Experimental studies on spalling characteristics of Indian lignite coal in context of underground coal gasification

Fuel 154 (2015) 326–337

Contents lists available at ScienceDirect

Fuel journal homepage: www.elsevier.com/locate/fuel

Experimental studies on spalling characteristics of Indian lignite coal in context of underground coal gasification Sminu Bhaskaran a, Ganesh Samdani b, Preeti Aghalayam c, Anuradda Ganesh a, R.P. Singh d, R.K. Sapru d, P.K. Jain d, Sanjay Mahajani b,⇑ a

Department of Energy Science and Engineering, Indian Institute of Technology Bombay, Mumbai 400076, India Department of Chemical Engineering, Indian Institute of Technology Bombay, Mumbai 400076, India Department of Chemical Engineering, Indian Institute of Technology Madras, Chennai 600036, India d UCG Group, IRS, ONGC, Chandkheda, Ahmedabad, Gujarat, India b c

a r t i c l e

i n f o

Article history: Received 11 December 2014 Received in revised form 24 March 2015 Accepted 25 March 2015 Available online 8 April 2015 Keywords: Underground coal gasification Lignite Spalling Crack pattern

a b s t r a c t Underground Coal Gasification (UCG) is considered to be a clean coal technology primarily intended to utilize deep underground (>300 m) coal deposits. In this process, a mixture of reactant gases like air/oxygen and steam are injected directly to an ignited portion of underground coal seam. UCG involves complex interactions of different processes like drying, pyrolysis, chemical reactions and spalling. Spalling is detachment of small coal particles from the coal seam due to interconnection of cracks developed in it. It plays an important role by offering higher surface area to give improved performance. The mechanism of spalling and its characterization are not well understood. Furthermore, there are no well established experimental techniques to measure the spalling rates. This paper studies spalling behavior of a lignite coal, which is characterized by high moisture and volatile matter, and suggests a possible underlying mechanism. The rate of spalling was measured using an experimental setup under the UCG-like conditions. In this setup, a reacting coal block was attached to a load cell and suspended in a UCG-like environment. When the experiments were repeated under similar conditions with different blocks of same coal, it was found that there were variations in the rates of spalling. This might be due to the heterogeneity in coal blocks in the form of originally present fissures or weak regions. A UCG process model was used to explain these experimental results and also to investigate the effect of spalling rate on product gas calorific value. We believe that spalling happens due to formation and extension of cracks. Hence a microscopic crack pattern on a heated coal monolith was examined in different stages of heating to understand the mechanism of spalling. It is concluded that cracks are first formed during the initial stage of drying due to the capillary stresses developed due to removal of moisture from the pores and were further extended due to shrinkage of coal during pyrolysis. The detachment of coal particles happens due to horizontal linking of vertical cracks, which might result out of either horizontal cracks, if any, or available fissures and weak regions or relatively weak interlayer bonding at the bedding planes. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction The importance of coal as a fuel for economic growth is going to continue due to its abundance and technological developments in its downstream processing. Hence, it is necessary to utilize untouched coal deposits, most of which are available deep underground. Underground coal gasification is such a technology that provides access to coal deposits which are available at a depth of more than 300 m. It is also considered as a clean coal technology

⇑ Corresponding author. Tel.: +91 (0) 22 25767246. E-mail address: [email protected] (S. Mahajani). http://dx.doi.org/10.1016/j.fuel.2015.03.066 0016-2361/Ó 2015 Elsevier Ltd. All rights reserved.

worldwide due to its lesser polluting and highly efficient nature. UCG, like surface gasification, extends the applicability of coal to transportation sector and chemical synthesis when it is combined with technologies like gas-to-liquid and hydrogen generation from syn gas. UCG is a process of converting coal insitu into combustible product gas by injecting a mixture of reactant gases like air/oxygen and/or steam directly into an ignited portion of underground coal seam. The reactor in UCG process is a high temperature cavity formed by the consumption of coal during the ignition process and it grows three dimensionally into the coal seam during the process. The coal present at the roof of the cavity undergoes

S. Bhaskaran et al. / Fuel 154 (2015) 326–337

327

Nomenclature A Cg Cp Deff DH H M V R T aij as,ij hT hTcav keff ky,cav t v

er q r

area (m2) gas concentration (kmol/m3) specific heat (kJ/kmol/K or kJ/kg/K) effective diffusivity (m2/s) heat of reaction (kJ/kmol) enthalpy (kJ/kmol) molecular weight (kg/kmol) volume (m3) rate of reaction (kmol/m3/s) temperature (K) stoichiometric coeff of ith gas species in jth reaction stoichiometric coeff of ith solid species in jth reaction heat transfer coefficient in rubble zone (kW/m3/K) heat transfer coefficient between void and roof surface (kW/m2/K) effective conductivity (kW/m/K) mass transfer coefficient between void and roof surface (m/s) time (s) velocity (m/s) radiation emissivity solid density (kg/m3) Stefan boltzman constant (kW/m2/K4)

