804 ± 23 798 ± 18 695 ± 10

B-OC-L = Bursa lignite + olive cake + limestone, D-OC-L = Denizli lignite + olive cake + limestone.

of possible interaction between biomass and limestone on ash composition and agglomeration. Operational parameters for the tests are given in Table 2. From now on when ‘‘fuel mixture” is mentioned in the remaining part of this study, 50/50% by wt. of coal and olive cake mixture is meant. At the end of the combustion experiments, bottom ash from the combustor, fly ash samples from the 2nd cyclone and from the bag filter were collected. Ash samples are subjected to XRF, XRD, and SEM-EDS analyses to determine their compositions.

2.3. Ash analyses 2.3.1. X-ray Fluorescence (XRF) and X-ray Diffraction (XRD) analyses Two Turkish lignites and olive cake were sampled and ashes from these fuels were obtained by combusting these fuels in a furnace according to the ASTM standards (ASTM D1374-04 Standard Test Method for ash in the Analysis Sample of Coal and Coke from Coal for coal samples, ASTM E1755-01 Standard Test Method for Ash in Biomass for biomass fuels) for XRF analysis in order to characterize the elemental composition of their ashes. At the end of the co-combustion tests, bottom ash (BAsh), fly ash captured in the second cyclone (FAsh-C), and fly ash captured in the bag filter (FAsh-BF) were collected and samples from each ash were subjected to XRF and XRD analyses. XRF and XRD give semi-quantitative results. The XRF analyses were done with Philips PW-2404 XRF Spectrometer by an accredited laboratory of Marmara Research Center-Material Institute (MRC-MI) of TUBITAK. XRD analyses of BAsh, FAsh-C, and FAsh-BF for the co-combustion tests and fuel ash samples were conducted at MRC-MI.

2.3.2. Scanning Electron Microscopy (SEM) A Scanning Electron Microscope (SEM) (Quanta 400F Field Emission) at METU-Central Laboratory was used to analyze ash samples taken from the bottom ash at the end of co-combustion tests. Energy Dispersive X-ray spectroscopy (EDS) technique was applied to ash samples in order to determine the elemental composition of the samples. 2.3.3. Ash fusion temperatures Ash fusion temperatures (Initial deformation temperature (IT), softening temperature (ST), hemispherical temperature (HT), and fluid temperature (FT)) of fuel samples were determined by the accredited laboratory of in MRC-EI of TUBITAK according to the ASTM D 1857-04 (Standard Test Method for Fusibility of Coal and Coke Ash). The ash samples were analyzed in the LECO AF700 Ash Fusion Determinator. IT for Bursa-Orhaneli, DenizliKale and olive cake is 1281 °C, 1237 °C and 1121 °C, respectively. 3. Results and discussion 3.1. Co-combustion of Bursa-Orhaneli lignite and olive cake without and with limestone addition The XRF results of BAsh, FAsh-C and FAsh-BF for the cocombustion tests of Bursa-Orhaneli lignite and clive cake without (B-OC) and limestone (B-OC-L) are given in Table 3. XRF analyses of the fuel ashes are also given in Table 3 for comparison. When Table 3 is investigated, it can be seen that Al2O3, CaO, Fe2O3, MgO, and SiO2 are the major oxides in the Bursa-Orhaneli lignite ash. However, K2O, CaO, and SiO2 are the major oxides in the olive cake ash. When these results are compared with the

328

M. Varol, A.T. Atimtay / Bioresource Technology 198 (2015) 325–331

Table 3 XRF results of fuel ashes alone and ashes formed at the end of the co-combustion tests with Bursa-Orhaneli lignite and olive cake. Compound

Olive cake ash*

(% by wt.) Al2O3 CaO Cl Cr2O3 CuO Fe2O3 K2O MgO MnO2 Na2O P2O5 SO3 SiO2 Normalized to %

3.04 20.38 2.19 0.12 0.08 5.16 49.26 2.95 0.16 0.49 3.57 4.34 7.48 100.0

Combustion of Bursa-Orhaneli lignite + olive cake without Limestone (B-OC)

