- Email: [email protected]
Modelling and parametric investigations on spontaneous heating in coal pile
Fuel 176 (2016) 181–189
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Modelling and parametric investigations on spontaneous heating in coal pile Juwei Zhang a,⇑, Wonyoung Choi a, Takamasa Ito a, Katsumi Takahashi a, Masahiro Fujita b a b
Research Laboratory, IHI Corporation Co., Ltd., 1 Shin-Nakahara-cho, Isogo-ku, Yokohama, Kanagawa 235-8501, Japan IHI Transport Machinery Co., Ltd., 8-1, Akashicho, Chuo-ku, Tokyo 104-0044, Japan
a r t i c l e
i n f o
Article history: Received 28 October 2015 Received in revised form 16 February 2016 Accepted 19 February 2016 Available online 27 February 2016 Keywords: Coal Spontaneous Combustion Moisture Simulation
a b s t r a c t To predict the spontaneous heating of coal pile by computational fluid dynamics (CFD) is very meaningful for preventing the coal pile from self-ignition. A two-dimensional (2-D) numerical model, which could be simply implemented in FLUENT, was developed in this study. The chemical kinetic parameters of lowtemperature oxidation of coal, which were expressed by the outer surface area of particles, were well measured in experiments. The coal piles under different conditions were calculated by using the developed model and measured kinetic parameters, in order to make clear the effects of some important factors. The simulation results indicated the evaporation of moisture from coal played a critical role. The pile height, coal type, wind velocity, and heat loss from bottom have significant effects on the process of spontaneous heating of coal pile. According to these effects, in order to increase the heat loss of pile and then effectively inhibit the self-ignition, the low pile height, good ventilation surrounding the pile, and ground material with high thermal conductivity below the pile should be used. The model developed in this model is expected to become a reliable tool to predict the spontaneous heating of coal pile. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Spontaneous heating of low-rank coal stockpile can lead to the waste of coal source and even life-threatening fires. In order to predict the behavior of the spontaneous heating, many mathematical models have been developed. These models can be generally classified as one-dimensional (1-D) [1–5] or two-dimensional (2-D) [6–7] models. Although the 1-D models have been developed and gradually improved for many years, they are still far from the real conditions. For example, a uniform gas velocity throughout the pile is usually assumed in the 1-D model, which is obviously different from reality condition. Furthermore, it is difficult to well simulate the effects of wind in surrounding air by using the 1-D models. However, the 2-D model is considered to be capable of catching the main features of real spontaneous heating process. Moghtaderi et al. [6] focused on the effect of wind flow on self-heating of coal stockpiles by a 2-D model. The calculations show that the wind flow can significantly alter the dynamics of flow field inside the pile and thereby affects the self-heating process. Akgun and Essenhigh [7] investigated the effects of pile height, pile angle, particle diameter and coal moisture with a 2-D model. Their work shows
⇑ Corresponding author. Tel.: +81 45 759 2867; fax: +81 45 759 2207. E-mail address: [email protected] (J. Zhang). http://dx.doi.org/10.1016/j.fuel.2016.02.059 0016-2361/Ó 2016 Elsevier Ltd. All rights reserved.
that the spontaneous ignition can occur if the height exceeds some critical value with sufficient time. Nevertheless, it seems to be difficult to apply these models to the other cases. One of the reasons is that most of these models were developed by in-house code. Although some models can be implemented in commercial CFD software, they are still difficult to be used. For example, in these models, as an important chemical kinetic parameter describing the reaction rate of coal oxidation, the pre-exponential factor always has a unit of s1, which should be more reasonable to be a surface-basis unit for a gas–solid reaction. In this study, a reliable 2-D model for the spontaneous heating of coal pile was built combined with the accurate measurement of coal oxidation rate. This model can be easily implemented in commercial CFD software. In addition, detailed parametric investigations were performed by applying this model to explore the effects of some important factors.
2. Experimental and modelling 2.1. Experimental To obtain the accurate low-temperature oxidation rate of coal is very critical for the modelling of spontaneous heating. Fig. 1 shows the measuring system, which mainly consists of a heating reactor
182
J. Zhang et al. / Fuel 176 (2016) 181–189
Nomenclature a oxygen reaction order (s) A0, A1, A2 pre-exponential factor (consistent unit) C inertial resistance factor (m2) Cp thermal capacity (J kg1 K1) d particle diameter, or diameter of main oxidation zone (m) E activation energy (kJ/mol) h height from ground (m) H heat (J=kgH2 O for evaporation, or J=kmolO2 for oxidation) Hu Air humidity (%) J mass transfer rate (mol m2 s1) L length of pile bottom (m) m mass in a cell (kg); M evaporation rate (kgH2 O m3 s1) n speed exponent (–) Q transfer rate of heat (W/m2) R reaction rate (kmolO2 kgcoal s1) or gas constant (J mol1 K1) S specific surface area (kg/m2), source term (J m3 s1 for energy) or (kg m2 s2 for momentum) Dt time step (s) T temperature (K) DT temperature difference (K) U wind velocity (m/s) v gas velocity (m/s) ~ v vector of gas velocity (m/s)
for coal oxidation and a gas chromatograph (GC: GC3200, GL science) for measuring O2 concentration. One sub-bituminous coal (coal A) and one bituminous coal (coal B) were used. The properties for the two coals are given in Table 1. Before the measurement, about 1.2 g of coal sample with particle size of 45–75 lm (mean diameter dp = 53 lm) was put into a reacting tube (stainless steel, 3/8 in. in diameter), which was placed in the heating reactor. The mixture gas of O2 (21 vol.%) and He with the flow rate of 0.015 l/ min (STP) was passed through the reacting tube and reacted with the coal particles. The porosity of coal sample and residence time of mixture gas in the reacting tube were about 0.43 and 0.51 s, respectively. The heating reactor was heated electrically from 30 to 250 °C at a heating rate of 1 °C/min. Simultaneously, the O2 concentration variations were measured by the GC and recorded by a computer. One measurement was repeated for six times. Finally, the changes of O2 consumption rate or the oxidation rate of coal roxi
DV x X DX
a
b
e
k
q l
cell volume (m3) characteristic length (m) mole fraction (–) thickness for thermal conduction (m) permeability (m2) thermal expansion coefficient (–) porosity (–) thermal conductivity (W m1 K1) density (kg/m3) viscosity (kg m1 s1)
Subscripts and superscripts 0 atmosphere boundary layer, or ambient condition c coal eff effective evap evaporation of moisture f fluid moisture H2O i coordinate x, y, or z loss heat loss oxi oxidation O2 oxygen p particle rise temperature rise s solid sat saturation status
Then, the Arrhenius plot and the corresponding kinetic parameters (E and A) could be obtained. However, our previous study [8] showed that the particle size had a significant effect on the oxidation rate of coal, as shown in Fig. 2. It can be seen that the relationship between ln rO2 and dp at different temperatures for two bituminous coals can be described with a straight line with an inclination of about 1/3. In addition, the reaction order for oxygen concentration is around 0.5 [8]. Thus, the Eq. (1) can be rewritten as 1=3
roxi ¼ A1 dp
E X 0:5 exp O2 RT
ð3Þ
or be expressed with outer surface area of coal particles,
E 0:5 roxi ¼ A2 S1=3 X exp O2 p RT
ð4Þ
1
ðmolO2 kgcoal s1 Þ with temperature were obtained. The relationship between the O2 consumption rate and temperature can be expressed with an Arrhenius type of equation, as follows
E r oxi ¼ A exp RT
ð1Þ
Then
E 1 ln roxi ¼ þ ln A0 R T
ð2Þ
O2 (21%)/He
so that the pre-exponential factor (A1 or A2) can be unified for particles with various sizes. The obtained E and A2 for coal A and B are also given in Table 1. During the low-temperature oxidation of coal, physical adsorption and chemisorption occur together. However, physical adsorption is a reversible process which plays a major role only at very low temperature. In many studies, the physical adsorption was usually ignored. Furthermore, the heat of adsorption for physical adsorption is between 6.7 and 20.9 kJ=molO2 , while that of chemisorption is about 300 kJ=molO2 [9]. In this study, only the data in the temperature above 60 °C with a good linearity was used to calculate the kinetic parameters, which was considered to be the kinetics of only the chemisorption. 2.2. Modelling
Reacting tube for coal oxidization
Heating reactor
GC
Fig. 1. Measuring system for oxidation rate of coal.