different processes like drying, pyrolysis, gasification and combustion. These processes result in crack formation and further, its extension in the roof of the cavity may lead to mechanical disintegrity. This failure of the mechanical structure of coal present at the roof of the cavity may occur at different scales. They are a. subsidence or total collapse of roof due to falling of large chunks of coal (particle size greater than 10 cm) and b. falling of small pieces of coal or spalling. Fig. 1 shows a simplified schematic diagram of UCG depicting the spalling process. In this work, we present a laboratory experimental technique to approximately measure the spalling rate under the UCG-like conditions, and throw a light on the possible mechanism and its role in UCG. Indian lignites are chosen for this work as they are potentially important reserves for UCG [1] and available in large quantities on the north western coastal belt of India. Spalling is one of the cavity growth mechanisms during the process of underground coal gasification. The evidence of spalling comes from the post combustion excavation studies of UCG field trials as reported by Cena et al. [2] and also from observations made in small scale laboratory experiments as reported by Daggupati et al. [3]. During the early stages of UCG process, cavity growth mainly occurs due to the consumption of coal by chemical reaction [4]. Once the cavity gains a considerably large size, the cavity growth occurs due to spalling, mainly in vertical direction [5]. It is therefore necessary to incorporate the spalling behavior in the UCG process model that helps to predict the shape and size of a growing cavity and the product gas compositions at any given time. Even though different investigators [5–7] have used many ways to include spalling phenomenon in their process model, a more scientifically acceptable representation of spalling in UCG process model is not found due to its complexity and availability of limited experimental data [6]. Britten [5] considered spalling by using two failure parameters; a failure temperature Tf and a spalling length lf, in their process model. Spalling is simulated by redefining the coal cavity interface when the temperature at a length lf into the coal block exceeds Tf. This model assumes spalling to be a result of only a thermomechanical failure, and crack formation due to reaction is not

s / v

residence time (s) porosity gas flow rate (m3/s)

Subscripts 0

c cav df dry g i in j l s roof spall T vap void w wet

initial value cross-section void zone drying front dry zone gas phase species index inlet of a zone reaction index liquid phase solid phase cavity roof spalled particles in the rubble zone total vaporization void zone water wet zone

considered to be the possible cause. Furthermore, experimental validation for lf and Tf is not found in literature. Mortazavi et al. [7] proposed two hypothetical spalling mechanisms during the removal of moisture from a coal block i.e. strength loss model and tensile cracking model. In the strength loss model, spalling occurs due to extension of horizontal and vertical micro-cracks, which are formed due to the strength reduction of coal block caused by the destruction of equilibrium pore pressure on removal of moisture from the pores. The cracks can also be formed due to the extension of fissures already present in the coal block. The weakening of coal block is more predominant at the bedding planes causing horizontal cracks needed for detachment of coal pieces from the roof. In tensile cracking model, the vertical macro-cracks are formed by the shrinkage of coal on removal of moisture. The horizontal cracks, responsible for spalling, are either provided by gravity at the weakest portions such as bedding planes or the extension of already present fissures. It may be noted that their studies are restricted to only drying of coal and furthermore an experimental validation of the proposed spalling models is not available. In the present work, we use similar theoretical support to explain our results, in the context of UCG wherein, the coal block is subjected to pyrolysis and other chemical reactions, in addition to drying. In both the models, formation and extension of cracks play a vital role in the phenomenon of spalling. Su et al. [8] used acoustic emission and X-ray computed tomographic techniques to monitor and analyze the micro crack structure inside a heated coal specimen. Upadhye et al. [9] studied spalling experimentally by suspending a coal block in hot stream of nitrogen and monitoring the weight of the coal block. But they could not observe significant amount of spalling using their apparatus for the types of coal they have studied. According to Hettema et al. [10], spalling of sedimentary rocks happens due to the steam pressure at the pores on drying. However, for spalling to happen by this mechanism, the permeability of the rock has to be very low and hence this mechanism may not be applicable to coal as it acquires highly porous structure during the course of drying and pyrolysis of coal.

328

S. Bhaskaran et al. / Fuel 154 (2015) 326–337

Fig. 1. Schematic of UCG process [3].

From the forgoing discussion it is clear that there exists only limited work on mechanism and role of spalling in UCG. Furthermore, there is no information on spalling characteristics of Indian lignites, which are proved to be potentially important reserves for UCG [1,11]. In view of this, we undertake the present work and present a newly developed experimental technique and the related analytical procedures to determine spalling rate. The relevance, if any, of the spalling process is further identified by conducting gasification in laboratory under the conditions that are similar to UCG. These experimental results are explained by simulating a UCG process model and the simulation results are used to investigate the effect of spalling rate on UCG performance. An attempt is then made to throw a light on the possible mechanism of spalling by presenting the results on crack development and crack expansion in coal block at different stages of the heat treatment.

