Primary Fragmentation Behavior of Indian Coals and Biomass during

Apr 27, 2018 - Department of Chemical Engineering, National Institute of Technology Karnataka , Surathkal, Mangalore , India 575025. Energy Fuels ... ...
0 downloads 4 Views 4MB Size
Subscriber access provided by University of Glasgow Library

Combustion

Primary Fragmentation Behaviour of Indian Coals and Biomass during Chemical Looping Combustion K. Sekar Pragadeesh, and D. Ruben Sudhakar Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/acs.energyfuels.8b00640 • Publication Date (Web): 27 Apr 2018 Downloaded from http://pubs.acs.org on April 28, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Primary Fragmentation Behaviour of Indian Coals and Biomass during Chemical Looping Combustion K. Sekar Pragadeesh, D. Ruben Sudhakar* Department of Chemical Engineering, National Institute of Technology Karnataka, Surathkal, Mangalore, India - 575025 KEYWORDS: Chemical Looping Combustion, Primary fragmentation, coal, biomass.

ACS Paragon Plus Environment

1

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 45

ABSTRACT: Devolatilisation and fragmentation are important physical phenomena occurring during solid fuel Chemical Looping Combustion (CLC). Primary fragmentation during devolatilisation strongly affects the rate of fuel conversion, emissions, and fine particulates generation in a fuel reactor of a fluidized bed CLC unit, thus forming a critical design input. The present study focuses on investigating the primary fragmentation behaviour of large coal and biomass (wood) particles during devolatilisation phase of CLC. Three types of coals (2 Indian coals, 1 Indonesian coal) and one type of Casuarina wood of three sizes in the range of 8 to 25 mm, at different fuel reactor bed temperatures (800, 875 and 950 oC) are studied for primary fragmentation. Iron ore with 64% Fe is used as the oxygen carrier bed material, with steam as the fluidizing medium in the fuel reactor. The fragmentation behaviour is expressed in terms of number of fragments, fragmentation index, frequency of fragmentation, particle size distribution of fragments at different residence times of coal during devolatilisation in the fuel reactor. Under the conditions of study, the number of fragments increases with increase in particle size and temperature, for all fuels studied. Also, it is found that the number of fragments increases with the decrease in compressive strength of both coal and biomass particles. The Indian coals are found to fragment in the earlier stages of devolatilisation while the Indonesian coal and the biomass particles begin to fragment in the later stages of devolatilisation. Maximum fragmentation index is found with Indian coal - IC1 which has the highest fixed carbon content among the fuels studied and the least value is observed in biomass. Different modes of fragmentation exhibited by each fuel type is discussed. Indian coals do not show any volumetric changes as such, whereas Indonesian coal indicates some degree of volumetric expansion.

ACS Paragon Plus Environment

2

Page 3 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

1. INTRODUCTION Solid fuels, especially coal will stay as the mainstay source of energy for the next 100 years, considering the current reserves to production ratio 1. Since coal is the major fuel consumed for energy generation, the emission of carbon dioxide is inevitably increasing day by day. Developing carbon capture technologies and retrofitting them to the existing power generation units is becoming a very essential step now. Chemical Looping Combustion (CLC) is known to be an efficient technology for in-process carbon dioxide capture 2. The factors affecting the chemical looping process are oxygen carrier, fuel type, fuel size, fluidisation (gasifying) medium, operating parameters like bed temperature, pressure etc. The factors influencing the CLC process can be broadly classified into three groups as oxygen carrier, fuel, operating parameters. Studies on CLC using gaseous fuels have been widely reported in literature

2,3

. In

recent years, solid fuels like coal, biomass, petcoke, sewage sludge etc. are being studied with various oxygen carriers as reactive bed materials

4–8

. Among the solid fuels, coal is the major

fuel under study. In most of the works in chemical looping combustion, pulverised fuels (PF) and coarse fuel particles up to a few millimetres have been used. As the reserve levels of high grade coals (high calorific value and low ash) are limited, the future may demand the use of high ash, low grade fuels like Indian coals. But, pulverisation of these low grade coals and using them in CLC is not an attractive economic option. Generally, utilising coals in large particle sizes (in range of mm sizes) will limit the fuel conversion rates. However, in the case of insitu Gasification-CLC (iG-CLC) where fuel particles release the gasification products slowly, the use of large particles have advantages in terms of (i) reduction in cost of pulverisation, (ii) minimising loss of char particles while using pulverised coal 9, (iii) lesser oxygen carrier loading requirements compared to pulverised coal and also eliminating the need

ACS Paragon Plus Environment

3

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 45

of oxygen polishing step. Further, in case of high carbon coals (which have low tensile strength) 10

, particle comminution favourably influences the overall conversion of large fuel particles.

They generate smaller fuel fragments which then expose more surface area for conversion. During fuel conversion, comminution during devolatilisation is highly influential since the driving forces for size reduction during later stages of conversion are inferior to the pressures developed by volatile evolution. Particle comminution can be broadly classified into primary fragmentation, secondary fragmentation, percolative fragmentation and attrition 11. Among these fragmentation mechanisms, primary fragmentation is considered to be the critical size reduction phenomenon during the fuel conversion

12–14

, since it affects the particle size distribution in

reactor and the char conversion rates. The significance of primary fragmentation lies in the timing of fragmentation wherein the particles fragmenting earlier during devolatilisation are expected to have high conversion rates due to the generation of new surface area. Fragmentation of fuel particles immediately after the introduction into combustion environment has been attributed to the thermal shock or stress created by the temperature gradient

15–18

, which often result in large number of fragments.

Fragmentation events occurring in later stages of devolatilisation are due to the mechanical stresses created by the evolution of volatile matter at high temperatures

14,19,20

. Primary

fragmentation behaviour depends on the volatile content and the mechanical strength of the coal particles

11

, as well as the operating parameters such as particle size and temperature. The

compressive strength and the porosity of the individual coal particles help in understanding the initiation and the extent of fragmentation

15,21

. In literature, primary fragmentation of various

solid fuels have been widely studied in conventional air fluidised bed combustion environments 10,11,13,22–25

, using irregular shaped fuel particles and sand as the bed material

10,20,25–28

. As

ACS Paragon Plus Environment

4

Page 5 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

opposed to conventional combustion environments, the use of steam as fluidization medium in CLC is expected to influence the fragmentation and favourably the char reactivity

29

. Steam

could probably modify the pore structures, which could alter the fragmentation behaviour of fuel particles in the CLC environment. To the best of the authors’ knowledge, there is no study available in open literature on the fragmentation of large fuel particles in chemical looping conditions. The present study attempts to investigate the primary fragmentation behaviour of four different fuels, basically with different ash and volatile matter content in them, in the fuel reactor mode of chemical looping process. Also, the fragmentation behaviour of the fuels is related with their compressive strength and porosity (in terms of Pore Resistance Number), indicating their degree of influence. 2. EXPERIMENTAL 2.1.Experimental setup and materials

Figure 1. Schematic of the batch mode FB-CLC unit.