PC
ANSYS FLUENT 15.0 was employed to simulate the spontaneous heating of coal pile. The coal pile was modeled as a 2-D triangle, as shown in Fig. 3. One part of the surrounding space around the pile
183
J. Zhang et al. / Fuel 176 (2016) 181–189 Table 1 Properties of coals for measurement. Sample
Type
Proximate analysis (wt.%)
Coal A Coal B
Sub-bituminous Bituminous
Moisture
Volatiles
Fixed carbon
Ash
12.6 7.0
32.2 36.5
31.5 42.4
23.7 14.1
0.5
A2 ðmolO2 ðm kgcoal Þ
0.25 0.12
0.85 0.90
44.18 54.68
21.97 307.54
s1 Þ
dp e3 150 ð1 eÞ2
ð7Þ
C¼
3:5 ð1 eÞ dP e3
ð8Þ
-1.5
The natural convection was considered with the Boussinesq model which treats gas density as a constant value in all solved equations, except for the buoyancy term in the momentum equation:
-2
--- Bituminous coal 1 Bituminous coal 2 -1
-0.5
0
ðq q0 Þg ¼ q0 bðT T 0 Þg
0.5
The energy equation in the porous medium can be expressed as
dp=1 mm
@ ðð1 eÞqs C p;s þ eqf C p;f Þ þ eqf C p;f ð~ v rTÞ @t ¼ keff ðr rTÞ þ Soxi þ Sevap
Fig. 2. Effect of particle size on oxidation rate of coal.
was also included into the calculation domain, in order to take into account the effect of wind flow. The whole domain has the dimension of 10L 2L (L is the length of pile bottom). A vertical velocity profile for the wind was used according to the power law equation for the atmospheric boundary layer [10], as follows
n h h0
ð5Þ
In this study, h0 and n were chosen as 300 m and 0.15, respectively, which are the typical values for a terrain of open country with low scrub or scattered tress. A typical gas velocity distribution around the pile is also shown in Fig. 3. The velocity distributions before and after the pile are obviously asymmetrical, which is believed to have a significant effect on the flow and temperature fields in the pile. The coal pile and surrounding space were simulated as a porous media and a fluid domain, respectively. The equations for continuity, momentum, species, and energy were solved for both domains of the coal pile and surrounding space. The effects of natural convection and turbulence (realizable j-e turbulence model) were included in the model. The porous medium was modeled by adding a momentum source term to the standard fluid momentum equations, as follows
1 1 Si ¼ lv i þ C qf jv jv i 2 a
ð9Þ
ð6Þ
where the permeability a (1/a is called the viscous resistance coefficient) and inertial resistance coefficient C were calculated by the following equations, respectively.
keff ¼ ekf þ ð1 eÞks
ð11Þ
The energy source from coal oxidation was calculated by
Soxi ¼ r oxi qc ð1 eÞHoxi
ð12Þ
The energy source from the evaporation of moisture from coal was calculated by
Sevap ¼ Hevap M evap
ð13Þ
It is assumed that the heat from coal oxidation is preferentially used for the evaporation, until the moisture in the gas mixture reaches status of saturation. The rate of moisture evaporation Mevap was limited by both the available heat from coal oxidation and the moisture saturation extent in bulk gas mixture. The Mevap determined by the heat from coal oxidation is
Mevap;1 ¼ Soxi =Hevap
ð14Þ
Wall Wind
0 Gas velocity
Coal pile H
θ
L θ=40
Wall 10L
Fig. 3. Schematic diagram of geometry and typical gas velocity distribution.
2L
max
Inlet
ð10Þ
where it is assumed that gas and solid have the same temperature. The forced convection heat transfer on the surface of coal pile is not considered. The heat transfer coefficient of forced convection was calculated and compared with the heat transfer coefficient of thermal conduction, as shown in Fig. 4. It can be seen that the heat transfer coefficient of convection is far higher than that of thermal conduction even at low velocities. This indicates the resistance of heat transfer due to the convection on the surface of coal pile can be ignored. The heat generated in the porous medium is mainly transferred to the surrounding environment by thermal conduction. The effective thermal conductivity of the porous medium was calculated by
Outlet
lnrO2 (mlO2 g-1 min-1)
-1
lndp (mm)
U ¼ U0
E (kJ/mol)
a¼
-0.5
-3 -1.5
H/C
2
0
-2.5
2=3
O/C
184
J. Zhang et al. / Fuel 176 (2016) 181–189
humidity, heat loss from pile bottom, initial temperature of pile and wind velocity could be investigated.
Convection heat transfer coefficient hconv. (W/m2K)
20 Calculated with a horizontal plate with characteristic length of 1 m
16 12
3. Simulation results
8 3.1. Dimensionless analysis
4 Conduction
0
In order to analyze the roles of convection and diffusion in the heating transfer in the pile, Péclet number for heat transfer was used in this study. It can be defined as
-4 0
1
2 3 4 5 Free wind velocity U (m/s)
6
Fig. 4. Comparison of heat transfer coefficient between forced convection and thermal conduction at difference wind velocities.
The Mevap determined by the extent of moisture saturation extent in gas mixture is
M evap;2 ¼ ðmH2 O;sat mH2 O Þ=ðDt DVÞ
ð15Þ
where mH2 O;sat and mH2 O are mass of moisture in a calculated cell at saturation and present status, respectively. It should be noted that mH2 O;sat is a function of temperature. The minimum value of the Mevap,1 and Mevap,2 was used as the final Mevap. The moisture in the coal in every cell was updated in every time step after Mevap was determined. In this study, the grid size in the coal pile and time step were 20–60 mm and 300 s, respectively. The study of Moghtaderi et al. [6] has showed that the grid size of 250 mm and time step of several days were numerically accurate enough for the simulation of spontaneous heating of coal pile. In this study, the calculation with time step of 10 s was also performed and the result almost showed no difference to that with time step of 300 s. The 2nd order upwind discretization scheme was used for the equations of continuity, momentum, species, and energy. The main models and parameters used in the simulation are listed in Table 2. The settings for different cases are given in Table 3. It should be noticed that the practical objective of this simulation is to predict the spontaneous heating of coal piles, which is formed with the fallen fine coal particles from the belt conveyor, and then to determine how often the coal pile should be cleaned. Thus, coal piles with fine particles and small pile size were adopted for simulation. However, the models developed in this study can be used for predicting the spontaneous heating of any coal pile. Via the calculation of the cases listed in Table 3, the effects of pile height, coal type, coal particle size, air
Peh ¼
qg C pg V g x
ð16Þ
kc
The main heat transfer mechanism is convection when Peh 1, and thermal conduction when Peh 1. Peh in all cases are listed in Table 4. It can be seen that except case 5, all the other cases have the Peh obviously lower than 1, which indicates the thermal conduction is the main mechanism for heat transfer. The heat from coal oxidation is not only used for moisture evaporation (Qeva) and heating of coal and gases in the pile (Qrise), but also for the heat lost to the environment (Qloss), i.e.,
Q oxi ¼ Q eva þ Q rise þ Q loss
ð17Þ
If Peh 1, the main form of heat loss is the thermal conduction, which can be simply expressed by
Q loss ¼ k
DT DX
ð18Þ
According to Eq. (18), the heat loss Qloss is low when DX is large. The ignition in the pile always occurs at the position, where heat is accumulated easily and/or difficult to be dissipated. Thus, the ignition position generally has the largest DX and thereby lowest Qloss, if sufficient oxygen available for the oxidation exists in the pile. In this study, Damköhler number Da, expressing the ratio of the oxidation rate to the transportation rate by convection or diffusion, was also used to analyze the simulation results. Fig. 5 shows the vectors of gas velocity in coal piles. It can be seen that in all the cases using dp = 0.025 mm, the gas in the pile flows outward because of very high resistance (Fig. 11), while in the case 5 with dp = 2.5 mm, the gas flows inward. This indicates, in all the cases except case 5, the O2 can enter into pile only by diffusion rather than convection. The typical diffusion and consumption modes of O2 in coal pile with dp = 0.025 mm is showed in Fig. 6. In order to calculate Da simply, the main oxidation zone is simplified into a half circle with diameter of d and diffusion occurs with the O2
Table 2 Input parameters for simulation of the baseline case (case 3 in Table 3). Item Oxidation rate (coal A) Coal properties (coal A)
Heat generation Surrounding air
Frequency factor A2 Activation energy Heat capacity Mean diameter Apparent density Thermal conductivity Porosity Initial moisture content Oxidation Evaporation Temperature Humidity Free wind velocity
Unit 2=3
molO2 ðm kgcoal Þ kJ/mol J/(kg K) mm kg/m3 W m1 K1 – wt.% J=kmolO2 kJ=kgH2 O °C % m/s
s
1
Value
Source
21.97
Measurement in this study
44.18 1046 0.025 1000 0.337 0.43 12.6 3.0 108 2260 25 50 0.3
Constant Assumption
185
J. Zhang et al. / Fuel 176 (2016) 181–189 Table 3 Boundary conditions for different cases. Case
Coal
Pile height H (m)
Particle diameter dp (mm)
Wind velocity U (m/s)
Air humidity Hu (%)
Initial Temp. T0 (°C)a
Setting of pile bottom
1 2 3 4 5 6 7 8 9 10 11 12
A A A B A A A A A A A A
0.42 1.25 2.1 2.1 2.1 1.25 1.25 2.1 2.1 2.1 2.1 1.25
0.025 0.025 0.025 0.025 2.5 0.025 0.025 0.025 0.025 0.025 0.025 0.025
0.3 0.3 0.3 0.3 0.3 0.01 3 0.01 3 0.3 0.3 0.3
50 50 50 50 50 50 50 50 50 30 50 50
25 25 25 25 25 25 25 25 25 25 25 30
Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic With heat lossb Adiabatic
Note: The changed conditions are written in the font of bold and italic. a The initial temperature of pile is same with that of environment. b It is assumed that the ground temperature is 25 °C and heat transfer coefficient is 10 W/(m2 K).