2. Apparatus and procedure The lignite samples of interest were collected from the Vastan lignite mines near Surat, Gujarat, India. The fresh lignite samples were cut using the proclaim machines from the coal seam at a depth of around 50 m and were transported to the laboratory site. The samples were preserved against moisture loss and cracking, by placing them under water. Initial experiments were conducted

Flare

with an objective to only examine the presence of spalling, wherein, a coal block bottom surface was heated by placing it over a channel through which hot gas was flown continuously. The channel was observed at the end of the experiment for spalled coal particles, if any. The details of this experimental setup and the results are explained elsewhere [12]. Once it was established that the lignite of interest has a tendency to spall, we performed systematic experiments to study spalling quantitatively, using a specially designed apparatus. The schematic of the experimental setup is shown in Fig. 2. An attempt was made to mimic UCG condition in laboratory. The main part of the apparatus was a furnace with a plate type heater at the bottom surface of its heating chamber. The heater is equivalent to the burning coal particles on the floor of the cavity in the actual UCG process. The dimensions of the heating chamber were 380  280  215 mm. The inlet and outlet ports for the gas were provided in the heating chamber just above the heating plate. A box type coal block holder made up of stainless steel (SS-410) was placed in the chamber, with its bottom face exposed to the hot flowing gas. The dimension of the coal block holder was 300  200  140 mm. The holder contained thermal insulation at all its internal faces with a thickness of 15 mm. The coal block, which was cut into a dimension suitable to fit into the holder, was then placed inside the holder. While cutting the coal block, the exposed surface was kept parallel to the bedding plane. This exposed surface is equivalent to the roof of the cavity in the actual

Data Logger

Furnace Lid Load cell

GC for gas analysis

Coal block Sampling valve Coal Block Holder Inlet

Tar collection box

Thermal insulation Heater Fig. 2. Schematic of the apparatus used for spalling studies.

329

S. Bhaskaran et al. / Fuel 154 (2015) 326–337

UCG process. The purpose of the thermal insulation between the coal block and holder walls was to heat the coal block mainly from its exposed bottom face. This holder was hung from a load cell that was placed outside the furnace, using a stainless steel rod, to monitor the weight of the coal block. This hanging rod was passed through an opening made at the centre of the lid of the furnace. Care was taken to prevent the load cell coming in contact with any solid object so as to have error-free results. The load cell reading was recorded at regular time intervals of 10 s. A blank run was conducted before starting the experiments with ceramic bricks placed in the holder to examine the load cell fluctuations due to the buoyancy effects, if any. The product gas was sampled from the outlet tube and analyzed on gas chromatograph for its composition and the remaining gas was flared up.

Fig. 4. Weight difference vs time plot. Conditions are same as that of Fig. 3.

2.1. Identification of spalling events 10

7000

Weight of coal block (g)

6000

4785

4730

5000

5.65

5.69

3660

4000

3580 9.54

9.59

3000

2000

0

2

4

6

8

10

12

14

16

Time (hours) Fig. 3. Weight of coal block vs time plot: pyrolysis experiment under nitrogen atmosphere and heating chamber bottom temperature 700 °C. Two spalling events are shown in magnified plots.

0 0

2

4

-10

Δw (g)

The spalling events during the experiment were identified from the discontinuities in the weight of coal block vs time profile, which was obtained from the load cell. One such representative profile is shown in Fig. 3. The initial weight of the coal block was around 6.8 kg. The furnace temperature was increased from room temperature to 700 °C and maintained at that value, under the continuous flow of nitrogen. The coal block, while getting heated from the exposed surface inward, loses its weight due to two distinct processes that are: spalling which is intermittent in nature and loss due to drying, pyrolysis and self-gasification, which are continuous with time. Spalling events thus caused discontinuities in the weight vs time profile in Fig. 3 and two of such representative discontinuities are shown as magnified insets. For ease in identification of discontinuities, weight differences at regular time interval of 10 s, were plotted against time, as shown in Fig. 4. In Fig. 4, each point represents Dw ¼ wiþ1  wi , where wi and wi+1 are the weights of coal block at times ti and ti+1. Separate set of experiments were also conducted by covering the exposed surface of coal block by a mesh so as to avoid detachment of spalled particles from the holder. The corresponding Dw vs time plot is shown in Fig. 5. Comparison of Figs. 4 and 5 allows us to define spalling event as the one when the weight drops were equal to or more than 5 g in the given time interval. The overall rate of spalling was then calculated by dividing the total amount of spalled coal by the total duration of the experiment. It is important to study the crack formation and its extension to understand the mechanism of spalling and thereby explain the

6

8

Time (h)

-20 -30 -40 -50

Fig. 5. Plot of weight difference vs time when no spalling occurred. Conditions are same as that of Fig. 3.

results of spalling experiments. The surface crack structure of dried and pyrolyzed lignite samples were examined using optical microscope. The setup used for the sample preparation in this case is explained elsewhere [12].