ACS Paragon Plus Environment

5

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 45

Figure 1 shows the fluidised bed combustor unit made of a stainless-steel cylindrical reactor of inner diameter 130 mm and of height 600 mm employed in this study. The temperature of the reactor is maintained by electrical heating, where the fluidising media is preheated to 300 oC before entering the reactor. A stainless-steel mesh basket of 2 mm mesh opening is used to hold the fuel particles during devolatilisation and retrieve them at the end of devolatilisation for analysis. A polished stainless steel (SS) plate is used as a mirror to see the in-bed events. The reactor is operated in a bubbling mode with gas velocity of 2.5 times the minimum fluidisation velocity. The experiments are conducted at three different bed temperatures of 800, 875 and 950 o

C. The composition and the properties of the fuels and the oxygen carrier (iron ore) used in the

study are given in Table 1 and Table 2 respectively. Table 1. Properties and compositions of fuels used

Composition (%wt.)

Indian Coal 1 Indian Coal 2 Indonesian (IC1) (IC2) Coal (IDC)

Biomass (BM)

Moisture

8.2

4.5

6.3

8.9

Volatile Matter

33.5

23.2

40.4

79.8

Fixed Carbon

31.9

29.62

43.62

10.9

Ash

26.4

42.68

9.68

0.4

Carbon

53.53

43.62

60.74

43.76

Hydrogen

2.133

1.86

5.01

5.69

Nitrogen

0.64

1.04

0.89

0.16

Oxygen

8.98

5.88

17.18

41.02

702

939

647

780

Elements (%wt.)

Property Bulk density (kg/m3)

ACS Paragon Plus Environment

6

Page 7 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Particle density (kg/m3)

1480

1917

1274

-

Two types of bituminous Indian coals (IC1, IC2), mainly differing in volatile matter and ash content are chosen for this study, along with a low ash bituminous Indonesian coal (IDC) and a Casuarina equisetifolia wood (BM) to compare and elucidate the behavioural differences. Natural hematite sourced from SLR Metaliks Ltd., Bellary, India is used as the oxygen carrier, in a size range of +250-425 µm. The bed loading is maintained to a static bed height of 130 mm. Coal fuels of different average sieve diameters, namely, 9 mm (+8 mm -10 mm), 14.25 mm (+12.5 - 16 mm), 22.5 mm (+20 mm - 25 mm) and biomass particles of 10 mm, 15 mm and 20 mm sizes are used. Sizes of the individual particle samples are expressed in terms of mass equivalent diameters 30. Table 2. Characteristics of the oxygen carrier Physical properties Bulk density (kg/m3)

2186

Particle density (kg/m3)

4541

Mean particle diameter (µm)

337.5 #

Chemical composition

Loss on Ignition; +Total Iron

Composition

LoI#

Fe(T)+

SiO2

Al2O3

P

Mn

Mass %

4.80

62.43

1.20

2.25

0.151

1.01

2.2.Methodology 2.2.1. Determination of devolatilisation time Devolatilisation times of the fuel samples are determined by Colour Indistinction Method 31. In Colour Indistinction Method (CIM), devolatilisation time (τd) of coal or biomass is determined

ACS Paragon Plus Environment

7

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 45

based on the colour change of the particle surface from black to reddish orange i.e. the end point of devolatilisation is marked when the fuel particle becomes indistinguishable with the hot fluidised bed. The devolatilisation time in CIM is defined as the time period between the introduction of the fresh fuel particle into the bed and the time of disappearance. 2.2.2. Fragmentation experiments For primary fragmentation experiments, the bed of oxygen carrier is fluidised initially with air in order to oxidise the oxygen carrier while being heated to the desired study temperature. Once the desired temperature is reached, air fluidisation is swapped by steam fluidisation. Based on the devolatilisation time of particular sized fuel particles obtained using CIM, the different residence times (i.e. quarters of total devolatilisation time) for fragmentation experiments are determined. A stainless-steel mesh basket is used for introducing fuel particles into the reactor and retrieving them back at predetermined time periods. Single particle experiments are focused in the present study and each experiment is repeated for a minimum of five times and sometimes, additional experiments are conducted in case of not obtaining consistent results in five repetitions. 2.2.3. Determination of Compressive strength of fuel particles Compressive strengths of the coal and biomass particle samples are tested using a computerised Universal Testing Machine (Shimadzu AG-X plus – 100 kN) at a loading rate of 1 mm/min. Since shaping the fuel particle creates failures in the microstructure, which most probably influences and alters the actual compressive strength, the compressive strength of the fuel particles are tested without shaping to a regular form, which is a more realistic approach

32

.

Except wood particles, particles of other fuels are used in irregular forms. The compressive strengths of coal samples are tested in the direction perpendicular to the bedding plane, whereas

ACS Paragon Plus Environment

8

Page 9 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

for the biomass, strength have been tested along the grain direction. Compressive strength (stress) has been calculated following a similar procedure reported by Zhong et al. 32. 3. RESULTS AND DISCUSSION The results of the primary fragmentation behaviour of four types of fuels are expressed in terms of percentage of fragmentation, number of fragments, frequency of fragmentation, timing interval of fragmentation events, fragmentation index, and particle size distribution of fuel particles. Representative plots are alone provided in the following main section and the remaining data are provided as Supporting Information. However, the discussion presented here includes the analysis of data in the figures in Supporting Information too. 3.1.Percentage of fragmentation events The percentage of fragmentation events gives the information on the probability of the fragmentation event to occur. Percentage of fragmentation (PF) is defined as, PF =

                    

* 100 %

(1)

Figure 2. Percentage of fragmentation events of (a) IC1 and (b) IC2.

ACS Paragon Plus Environment

9

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 45

Figure 3. Percentage of fragmentation events of (a) IDC and (b) BM. Figures 2 and 3 show the percentage of fragmentation events for different fuels studied at different operating conditions. The percentage of particles that undergo fragmentation during devolatilisation increases with the increase in temperature for all the fuel types studied. This behaviour may be attributed to the thermal shock experienced by the particle. Fragmentation percentage is found to raise with increase in particle diameter 15, but with few exceptions.