Vg (10
1 2 3 4 5 6 7 8 9 10 11 12
0.21 2.8 19 1.1 73 3.0 2.0 13 4.1 17 3.9 3.5
m/s)
x (m)
Peh
0.32 0.96 1.6 1.6 1.6 0.96 0.96 1.6 1.6 1.6 1.6 0.96
0.00062 0.024 0.28 0.016 1.1 0.026 0.018 0.18 0.06 0.25 0.057 0.031
3e-5
0 Velocity (m/s) (a) Case 3, dp=0.025 mm, 14 days
(a) 6
CO2,0 d W
4
XO2,0 = 10 vol.% XO2,s = 21vol.% d = 0.1W
2 0
0
4
Da=1
8
12
16
8
(b)
H=0.42 m H=1.25 m H=2.1 m
6 4
XO2,0 = 10 vol.% d = 0.1W
2
Da=1
0 10
13
16
19
22
O2 fraction on pile surface XO2,s (vol.%) Heat from oxidation Qoxi (W)
(Vectors in other cases with dp=0.025 mm are similar with case 3)
1.6e-4
0 Velocity (m/s)
8
Pile height H (m) Damköhler number Da
Case
6
Damköhler number Da
Table 4 Péclet numbers in different cases.
(b) Case 5: dp=2.5 mm, 14 days
Fig. 5. Vectors of gas velocity in coal piles.
3
(c) H=0.42 m H=1.25 m H=2.1 m
2
1
0 10
13
16
19
22
O2 fraction on pile surface XO2,s (vol.%)
21
Diffusion
Wind velocity U (m/s)
XO2,0 13 O2 (vol.%) Consumption
Fig. 7. Changes of Damköhler number (a) and heat from oxidation (b) with O2 concentration on surface of coal pile with the assumption of C O2;o ¼ 10%.
d W
Fig. 6. Typical manners of diffusion and consumption of O2 in coal pile.
concentration difference between pile surface and center of main oxidation zone (Fig. 6, right), based on a typical O2 distribution in the pile (Fig. 6, left). Thus, Da can be calculated by
r oxi pd =8 r oxi d=4 ¼ JO2 pd=2 DO2 ðC O2 ;s C O2 ;0 Þ=x 2
Da ¼ ¼
r oxi xdRT 4DO2 PðX O2 ;s X O2 ;0 Þ
ð19Þ
The oxidation rate is limited by chemical kinetics if Da < 1 and by O2 diffusion if Da > 1. Meanwhile, the heat generated from the main oxidation zone Qoxi can be calculated by (roxi pd2/8) Hevap if Da < 1 and by ðJ O2 pd=2Þ Hevap if Da > 1, respectively.
J. Zhang et al. / Fuel 176 (2016) 181–189
Maximum temperature ( )
186
100 75
80
Stage 2: constant temp.
60 40 20
Case 3
0
0
4
8
12
16
20
Time (day) Fig. 8. Maximum temperature in coal pile as a function of time in baseline case.
It is reasonable to assume that x and d are in direct proportion to the pile size (W or H), so that Da changes with W2. Under the assumption of d = 0.1W, X O2;0 ¼ 10%, and X O2 ;s ¼ 21%, the change of Da with pile height can be calculated and the result is shown in Fig. 7(a). It can be seen that Da increases monotonously with the increasing of pile height and the oxidation rate is limited by O2 diffusion (Da > 1) when the pile is higher than 6 m. In fact, the O2 concentration at the surface of pile ðX O2 ;s Þ maybe lower than 21% due to the limitation of mass transfer rate of O2 in the bulk gas around the pile. The increasing of wind velocity U must lead to the increasing of X O2 ;s until X O2 ;s reaches 21%. Under the assumption of d = 0.1W and X O2 ;0 ¼ 10%, the change of Da and Qoxi in the main oxidation zone for three piles with different heights with the increasing of X O2 ;s can be calculated and the results are shown in Fig. 7(b) and (c), respectively. It can be seen that for the pile of H = 0.42 m always has a Da lower than 1, the pile of H = 1.25 m has a Da lower than 1 if X O2 ;s > 11%, and the pile of H = 2.1 m has a Da lower than 1 if X O2 ;s > 13%. As shown in Fig. 7 (c), for the pile of H = 2.1 m, Qoxi is first increased (Da > 1) and then keeps constant (Da < 1), as the increasing of X O2 ;0 (or U). The former stage with increasing Qoxi gradually shortens as the deceasing of pile height until completely disappears for the pile of H = 0.42 m. Although the heat loss Qloss is increased as the increasing of U (the surface of pile is cooled), the temperature in the main oxidation zone may be increased due to the increasing of Qoxi in the former stage. However, the temperature in the main oxidation zone deceases as the increasing of U, because Qloss is increased and Qoxi has no change in the later stage. 3.2. Results of baseline case The simulation results for the baseline case (case 3 in Table 3) are shown in Figs. 8 and 9. According to Eq. (18), the ignition occurs
at the middle position which has the largest DX and lowest Qloss in the horizontal direction, as shown in Fig. 9. In the vertical direction, the ignition position must be at the adiabatic pile bottom, where Qloss is minimal (zero) along the vertical direction. A typical curve of maximum temperature rise in the spontaneous heating of coal pile was obtained, as shown in Fig. 8. According to this curve, three different stages exist in the process of spontaneous heating. In the 1st stage (temperature-rise stage), the maximum temperature is gradually increased, Qoxi obviously exceeds Qeva and Qloss, and the excessive heat can be used for heating the pile. In this stage, the Mevap,2 calculated by Eq. (15) is obviously lower than the Mevap,1 calculated by Eq. (14), because of low temperature, thus, the Mevap,2 is adopted as the evaporation rate. The amount of evaporated moisture in the first stage is very small, which can be seen from the 4th days moisture content in coal in the 4th day, as shown in Fig. 9. As the temperature gradually rises, the moisture content in saturation mH2 O;sat and Mevap,2 are rapidly increased. When the Mevap,2 exceeds Mevap,1, Mevap,1 is used as the evaporation rate, and this point is defined as the beginning of the 2nd stage (constant-temperature stage). In this stage, all of the heat from oxidation is used for evaporation and offsetting the heat loss, so that no heat can be available for heating the pile, therefore the maximum temperature is constant at this stage (around 75 °C in Fig. 8). As the evaporation further proceeds, the moisture content in the coal is gradually decreased and the position with the lowest Qloss and highest temperature in the pile is dried firstly (see moisture content in coal in the 14th day shown in Fig. 9). At this dried position, all of the heat from oxidation can be used for heating the pile and offsetting the heat loss. The temperature is increased rapidly, and finally spontaneous ignition occurs, which is the 3rd stage: ignition stage. From Fig. 9, it can be also seen that the temperature distribution in the coal pile is asymmetric and the position with the maximum temperature (ignition position) is a little closer to the left side because of the asymmetrical distribution of wind velocity around the coal pile, as shown in Fig. 3. 3.3. Parametric investigations The effects of pile height, coal type, coal particle size, wind velocity, air humidity, heat loss from pile bottom, and wind velocity on the temperature rise in coal pile are shown in Fig. 10(a)–(h), respectively. According to Fig. 10, it can be seen that pile height, coal type, heat loss, initial temperature of pile and wind velocity have obvious effects on the spontaneous heating process of coal piles, while the effects of particle size and air humidity are not so obvious.
80
25 12.6
0 Moisture in coal (wt.%)
4 days (in stage 1)
12 days (in stage 2)
Fig. 9. Changes of temperature and coal moisture content distributions.