3. UCG process model for spalling apparatus The UCG process model used in the present work is first developed by Samdani et al. [13] for the growth of outflow channel in UCG process. It is based on compartment modeling approach i.e., the computational domain is divided into different sections and each section has three sub zones such as rubble zone representing the spalled coal particles on the heater plate, void zone representing the vacant space inside the heating chamber and roof zone representing the hanging coal block. The proposed compartment model for the spalling apparatus is shown in Fig. 6. This pattern of compartments is decided based on RTD studies by CFD [13,14]. The rubble zone is modeled as a PFR with number of CSTRs connected in series. It contains both gas phase and solid phase with gas phase consisting of six different gas species such as O2, H2O, H2, CH4, CO and CO2 and solid phase consisting of three different species such as coal, char and ash. Different reactions such as pyrolysis, heterogeneous char reactions and homogeneous gas phase reactions take place in the rubble zone. The volume of the rubble zone increases when a spalling event happens. In this model, rate of spalling is considered as an input parameter. The void zone is modeled as a CSTR consisting of only gas phase. Different gas phase reactions are considered here. Inside the coal block or in the roof zone there exist two distinct regions: dry coal and wet coal separated by a moving drying front. Heat transfer to the coal block is by radiation from the rubble zone and by convection from the void zone. The heat transfer inside the coal block is by conduction. The details of the governing equations in different

330

S. Bhaskaran et al. / Fuel 154 (2015) 326–337

Fig. 6. Proposed compartment model for spalling apparatus [13].

4. Results and discussion The proximate analysis of the lignite sample of interest is given in Table 1. This lignite contains large amount of moisture (43.3%) and considerable amount of volatile matter (25.5%). In the fixed coal block experiments, the coal block was heated through the exposed surface, in inert atmosphere, to examine its tendency to spall. When the heater temperature was maintained at 150 °C to conduct drying experiment, no significant spalling was observed from the coal block even after a long time of around 20 h. On the contrary, significant spalling was observed when pyrolysis or chemical reactions were conducted on the coal block at temperatures in the range of 700–900 °C during the post experimental examinations. This proved that the lignite used for the present study was prone to spalling. The heated surface of coal block after spalling and the spalled particles observed on the channel are shown in Fig. 7. Also, burning coal particles from the roof falling on the channel were observed through a glass window.

stream was a mixture of oxygen and steam with steam to oxygen molar ratio of 2.04 and the bottom heater temperature was raised from room temperature to 900 °C and maintained at that temperature throughout the experiment. In Fig. 8, carbon dioxide content decreases and hydrogen and carbon monoxide contents increase initially and then attain steady state conditions. This is due to

80

Gas composion (%)

zones, their boundary conditions, chemical reactions and their rate expressions and the simulation algorithm are given in Appendix A.

70

H2

CO

60

CH4

CO2

50 40 30 20 10 0 0

50

100

150

200

250

300

Time (min) Fig. 8. Product gas compositions for a gasification experiment at 900 °C and steam to oxygen molar ratio 2.04; Composition of oxygen is not shown in figure.

4.1. Product gas analysis 180

Table 1 Proximate analysis of lignite samples.

160 140

Spalling (g)

In the experiments described before, along with quantification of spalling rate, we also analyzed the product gas at regular time intervals. Fig. 8 shows the typical product gas composition in gasification experiments conducted in this equipment. The inlet

120 100 80 60 40 20

Value (%)

Moisture Volatile matter Fixed carbon Ash

43.3 25.5 26.5 4.7

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250

Components

Time (min) Fig. 9. Amounts of spalled coal particles accumulated in every 10 min interval. Experimental conditions same as that of Fig. 8.

Spalled particles on heater plate

Fig. 7. Coal block exposed surface and spalled particles on the channel after 18 h for a pyrolysis experiment at 700 °C. (Arrows show the gas flow directions.)

331

S. Bhaskaran et al. / Fuel 154 (2015) 326–337

Table 2 Model input parameters. No.

Parameter

Value

1 2 3 4 5 6

Number of compartments Number of CSTRs in rubble zone per compartment Step size in time Total simulation time Size of a volume element Initial permeability

10 1 2.5 s 10 h 0.125 cm3 10 Darcy

Spalling rate (kg/m2/hr)

30 25 20 15 10 5 0 1

2

3

4

5

6

7

8

9

10

11

12

13

Experiments Fig. 12. Rates of spalling for different coal blocks during combustion experiments.