Figure 4. Heterogenous composition in a coal fuel (IC2). The effect of particle size on PF can be described by the fact that a larger particle has higher temperature gradient compared to that of a smaller one 21. Thus, a larger fuel particle experiences more thermal shock

15

and this phenomenon is closely related to the fuel’s homogeny. Figure 4

ACS Paragon Plus Environment

10

Page 11 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

illustrates the differences in heterogeny between a small and a large particle using IC2. Cross section of the sample shows that the number of stripes (layers of coal formation) are more in a large particle. Another side of the sample shows interspersed pattern of grey and white matter. The stratified structure of coal offers thermal resistance to heat transfer and at each layer, the thermal conductivity and thermal expansion co-efficient continuous to change. Thus, it may result in increased thermal stress, making the particle susceptible to failure at high temperatures.

Figure 5. Morphology of cross sections of coal particles. It means if a fuel particle has irregularities in structure and composition, then with the increase in particle size, the non-homogeneity gets compounded, resulting in increased thermal stress

33

.

Finally, the particle tends to crack severely and produce daughter fragments. If the particle is more homogenous, the effect of particle size is less pronounced. This can be substantiated by the results of homogenous and non-homogenous fuel particles i.e. biomass particle and coal particle respectively. As known, the coal is a complex formation whereas the woody biomass is more consistent in composition and it can be witnessed from the Figures 2 and 3 that coal particles tend to fragment more. Particles of all coal types have a minimum of 20% and a maximum of 60% chances for fragmentation (mostly at 950 oC). Table 3. Pore resistance number (PRN) of fuels used in this study

ACS Paragon Plus Environment

11

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 45

Pore Resistance Number

Critical diameter (mm) for fragmentation by definitiona

Minimum particle size (mm) to get fragmented (Present experiments)

Indian Coal 1

4.08

7.41

8

Indian Coal 2

5.15

6.19

8

Indonesian Coal

6.412

5.48

8

Biomass (Wood)

8.96

4.46

20

Fuel

a

based on critical diameter of fragmentation defined by Dakic et al.20 Dacombe et al.

15

noted that the particle’s propensity to fragment increases with increase in

fixed carbon and mineral matter of the fuel. Indian coals behave in line with this observation as evidenced from Figure 2. They fragmented more severely than other fuels in this study. Among the coal types, IC1 is found to have the maximum probability of fragmentation, since particles in all sizes fragment. Figure 5 shows the surface morphology of coal particles, where the homogeneity is mapped to even and shining texture. In that way, IC1 particle appears to be shiny but has a rough texture (showing weakness in structure), IC2 has even texture but shiny and also patches of white and grey spots showing the heterogenous nature. The IDC particle is both shiny and has an even texture (possibly due to its high vitrinite). IDC particles are relatively less likely to fragment and this property is owing to the fuel’s composition and homogenous nature (Figure 5). Similar kind of observation is made with biomass particles, where 10 and 15 mm particles do not fragment at low temperatures. Dakic et al.

20

relates the pore resistance number (PRN) with

the critical diameter of fragmentation. Critical diameter (Dcrit) is defined as the diameter of the largest particle for which probability of fragmentation (PF) is least or nil. In the present study, PF corresponding to Dcrit are experimentally found to be 10% or even less for IC2, IDC and it is nil for IC1. With no exceptions to PRN relationship with critical diameter for fragmentation, all fuel

ACS Paragon Plus Environment

12

Page 13 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

particles of sizes larger than the defined critical diameter (Table 3) are seen to have fragmented. However, PRN relation may not always assure 100 % chances of fragmentation. 3.2.Number of fragments The number of fragments (NF) or particle multiplication factor or fragmentation ratio can be defined as the ratio of the number of particles at the end of devolatilisation or at specified residence times to the initial number of particles.           

NF=             

(2)

Since the present study deals with single particle experiments, NF is directly the particle count at a particular residence time. If a particle does not fragment, the number of fragments remains one; when a particle fragments, the number of fragments would be a minimum of two. The effect of fuel types on the number of fragments produced by fuel particles at different residence times during devolatilisation at a bed temperature of 950 oC is presented in Figure 6. The data points represent the average number of fragments at that specified residence time. Also, considering the stochastic nature of the fragmentation phenomenon, each of the data points is marked with bars indicating the minimum and maximum NF obtained at that particular residence time. Among the different kinds of fuels, namely coal and biomass, it can be observed that at all bed temperatures and particle sizes, the NF generated by coals are higher than biomass. Coals used in this study typically generated up to 6 fragments on an average, during the course of devolatilisation, while the biomass particles generated less than 3 fragments under various conditions of operation (Figures S1 - S4). Biomass, compared to coal exhibited a very consistent fragmentation behaviour, as can be seen from Figure 6 and S4. The biomass particles hardly fragment and only the 20 mm particles are found to produce fragments counting an average of 3. This considerable difference in the behaviour between coal and biomass can be

ACS Paragon Plus Environment

13

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 45

attributed to their difference in fuel properties namely porosity, thermal conductivity and particle density (Coal: 1300 to 1900 kg/m3, Wood: 800 kg/m3). Coal with relatively higher density (1900 kg/m3) and lower porosity (5-10 %), tend to fragment severely 34 than biomass. Also, the thermal diffusivity of the coals (about 1.5*10-7 m2/s) differs considerably from biomass (about 4*10-7 m2/s), which is an important reason for their different fragmentation behaviour. A comparable fragment count is observed in case of air environment for the same wood fuel, devolatilised even at a lower bed temperature of 850 oC

35

. This indicates a noticeable effect of the fluidisation

medium/combustion environment on the fragmentation phenomenon.

ACS Paragon Plus Environment

14

Page 15 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Figure 6. Number of fragments of all fuel types in three particle size ranges (a) 8-10 mm, (b) 12.5-16 mm and (c) 20-25 mm at 950 oC. Among the different coals studied, IC1 displayed a consistent trend of increase in number of fragments with increase in residence time in the combustor. While IC2 and IDC occasionally deviated from their behaviour slightly at certain residence times. For instance, consider the case of IC2 of 20-25 mm at 950 oC in Figure 6, the NF decrease gradually with increasing residence time, which appears counter intuitive. However, this gradual decrease can happen due to the following reasons. (i)

Fragments generated in a previous quarter may generate relatively smaller fragments in the following quarter, of sizes small enough that elutriates out of the basket/bed, resulting in complete loss of original small fragments.

(ii)

Coals in specific, during different quarters of devolatilisation, especially in the later quarters (3rd,4th) generate large number of very small fragments (flakes ≈ 2 mm) which are not considered for NF, thus decreasing the NF in the later quarters of devolatilisation. However, mass of these very small fragments generated is taken into account for other relevant parameters, for e.g. conversion calculations.