Combustion of Bursa-Orhaneli lignite + olive cake with limestone (B-OC-L)

Bursa-Orhaneli ash*

Bursa-Orhaneli ash*

BAsh

FAsh-C

FAsh-BF

(% by wt.) 10.14 19.73 0.00 0.06 0.22 9.97 0.58 7.87 0.15 0.27 0.22 29.33 20.82

2.73 10.47 0.00 0.17 0.07 5.88 4.44 11.20 0.10 0.11 0.30 8.12 55.96

100.0

BAsh

FAsh-C

FAsh-BF

6.51 29.99 0.28 0.04 0.11 10.56 5.86 4.84 0.16 0.15 1.53 17.99 21.37

6.05 31.95 0.88 0.02 0.09 9.60 9.89 4.81 0.24 0.36 1.33 23.33 10.89

(% by wt.)

100.0

7.33 22.63 0.13 0.06 0.08 9.43 5.15 4.68 0.20 0.18 1.75 13.60 34.21 100.0

7.15 24.53 0.24 0.02 0.09 10.16 9.72 5.28 0.30 0.35 1.27 24.93 15.27 100.0

6.80 16.83 0.00 0.11 0.01 10.11 0.28 9.72 0.15 0.13 0.16 29.59 25.66

1.73 35.48 0.02 0.18 0.05 3.43 4.23 5.46 0.08 0.07 0.30 26.37 22.35

99.9

100.0

100.0

100.0

B-OC: Bursa-Orhaneli + olive cake test without limestone, B-OC-L: Bursa-Orhaneli + olive cake test with limestone, Bash: Bottom ash, FAsh-C: Fly ash from cyclone, FAsh-BF: Fly ash from bag filter. * Represents the analysis of coal/olive cake ash obtained by combusting coal/olive cake sample (taken before the co-combustion test) in a furnace according to the ASTM standards. The analysis of olive cake ash did not change. Therefore, it was reported only once.

results obtained from CFBC ash, namely BAsh, FAsh-C, and FAsh-BF, a considerable change has been observed. Al2O3 content has decreased considerably in the bottom ash of CFBC. When limestone is added to the combustor in the B-OC-L test, CaO content in the BAsh increased to 35.5% from 10.5%. In order to understand in which phase these elements are, XRD analysis of each sample were performed. The results are given in Table 4. As can be seen from Table 4, mostly Anhydrite (CaSO4) phase was determined in the ash of Bursa-Orhaneli lignite. This shows that Ca and S elements seen in XRF result are probably in the form of CaSO4. Si and Fe elements may be in the form of SiO2 and Fe2O3, respectively. In olive cake ash, K element which is the major element in the ash, is in the Potassium Carbonate Hydrate (K2CO31,5H2O) and Fairchildite (K2Ca(CO3)2) phase. A small amount of K element is in Arcanite (K2SO4) phase. 2.2% Cl element in XRF result appeared in the Sylvite (KCl) phase. In the XRD analysis of bottom ash of B-OC test, different phases were determined but these phases could not be quantified according to the Rietveld method. Quartz, Lime, Anhydrite, Arcanite and Forsterite phases were encountered in the bottom ash. The Quartz (sand) used as bed material, was the dominant phase in the bottom ash. Similarly, Quartz was the dominant phase in the fly ash collected in the second cyclone.

Limestone addition to the fuel mixture might transfer Potassium (K) element from Arcanite to Potassium Calcium Sulfate in the bottom ash. While K was in the form of Arcanite in the B-OC test, it mostly appeared in Potassium Calcium Sulfate phase in the bottom ash with limestone addition to the fuel mixture. Quartz, Calcite, Lime, Anhydrite, Potassium Calcium Sulfate and Forsterite phases were determined in the bottom ash obtained from B-OC-L test. Calcite in the limestone reacts with sulfur in coal and forms Anhydrite and Potassium Calcium Sulfate phases. According to the XRD analysis in Table 4, the phases containing K element in olive cake ash are encountered in the fly ash sample from the bag filter. These phases are Arcanite and Potassium Calcium Sulfate. Encountering of potassium salts in the bag filter might be an indication that potassium leaves the system in gas phase, and then it turns into potassium salts and finally caught in the bag filter since the gas temperature is below 200 °C in the filter. It is very interesting to note that there is no K2SiO3 (l) formation found in the XRD analysis. This is a good point that all the available K in the olive cake has reacted with S compounds and formed potassium sulfate, which does not cause a sticky layer formation and consequently no agglomeration was seen during the tests. A sample taken from the bottom ash of B-OC-L was analyzed by SEM. According to the SEM-EDS, Ca, K, S and O elements were the