Probability of spalling rate, %

the initial stabilization period during which temperature inside the heating chamber increases and more heat penetrates into the coal block enhancing the rates of pyrolysis, gasification and spalling processes. In other words, the system shifts from the initial combustion dominant phase to the one in which combustion, pyrolysis, gasification and spalling occur simultaneously. Towards the end of the experiment, carbon dioxide content increases again showing a decrease in the availability of coal inside the chamber. The amounts of coal particles spalled in every 10 min intervals are shown in Fig. 9. After comparing Figs. 8 and 9, it is clear that variations in amounts of spalled coal particles with time did not produce any significant fluctuations in product gas quality. The results are reproducible. The process model developed for the spalling apparatus was simulated by using the similar experimental and spalling conditions as input parameters. Other model specific input parameters

25 20 15 10 5 0 0-5

5-10

10-15

15-20

20-25

25-30

Rate of spalling (kg/m2/hr) Fig. 13. Variation in spalling rates over the different experimental runs.

Fig. 10. Comparison of experimental and simulation results for product gas calorific value.

Fig. 11. Product gas composition predicted by process model.

332

S. Bhaskaran et al. / Fuel 154 (2015) 326–337

Avg. CV (kJ/mol)

200

time-averaged product gas compositions observed in experiment (Fig. 8). A more detailed analysis of the model simulation results is available elsewhere [13].

180 160

4.2. Rate of spalling

140 120 100

5

10

15

20

Spalling rate (kg/m2/hr) Fig. 14. Simulation results showing variation in calorific value of product gas as a function of spalling rate.

are given in Table 2. Fig. 10 shows the comparison of experimental and simulation results for the product gas calorific values. It shows

ProbabilityðX1; X2Þ ¼

Several experiments were conducted with different coal blocks collected from the same field. The spalling rates observed are shown in Fig. 12. It is clear that there was a significant variation in the spalling rate measured for different coal blocks, under otherwise similar conditions. This variation is attributed to the heterogeneity in the coal seam. The rate of spalling in an actual UCG cavity is thus the net effect of spalling taking place over the entire surface exposed to the hot environment. Hence, it was thought appropriate to consider a distribution of rate over the surface of UCG cavity. The distribution in the rate of spalling over the surface is thus captured in probability of given spalling rate, which is calculated as,

number of experiments where the observed spalling rate is inðX1; X2Þ Total number of experiments

that the proposed model is able to predict the product gas calorific values in an UCG experiment. Fig. 10 also shows results of the simulation performed by imposing a hypothetical ‘‘no spalling’’ condition which corresponds to the absence of rubble zone on the floor of the heating chamber. It is clear from these results that spalling enhances the calorific value of the product gas to a large extent. The inferior performance in ‘‘no spalling’’ case is mainly due to high diffusional and conductive resistances for reacting gas molecules and heat, respectively, at the interior of the coal block when it is intact. The product gas composition predicted by the process model is shown in Fig. 11. The prediction of H2 composition is within ±10% deviation but CO composition is over-predicted and CO2 and CH4 compositions are under-predicted with respect to

where X1 and X2 are the lower and upper limits of the range of spalling rate under consideration. The distribution of rate over the surface is shown in Fig. 13. A relatively simple way to incorporate spalling in the process model is to consider the exposed surface averaged spalling rate while modeling the dynamic behavior of UCG cavity. Simulations of the spalling apparatus model were performed by varying the rate of spalling. The results of average calorific values of product gas at different rates of spalling are shown in Fig. 14. It shows that higher spalling rate enhances the gas calorific value till a critical spalling rate after which it does not have an effect. The reason is that, after this critical spalling rate, the surface available for reaction is large enough and the availability of reacting gas species becomes the limiting factor. It is this

(a)

(b)

(c)

(d)

Fig. 15. Microscopic images of exposed surfaces of coal monoliths, (a) fresh coal, (b) dried coal, (c) pyrolyzed coal, and (d) reacted coal.

333

S. Bhaskaran et al. / Fuel 154 (2015) 326–337

other cases spalling rate may control the performance thereby making impact on the product gas quality.

Table 3 Shrinking behavior of coal after drying and pyrolysis.

After drying After pyrolysis

Linear shrinkage (%)

Volumetric shrinkage (%)

8 30

27 65

4.3. Crack patterns

reason, why in the present case we observe composition and the calorific value of the outgoing gas not to be so sensitive to the fluctuations in the spalling rate. The lignite coal of interest is thus prone to giving sufficiently high spalling rate. However, in some

The microscopic analysis of crack patterns on a heated coal monolith was performed to understand the mechanism of spalling and thereby provide a theoretical support to the experimental results obtained in this work. Fig. 15 shows the crack pattern on a coal monolith after drying, pyrolysis and partial combustion. Numerous micro-cracks were evident after drying. Upon

Fresh coal block Hot cavity Gas in

Gas out

Vertical micro-cracks are formed due to capillary forces on initial stages of drying Gas out

Gas in Widening of cracks due to tensile stresses of shrinkage on further drying and pyrolysis Gas in

Gas out

Initiation of horizontal cracks.

Gas in

Gas out

Spalling by crosslinking of vertical cracks by horizontal cracks under gravity field. Gas in

Spalling

Fig. 16. Schematic of processes leading to spalling.