14 days (in stage 3)
J. Zhang et al. / Fuel 176 (2016) 181–189
187
Fig. 10. Effects of pile height (a), coal type (b), particle size (c) air humidity (d), heat loss of bottom face (e), initial temperature (f) and wind velocity (g and h) on temperature rise in coal piles.
The effect of pile height is shown in Fig. 10(a). As the pile height is increased, the time needed to reach ignition is increased. At the pile height of 0.42 m, ignition doesn’t happen. According to Eq. (18), this is mainly attributed to the fact that DX at the potential position of ignition of pile becomes larger and Qoxi becomes lower, as pile height is increased. In order words, the heat is easier to accumulate in the larger pile. Although increasing the pile height can increase the O2 diffusion rate, it has little effect on Qoxi, because the oxidation reaction is tends to be limited by chemical kinetics for piles with heights 0.422.1 m (Da < 1, see Fig. 7). The effect of pile height has a very important practical meaning. As mentioned above, one of the objectives of this study is to determine the frequency of cleaning the coal pile. The result of Fig. 10(a)
indicates that the pile height of coal A should be maintained lower than 1.25 m and the pile should be cleaned in about two months. The effect of coal type is shown in Fig. 10(b). The spontaneous ignition occurs in pile of coal A within 15 days, but doesn’t occur in the pile of coal B. As expected, this indicates that the spontaneous ignition is easier to occur for the coal with lower rank. The effect of particle size is shown in Fig. 10(c). To increasing the mean particle diameter with 100 times (from 0.025 mm to 2.5 mm) leads to little change of temperature-rise. The effects of particle size on the viscous resistance coefficient and specific surface area of particles are shown in Fig. 11. It can be seen that although the increase of particle diameter from 0.025 mm to 2.5 mm decreases the resistance, but it also significantly decreases
80
1.E+12 1.E+10
60
1.E+08 40
1.E+06 1.E+04
20
1.E+02 1.E+00 0
3
5
8
10
0
Specific surface area m2/kg
J. Zhang et al. / Fuel 176 (2016) 181–189
Resistance coefficient 1/α (m-2)
188
Particle diameter mm Fig. 11. Changes of resistance coefficient and specific outer surface area with particle diameter.
Case 11: with heat loss of pile bottom, 14 days 80
25 Temp. (
time needed to reach the ignition is obviously increased. Fig. 12 shows the distributions of temperature and moisture content in the coal for case 11 considering the heat loss. It can be seen that the ignition position with dry coal moves upside, compared with the adiabatic case shown in Fig. 9, because the position of the lowest Qloss also moves upside, when heat loss from the bottom exists. The effect of initial temperature of pile (or the environment temperature) in the pile is shown in Fig. 10(f). It can be seen that as the initial temperature increases from 25 to 30 °C, the time needed to reach ignition is decreased. Similar result was also obtained by Schmal et al. [1]. The delay of temperature rise at low initial temperature is mainly due to the very low reaction rate of coal oxidation. The effect of wind velocity in the pile with height of 1.25 and 2.1 m are shown in Fig. 10(g) and (h), respectively. It can be seen that the effect of wind velocity is different for piles with different height. It can be easily explained by the Fig. 7(c) and the discussion in section ‘‘dimensionless analysis”. In the smaller pile (H = 1.25 m), O2 diffusion in the pile is always high enough for supplying O2 for oxidation under different velocities (Da < 1, Fig. 13), and the pile is in the former stage in which increasing U leads to little change of Qoxi but only the increasing of Qloss (Fig. 7c). Therefore, the time needed to reach ignition is monotonously increased as wind velocity is increased (Fig. 10g). However, in the larger pile (H = 2.1 m), at the low wind velocity (U = 0.01–0.3 m/s), the O2 diffusion in the pile is relatively not sufficient (Da > 1, Fig. 13), and the
)
12.6
0 Moisture in coal (wt.%) Fig. 12. Distributions of temperature and moisture content in coal pile (14 days).
pile is in the former stage in which Qoxi increases with the increasing of U. Therefore, the time needed to reach ignition is shortened (Fig. 10h) by increasing U from 0.01 to 0.3 m/s. Nevertheless, to some extent, after the O2 diffusion rate is sufficient, if the wind velocity is further increased from 0.3 to 3 m/s, the pile goes into the later stage in which increasing U leads to little change of Qoxi
the available surface area for reaction. The former effect leads to more O2 available in the pile and hereby could cause an increase of the oxidation rate, but the later leads to the decrease of oxidation rate. As a result, the change of particle size has little effect on the oxidation rate and temperature-rise history. The effect of air humidity is shown in Fig. 10(d). The higher air humidity results in longer time to be ignited, because at the initial time, the evaporation rate of moisture Mevap,2 is higher, which absorbs more heat. However, the effect is not as evident as that of the pile height or wind velocity. In the real condition, the pile bottom which contacts with the ground is not adiabatic and must have heat loss. The effect of this heat loss is shown in Fig. 10(e). From Fig. 10(e), as expected, the
but only the increasing of Qloss (Fig. 7c) and thereby the time needed to reach the ignition is increased (Fig. 10h). In a word, the high wind velocity is desired to prevent the pile from ignition, if the pile can be maintained below a certain height. In practical conditions, pile height, ventilation surrounding the pile, and ground material can be controlled to some extent. According to the above analyses, the feasible approaches to prevent the coal pile from ignition is to maintain the coal pile as small as possible (e.g., lower than 1.25 m in this study), at the same time, to keep good ventilation surrounding the pile to form a certain wind velocity, and to adopt some ground material with high thermal conductivity below the pile. All these approaches promote heat loss from the pile to the environment.
21 U=0.01 m/s
U=0.3 m/s
13 U=3 m/s O2 (vol.%) 4 days, H=1.25 m
14 days, H=1.25 m
4 days, H=2.1 m
Fig. 13. O2 concentration in coal piles with height of 1.25 m and 2.1 m.
14 days, H=2.1 m
J. Zhang et al. / Fuel 176 (2016) 181–189
4. Conclusions A 2-D numerical model for predicting the spontaneous heating of coal pile was developed in this study. In the model, the used chemical kinetic parameters (activation energy E and preexponential factor A2) for low-temperature oxidation of coal were well measured with a tube reactor system, which could make sure the accuracy of prediction. Moreover, the effect of particle size (or reaction surface area) was integrated into the A2, thereby the kinetics parameters once measured could be used for other coal particles with different sizes. In the model, the evaporation rate of moisture from coal was limited by both the heat from coal oxidation and the extent of moisture saturation in the bulk gas mixture. According to the simulation results, the typical process of spontaneous heating could be divided into three stages: temperature-rise stage, constanttemperature stage, and ignition-stage. In this process, the evaporation of moisture from coal plays a critical role. 12 cases were calculated in this study to investigate the effects of different factors. Among these factors, pile height, coal type, heat loss from bottom, initial temperature of pile and wind velocity obviously affect the temperature-rise history, while particle size and air humidity have small effects. In summary, low pile height, good ventilation, and ground material with high thermal conduc-
189
tivity should be used to enhance the heat loss and thereby prevent the coal pile from spontaneous ignition. References [1] Schmal D, Duyzer JH, van Heuven JW. A model for the spontaneous heating of coal. Fuel 1985;1985(64):963–72. [2] Brooks K, Glasser D. A simplified model of spontaneous combustion in coal stockpiles. Fuel 1986;65:1035–41. [3] Brooks K, Balakotaiah V, Luss D. Effect of natural convection on spontaneous combustion of coal stockpiles. AIChE J 1988;34:353–65. [4] Arisoy A, Akgün F. Modelling of spontaneous combustion of coal with moisture content included. Fuel 1994;73:281–6. [5] Krishnaswamy S, Agarwal PK, Gunn RD. Low-temperature oxidation of coal. 3. Modelling spontaneous combustion in coal stockpiles. Fuel 1996;75:353–62. [6] Moghtaderi B, Dlugogorski BZ, Kennedy EM. Effects of wind flow on selfheating characteristics of coal stockpiles. Process Saf Environ Prot 2000;78:445–53. [7] Akgun F, Essenhigh RH. Self-ignition characteristics of coal stockpiles: theoretical prediction from a two-dimensional unsteady-state model. Fuel 2001;80:409–15. [8] Takahashi K, Uchida H, Watanabe T, Takahashi K. Spontaneous combustion of bituminous coals-low temperature oxidation and simulation of temperature rise in coal storage (in Japanese). Ishikawajima-Harima Eng Rev 1983;23:1–6. [9] Wang H, Dlugogorski BZ, Kennedy EM. Coal oxidation at low temperatures: oxygen consumption, oxidation products, reaction mechanism and kinetic modelling. Prog Energy Combust Sci 2003;29:487–513. [10] Aynsley RM, Melbourne W, Vickery BJ. Architectural aerodynamics. London: Applied Science Publisher Ltd; 1977. p. 89.