Table 4 XRD (Rietveld method) results of fuel ashes and ashes formed at the end of the co-combustion tests with Bursa-Orhaneli lignite and olive cake. Phase

Olive Cake ash

Bursa-Orhaneli ash

(% by wt.) Quartz, SiO2 Calcite, CaCO3 Lime, CaO Anhydrite, CaSO4 Hematite, Fe2O3 Arcanite, K2SO4 Potassium Calcium Sulfate, K2Ca2(SO4)3 Fairchildite, K2Ca(CO3)2 Potassium Carbonate Hydrate, K2CO3.1,5H2O Sylvite, KCl Forsterite, Mg2(SiO4)

2.7 9.9

24.1 2.2 61.3 12.4

9.3 35.8 28 4.4

BAsh

FAsh-C

(B-OC) (% by wt.) p 73.7 p p p

BAsh

FAsh-C

23.6

24.5 3.3 18.4 18.7

47.6 9.1 7.9 10.6 12.4

21.8

12.4

FAsh-BF

(B-OC-L) (% by wt.)

4.4 4 17.9

5.4 16 21.9 17.5 15.6

p

B-OC: Bursa-Orhaneli + olive cake test without limestone, B-OC-L: Bursa-Orhaneli + olive cake test with limestone, quantified. Bash: Bottom ash, FAsh-C: Fly ash from cyclone, FAsh-BF: Fly ash from bag filter.

FAsh-BF

8.1 6.2 12.1 17.3 21.2 12.1 23

13.4 p

: It means that phase was detected but could not be

329

M. Varol, A.T. Atimtay / Bioresource Technology 198 (2015) 325–331

main elements on a particle investigated. Si was very low. When XRD result of bottom ash in Table 4 was investigated, it can be said that the particle was probably composed of CaSO4 and/or K2Ca2(SO4)3 phases. No phase for K2SiO3 was found. 3.2. Co-combustion of Denizli-Kale lignite and olive cake without and with limestone addition The XRF results of BAsh, FAsh-C and FAsh-BF for the cocombustion tests of Denizli-Kale lignite and olive cake without (D-OC) and with (D-OC-L) limestone is given in Table 5. As can be seen in Table 5, SiO2, Al2O3, CaO, Fe2O3, and MgO are the major oxides in the Denizli-Kale ash. However, K2O, CaO, and SiO2 are the major oxides in the olive cake ash. SiO2 is the major oxide present in all ash samples. It is followed by SO3 and Al2O3. Ash content of Denizli-Kale lignite (17.16% by wt. in Table 1) is higher than that of olive cake (7.00% by wt. in Table 1). Although about half of the olive cake ash was K2O, K2O in ashes of the co-combustion tests did not exceed 6% by wt. Ca and S elements seen in XRF result are probably in the form of Anhydrite. Anhydrite was determined in all ashes from the CFBC tests and also in

Denizli-Kale ash as can be seen in Table 6. The phase composition of Denizli-Kale ash could not be quantified according to the Rietveld method. It should be always kept in mind that the numeric values presented with XRD results are semiquantitative. Therefore, the results give the presence of a phase in an ash sample, but the percentage of the phase in the ash sample is not very certain. Cl coming from olive cake could not be detected in the bottom and fly ashes. Si element seen in XRF results was in the form of Quartz and Muscovite. While K element was in the form of Arcanite, Fairchildite, Sylvite, Kalicinite, and Muscovite in the ash of olive cake, it seems most of the K was transformed into Muscovite after combustion. It is again very interesting to note here that there is no K2SiO3 (l) formation found in the XRD analysis. This is a good point that all the available K in the olive cake has reacted with S compounds and formed Muscovite or potassium sulfate, which does not cause a sticky layer formation and consequently no agglomeration was seen during the tests in CFBC. Potassium was also seen as Potassium Calcium Sulfate in bag filter ash. When Table 5 is studied, it can be seen that SiO2, Al2O3, CaO, Fe2O3, and MgO are the major oxides in the Denizli-Kale ash. The presence of Al2O3 and SiO2 in the coal ash is an indication of