Gas out

334

S. Bhaskaran et al. / Fuel 154 (2015) 326–337

subsequent pyrolysis of the same monolith, it was found that new cracks were not formed but the existing cracks were widened due to the shrinkage of coal upon pyrolysis. The shrinkage factor of this coal was measured separately and is given in Table 3. It showed a higher shrinking behavior especially after pyrolysis. After partial combustion of the same monolith, it was found that chemical reactions removed the coal materials present in a crack and thus helped in complete separation of two portions of coal. Thus chemical reactions may enhance spalling, but its effects on spalling rate need to be studied further. The experiments were repeated and the same qualitative behavior was observed. From the analysis of crack pattern, it is clear that micro-cracks were formed during drying and same cracks were widened during pyrolysis of lignite. The moisture present in the coal block plays a major role in crack formation. Since this coal contains large amount of moisture (>40%), the initial micro crack development during drying process can be related to the crack development in clay type soil upon drying. It is attributed to the capillary stresses in the pores, which are partially filled with water during the initial stages of drying [15]. The nature of the stress is characterized as tensile [16]. The invading air exerts the pressure force, which depends on the pore dimension. All these stresses cause the loss of strength during drying and thus the crack development. The similar behavior is also observed in drying colloids and film formation [17,18]. The shrinkage factor shows that few macro cracks can also be formed due to shrinkage during drying and the corresponding failure of tensile strength. During pyrolysis the cracks are widened in accordance with the tensile cracking model due to higher shrinking tendency of lignite. Initially, vertical cracks are formed perpendicular to the surface exposed to high temperature. When the vertical cracks are sufficiently large, there is free convective flow of gas and this initiates the formation of horizontal cracks [19]. For spalling to happen, there should be interlinking of vertical cracks by horizontal cracks. The rate of spalling depends on the dynamics of this interlinking of vertical cracks. The schematic diagram shown in Fig. 16 describes the step by step process for the formation of cracks leading to an event of spalling. The horizontal cracks can also be provided by the extension of already existing weak regions or fissures or the weakening at the interlayer bonding of bedding planes in the coal block due to thermal degradation. But the availability of these weak regions or fissures may not be uniform over a large coal seam. This might be the reason for the variations in the rate of spalling when the experiment was conducted with different coal blocks. When a coal block is found to be heavily spalling, it indicates that the particular block is weaker. Another possibility is that one spalling event may induce another spalling, and as a result of this self accelerating process the overall spalling rate for that particular coal block remains higher.

5. Conclusion The qualitative and quantitative characterization of a lignite type coal was performed through laboratory scale experiments to determine its spalling tendency. Spalling in context of UCG was demonstrated by exposing one face of coal block to a hot stream of reactant gases. The rate of spalling under UCG-like conditions was estimated and a wide variation was observed for different coal blocks, which can be attributed to the heterogeneity in coal seam in terms of available fissure or weak regions. A compartment process model was proposed to explain the experimental results. The simulation results proved that the spalling influences the product gas calorific value only if the spalling rate is less than a critical value, but above which it does not have an influence on UCG performance. The microscopic analysis of the evolution of crack pattern helped understand the mechanism of spalling of a coal type with high moisture and volatile matter contents. The cracks were developed during the drying process by the destruction of equilibrium pore pressure (strength loss model) and these cracks were extended due to shrinkage of coal (tensile cracking model) during the pyrolysis process. We believe that the interlinking of these cracks, which was necessary for spalling to happen, was provided by the formation of horizontal cracks by the same way of that of vertical cracks or originally available fissures or weak regions or the bonding of different layers at the bedding planes. Appendix A This appendix provides the details of governing equations present in the model, chemical reactions and their rate parameters and the solution procedure. The Governing equations used in different sub zones of the model are following, A.1. Rubble zone gas phase The rubble zone gas phase was considered as CSTRs in series. The governing equations for a single CSTR is given below, Species mass balance:

  X n @C gi 1 m þ ¼ C gi;in  C gi aij Rj @t sspall min j¼1

ðA:1Þ

Initial condition, at t = 0, Cgi = Cgi,0. Energy balance:

  n n X X @T g 1 m C gi C pi ¼ C gi;in Hi;in  C gi Hi DH j R j  hT ðT g  T s Þ  @t s min i¼1 j¼1 ðA:2Þ

Table A.1 Chemical reaction kinetic parameters.

a b c

No.

Reaction

Representation

Ea

1 2 3 4 5 6

Char combustion Steam gasification CO2 gasification Methanation Water–gas-shift Gas phase oxidation

C + O2 ? CO2 C + H2O ? CO + H2 C + CO2 ? 2CO C + 2H2 ? CH4 CO + H2O ? CO2 + H2 H2 + 0.5O2 ? H2O CO + 0.5O2 ? CO2 CH4 + 2O2 ? CO2 + 2H2O

140,000 1.69E+7 0, 1 RPM 207,013 1.261E+8 0, 1 RPM 237,020 4.71E+7 0, 0.23 RPM 150,000 2.337E6 1 Volumetric 60,272 3E+7 0, 1 Volumetric Approximate rates as a function of gas temperature and oxygen concentration

Activation energy (J/mol). Pre-exponential factor (s1 or s1 atmN). Temperature and concentration exponents.

k0b

a , Nc

Kinetic model

Ref.