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Modelling and parametric investigations on spontaneous heating in coal pile Juwei Zhang a,⇑, Wonyoung Choi a, Takamasa Ito a, Katsumi Takahashi a, Masahiro Fujita b a b
Research Laboratory, IHI Corporation Co., Ltd., 1 Shin-Nakahara-cho, Isogo-ku, Yokohama, Kanagawa 235-8501, Japan IHI Transport Machinery Co., Ltd., 8-1, Akashicho, Chuo-ku, Tokyo 104-0044, Japan
a r t i c l e
i n f o
Article history: Received 28 October 2015 Received in revised form 16 February 2016 Accepted 19 February 2016 Available online 27 February 2016 Keywords: Coal Spontaneous Combustion Moisture Simulation
a b s t r a c t To predict the spontaneous heating of coal pile by computational fluid dynamics (CFD) is very meaningful for preventing the coal pile from self-ignition. A two-dimensional (2-D) numerical model, which could be simply implemented in FLUENT, was developed in this study. The chemical kinetic parameters of lowtemperature oxidation of coal, which were expressed by the outer surface area of particles, were well measured in experiments. The coal piles under different conditions were calculated by using the developed model and measured kinetic parameters, in order to make clear the effects of some important factors. The simulation results indicated the evaporation of moisture from coal played a critical role. The pile height, coal type, wind velocity, and heat loss from bottom have significant effects on the process of spontaneous heating of coal pile. According to these effects, in order to increase the heat loss of pile and then effectively inhibit the self-ignition, the low pile height, good ventilation surrounding the pile, and ground material with high thermal conductivity below the pile should be used. The model developed in this model is expected to become a reliable tool to predict the spontaneous heating of coal pile. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Spontaneous heating of low-rank coal stockpile can lead to the waste of coal source and even life-threatening fires. In order to predict the behavior of the spontaneous heating, many mathematical models have been developed. These models can be generally classified as one-dimensional (1-D) [1–5] or two-dimensional (2-D) [6–7] models. Although the 1-D models have been developed and gradually improved for many years, they are still far from the real conditions. For example, a uniform gas velocity throughout the pile is usually assumed in the 1-D model, which is obviously different from reality condition. Furthermore, it is difficult to well simulate the effects of wind in surrounding air by using the 1-D models. However, the 2-D model is considered to be capable of catching the main features of real spontaneous heating process. Moghtaderi et al. [6] focused on the effect of wind flow on self-heating of coal stockpiles by a 2-D model. The calculations show that the wind flow can significantly alter the dynamics of flow field inside the pile and thereby affects the self-heating process. Akgun and Essenhigh [7] investigated the effects of pile height, pile angle, particle diameter and coal moisture with a 2-D model. Their work shows
⇑ Corresponding author. Tel.: +81 45 759 2867; fax: +81 45 759 2207. E-mail address: [email protected] (J. Zhang). http://dx.doi.org/10.1016/j.fuel.2016.02.059 0016-2361/Ó 2016 Elsevier Ltd. All rights reserved.
that the spontaneous ignition can occur if the height exceeds some critical value with sufficient time. Nevertheless, it seems to be difficult to apply these models to the other cases. One of the reasons is that most of these models were developed by in-house code. Although some models can be implemented in commercial CFD software, they are still difficult to be used. For example, in these models, as an important chemical kinetic parameter describing the reaction rate of coal oxidation, the pre-exponential factor always has a unit of s1, which should be more reasonable to be a surface-basis unit for a gas–solid reaction. In this study, a reliable 2-D model for the spontaneous heating of coal pile was built combined with the accurate measurement of coal oxidation rate. This model can be easily implemented in commercial CFD software. In addition, detailed parametric investigations were performed by applying this model to explore the effects of some important factors.
2. Experimental and modelling 2.1. Experimental To obtain the accurate low-temperature oxidation rate of coal is very critical for the modelling of spontaneous heating. Fig. 1 shows the measuring system, which mainly consists of a heating reactor
182
J. Zhang et al. / Fuel 176 (2016) 181–189
Nomenclature a oxygen reaction order (s) A0, A1, A2 pre-exponential factor (consistent unit) C inertial resistance factor (m2) Cp thermal capacity (J kg1 K1) d particle diameter, or diameter of main oxidation zone (m) E activation energy (kJ/mol) h height from ground (m) H heat (J=kgH2 O for evaporation, or J=kmolO2 for oxidation) Hu Air humidity (%) J mass transfer rate (mol m2 s1) L length of pile bottom (m) m mass in a cell (kg); M evaporation rate (kgH2 O m3 s1) n speed exponent (–) Q transfer rate of heat (W/m2) R reaction rate (kmolO2 kgcoal s1) or gas constant (J mol1 K1) S specific surface area (kg/m2), source term (J m3 s1 for energy) or (kg m2 s2 for momentum) Dt time step (s) T temperature (K) DT temperature difference (K) U wind velocity (m/s) v gas velocity (m/s) ~ v vector of gas velocity (m/s)
for coal oxidation and a gas chromatograph (GC: GC3200, GL science) for measuring O2 concentration. One sub-bituminous coal (coal A) and one bituminous coal (coal B) were used. The properties for the two coals are given in Table 1. Before the measurement, about 1.2 g of coal sample with particle size of 45–75 lm (mean diameter dp = 53 lm) was put into a reacting tube (stainless steel, 3/8 in. in diameter), which was placed in the heating reactor. The mixture gas of O2 (21 vol.%) and He with the flow rate of 0.015 l/ min (STP) was passed through the reacting tube and reacted with the coal particles. The porosity of coal sample and residence time of mixture gas in the reacting tube were about 0.43 and 0.51 s, respectively. The heating reactor was heated electrically from 30 to 250 °C at a heating rate of 1 °C/min. Simultaneously, the O2 concentration variations were measured by the GC and recorded by a computer. One measurement was repeated for six times. Finally, the changes of O2 consumption rate or the oxidation rate of coal roxi
DV x X DX
a
b
e
k
q l
cell volume (m3) characteristic length (m) mole fraction (–) thickness for thermal conduction (m) permeability (m2) thermal expansion coefficient (–) porosity (–) thermal conductivity (W m1 K1) density (kg/m3) viscosity (kg m1 s1)
Subscripts and superscripts 0 atmosphere boundary layer, or ambient condition c coal eff effective evap evaporation of moisture f fluid moisture H2O i coordinate x, y, or z loss heat loss oxi oxidation O2 oxygen p particle rise temperature rise s solid sat saturation status
Then, the Arrhenius plot and the corresponding kinetic parameters (E and A) could be obtained. However, our previous study [8] showed that the particle size had a significant effect on the oxidation rate of coal, as shown in Fig. 2. It can be seen that the relationship between ln rO2 and dp at different temperatures for two bituminous coals can be described with a straight line with an inclination of about 1/3. In addition, the reaction order for oxygen concentration is around 0.5 [8]. Thus, the Eq. (1) can be rewritten as 1=3
roxi ¼ A1 dp
E X 0:5 exp O2 RT
ð3Þ
or be expressed with outer surface area of coal particles,
E 0:5 roxi ¼ A2 S1=3 X exp O2 p RT
ð4Þ
1
ðmolO2 kgcoal s1 Þ with temperature were obtained. The relationship between the O2 consumption rate and temperature can be expressed with an Arrhenius type of equation, as follows
E r oxi ¼ A exp RT
ð1Þ
Then
E 1 ln roxi ¼ þ ln A0 R T
ð2Þ
O2 (21%)/He
so that the pre-exponential factor (A1 or A2) can be unified for particles with various sizes. The obtained E and A2 for coal A and B are also given in Table 1. During the low-temperature oxidation of coal, physical adsorption and chemisorption occur together. However, physical adsorption is a reversible process which plays a major role only at very low temperature. In many studies, the physical adsorption was usually ignored. Furthermore, the heat of adsorption for physical adsorption is between 6.7 and 20.9 kJ=molO2 , while that of chemisorption is about 300 kJ=molO2 [9]. In this study, only the data in the temperature above 60 °C with a good linearity was used to calculate the kinetic parameters, which was considered to be the kinetics of only the chemisorption. 2.2. Modelling
Reacting tube for coal oxidization
Heating reactor
GC
Fig. 1. Measuring system for oxidation rate of coal.