IC1 particles of 22.5 mm generated a maximum of 13 fragments at 950 oC and the average NF is 6. The smaller sized particles of 9 mm and 14.25 mm yielded an average of around 2 and 3 fragments respectively. The IC2 particles of 22.5 mm size produced 9 fragments at the maximum, while the averages lie in the range of 2 to 3. Indonesian coal (IDC) particles of largest size generated highest count of 8 fragments with averages between 2 and 4, and the smaller particles have 1 to 2 particle fragments. The extent of fragmentation is found to increase with increase in particle size 15, which can be noticed from all the fuel types (Figures S1 - S4). Also, it is to be noted that high-volatile coals (IDC) yield lesser number of fragments compared to the

ACS Paragon Plus Environment

15

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

medium-volatile types (IC1 and IC2)

15

. In a study by Kosowska-golachowska and Lukos

Page 16 of 45

34

in

air environment, 10-12 mm particles of a Polish coal with similar properties that of IDC exhibited a comparable degree of fragmentation (about 7) to that of 20-25 mm IDC particles in present study. It may be the steam environment in CLC which favoured a lower degree of fragmentation for this type of coal in smaller particle sizes.

Figure 7. Number of fragments of IC1 in three particle size ranges [(a) 8-10 mm, (b) 12.5-16 mm, (c) 20-25 mm] at 800, 875 and 950 oC. The number of fragments is found to increase with the increase in particle size for all fuels at all temperatures studied. Out of different sized particles, the largest particles used (20 - 25 mm)

ACS Paragon Plus Environment

16

Page 17 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

produced relatively more number of fragments as noticed in earlier works conducted in air-FB experiments 15,22,36. The maximum average NF of 8-10, 12.5-16 and 20-25 mm particles are 2, 3 and 6 particles respectively. The number of fragments increases by a factor of 2 to 3 when the particle size increases from 8 to 25 mm.

Figure 8. Comparison of number of fragments obtained in air

23

at 850 oC and in CLC

environments at 875 oC. Figure 7 shows the number of fragments generated by IC1 at different residence times during devolatilisation at bed temperatures 800, 875 and 950 oC. For a particular sized particle, the largest number of fragments are found with samples devolatilised at the highest temperature studied i.e. 950 oC. Also, relatively increasing trends of NF is observed with the rise in bed temperature, which is more likely in conventional air combustion experiments too 37. Moreover, it can be noticed that the NF increases with increase in residence time at 950 oC for all fuel types

ACS Paragon Plus Environment

17

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 45

of different sizes, with few exceptions (Figures S1 - S4). But, at lower temperatures, the NF does not show a consistent pattern. For all fuel types except IC2, the maximum number of fragments is found in the third or fourth quarters. The increasing trend may be because of the propagation of the crack resulting in formation of multiple fragments. Among the fuels, the Indian coals are found to generate more number of fragments in the second and third quarter, whereas the IDC and biomass particles produce more fragments in the last quarter. In Figure 8, comparison of the number of fragments (particle multiplication factor) obtained in air

23

and CLC combustion environments shows that the air environment promotes fuel

fragmentation. In CLC environment, the degree of fragmentation is relatively lower (about 5 times) than that during air combustion. However, for the fuel with higher fixed carbon content (IDC), the number of fragments is noticeably higher in CLC experiments. 3.2.1. Relationship of physical properties of fuels to their fragmentation characteristics A fuel’s fragmentation characteristic can always be related to its physical properties such as strength, porosity, density, size etc. In literature, the number of fragments is related to diameter15, compressive strength15, porosity38 etc. individually to NF. Equations 3 and 4 are defined to express the fragmentation characteristic of coal and biomass respectively in terms of number of fragments, with the physical properties such as diameter and compressive strength. NFcoal = 0.81d0.379σc-0.116

(3)

NFbiomass = 1.2d0.924σc-0.559

(4)

where ‘d’ is the fuel particle diameter (mass equivalent) in mm, ‘σc’ is the compressive strength in MPa. The correlation coefficients (R2) of equations 3 and 4 are 0.77 and 0.84 respectively. Table 4. Compressive strength of fuel particles

ACS Paragon Plus Environment

18

Page 19 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Fuel

Compressive Strength (MPa)

Average number of fragments

8 – 10 mm

12 – 16 mm

20 - 25 mm

8 – 10 mm

12 – 16 mm

20 - 25 mm

IC1

5.73 ± 0.51

3.27 ± 0.37

1.11 ± 0.44

1.6

2.2

2.93

IC2

5.55 ± 1.23

2.74 ± 0.65

1.05 ± 0.25

1.2

1.7

2.4

IDC

1.73 ± 0.77

1.30 ± 0.45

0.56 ± 0.03

1.7

1.9

3

BM

71.78 ± 2.46

52.24 ± 7.17

32.46 ± 5.24

1.07

1.07

2.33

A parity plot showing the quality of the correlations is given as Figure 9. It can be inferred that particles size is the dominant influencing factor of NF, followed by compressive strength. It can be observed that the number of fragments increases with the increase in particle diameter, whereas it is inversely related to the compressive strength, as also can be noticed from Table 4.

Figure 9. Parity plot between experimental NF and predicted NF (using equations 3 and 4). On the fuel side, biomass has the least number of fragments as expected, whose compressive strength is the highest among all these fuels. Among the coals, the compressive strength decreases with increase in coal’s volatile matter 17. On this basis, IDC has the least compressive

ACS Paragon Plus Environment

19

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 45

strength as it has the highest volatile matter among the coal types and thus has highest average number of fragments at the end of devolatilisation. 3.3.Frequency of fragmentation The frequency of fragmentation events of all the fuels of different particle sizes studied can be found from Figures 10 and 11. The frequency of fragmentation (FF) is defined as the number of instances a fuel particle fragments in the entire devolatilisation period. It is a significant characteristic of a fuel getting converted at certain operating conditions, which influence the devolatilisation times and conversion rates i.e. if a fuel has high FF, it will apparently have lesser devolatilisation time and the overall conversion rate increases.

Figure 10. Frequency of fragmentation of IC1 (a) and IC2 (b) particles. Frequency of fragmentation is counted only on instances when a particle fragments afresh in that particular quarter of devolatilisation and is mainly based on the number of fragments in a quarter. For example, say the NF of a fuel particle in the first, second, third and fourth quarters are 2,2,3,2. In this case, frequency is counted in the first and third quarter, making the frequency value to two.

ACS Paragon Plus Environment

20

Page 21 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

The highest frequency of fragmentation events is found in IC1 and IDC fuel particles, which fragmented 3 times during devolatilisation. This behaviour may be due to the higher volatile content of these two fuels. The least frequency of fragmentation is found in the case of biomass particles, with almost no fragmentation at lower temperatures. The fragmentation frequency of the same wood particles studied under conventional air-FBC 35 is higher than in FB-CLC, hinting the probable influence of steam atmosphere. A study by Molina-Sabio et al.