[20] [21] [5] [5]

S. Bhaskaran et al. / Fuel 154 (2015) 326–337

Fig. A.1. Algorithm for simulation of model.

335

336

S. Bhaskaran et al. / Fuel 154 (2015) 326–337

Initial condition: at t = 0, Tg = Tg,0. Volumetric flow rate:



sspall v ¼ v0 where

P Pn

j¼1 aij Rj M i

i

sspall ¼

P

V spall

m0

þ

P

i C gi;in M i

Boundary conditions : at V ¼ V roof ; Deff Aroof 

i C gi M i

¼ ky;cav ðC gi;v oid  C gi;roof Þ ðA:3Þ

.

Since thermal equilibrium was assumed between gas and solid phases, energy balance equation was solved only for solid phase. Roof dry zone solid phase: Species balance:

Species balance: n X @ qi ¼ M i as;ij Rj @t j¼1

ðA:4Þ

Initial condition: at t = 0, qi = qi,0. Energy balance:

qi C ps;i

i

@T s ¼ hT ðT g  T s Þ  @t 

@C gi @C steam ¼ 0 and  Deff Ad @V @V ¼ v df /ðqw;l  qw;g Þ=Mw

at V ¼ V d ; Deff Ad

A.2. Rubble zone solid phase

X

n X @ qi ¼ Mi as;ij Rj @t j¼1

DH j R j

X

qi C ps;i

j¼1

rer 4 4 ðT  T roof ÞAspall =V spall 2  r s

i

ðA:5Þ

Initial condition: at t = 0, Ts = Ts,0

  X n @T s @ @T s ¼ DH j R j  keff A2c @V @t @V j¼1

Boundary conditions : at V ¼ V 0 ; keff Aroof ¼

The void was considered as CSTR. Species mass balance:

ðA:6Þ Initial condition: at t = 0, Cgi = Cgi,0. Energy balance:

þ ky;cav ðC gi;roof  C gi ÞHi;roof Aroof =V cav ðA:7Þ

Initial condition: at t = 0, Tg = Tg,0. Volumetric flow rate: j¼1 aij Rj M i

P

þ

P

i C gi M i

V v oid

m0

i C gi;in M i Þ

ðA:8Þ

.

A.4. Roof model The roof zone consisted of a dry region and wet region separated by a moving drying front. In the wet zone, only energy balance equation was solved with conduction as the only mode of heat transfer. Roof dry zone gas phase: Since the characteristic time related to the variations in gas species composition was larger compared to that of solid species densities, a pseudo-steady state condition was assumed for gas species balance as given below.

  X n @C gi @ þ Deff A2c aij Rj ¼ 0 @V @V j¼1

Roof wet zone: In the wet zone, only energy balance equation was solved with conduction as the only mode of heat transfer.

qs C ps

v df

þ hT;cav ðT s;roof  T g ÞAroof =V cav

i

ðT 4s;spall  T 4s Þ þ hT;cav ðT v oid  T s Þ

  @T s @ 2 @T s kAc ¼ @V @t @V

ðA:12Þ

Boundary conditions: at V = VT, Ts = TT and at V = Vdf, Ts = Td. The velocity of the drying front, at which the water vaporization takes place, was calculated based on the heat balance across the drying front as given below,

  X n n X @T g 1 m  C gi C pi ¼ C gi;in Hi;in  C gi Hi DH j R j @t s min i¼1 j¼1

P Pn

rer 2  er

@T s @V

at V ¼ V df ; T s ¼ T df :

  X n @C gi 1 m þ ¼ C gi;in  C gi aij Rj þ ky;cav ðC gi;roof  C gi ÞAroof =V cav @t sv oid min j¼1

þ hT;cav ðT s;spall  T g ÞAspall =V spall

ðA:11Þ

Initial condition : at t ¼ 0; T s ¼ T s;0 :