PC
ANSYS FLUENT 15.0 was employed to simulate the spontaneous heating of coal pile. The coal pile was modeled as a 2-D triangle, as shown in Fig. 3. One part of the surrounding space around the pile
183
J. Zhang et al. / Fuel 176 (2016) 181–189 Table 1 Properties of coals for measurement. Sample
Type
Proximate analysis (wt.%)
Coal A Coal B
Sub-bituminous Bituminous
Moisture
Volatiles
Fixed carbon
Ash
12.6 7.0
32.2 36.5
31.5 42.4
23.7 14.1
0.5
A2 ðmolO2 ðm kgcoal Þ
0.25 0.12
0.85 0.90
44.18 54.68
21.97 307.54
s1 Þ
dp e3 150 ð1 eÞ2
ð7Þ
C¼
3:5 ð1 eÞ dP e3
ð8Þ
-1.5
The natural convection was considered with the Boussinesq model which treats gas density as a constant value in all solved equations, except for the buoyancy term in the momentum equation:
-2
--- Bituminous coal 1 Bituminous coal 2 -1
-0.5
0
ðq q0 Þg ¼ q0 bðT T 0 Þg
0.5
The energy equation in the porous medium can be expressed as
dp=1 mm
@ ðð1 eÞqs C p;s þ eqf C p;f Þ þ eqf C p;f ð~ v rTÞ @t ¼ keff ðr rTÞ þ Soxi þ Sevap
Fig. 2. Effect of particle size on oxidation rate of coal.
was also included into the calculation domain, in order to take into account the effect of wind flow. The whole domain has the dimension of 10L 2L (L is the length of pile bottom). A vertical velocity profile for the wind was used according to the power law equation for the atmospheric boundary layer [10], as follows
n h h0
ð5Þ
In this study, h0 and n were chosen as 300 m and 0.15, respectively, which are the typical values for a terrain of open country with low scrub or scattered tress. A typical gas velocity distribution around the pile is also shown in Fig. 3. The velocity distributions before and after the pile are obviously asymmetrical, which is believed to have a significant effect on the flow and temperature fields in the pile. The coal pile and surrounding space were simulated as a porous media and a fluid domain, respectively. The equations for continuity, momentum, species, and energy were solved for both domains of the coal pile and surrounding space. The effects of natural convection and turbulence (realizable j-e turbulence model) were included in the model. The porous medium was modeled by adding a momentum source term to the standard fluid momentum equations, as follows
1 1 Si ¼ lv i þ C qf jv jv i 2 a
ð9Þ
ð6Þ
where the permeability a (1/a is called the viscous resistance coefficient) and inertial resistance coefficient C were calculated by the following equations, respectively.
keff ¼ ekf þ ð1 eÞks
ð11Þ
The energy source from coal oxidation was calculated by
Soxi ¼ r oxi qc ð1 eÞHoxi
ð12Þ
The energy source from the evaporation of moisture from coal was calculated by
Sevap ¼ Hevap M evap
ð13Þ
It is assumed that the heat from coal oxidation is preferentially used for the evaporation, until the moisture in the gas mixture reaches status of saturation. The rate of moisture evaporation Mevap was limited by both the available heat from coal oxidation and the moisture saturation extent in bulk gas mixture. The Mevap determined by the heat from coal oxidation is
Mevap;1 ¼ Soxi =Hevap
ð14Þ
Wall Wind
0 Gas velocity
Coal pile H
θ
L θ=40
Wall 10L
Fig. 3. Schematic diagram of geometry and typical gas velocity distribution.
2L
max
Inlet
ð10Þ
where it is assumed that gas and solid have the same temperature. The forced convection heat transfer on the surface of coal pile is not considered. The heat transfer coefficient of forced convection was calculated and compared with the heat transfer coefficient of thermal conduction, as shown in Fig. 4. It can be seen that the heat transfer coefficient of convection is far higher than that of thermal conduction even at low velocities. This indicates the resistance of heat transfer due to the convection on the surface of coal pile can be ignored. The heat generated in the porous medium is mainly transferred to the surrounding environment by thermal conduction. The effective thermal conductivity of the porous medium was calculated by
Outlet
lnrO2 (mlO2 g-1 min-1)
-1
lndp (mm)
U ¼ U0
E (kJ/mol)
a¼
-0.5
-3 -1.5
H/C
2
0
-2.5
2=3
O/C
184
J. Zhang et al. / Fuel 176 (2016) 181–189
humidity, heat loss from pile bottom, initial temperature of pile and wind velocity could be investigated.
Convection heat transfer coefficient hconv. (W/m2K)
20 Calculated with a horizontal plate with characteristic length of 1 m
16 12
3. Simulation results
8 3.1. Dimensionless analysis
4 Conduction
0
In order to analyze the roles of convection and diffusion in the heating transfer in the pile, Péclet number for heat transfer was used in this study. It can be defined as
-4 0
1
2 3 4 5 Free wind velocity U (m/s)
6
Fig. 4. Comparison of heat transfer coefficient between forced convection and thermal conduction at difference wind velocities.
The Mevap determined by the extent of moisture saturation extent in gas mixture is
M evap;2 ¼ ðmH2 O;sat mH2 O Þ=ðDt DVÞ
ð15Þ
where mH2 O;sat and mH2 O are mass of moisture in a calculated cell at saturation and present status, respectively. It should be noted that mH2 O;sat is a function of temperature. The minimum value of the Mevap,1 and Mevap,2 was used as the final Mevap. The moisture in the coal in every cell was updated in every time step after Mevap was determined. In this study, the grid size in the coal pile and time step were 20–60 mm and 300 s, respectively. The study of Moghtaderi et al. [6] has showed that the grid size of 250 mm and time step of several days were numerically accurate enough for the simulation of spontaneous heating of coal pile. In this study, the calculation with time step of 10 s was also performed and the result almost showed no difference to that with time step of 300 s. The 2nd order upwind discretization scheme was used for the equations of continuity, momentum, species, and energy. The main models and parameters used in the simulation are listed in Table 2. The settings for different cases are given in Table 3. It should be noticed that the practical objective of this simulation is to predict the spontaneous heating of coal piles, which is formed with the fallen fine coal particles from the belt conveyor, and then to determine how often the coal pile should be cleaned. Thus, coal piles with fine particles and small pile size were adopted for simulation. However, the models developed in this study can be used for predicting the spontaneous heating of any coal pile. Via the calculation of the cases listed in Table 3, the effects of pile height, coal type, coal particle size, air
Peh ¼
qg C pg V g x
ð16Þ
kc
The main heat transfer mechanism is convection when Peh 1, and thermal conduction when Peh 1. Peh in all cases are listed in Table 4. It can be seen that except case 5, all the other cases have the Peh obviously lower than 1, which indicates the thermal conduction is the main mechanism for heat transfer. The heat from coal oxidation is not only used for moisture evaporation (Qeva) and heating of coal and gases in the pile (Qrise), but also for the heat lost to the environment (Qloss), i.e.,
Q oxi ¼ Q eva þ Q rise þ Q loss
ð17Þ
If Peh 1, the main form of heat loss is the thermal conduction, which can be simply expressed by
Q loss ¼ k
DT DX
ð18Þ
According to Eq. (18), the heat loss Qloss is low when DX is large. The ignition in the pile always occurs at the position, where heat is accumulated easily and/or difficult to be dissipated. Thus, the ignition position generally has the largest DX and thereby lowest Qloss, if sufficient oxygen available for the oxidation exists in the pile. In this study, Damköhler number Da, expressing the ratio of the oxidation rate to the transportation rate by convection or diffusion, was also used to analyze the simulation results. Fig. 5 shows the vectors of gas velocity in coal piles. It can be seen that in all the cases using dp = 0.025 mm, the gas in the pile flows outward because of very high resistance (Fig. 11), while in the case 5 with dp = 2.5 mm, the gas flows inward. This indicates, in all the cases except case 5, the O2 can enter into pile only by diffusion rather than convection. The typical diffusion and consumption modes of O2 in coal pile with dp = 0.025 mm is showed in Fig. 6. In order to calculate Da simply, the main oxidation zone is simplified into a half circle with diameter of d and diffusion occurs with the O2
Table 2 Input parameters for simulation of the baseline case (case 3 in Table 3). Item Oxidation rate (coal A) Coal properties (coal A)
Heat generation Surrounding air
Frequency factor A2 Activation energy Heat capacity Mean diameter Apparent density Thermal conductivity Porosity Initial moisture content Oxidation Evaporation Temperature Humidity Free wind velocity
Unit 2=3
molO2 ðm kgcoal Þ kJ/mol J/(kg K) mm kg/m3 W m1 K1 – wt.% J=kmolO2 kJ=kgH2 O °C % m/s
s
1
Value
Source
21.97
Measurement in this study
44.18 1046 0.025 1000 0.337 0.43 12.6 3.0 108 2260 25 50 0.3
Constant Assumption
185
J. Zhang et al. / Fuel 176 (2016) 181–189 Table 3 Boundary conditions for different cases. Case
Coal
Pile height H (m)
Particle diameter dp (mm)
Wind velocity U (m/s)
Air humidity Hu (%)
Initial Temp. T0 (°C)a
Setting of pile bottom
1 2 3 4 5 6 7 8 9 10 11 12
A A A B A A A A A A A A
0.42 1.25 2.1 2.1 2.1 1.25 1.25 2.1 2.1 2.1 2.1 1.25
0.025 0.025 0.025 0.025 2.5 0.025 0.025 0.025 0.025 0.025 0.025 0.025
0.3 0.3 0.3 0.3 0.3 0.01 3 0.01 3 0.3 0.3 0.3
50 50 50 50 50 50 50 50 50 30 50 50
25 25 25 25 25 25 25 25 25 25 25 30
Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic Adiabatic With heat lossb Adiabatic
Note: The changed conditions are written in the font of bold and italic. a The initial temperature of pile is same with that of environment. b It is assumed that the ground temperature is 25 °C and heat transfer coefficient is 10 W/(m2 K).