Table 5 XRF results of fuel ashes alone and ashes formed at the end of the co-combustion tests with Denizli-Kale lignite and olive cake. Compound

Olive Cake Ash*

(% by wt.) Al2O3 CaO Cl Cr2O3 CuO Fe2O3 K2O MgO MnO2 Na2O P2O5 SO3 SiO2 Normalized to, %

3.06 18.65 2.30 0.03 0.07 5.09 52.10 2.93 0.13 0.46 3.18 3.93 7.35 100.0

Combustion of Bursa-Orhaneli lignite + olive cake without Limestone (D-OC)

Combustion of Bursa-Orhaneli lignite + olive cake with limestone (D-OC-L)

Bursa-Orhaneli Ash*

Bursa-Orhaneli Ash*

BAsh

FAsh-C

FAsh-BF

(% by wt.)

BAsh

FAsh-C

FAsh-BF

5.51 34.60 0.05 0.07 0.02 4.50 3.93 5.99 0.05 0.21 0.66 33.01 10.99 100.0

11.14 25.72 0.15 0.20 0.09 17.92 6.13 7.61 0.10 0.38 0.68 8.81 19.97 100.0

10.41 23.35 0.52 0.10 0.08 16.15 7.79 7.41 0.16 0.46 0.69 14.15 17.55 100.0

(% by wt.)

16.12 10.75

14.86 13.18

0.08 0.01 13.01 1.90 8.06 0.08 0.42 0.19 19.06 29.34 100.0

0.07 0.04 6.91 5.05 4.94 0.07 0.36 0.18 19.81 33.66 100.0

16.36 10.59 0.07 0.07 0.05 12.59 4.86 7.77 0.08 0.51 0.55 11.56 33.98 100.0

14.92 13.00 0.09 0.08 0.07 14.09 5.39 8.78 0.10 0.51 0.57 15.30 26.12 100.0

9.72 15.25 0.16 0.01 14.19 0.76 12.85 0.04 0.71 0.20 31.72 13.71 100.0

D-OC: Denizli-Kale + olive cake test without limestone, D-OC-L: Denizli-Kale + olive cake test with limestone, Bash: Bottom ash, FAsh-C: Fly ash from cyclone, FAsh-BF: Fly ash from bag filter. * Represents the analysis of coal/olive cake ash obtained by combusting coal/olive cake sample (taken before the co-combustion test) in a furnace according to the ASTM standards. The analysis of olive cake ash did not change. Therefore, it was reported only once.

Table 6 XRD (Rietveld method) results of fuel ashes alone and ashes formed at the end of the co-combustion tests with Denizli-Kale lignite and olive cake. Phase

Olive cake ash

Denizli-Kale ash

(% by wt.) Quartz, SiO2 Calcite, CaCO3 Lime, CaO Anhydrite, CaSO4 Hematite, Fe2O3 Magnetite, Fe3O4 Arcanite, K2SO4 Potassium calcium sulfate, K2Ca2(SO4)3 Fairchildite, K2Ca(CO3)2 Calcium magnesium sulfate, CaMg3(SO4)4 Sylvite, KCl Kalicinite, KAl(SO4)211H2O Muscovite, KAl2(Si,Al)4O10(OH)2 Magnesioferrite, MgFe2O4 Calcium magnesium aluminum oxide, CaMg2Al16O27

15.6 17.7

BAsh

FAsh-C

FAsh-BF

(D-OC) (% by wt.) p p p p p

BAsh

FAsh-C

FAsh-BF

(D-OC-L) (% by wt.)