39

shows that the

porosity of char has been improved by using steam for activation. The steam may probably enlarge the pores of the particle, leading to reduction in pressure build-up within the particle. The improved porous nature of the fuel may aid easier volatile release, thus reducing the fragmentation events.

Figure 11. Frequency of fragmentation of IDC (a) and biomass (b) particles. In all the fuel types, frequency of fragmentation events either increases or stays as such with increase in particle size, except for fuel IC1. The particles of largest sizes tend to fragment many times, which might be due to the thermal stress they experience 22. A different behaviour can be observed in case of IC1, whose 22.5 mm (average) particles do not exhibit a second instance of fragmentation till the bed temperature is raised to 950 oC. This shows that additional thermal

ACS Paragon Plus Environment

21

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 45

energy is needed for a larger particle of such fuel to undergo thermal stress and break apart. With the increase in temperature, the frequency of fragmentation increases and the highest frequencies are observed for all fuels at 950 oC, except IC2. The reason for this exceptional behaviour of IC2 may be due to a condition in which the first instance of fragmentation itself have developed enough room to vent the pressure created by volatile evolution (Figure 25). On an average, all fuel types other than biomass have fragmentation frequency of 2. Biomass particles fragment only at 950 oC but with a maximum frequency of 1. IDC particles of all sizes have relatively higher frequency of fragmentation at 875 oC compared to other temperatures of study. 3.4.Timing of fragmentation Timing of fragmentation (TF) is an important information about the occurrence of fragmentation event, since it influences the rate of fuel conversion, fine generation and the other parameters like gas species concentration etc. TF signifies the degree influence of fragmentation on the rate of fuel conversion, for e.g. if a particle fragments in first two quarters, the fragmentation event has high influence over the fuel conversion rate and when it fragments in later quarters, the fragmentation become less influential. Timing of fragmentation can be defined as the time slot or simply as one or more quarters of total devolatilisation time, in which the fragmentation events occur. It is defined based on the frequency of fragmentation i.e. whenever a frequency is counted, that particular time interval is considered.

ACS Paragon Plus Environment

22

Page 23 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Figure 12. Timing interval of fragmentation for IC1.

Figure 13. Timing interval of fragmentation for IC2.

ACS Paragon Plus Environment

23

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 45

Figure 14. Timing interval of fragmentation for IDC. Timing intervals of fragmentation of various fuels devolatilised at three bed temperatures are shown in Figures 12 to 15. IC1 particles fragment in second and third quarters of devolatilisation at 800 oC. However, they start to fragment in the very first quarters at higher temperatures. IC2 particles tend to fragment in the first and second quarters, with some exceptions. Indonesian coal (IDC) fragments mostly in the third and fourth quarters but it fragments even in second quarter at higher temperatures. Biomass particles of 20 mm fragment in the fourth quarter at all temperatures of study. However, the smaller sized particles fragment only at 950 oC in the final quarter of devolatilisation.

ACS Paragon Plus Environment

24

Page 25 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Figure 15. Timing interval of fragmentation for BM. Overall, it can be noticed that IC1 particles start fragmenting in very early stages (first and second quarters) of the devolatilisation process, whereas the biomass particles are the least fragmenting, which fragment only in the last quarter. The compressive strengths of coal particles are also found to influence the timing of fragmentation21, where the coal with a higher strength fragments quicker. This means that stronger coals tend to develop cracks or fragments in the earlier stages, due to the mounting of pressure build up during volatile release process. If the particles fragment in the first and second quarters, size reduction helps increase in surface area and subsequent fuel conversion. Fuel particle that generates fragments in the last quarter does not have much influence on the devolatilisation time as well as the overall conversion time. The increase in temperature is found to shift the fragmentation time interval to an earlier quarter of devolatilisation in all fuel types, except for a very few cases. 3.5.Fragmentation index

ACS Paragon Plus Environment

25

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 45

Fragmentation index is defined as the ratio of the number of fragments to the variation factor of the feed particles 17. Fragmentation index, FI = Fd = ∑"#

%$.    ()          ( )

 

(5) (6)

!

where Xi is the mass fraction of particle fragment of diameter, di; da = average particle diameter of the particles subjected to devolatilisation. The fragmentation index (FI) for one fuel (IC1) at three bed temperatures and of different particle sizes is shown in Figure 16. FI of all fuel types at 950 oC for 3 particle sizes is presented in Figure 17. The FI value depends mainly on the Fd value. Fd would always be less than 1 if the fuel is of non-swelling type and when the fuel is of swelling nature, the Fd will be ≥1 40. Since FI is inversely proportional to Fd, the FI value will be ≤ 1 for swelling types and ≥1 for nonswelling and shrinking types. In the Figure 16, it can be seen that the maximum changes in the particle size occur in the third quarter for the fuel type IC1. During this period, the particle is either losing its mass drastically or the number of fragments increases hugely or both. Also, it is noticed that IC1 has its least devolatilisation time at 950 oC because of increased surface area which is reflected from the relatively higher FI (Figure S5). IC2 particles of 22.5 mm always have their peak FI values in the second quarter at all temperatures studied. Whereas, for the smaller particles, the occurrence of maximum FI prepones one quarter with increase in temperature from 800 to 950 oC; at the highest temperature, maximum FI value is recorded in the second quarter (Figure S6). Indonesian coal (IDC) of all particle sizes have maximum size variations in the last quarter, which indicates the occurrence of more than one instance of fragmentation event (Figure S7). But smaller particles of 9 and 14.25 mm have their maximum FI in second quarter at 800 oC.

ACS Paragon Plus Environment

26

Page 27 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Biomass particles of all sizes have maximum FI in the last quarter, since the onset of fragmentation occurs just before the endpoint of devolatilisation (Figure S8).

Figure 16. Fragmentation index of IC1 of three particle sizes [(a) 8-10 mm, (b) 12.5-16 mm, (c) 20-25 mm] at bed temperatures 800, 875 and 950 oC.

ACS Paragon Plus Environment

27

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 45

Figure 17. Fragmentation index of all fuel types of particle sizes (a) 8-10 mm, (b) 12.5-16 mm and (c) 20-25 mm at 950 oC. Overall, volatile-rich fuel particles (IDC and BM) have the peak size/mass variation in the last quarter and it is the second/third quarter for other fuels (Indian coals) where the maximum FI is found. Similarly, in air combustion experiments, Zhang et al.