A.3. Void zone

where sv oid ¼

ðA:10Þ

Initial condition: at t = 0, qi = qi,0. Energy balance:

n X

 hT;cav ðT s  T v oid ÞAspall =V spall

v ðsv oid ¼ v0

@C gi @V

ðA:9Þ

¼

1 /ðqw;l  qw;g Þ=M w

kd

@T @x xdf 

þ kdþ

DHv ap

@T @x xdf þ

! ðA:13Þ

The details of the chemical reactions considered in this model and their kinetic parameters are given in Table A.1. The algorithm for numerically solving the model was given in Fig. A.1. References [1] Khadse A, Qayyumi M, Mahajani S, Aghalayam P. Underground coal gasification: a new clean coal utilization technique for India. Energy 2007;32:2061–71. http://dx.doi.org/10.1016/j.energy.2007.04.012. [2] Cena RJ, Britten JA, Thorsness CB. Excavation of the partial seam CRIP underground coal gasification test site. 13th Undergr Coal Gasif Symp, Wyoming: Lawrence Livermore National Lab., CA (USA); 1987. [3] Daggupati S, Mandapati RN, Mahajani SM, Ganesh A, Sapru RK, Sharma RK, et al. Laboratory studies on cavity growth and product gas composition in the context of underground coal gasification. Energy 2011;36:1776–84. http:// dx.doi.org/10.1016/j.energy.2010.12.051. [4] Park KY, Edgar TF. Modeling of early cavity growth for underground coal gasification. Ind Eng Chem Res 1987;26:237–46. http://dx.doi.org/10.1088/ 1748-6041/6/2/025001. [5] Britten JA. Recession of a coal face exposed to a high temperature. Int J Heat Mass Transf 1986;29:965–78. http://dx.doi.org/10.1016/00179310(86)90196-1. [6] Perkins G, Sahajwalla V. A numerical study of the effects of operating conditions and coal properties on cavity growth in underground coal gasification. Energy Fuel 2006;20:596–608. http://dx.doi.org/10.1021/ ef050242q.

S. Bhaskaran et al. / Fuel 154 (2015) 326–337 [7] Mortazavi HR, Emery AF, Corlett RC, Lockwood WR. The effect of moisture on the structural stability of a coal cavity. J Energy Resour Technol 1986;108:246–53. http://dx.doi.org/10.1115/1.3231273. [8] Su F, Nakanowataru T, Itakura K, Ohga K, Deguchi G. Evaluation of structural changes in the coal specimen heating process and UCG Model experiments for developing efficient UCG systems. Energies 2013;6:2386–406. http:// dx.doi.org/10.3390/en6052386. [9] Upadhye RS, Field JE, Fields DB, Britten JA, Thorsness CB. Experimental investigation of coal spalling. 13th Undergr. Coal Gasif. Symp., Wyoming: Lawrence Livermore National Lab., CA (USA); 1986. [10] Hettema MHH, Wolf KHAA, De Pater CJ. The influence of steam pressure on thermal spalling of sedimentary rock: theory and experiments. Int J Rock Mech Min Sci 1998;35:3–15. http://dx.doi.org/10.1016/S01489062(97)00318-5. [11] Patro DN, Sharma DK, Singh RP. Seismic for exploration of lignites being targeted for UCG. 7th Int Conf Expo Pet Geophys, Hyderabad: Society of Petroleum Geophysicists, India; 2008, p. 383. [12] Bhaskaran S, Ganesh A, Mahajani S, Aghalayam P, Sapru RK, Mathur DK. Comparison between two types of Indian coals for the feasibility of underground coal gasification through laboratory scale experiments. Fuel 2013;113:837–43. http://dx.doi.org/10.1016/j.fuel.2013.05.080. [13] Samdani G, Aghalayam P, Ganesh A, Sapru RK, Lohar BL, Mahajani S. A process model for Underground coal gasification- Part-II Growth of outflow channel.

[14]

[15] [16] [17]

[18]

[19] [20]

[21]

337

Unpublished results 2015. . Samdani G, Aghalayam P, Ganesh A, Sapru RK, Lohar BL, Mahajani S. A process model for Underground coal gasification- Part-I Cavity growth. Unpublished results 2015. . Mitchell JK, Soga K. Fundamentals of soil behaviour. 3rd ed. New York: John Wiley and Sons Inc; 2005. Roberts CC, Francis LF. Drying and cracking of soft latex coatings. J Coatings Technol Res 2012;10:441–51. http://dx.doi.org/10.1007/s11998-012-9425-7. Schubert H. Capillary forces – modeling and application in particulate technology. Powder Technol 1984;37:105–16. http://dx.doi.org/10.1016/ 0032-5910(84)80010-8. Martinez CJ, Lewis JA. Shape evolution and stress development during LatexSilica film formation. Langmuir 2002;18:4689–98. http://dx.doi.org/ 10.1021/la0114833. Bouma J. Soil morphology and preferential flow along macropores. Agric Water Manag 1981;3:235–50. http://dx.doi.org/10.1016/0378-3774(81)90009-3. Mandapati RN, Daggupati S, Mahajani SM, Aghalayam P, Sapru RK, Sharma RK, et al. Experiments and kinetic modeling for CO2 gasification of indian coal chars in the context of underground coal gasification. Ind Eng Chem Res 2012;51:15041–52. http://dx.doi.org/10.1021/ie3022434. Tomita A, Mahajan OP, Walker PL. Reactivity of heat-treated coals in hydrogen. Fuel 1977;56:137–44. http://dx.doi.org/10.1016/0016-2361(77)90133-8.