19.2

32.2

24

26.6

12.8 1.8

17.6 2.9

7.9

2.9

19 4.1 1.5 34 8 7

0.4

10

14 72

18.7 5.4 6.9 17.8 2.3

12.1 3.1 27.8 1.8 15.6 9.4

p

p

43.5 10.7

46.6 6.8

D-OC: Denizli-Kale + olive cake test without limestone, D-OC-L: Denizli-Kale + olive cake test with limestone, quantified. Bash: Bottom ash, FAsh-C: Fly ash from cyclone, FAsh-BF: Fly ash from bag filter.

p

50.1 2.3

p p

37.4 11.6

6 11

: It means that phase was detected but could not be

330

M. Varol, A.T. Atimtay / Bioresource Technology 198 (2015) 325–331

aluminum silicate contamination from soil (Hiltunen et al., 2008). There is also sulfur remained in the Denizli-Kale ash. This result is consistent with the sulfur content of coal ash (Sash (Stotal  Scombustible) = 1.85%) according to the ultimate analysis of Denizli-Kale lignite given in Table 1. K2O, CaO, and SiO2 are the major oxides in the olive cake ash. Because limestone is used in this test, CaO is the major oxide in all ash samples. With the addition of limestone into the combustor, Lime and Calcite phases were detected in ash samples. While Calcite was not seen in the bottom ash, it appears in both fly ash samples. This may be due to some small limestone particles, which are not calcined, escaped to the cyclone and bag filter. Unreacted lime can be seen particularly in the bottom ash. Most of the Ca element was present in the form of Anhydrite. This is due to the sulfation of CaO coming from both limestone and fuel ash. Presence of Sulfur in the combustion medium is very advantageous while burning high-K content olive cake to minimize the risk of agglomeration. K was detected in Arcanite or Potassium Calcium Sulfate phases and in Muscovite phase in the bottom ashes for the co-combustion tests of olive cake with Bursa-Orhaneli and DenizliKale lignites, respectively. Moreover, potassium silicates that are the main reason for the agglomeration problem could not be found in bottom ashes. Therefore, according to the results of this study it is advantageous to add high sulfur coals to the combustor during the combustion of high-alkali content olive cake or high-alkali content biomass. Also, instead of burning olive cake alone, always co-combustion of olive cake with a high-sulfur coal should be considered to prevent agglomeration in a fluidized bed. Limestone addition to bed for sulfur capture also helps to prevent agglomeration in the bed. SEM micrograph of a particle from the bottom ash of D-OC-L test showed that the particle was most probably made up of CaSO4. This is consistent with the XRD result of the bottom ash in Table 6. Lime phase was also observed which means that full sulfation of CaO could not be done and some CaO remained in the limestone particles. 3.3. Bed agglomeration There are mainly two important factors on the formation of bed agglomeration. They are: operation temperature and composition of fuel ash. Operation temperature is very important when compounds of the fuel ash have lower melting points than operation temperature. The operational temperature of the combustor in this study was kept at about 850 °C, which is nearly 270 °C less than IT of olive cake (1121 °C). IT of coals is even higher (>1230 °C). Therefore, the risk of agglomeration was minimum during experiments. Olive cake, with its high K2O content, could potentially cause agglomeration in the combustor for long-term tests. However, any agglomeration was not observed during co-combustion tests lasted between 25 h and 50 h, which meant that eutectics did not form during the tests. When XRD analyses in Table 4 and Table 6 are investigated, it can be seen that these eutectics were not detected. This may be explained by the effect of high sulfur content of the lignites used in the experiments to prevent agglomeration. Potassium (K), for example, can easily volatilize during char combustion and form gaseous compounds such as potassium chloride (KCl) or potassium hydroxide (KOH). These gaseous compounds can react with silica in the bed to form ‘‘alkali silicates” which have melting temperatures in the range of 930–980 °C (Hupa, 2008). Using high-sulfur lignite with olive cake for the cocombustion tests might minimize the risk of agglomeration because sulfur in the lignite reacted with alkali oxides and formed alkali sulfates. Therefore, sulfur being an alternative reactant for alkali metals; it generates a competitive reaction to the reactions