17

found that the fragmentation

index is maximum for the coal with the highest volatile content. The maximum FI values noticed in IC1, IC2, IDC and BM are 10.57, 4.87, 4.96 and 4.96 respectively (Figures S5 – S8), all observed for the largest particle sizes studied (except few cases where 12.5-16 mm particles have maximum FI). Among all the fuels, IC1 is found to undergo much size variations, which can be

ACS Paragon Plus Environment

28

Page 29 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

seen from Figure 17. Biomass registers a considerable FI (3-5) only for the highest size (20 mm) studied, that too at the end of devolatilisation time (implying less significance). But for the other sizes, regardless of the operating conditions, BM exhibits the lowest FI (viz. 1 to 1.5) compared to all other fuels studied, indicating a relatively lesser size changes due to fragmentation. From the particle size point of view, the fragmentation index is always higher for larger particle sizes, as observed in conventional combustion systems 17,36. Similarly, temperature has a proportional influence on FI and its effect on FI can be seen from Figure 16. Observations in line with the present study are noticed in literature findings on air-FBC 36,37,41. In the order of influence on FI, the operating parameters can be arranged as, Fuel type > Particle size > Temperature 3.6.Particle size distribution If a particle fragments in the combustor, the particle size distribution (PSD) gives the entire picture of the particle sizes formed out of fragmentation. For every quarter of the devolatilisation time, PSD is represented in terms of cumulative mass fraction. The PSD of all fuels of three sizes at 950 oC are given as Figures 18 – 21 and the PSD at temperatures 800 and 875 oC are provided in Figures S9 – S16. As a common trend, two kinds of patterns emerge from PSD observations, as follows. a) If a particle fragments in the first two quarters, the smaller particles generated tend to get converted and escape out of basket/reactor and only the large particle fragments are left behind in the combustor, in the later quarters [More likely in Indian coals]. Another case is, if a particle fragments in third/fourth quarter, the PSD shows the presence of smaller particle fragments near the end of devolatilisation [mostly in IDC and biomass particles].

ACS Paragon Plus Environment

29

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 45

Figure 18. Particle size distribution of IC1 particles of (a) 8-10 mm, (b) 12.5-16 mm and (c) 2025 mm at the end of various quarters of devolatilisation at 950 oC. During fragmentation experiments, processes such as coarse particle undergoing only conversion without fragmentation, coarse particle fragmenting into smaller particles and complete conversion of smaller fines, followed by elutriation are observed in the present study. Similar observations are found in conventional air combustion of biomass as well

35

. The

originating location of fractures or fissures in the particle determines the particle size distributions in a fragmentation process. If the crack develops in the inner radial positions of the fuel structure (due to the high volatile content of the fuel), it results in the formation of a larger

ACS Paragon Plus Environment

30

Page 31 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

daughter particle

41

. The fracture originating in the outer radial locations due to the thermal

stresses leads to exfoliation of outer layers, resulting in generation of fine particles

40

. Thus, it

can be understood that if most of the resulting fragments lie in the smaller particle size range, it denotes that the thermal stresses are predominant in creating fragments. If mass fraction is higher for larger particle sizes, the fragmentation cause could be the outburst due to excessive internal pressure developed at high temperatures.

Figure 19. Particle size distribution of IC2 particles of (a) 8-10 mm, (b) 12.5-16 mm and (c) 2025 mm at the end of various quarters of devolatilisation at 950 oC.

ACS Paragon Plus Environment

31

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 45

Figure 20. Particle size distribution of IDC particles of (a) 8-10 mm, (b) 12.5-16 mm and (c) 2025 mm at the end of various quarters of devolatilisation at 950 oC. For all of the fuel types, at all bed temperatures, the larger fuel particles generated more fragments as well as fines. As a general pattern, fragments generated in the earlier quarters are not retained until the end of devolatilisation, because of subsequent fragmentation into very small fines which may either escape the basket or due to further conversion and elutriation. Sometimes, particles which fragment in earlier quarters may show a trend of reduced number of particles in the next quarter, and again have their fragments increase in the subsequent quarter.

ACS Paragon Plus Environment

32

Page 33 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

This is because of another instance of fragmentation event witnessed by a larger particle (fragment) from the earlier quarter.

Figure 21. Particle size distribution of biomass particles of 10 (a), 15 (b) and 20 mm (c) at the end of various quarters of devolatilisation at 950 oC. At any given temperature for any particle size of a fuel, a concurrent increase in the mass loss with residence time is reflected from the decreasing mass equivalent diameter of the largest fragment, which can be noticed from all PSD figures. Compared to other coal types, IC1 generated relatively larger fraction of smaller fragments and fines, which also lead to the increase in the particle fragment count (Figures S9, S10).

ACS Paragon Plus Environment

33

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 45

Meanwhile, IDC produced relatively lesser fractions of smaller particles (Figures S13, S14), which may be attributed to the coal’s plastic nature and polymerisation reactions during the devolatilisation that prevented the particle from fragmenting. Though the cracks are fully developed, they are held together by thin joints (Figure 26). The biomass particles fragment rarely and most of the particles stay as such from the time of introduction till the end of devolatilisation. In cases where BM fragments, the daughter particles are usually of sizes above half of the original diameter of the parent particle (Figures S15, S16). PSD of all coal types show decrements in the mass fractions of smaller sized fragments with the increase in temperature. The probable cause for this type of phenomena could be the conversion of the smaller sized fragments due to fines generation and their conversion followed by elutriation/ basket escape incidents, as discussed earlier. 3.7.Modes of fragmentation For generalising the qualitative pattern of the fragmentation phenomenon, fuel particles retrieved from the bed at different residences times during devolatilisation are further analysed. The crack patterns can be observed from the images of various fuels given in Figures 22 to 26. The biomass particles seem to form initially sub-surface cracks 42 that later develop into bigger cracks leading to fragmentation. The transverse sides of biomass particles show cracks parallel (and continuous) to the longitudinal axis i.e. along the grain direction (Figure 23). Whereas, the top and bottom faces (plane perpendicular to the longitudinal axis) have cracks propagating radially in a discontinuous manner. These observations match the literature findings by Dhanarathinam and Kolar (2013) in air-FB combustion environment. Figure 23 shows the particle shrinkage leading to crack propagation (due to thermal strain). The particle at the end of devolatilisation time

ACS Paragon Plus Environment

34

Page 35 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

appears to be on a verge of failure by fragmentation, which is a combined effect of crack development, shrinkage and weakening of the fuel particle. In case of coal particles, they fragment mostly in a systematic manner of crack formation, crack propagation followed by particle breakage. At some instances, Indian coal particles exhibited stochastic fragmentation behaviour as discussed earlier in the number of fragments and fragmentation index sections. This may be characterised by the fragmentation events occurring in the earlier quarters of devolatilisation. In such particles, the degree of non-homogeneity may be relatively higher than other particles of the same fuel, which results in increased thermal stress within the particle leading to random failures.