Vg (10
1 2 3 4 5 6 7 8 9 10 11 12
0.21 2.8 19 1.1 73 3.0 2.0 13 4.1 17 3.9 3.5
m/s)
x (m)
Peh
0.32 0.96 1.6 1.6 1.6 0.96 0.96 1.6 1.6 1.6 1.6 0.96
0.00062 0.024 0.28 0.016 1.1 0.026 0.018 0.18 0.06 0.25 0.057 0.031
3e-5
0 Velocity (m/s) (a) Case 3, dp=0.025 mm, 14 days
(a) 6
CO2,0 d W
4
XO2,0 = 10 vol.% XO2,s = 21vol.% d = 0.1W
2 0
0
4
Da=1
8
12
16
8
(b)
H=0.42 m H=1.25 m H=2.1 m
6 4
XO2,0 = 10 vol.% d = 0.1W
2
Da=1
0 10
13
16
19
22
O2 fraction on pile surface XO2,s (vol.%) Heat from oxidation Qoxi (W)
(Vectors in other cases with dp=0.025 mm are similar with case 3)
1.6e-4
0 Velocity (m/s)
8
Pile height H (m) Damköhler number Da
Case
6
Damköhler number Da
Table 4 Péclet numbers in different cases.
(b) Case 5: dp=2.5 mm, 14 days
Fig. 5. Vectors of gas velocity in coal piles.
3
(c) H=0.42 m H=1.25 m H=2.1 m
2
1
0 10
13
16
19
22
O2 fraction on pile surface XO2,s (vol.%)
21
Diffusion
Wind velocity U (m/s)
XO2,0 13 O2 (vol.%) Consumption
Fig. 7. Changes of Damköhler number (a) and heat from oxidation (b) with O2 concentration on surface of coal pile with the assumption of C O2;o ¼ 10%.
d W
Fig. 6. Typical manners of diffusion and consumption of O2 in coal pile.
concentration difference between pile surface and center of main oxidation zone (Fig. 6, right), based on a typical O2 distribution in the pile (Fig. 6, left). Thus, Da can be calculated by
r oxi pd =8 r oxi d=4 ¼ JO2 pd=2 DO2 ðC O2 ;s C O2 ;0 Þ=x 2
Da ¼ ¼
r oxi xdRT 4DO2 PðX O2 ;s X O2 ;0 Þ
ð19Þ
The oxidation rate is limited by chemical kinetics if Da < 1 and by O2 diffusion if Da > 1. Meanwhile, the heat generated from the main oxidation zone Qoxi can be calculated by (roxi pd2/8) Hevap if Da < 1 and by ðJ O2 pd=2Þ Hevap if Da > 1, respectively.
J. Zhang et al. / Fuel 176 (2016) 181–189
Maximum temperature ( )
186
100 75
80
Stage 2: constant temp.
60 40 20
Case 3
0
0
4
8
12
16
20
Time (day) Fig. 8. Maximum temperature in coal pile as a function of time in baseline case.
It is reasonable to assume that x and d are in direct proportion to the pile size (W or H), so that Da changes with W2. Under the assumption of d = 0.1W, X O2;0 ¼ 10%, and X O2 ;s ¼ 21%, the change of Da with pile height can be calculated and the result is shown in Fig. 7(a). It can be seen that Da increases monotonously with the increasing of pile height and the oxidation rate is limited by O2 diffusion (Da > 1) when the pile is higher than 6 m. In fact, the O2 concentration at the surface of pile ðX O2 ;s Þ maybe lower than 21% due to the limitation of mass transfer rate of O2 in the bulk gas around the pile. The increasing of wind velocity U must lead to the increasing of X O2 ;s until X O2 ;s reaches 21%. Under the assumption of d = 0.1W and X O2 ;0 ¼ 10%, the change of Da and Qoxi in the main oxidation zone for three piles with different heights with the increasing of X O2 ;s can be calculated and the results are shown in Fig. 7(b) and (c), respectively. It can be seen that for the pile of H = 0.42 m always has a Da lower than 1, the pile of H = 1.25 m has a Da lower than 1 if X O2 ;s > 11%, and the pile of H = 2.1 m has a Da lower than 1 if X O2 ;s > 13%. As shown in Fig. 7 (c), for the pile of H = 2.1 m, Qoxi is first increased (Da > 1) and then keeps constant (Da < 1), as the increasing of X O2 ;0 (or U). The former stage with increasing Qoxi gradually shortens as the deceasing of pile height until completely disappears for the pile of H = 0.42 m. Although the heat loss Qloss is increased as the increasing of U (the surface of pile is cooled), the temperature in the main oxidation zone may be increased due to the increasing of Qoxi in the former stage. However, the temperature in the main oxidation zone deceases as the increasing of U, because Qloss is increased and Qoxi has no change in the later stage. 3.2. Results of baseline case The simulation results for the baseline case (case 3 in Table 3) are shown in Figs. 8 and 9. According to Eq. (18), the ignition occurs
at the middle position which has the largest DX and lowest Qloss in the horizontal direction, as shown in Fig. 9. In the vertical direction, the ignition position must be at the adiabatic pile bottom, where Qloss is minimal (zero) along the vertical direction. A typical curve of maximum temperature rise in the spontaneous heating of coal pile was obtained, as shown in Fig. 8. According to this curve, three different stages exist in the process of spontaneous heating. In the 1st stage (temperature-rise stage), the maximum temperature is gradually increased, Qoxi obviously exceeds Qeva and Qloss, and the excessive heat can be used for heating the pile. In this stage, the Mevap,2 calculated by Eq. (15) is obviously lower than the Mevap,1 calculated by Eq. (14), because of low temperature, thus, the Mevap,2 is adopted as the evaporation rate. The amount of evaporated moisture in the first stage is very small, which can be seen from the 4th days moisture content in coal in the 4th day, as shown in Fig. 9. As the temperature gradually rises, the moisture content in saturation mH2 O;sat and Mevap,2 are rapidly increased. When the Mevap,2 exceeds Mevap,1, Mevap,1 is used as the evaporation rate, and this point is defined as the beginning of the 2nd stage (constant-temperature stage). In this stage, all of the heat from oxidation is used for evaporation and offsetting the heat loss, so that no heat can be available for heating the pile, therefore the maximum temperature is constant at this stage (around 75 °C in Fig. 8). As the evaporation further proceeds, the moisture content in the coal is gradually decreased and the position with the lowest Qloss and highest temperature in the pile is dried firstly (see moisture content in coal in the 14th day shown in Fig. 9). At this dried position, all of the heat from oxidation can be used for heating the pile and offsetting the heat loss. The temperature is increased rapidly, and finally spontaneous ignition occurs, which is the 3rd stage: ignition stage. From Fig. 9, it can be also seen that the temperature distribution in the coal pile is asymmetric and the position with the maximum temperature (ignition position) is a little closer to the left side because of the asymmetrical distribution of wind velocity around the coal pile, as shown in Fig. 3. 3.3. Parametric investigations The effects of pile height, coal type, coal particle size, wind velocity, air humidity, heat loss from pile bottom, and wind velocity on the temperature rise in coal pile are shown in Fig. 10(a)–(h), respectively. According to Fig. 10, it can be seen that pile height, coal type, heat loss, initial temperature of pile and wind velocity have obvious effects on the spontaneous heating process of coal piles, while the effects of particle size and air humidity are not so obvious.
80
25 12.6
0 Moisture in coal (wt.%)
4 days (in stage 1)
12 days (in stage 2)
Fig. 9. Changes of temperature and coal moisture content distributions.
14 days (in stage 3)
J. Zhang et al. / Fuel 176 (2016) 181–189
187
Fig. 10. Effects of pile height (a), coal type (b), particle size (c) air humidity (d), heat loss of bottom face (e), initial temperature (f) and wind velocity (g and h) on temperature rise in coal piles.