of SiO2 with alkali oxides or salts. Sulfur reacts with alkalies and thus, there is not enough alkali metal left to react with Si in the combustor. As a consequence, alkali silicates, which are the main reasons for the agglomeration problem, could not be formed. The total sulfur content of Bursa-Orhaneli and Denizli-Kale lignites were on the average 2.52% and 3.74% by wt. (on dry basis) (Table 1), respectively. Although Bursa-Orhaneli lignite was selected to represent low-sulfur lignite, its sulfur content may still be considered high enough to prevent agglomeration. Mixing high-alkali biomass with high-sulfur and high-ash coals is also a good application to dilute the alkali content of the fuel mixture (Werther et al., 2000). If there is enough Fe2O3 in the ash, Fe2O3 can react with alkali oxides or salts by competing with the reactions of SiO2 with alkali salts. Thus, the eutectic mixtures formed can reduce the tendency of formation of agglomeration because these mixtures have melting temperatures higher than 1135 °C (Grubor et al., 1995). The reactions of Fe2O3 with alkali oxides or salts could be the reason why agglomeration did not occur in the combustor. Fe2O3 contents in the ash of Bursa-Orhaneli lignite for B-OC and B-OC-L tests were 9.97% and 10.11% by wt., respectively (see Table 3). As can be seen from Table 3 for co-combustion of Bursa-Orhaneli lignite and olive cake, bottom ash contains 5.88% and 3.43% Fe2O3 for B-OC and BOC-L tests, respectively. Fe2O3 contents in the ash of Denizli-Kale lignite for D-OC and D-OC-L tests were 13.01% and 14.19% by wt., respectively (see Table 5). Bottom ash has 6.91% and 4.50% Fe2O3 for D-OC and D-OC-L tests, respectively. Although these amounts are enough to react with X2O or X2CO3 (X can be Na or K), X2Fe2O4 was not determined in the bottom ash by XRD analysis. On the contrary, Fe2O3 is the only phase that is detected in the bottom ash. But, it is not the dominant phase. It is also important to note that K was kept in the ash as Potassium Calcium Sulfate, Arcanite and Anhydrite phases. Si was detected in Forsterite (Mg2(SiO4)) for the co-combustion tests of Bursa-Orhaneli lignite and olive cake. Therefore, it can be said that these two elements, K and Si, could not react to form alkali silicates that are the leading compounds for agglomeration. The presence of Mg oxide in the fuel ashes (7.9–9.7% by wt. in Bursa-Orhaneli ash and 2.9% by wt. in olive cake ash (see Table 3) may have a critical role to prevent the agglomeration. This may be another reason why agglomeration did not happen for the tests carried out in this study. One of the reasons why there was no agglomeration for DOC and D-OC-L tests might be the presence of Al in fuel ashes. Al oxide content of fuel ashes (9.7–16.1% by wt. for Denizli-Kale ash and 3.1% by wt. for olive cake ash (see Table 5 for XRF analyses) appeared as Muscovite phase in the bottom and fly ashes at high levels. Muscovite contains K and Si elements in its structure preventing them to form alkali silicates. BAI is a good indication to predict the agglomeration tendency of fuel ash before they are used in a combustor. It is given in Eq. (1). If the ratio is less than 0.15, agglomeration is possible (Bapat et al., 1997). The BAI of the olive cake was calculated as 0.10. It is less than 0.15 and this shows that there is a risk of agglomeration for the combustion of olive cake alone. However, the percentage of olive cake in the fuel mixture was 50% by wt. for the cocombustion tests. BAI was calculated for each test. It was 0.49, 0.38, 0.50 and 0.55 for B-OC, B-OC-L, D-OC, D-OC-L, respectively. These can be an evidence to explain why agglomeration was not observed in the co-combustion tests.