Figure 22. Crack patterns over the transverse face of biomass particles (15 mm) during various stages of devolatilisation.

Figure 23. Crack patterns over the longitudinal face of biomass particles (15 mm) during various stages of devolatilisation.

ACS Paragon Plus Environment

35

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 45

Figure 24. Crack formation and fragmentation pattern of IC1 particles (12.5 - 16 mm) during various stages of devolatilisation.

Figure 25. Crack formation and fragmentation pattern of IC2 particles (12.5 - 16 mm) during various stages of devolatilisation.

Figure 26. Crack formation and fragmentation pattern of IDC particles (12.5 - 16 mm) during various stages of devolatilisation. IC1 samples (Figure 24) develop multiple sub-surface cracks in the first quarter and then they split into larger fragments and few smaller fragments. In the subsequent quarter, larger fragments may undergo second instance of fragmentation, resulting in smaller fragments. In the last quarter,

ACS Paragon Plus Environment

36

Page 37 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

the fines are lost and larger fragments stay as such, only undergoing further conversion. IC2 samples as seen in Figure 25 undergo fragmentation in a very methodical way. The crack initiation starts along the bedding plane (plane of organic matter deposition during coal formation). As devolatilisation continues, samples develop into layered wafer-like or finger-like structures as a result of crack propagation. Then, they produce fragments along a well-developed crack, in the form of thin sheets and flakes. A flat surface is usually noticed in the interface of crack, which denotes the crack propagation in a parallel direction to the bedding plane 19. IDC samples show a stochastic crack formation pattern, but the crack propagation and particle split-up process is systematic and sequential. As expected from the PRN relation with fragmentation, the IDC particles with the highest PRN among coals have the maximum crack formations. Even though the cracks are well developed to result in a fragmentation event, it is observed that they are held together by thin bridges for a longer time. Concurrent expansion and crack propagation in IDC gives the particle, a flower-like structure (Figure 26). At the end of devolatilisation, the fuel fragments cleave apart once the carbon bridges between them weaken. Except IC1, all other fuel types show a systematic fragmentation process. Thus, depending on the fuel nature, the process of crack formation, propagation and fragmentation varies. 3.8.Conceptual pathways of fuel fragmentation To summarise the ways in which the fuel particles undergo primary fragmentation, conceptual pathways of fragmentation are presented in Figures 27 to 30 (from time, t = 0 to t = τd for each quarter of devolatilisation). As a general trend, IC1 particles start to fragment in the second quarter and undergo further fragmentation in the later quarters, producing large number of smaller fragments. IC2 particles like IC1 start to fragment in the second quarter, larger particles

ACS Paragon Plus Environment

37

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 45

stay as such till the third quarter but the smaller ones fragments into fines and in the last quarter, larger particles weaken, generating fragments. IDC starts expanding and forms cracks till the second quarter and generates very few smaller fragments in the subsequent quarter. In the final quarter, more particles are generated and the mass fraction of bigger particles comes down drastically. Biomass particles tend to develop cracks till the last quarter with simultaneous shrinkage. Just before the end of devolatilisation, they start to fail at places where the stresses created by shrinkage supersedes the particle strength, along the crack path. The fragments are usually bent flakes which defoliates from the surface. Changes to the particle geometry for all tested solid fuels are listed in Table 5.

Figure 27. Fragmentation pathway of IC1.

Figure 28. Fragmentation pathway of IC2.

ACS Paragon Plus Environment

38

Page 39 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Figure 29. Fragmentation pathway of IDC.

Figure 30. Fragmentation pathway of biomass particles. Table 5. Summary of changes to particle geometry during primary fragmentation Parametersb

IC1

IC2

IDC

BM

Fragmentation At particle centre mode

Exfoliation into At particle centre Exfoliation flakes with plasticity

Sphericity

Decrease

Decrease heavily

Decrease moderately

Decrease

Symmetry

Decrease moderately

Maintained

Decrease

constant

moderately

Maintained constant

B/L ratio

Decrease moderately

Nearly constant

Nearly constant

Convexity

Nearly constant

Decrease heavily

Expands concave structure

b

to

Nearly constant a Cube shrinks to form concave sides

adapted from Cui et al.24

ACS Paragon Plus Environment

39

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 40 of 45

4. CONCLUSION The primary fragmentation behaviour of four different fuel particles at different operating conditions is investigated during different residence times (four quarters) of devolatilisation process in CLC. Following is the summary of findings. i.

Out of all the studied fuel particles, 20 to 60 percent of the particles tend to fragment, in all fuel types.

ii.

Among the studied fuels, Indian coals are more likely to fragment while considering all operating conditions, especially IC1 (also produces the highest number of fragments).

iii.

The Indian coals are found to fragment very early in the initial quarters of devolatilisation, whereas IDC and BM fragment in the later quarters of devolatilisation, indicating a stronger influence and higher importance of fragmentation in the case of Indian coals on its conversion.

iv.

The number of fragments is found to be high for the large sized fuel particles and also at higher temperatures.

v.

Coal particles have a fragmentation frequency of 2 to 3 while the biomass particles have a maximum frequency of one.

vi.

IC1 has the maximum fragmentation index (FI) and the least FI is observed with biomass particles, reflecting the difference in the nature of fuel. It is also established that the compressive strength proportionally influences the number of

fragments. The modes in which each fuel type undergo fragmentation are discussed in detail, varying from structured to unstructured crack formation and propagation leading to failures. It is understood that all the four fuel types exhibit different modes of fragmentation which would significantly affect the rate-controlling char gasification phase during CLC.

ACS Paragon Plus Environment

40

Page 41 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

ASSOCIATED CONTENT *S Supporting Information Number of fragments generated at various operating conditions by IC1, IC2, IDC and BM are given in Figures S1 - S4 respectively. Fragmentation Index of IC1, IC2, IDC and BM are shown correspondingly in Figures S5 – S8. Figures S9 - S16 represents the Particle Size Distribution of IC1, IC2, IDC and BM at 800 and 875 oC.

AUTHOR INFORMATION Corresponding Author *E-mail: [email protected] Notes The authors declare no competing financial interest.

ACS Paragon Plus Environment

41

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 42 of 45

REFERENCES (1)

BP. BP Statistical Review of World Energy; 2017.

(2)

Lyngfelt, A. Appl. Energy 2014, 113, 1869–1873.

(3)

Nandy, A.; Loha, C.; Gu, S.; Sarkar, P.; Karmakar, M. K.; Chatterjee, P. K. Renew. Sustain. Energy Rev. 2016, 59, 597–619.