The effect of pile height is shown in Fig. 10(a). As the pile height is increased, the time needed to reach ignition is increased. At the pile height of 0.42 m, ignition doesn’t happen. According to Eq. (18), this is mainly attributed to the fact that DX at the potential position of ignition of pile becomes larger and Qoxi becomes lower, as pile height is increased. In order words, the heat is easier to accumulate in the larger pile. Although increasing the pile height can increase the O2 diffusion rate, it has little effect on Qoxi, because the oxidation reaction is tends to be limited by chemical kinetics for piles with heights 0.422.1 m (Da < 1, see Fig. 7). The effect of pile height has a very important practical meaning. As mentioned above, one of the objectives of this study is to determine the frequency of cleaning the coal pile. The result of Fig. 10(a)
indicates that the pile height of coal A should be maintained lower than 1.25 m and the pile should be cleaned in about two months. The effect of coal type is shown in Fig. 10(b). The spontaneous ignition occurs in pile of coal A within 15 days, but doesn’t occur in the pile of coal B. As expected, this indicates that the spontaneous ignition is easier to occur for the coal with lower rank. The effect of particle size is shown in Fig. 10(c). To increasing the mean particle diameter with 100 times (from 0.025 mm to 2.5 mm) leads to little change of temperature-rise. The effects of particle size on the viscous resistance coefficient and specific surface area of particles are shown in Fig. 11. It can be seen that although the increase of particle diameter from 0.025 mm to 2.5 mm decreases the resistance, but it also significantly decreases
80
1.E+12 1.E+10
60
1.E+08 40
1.E+06 1.E+04
20
1.E+02 1.E+00 0
3
5
8
10
0
Specific surface area m2/kg
J. Zhang et al. / Fuel 176 (2016) 181–189
Resistance coefficient 1/α (m-2)
188
Particle diameter mm Fig. 11. Changes of resistance coefficient and specific outer surface area with particle diameter.
Case 11: with heat loss of pile bottom, 14 days 80
25 Temp. (
time needed to reach the ignition is obviously increased. Fig. 12 shows the distributions of temperature and moisture content in the coal for case 11 considering the heat loss. It can be seen that the ignition position with dry coal moves upside, compared with the adiabatic case shown in Fig. 9, because the position of the lowest Qloss also moves upside, when heat loss from the bottom exists. The effect of initial temperature of pile (or the environment temperature) in the pile is shown in Fig. 10(f). It can be seen that as the initial temperature increases from 25 to 30 °C, the time needed to reach ignition is decreased. Similar result was also obtained by Schmal et al. [1]. The delay of temperature rise at low initial temperature is mainly due to the very low reaction rate of coal oxidation. The effect of wind velocity in the pile with height of 1.25 and 2.1 m are shown in Fig. 10(g) and (h), respectively. It can be seen that the effect of wind velocity is different for piles with different height. It can be easily explained by the Fig. 7(c) and the discussion in section ‘‘dimensionless analysis”. In the smaller pile (H = 1.25 m), O2 diffusion in the pile is always high enough for supplying O2 for oxidation under different velocities (Da < 1, Fig. 13), and the pile is in the former stage in which increasing U leads to little change of Qoxi but only the increasing of Qloss (Fig. 7c). Therefore, the time needed to reach ignition is monotonously increased as wind velocity is increased (Fig. 10g). However, in the larger pile (H = 2.1 m), at the low wind velocity (U = 0.01–0.3 m/s), the O2 diffusion in the pile is relatively not sufficient (Da > 1, Fig. 13), and the
)
12.6
0 Moisture in coal (wt.%) Fig. 12. Distributions of temperature and moisture content in coal pile (14 days).
pile is in the former stage in which Qoxi increases with the increasing of U. Therefore, the time needed to reach ignition is shortened (Fig. 10h) by increasing U from 0.01 to 0.3 m/s. Nevertheless, to some extent, after the O2 diffusion rate is sufficient, if the wind velocity is further increased from 0.3 to 3 m/s, the pile goes into the later stage in which increasing U leads to little change of Qoxi
the available surface area for reaction. The former effect leads to more O2 available in the pile and hereby could cause an increase of the oxidation rate, but the later leads to the decrease of oxidation rate. As a result, the change of particle size has little effect on the oxidation rate and temperature-rise history. The effect of air humidity is shown in Fig. 10(d). The higher air humidity results in longer time to be ignited, because at the initial time, the evaporation rate of moisture Mevap,2 is higher, which absorbs more heat. However, the effect is not as evident as that of the pile height or wind velocity. In the real condition, the pile bottom which contacts with the ground is not adiabatic and must have heat loss. The effect of this heat loss is shown in Fig. 10(e). From Fig. 10(e), as expected, the
but only the increasing of Qloss (Fig. 7c) and thereby the time needed to reach the ignition is increased (Fig. 10h). In a word, the high wind velocity is desired to prevent the pile from ignition, if the pile can be maintained below a certain height. In practical conditions, pile height, ventilation surrounding the pile, and ground material can be controlled to some extent. According to the above analyses, the feasible approaches to prevent the coal pile from ignition is to maintain the coal pile as small as possible (e.g., lower than 1.25 m in this study), at the same time, to keep good ventilation surrounding the pile to form a certain wind velocity, and to adopt some ground material with high thermal conductivity below the pile. All these approaches promote heat loss from the pile to the environment.
21 U=0.01 m/s
U=0.3 m/s
13 U=3 m/s O2 (vol.%) 4 days, H=1.25 m
14 days, H=1.25 m
4 days, H=2.1 m
Fig. 13. O2 concentration in coal piles with height of 1.25 m and 2.1 m.
14 days, H=2.1 m
J. Zhang et al. / Fuel 176 (2016) 181–189
4. Conclusions A 2-D numerical model for predicting the spontaneous heating of coal pile was developed in this study. In the model, the used chemical kinetic parameters (activation energy E and preexponential factor A2) for low-temperature oxidation of coal were well measured with a tube reactor system, which could make sure the accuracy of prediction. Moreover, the effect of particle size (or reaction surface area) was integrated into the A2, thereby the kinetics parameters once measured could be used for other coal particles with different sizes. In the model, the evaporation rate of moisture from coal was limited by both the heat from coal oxidation and the extent of moisture saturation in the bulk gas mixture. According to the simulation results, the typical process of spontaneous heating could be divided into three stages: temperature-rise stage, constanttemperature stage, and ignition-stage. In this process, the evaporation of moisture from coal plays a critical role. 12 cases were calculated in this study to investigate the effects of different factors. Among these factors, pile height, coal type, heat loss from bottom, initial temperature of pile and wind velocity obviously affect the temperature-rise history, while particle size and air humidity have small effects. In summary, low pile height, good ventilation, and ground material with high thermal conduc-
189
tivity should be used to enhance the heat loss and thereby prevent the coal pile from spontaneous ignition. References [1] Schmal D, Duyzer JH, van Heuven JW. A model for the spontaneous heating of coal. Fuel 1985;1985(64):963–72. [2] Brooks K, Glasser D. A simplified model of spontaneous combustion in coal stockpiles. Fuel 1986;65:1035–41. [3] Brooks K, Balakotaiah V, Luss D. Effect of natural convection on spontaneous combustion of coal stockpiles. AIChE J 1988;34:353–65. [4] Arisoy A, Akgün F. Modelling of spontaneous combustion of coal with moisture content included. Fuel 1994;73:281–6. [5] Krishnaswamy S, Agarwal PK, Gunn RD. Low-temperature oxidation of coal. 3. Modelling spontaneous combustion in coal stockpiles. Fuel 1996;75:353–62. [6] Moghtaderi B, Dlugogorski BZ, Kennedy EM. Effects of wind flow on selfheating characteristics of coal stockpiles. Process Saf Environ Prot 2000;78:445–53. [7] Akgun F, Essenhigh RH. Self-ignition characteristics of coal stockpiles: theoretical prediction from a two-dimensional unsteady-state model. Fuel 2001;80:409–15. [8] Takahashi K, Uchida H, Watanabe T, Takahashi K. Spontaneous combustion of bituminous coals-low temperature oxidation and simulation of temperature rise in coal storage (in Japanese). Ishikawajima-Harima Eng Rev 1983;23:1–6. [9] Wang H, Dlugogorski BZ, Kennedy EM. Coal oxidation at low temperatures: oxygen consumption, oxidation products, reaction mechanism and kinetic modelling. Prog Energy Combust Sci 2003;29:487–513. [10] Aynsley RM, Melbourne W, Vickery BJ. Architectural aerodynamics. London: Applied Science Publisher Ltd; 1977. p. 89.