4. Conclusion Sulfur in coal is very advantageous while co-combusting coal with biomass having high-K content (like olive cake) to minimize the agglomeration in a fluidized bed combustor. The results showed that K was in the Arcanite phase in the bottom ash for

M. Varol, A.T. Atimtay / Bioresource Technology 198 (2015) 325–331

the co-combustion of olive cake with Orhaneli lignite. On the other hand K was found in the Muscovite phase for the co-combustion with Kale lignite. The presence of K and Si in Muscovite phase in the bottom ash was confirmed with XRD analysis. Therefore, it is believed that this was the reason why agglomeration did not happen during co-combustion tests because the formation of potassium silicates, main cause of agglomeration, was prevented. Limestone addition to the fuel mixture transferred K element from Arcanite to Potassium Calcium Sulfate phase in the bottom ash. Acknowledgement The financial support provided to this project by the Turkish Scientific and Technical Research Council-TUBITAK (Project Code: KAMAG-105G023) is greatly appreciated. References Anthony, E., Jia, L., 2000. Agglomeration and strength development of deposits in CFBC boilers firing high-sulfur fuels. Fuel 79, 1933–1942. Arvelakis, S., Vourliotis, P., Kakaras, E., Koukios, E.G., 2001. Effect of leaching on the ash behavior of wheat straw and olive residue during fluidized bed combustion. Biomass Bioenergy 20, 459–470. Bapat, D.W., Kulkarni, S.V., Bhandarkar, V.P., 1997. Design and operating experience on fluidised bed boiler burning biomass fuels with high alkali ash. In: Preto, F.D. S. (Ed.), Proceedings of 14th International Conference on Fluidized Bed Combustion. ASME, Vancouver, pp. 165–174. Brus, E., Öhman, M., Nordin, A., 2005. Mechanisms of bed agglomeration during fluidized-bed combustion of biomass fuels. Energy Fuels 19, 825–832. Davidsson, K.O., Åmand, L.E., Steenari, B.M., Elled, A.L., Eskilsson, D., Leckner, B., 2008. Countermeasures against alkali-related problems during combustion of biomass in a circulating fluidized bed boiler. Chem. Eng. Sci. 63, 5314–5329. Duan, F., Chyang, C.-S., Zhang, L., Yin, S.-F., 2015. Bed agglomeration characteristics of rice straw combustion in a vortexing fluidized-bed combustor. Bioresour. Technol. 183, 195–202. Grubor, B.D., Oka, S.N., Ilic, M.S., Dakic, D. V., Arsic, B.T., 1995. Biomass FBC combustion-bed agglomeration problems. In: Heinschel, K.J. (Ed.), Proceedings of the 13th International Conference on Fluidized Bed Combustion. ASME, Orlando, FL, pp. 515–522. Hiltunen, M., Barišic´, V., Zabetta, E.C., 2008. Combustion of Different Types of Biomass in CFB Boilers-Foster Wheeler, In: 16th European Biomass Conference, Valencia, Spain. Hupa, M., 2008. Ash Behavior in Fluidized Bed Combustion – Recent Research Highlights, In: 9th International Conference on Circulating Fluidized Bed. Hamburg. Hupa, M., 2012. Ash-related issues in fluidized-bed combustion of biomasses: recent research highlights. Energy Fuels 26, 4–14. Khan, A.A., 2007. Combustion and Co-combustion of Biomass in a Bubbling Fluidized Bed Boiler (Ph.D. thesis). Delft University of Technology, Delft, Netherlands. Lin, W., Dam-Johansen, K., Frandsen, F., 2003. Agglomeration in bio-fuel fired fluidized bed combustors. Chem. Eng. J. 96, 171–185. Lin, W., Krusholm, G., Dam-Johansen, K., Musahl, E., Bank, L., 1997. Agglomeration phenomena in fluidized bed combustion of straw, In: Preto, F.D.S. (Ed.), Proceedings of the 14th International Conference on Fluidized Bed Combustion V1. ASME, Vancouver, pp. 831–837.