(4)

Niu, X.; Shen, L.; Jiang, S.; Gu, H.; Xiao, J. Chem. Eng. J. 2016, 294, 185–192.

(5)

Schmitz, M.; Linderholm, C.; Hallberg, P.; Sundqvist, S.; Lyngfelt, A. Energy & Fuels 2016, 30 (2), 1204–1216.

(6)

Linderholm, C.; Lyngfelt, A.; Cuadrat, A.; Jerndal, E. Fuel 2012, 102, 808–822.

(7)

Vilches, T. B.; Lind, F.; Ryden, M.; Thunman, H.; Berdugo Vilches, T.; Lind, F.; Rydén, M.; Thunman, H. Appl. Energy 2017, 190, 1174–1183.

(8)

Linderholm, C.; Schmitz, M.; Knutsson, P.; Kallen, M.; Lyngfelt, A.; Källén, M.; Lyngfelt, A. Energy & Fuels 2014, 28 (9), 5942–5952.

(9)

Ohlemueller, P.; Busch, J.-P.; Reitz, M.; Stroehle, J.; Epple, B.; Ohlemüller, P.; Busch, J.P.; Reitz, M.; Ströhle, J.; Epple, B. J. Energy Resour. Technol. 2016, 138 (4), 42203-1042203–042207.

(10)

Chirone, R.; Massimilla, L. Powder Technol. 1991, 64 (3), 249–258.

(11)

Chirone, R.; Massimilla, L.; Salatino, P. Prog. Energy Combust. Sci. 1991, 17 (4), 297– 326.

ACS Paragon Plus Environment

42

Page 43 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

(12)

Senneca, O.; Russo, S.; Chirone, R. Chemical Engineering Transactions. 2009, pp 569– 574.

(13)

Stubington, J. F.; Moss, B. Can. J. Chem. Eng. 1995, 73, 505–509.

(14)

Paprika, M. J.; Komatina, M. S.; Dakić, D. V.; Nemoda, S. Đ. Energy & Fuels 2013, 27 (9), 5488–5494.

(15)

Dacombe, P.; Pourkashanian, M.; Williams, A.; Yap, L. Fuel 1999, 78 (15), 1847–1857.

(16)

Senneca, O.; Scala, F.; Chirone, R.; Salatino, P. Fuel 2017, 201, 65–80.

(17)

Zhang, H.; Cen, K.; Yan, J.; Ni, M. Fuel 2002, 81 (14), 1835–1840.

(18)

Krishnamoorthy, V.; Tchapda, A. H.; Pisupati, S. V. Fuel 2017, 208, 247–259.

(19)

Chirone, R.; Massimilla, L. Powder Technol. 1989, 57 (3), 197–212.

(20)

Dakic, D.; Honing, G. van der; Valk, M. Fuel 1989, 68, 911–916.

(21)

Syred, N.; Kurniawan, K.; Griffiths, T.; Gralton, T.; Ray, R. Fuel 2007, 86 (14 SPEC. ISS.), 2221–2231.

(22)

Kosowska-golachowska, M. In 27th Annual International Pittsburgh Coal Conference, PCC 2010; 2010.

(23)

Chirone, R.; Massimilla, L. In Twenty-Second Symposium (International) on Combustion; 1988; pp 267–277.

(24)

Cui, T.; Zhou, Z.; Dai, Z.; Li, C.; Yu, G.; Wang, F. Energy and Fuels 2015, 29 (10), 6231–6241.

ACS Paragon Plus Environment

43

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(25)

Page 44 of 45

Prins, W.; Siemons, R.; Van Swaaij, W. P. M.; Radovanovic, M. Combust. Flame 1989, 75 (1), 57–79.

(26)

Chirone, R.; Cammarota, A.; D’amore, M.; Massimilla, L. Symp. Combust. 1982, 1213– 1221.

(27)

Sundback, C. A.; Beer, J. M.; Sarofim, A. F. In Twentieth Symposium (International) on Combustion; 1984; pp 1495–1503.

(28)

Stubington, J. F.; Linjewile, T. M. Fuel 1989, 68 (2), 155–160.

(29)

Khan, M. R.; Hshieh, F. Y. In Preprints of Papers of the American Chemical Society, Division of Fuel Chemistry; 1989; Vol. 34, pp 1245–1255.

(30)

Sudhakar, D. R.; Kolar, A. K. Energy & Fuels 2010, 24 (2), 4820–4832.

(31)

Pragadeesh, K. S.; Sudhakar, D. R. Unpubl. results (Submitted to Fuel).

(32)

Zhong, S.; Baitalow, F.; Nikrityuk, P.; Gutte, H.; Meyer, B. Fuel 2014, 125, 200–205.

(33)

Stanmore, B. R.; Brillard, A.; Gilot, P.; Delfosse, L. Symp. Combust. 1996, 26 (2), 3269– 3275.

(34)

Kosowska-golachowska, M.; Luckos, A.; Kosowska-Galachowska, M.; Luckos, A. In Proceedings of the 20th International Conference on Fluidized Bed Combustion; 2009; pp 330–334.

(35)

Sudhakar, D. R.; Reddy, K. S.; Kolar, A. K.; Leckner, B. Fuel Process. Technol. 2008, 89 (11), 1121–1134.

ACS Paragon Plus Environment

44

Page 45 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

(36)

Li, C.; Xia, Z. X.; Qiao, X. L.; Li, W. B.; Fang, M. X. Adv. Mater. Res. 2014, 953–954, 1254–1260.

(37)

Baizhong, S.; Hong, Q.; Qing, W.; Jingru, B.; Hongpeng, L.; Chunxia, J. 2009 Int. Conf. Energy Environ. Technol. 2009, 640–643.

(38)

Sasongko, D.; Rasrendra, C. B.; Indarto, A. Int. J. Ambient Energy 2017, 1–8.

(39)

Molina-Sabio, M.; Sanchez-Montero, M. J.; Juarez-Galan, J. M.; Salvador, F.; Rodrı´guez-Reinoso, F.; Salvador, A. J. Phys. Chem. B 2006, 110 (25), 12360–12364.

(40)

Paprika, M.; Komatina, M.; Dakić, D.; Živković, G.; Mladenović, M. Energy & Fuels 2015, 29 (5), 3394–3398.

(41)

Senneca, O.; Urciuolo, M.; Chirone, R.; Cumbo, D. Fuel 2011, 90 (9), 2931–2938.

(42)

Dhanarathinam, R. S.; Kolar, A. K. Fuel 2013, 112, 208–223.

ACS Paragon Plus Environment

45