331

Liu, R., Jin, B., Zhong, Z., Zhao, J., 2007. Reduction of bed agglomeration in CFB combustion biomass with aluminium-contain bed material. Process Saf. Environ. Prot. 85, 441–445. Liu, H., Feng, Y., Wu, S., Liu, D., 2009. The role of ash particles in the bed agglomeration during the fluidized bed combustion of rice straw. Bioresour. Technol. 100, 6505–6513. Lokare, S.S., 2008. A Mechanistic Investigation of Ash Deposition in Pulverized-Coal and Biomass Combustion (Ph.D. thesis). Brigham Young University, USA. Loo, S.V., Koppejan, J. (Eds.), 2008. The Handbook of Biomass Combustion and Cofiring, second ed. Earthscan, London. Ninduangdee, P., Kuprianov, V.I., 2015. Combustion of an oil palm residue with elevated potassium content in a fluidized-bed combustor using alternative bed materials for preventing bed agglomeration. Bioresour. Technol. 182, 272–281. Reddy, G.V., Mahapatra, S.K., 1999. Effect of coal particle size distribution on agglomerate formation in a fluidized bed combustor (FBC). Energy Convers. Manage. 40, 447–458. Scala, F., Chirone, R., 2006. Characterization and early detection of bed agglomeration during the fluidized bed combustion of olive husk. Energy Fuels 20, 120–132. Scala, F., Chirone, R., 2008. An SEM/EDX study of bed agglomerates formed during fluidized bed combustion of three biomass fuels. Biomass Bioenergy 32, 252– 266. Shimizu, T., Han, J., Choi, S., Kim, L., Kim, H., 2006. Fluidized-bed combustion characteristics of cedar pellets by using an alternative bed material. Energy Fuels 20, 2737–2742. Silvennoinen, J., Hedman, M., 2013. Co-firing of agricultural fuels in a full-scale fluidized bed boiler. Fuel Process. Technol. 105, 11–19. Sun, Z.-A., Jin, B.-S., Zhang, M.-Y., Liu, R.-P., Zhang, Y., 2008. Experimental study on cotton stalk combustion in a circulating fluidized bed. Appl. Energy 85, 1027– 1040. } cs, I., Palotás, Á.B., Winkler, L., Eddings, E.G., 2009. Szemmelveisz, K., Szu Examination of the combustion conditions of herbaceous biomass. Fuel Process. Technol. 90, 839–847. Toscano, G., Corinaldesi, F., 2010. Ash fusibility characteristics of some biomass feedstocks and examination of the effects of inorganic additives. J. Agric. Eng. 41, 13–19. Vamvuka, D., Zografos, D., Alevizos, G., 2008. Control methods for mitigating biomass ash-related problems in fluidized beds. Bioresour. Technol. 99, 3534– 3544. Varol, M., Atimtay, A.T., Olgun, H., Atakül, H., 2014a. Emission characteristics of cocombustion of a low calorie and high sulfur–lignite coal and woodchips in a circulating fluidized bed combustor: Part 1. Effect of excess air ratio. Fuel 117, 792–800. Varol, M., Atimtay, A.T., Olgun, H., 2014b. Emission characteristics of co-combustion of a low calorie and high-sulfur-lignite coal and woodchips in a circulating fluidized bed combustor: Part 2. Effect of secondary air and its location. Fuel 130, 1–9. Vuthaluru, H.B., Zhang, D., 1999. Control methods for remediation of ash-related problems in fluidised-bed combustors. Fuel Process. Technol. 60, 145–156. Werther, J., Saenger, M., Hartge, E.-U., Ogada, T., Siagi, Z., 2000. Combustion of agricultural residues. Prog. Energy Combust. Sci. 26, 1–27. Yang, T., Kai, X., Sun, Y., He, Y., Li, R., 2011. The effect of coal sulfur on the behavior of alkali metals during co-firing biomass and coal. Fuel 90, 2454–2460. Zevenhoven-Onderwater, M., Öhman, M., Skrifvars, B.-J., Backman, R., Nordin, A., Hupa, M., 2006. Bed agglomeration characteristics of wood-derived fuels in FBC. Energy Fuels 20, 818–824. Zheng, Y., Jensen, P., Jensen, A., Sander, B., Junker, H., 2007. Ash transformation during co-firing coal ash transformation during co-firing coal and straw. Fuel 86, 1008–1020.