Effects of particle size on flame structures through corn starch dust explosions

Accepted Manuscript Effects of particle size on flame structures through corn starch dust explosions Hongming Zhang, Xianfeng Chen, Ying Zhang, Yi Niu, Bihe Yuan, Huaming Dai, Song He PII:

S0950-4230(17)30590-9

DOI:

10.1016/j.jlp.2017.09.002

Reference:

JLPP 3584

To appear in:

Journal of Loss Prevention in the Process Industries

Received Date: 26 June 2017 Revised Date:

2 September 2017

Accepted Date: 2 September 2017

Please cite this article as: Zhang, H., Chen, X., Zhang, Y., Niu, Y., Yuan, B., Dai, H., He, S., Effects of particle size on flame structures through corn starch dust explosions, Journal of Loss Prevention in the Process Industries (2017), doi: 10.1016/j.jlp.2017.09.002. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Effects of particle size on flame structures through corn starch dust explosions Hongming Zhang, Xianfeng Chen∗, Ying Zhang, Yi Niu, Bihe Yuan, Huaming Dai, Song He

RI PT

(School of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, China)

SC

Abstract

The effects of particle size on the structures of flame reaction zone and preheat

M AN U

zone of corn starch dust were investigated. Corn starch dust with three different average particle sizes were chosen to conduct explosion experiments in a half-closed vertical duct. A high-speed camera was used to record the flame propagation process

TE D

and direct light emission images of flame propagation in duct were obtained. The structures of flame reaction zone and preheat zone were detected by an ion current probe and a fine thermocouple. The experimental results show that the structures of

EP

reaction zone and preheat zone vary with the particle size. There is a positive

AC C

correlation between the particle size and the thickness of reaction zone and preheat zone. Meanwhile, the thickness of the two flame structures are determined under these experimental conditions. Additionally, the mass concentrations of dust clouds in duct are founded as a critical common factor for flame structure variation. Keywords: dust explosion; particle size distribution; flame structure; ion current probe

∗

Corresponding author. Tel./fax:+86 13627212572 E-mail address: [email protected]

1

ACCEPTED MANUSCRIPT 1. Introduction Dust explosion is one of the most serious disasters in the process industries. Nowadays, with the advancement of powder technology and the increase of powder

RI PT

handling processes, dust explosions occur frequently, especially in grain dust, metal dust and chemical production dust. Corn starch dust is one of the most explosive materials in grain dust and many serious dust explosions accidents are triggered by

SC

corn starch. A large number of experiments have been done to study the macroscopic

M AN U

properties of corn starch dust explosions and the achievements in research have been used to guide industrial production (Zhong and Deng, 2000; Radandt et al., 2001; Ramírez et al., 2009; Chao and Dorofeev, 2015). However, the microscopic characteristics of flame propagation are also important to the initial deflagration

TE D

process of corn starch explosion. Thus, sufficient studies on the fundamental mechanisms of flame propagation have become more important for corn starch dust explosion prevention and mitigation (Eckhoff, 2003). Numerous efforts have been

EP

devoted to dust explosions (Abbasi and Abbasi, 2007; Amyotte et al., 2009; Amyotte

AC C

and Eckhoff, 2010). However, most of the previous studies focused their points on the macroscopic characteristic parameters of different dusts, such as the maximum pressure, rate of pressure rise, ignition energy and explosion concentration limits (Eckhoff, 2002; Amyotte et al., 2006; Cloney et al., 2013; Yuan et al., 2014;). Recently, the investigations on the microscopic mechanism of dust explosion have become more and more in-depth and comprehensive. Gao et al. (2014, 2013, 2015) studied the effect of dust particle characteristics on the flame microstructures and

2

ACCEPTED MANUSCRIPT propagation behavior during dust explosions and revealed the flame propagation mechanism. Han et al. (2000) conducted experiments to research the structure of the laminar flames through lycopodium dust clouds. The ball-shape flames and double

RI PT

flame structure with isolated individual burning particles were observed. Bidabadi et al. (2010) investigated the role of Lewis and Damköhler numbers on the combustion phenomenon of organic dust particles and divided the flame structure into four zones,

SC

preheat, vaporization, reaction and post flame zones. Proust (2006a, b) utilized

M AN U

experimental methods to measure the laminar burning velocities and maximum flame temperatures of the combustible dust-air mixtures. Some fundamental aspects on the flame propagation process has been studied. Benedetto et al. (2007) conducted a detailed theoretical study to establish a thermo-kinetic model of dust explosions

TE D

which linked the sensitivity of the deflagration index to turbulence intensity. However, due to the complexity of combustion processes of dust particle cloud, the current level understanding to the basic flame propagation mechanisms of dust

EP

explosion is still in a rudimentary state. Variation of preheat zone and movement of

AC C

reaction zone in a dust particles cloud in a vertical duct is still ambiguous (Han et al., 2000; Gao et al., 2013). It is well known that particle size distribution is an important factor affecting the flame microstructure of dust explosion. Thus, for the transition from combustion to explosion of dust clouds, it is significant to investigate the variation law of the particle size distribution on the flame structures. In the present study, the flame structures of preheat zone and reaction zone were analytically investigated in order to clarify the effects of particle size on flame

3

ACCEPTED MANUSCRIPT propagation process during starch dust explosions. Flame propagation processes in corn starch dust clouds with three different particle size in a vertical duct were recorded by a high speed camera system. Meanwhile, a fine thermocouple and an ion

RI PT

current probe were used to monitor the flame microstructures. Moreover, the flame propagation behaviors and variation characteristics of flame thickness during starch explosions in the duct were further studied.

SC

2. Experimental

M AN U

2.1 Experimental apparatus

The experimental apparatus schematically shown in Fig.1 consists of seven parts: a dust combustion duct, a gas supplying and powder spraying device, a high voltage ignition unit, a high-speed photography system, a fine thermocouple, an ion current

TE D

probe and a programmable logic controller. The small-scale combustion duct is 500 mm in height with a square cross-section of 80 mm × 80 mm. The top of the duct is open and the bottom is closed. Two pieces of optical glasses were installed in the two

EP

sides to observe the flame propagation process clearly. Dust clouds were ignited by

AC C

high voltage ignition unit which could generate a spark energy up to 80 J between the two electrodes.

The optical imaging system consists of a high-speed photography camera and CCD

lens which can capture and record flame propagation behaviors of corn starch dust explosion. In this study, a fine thermocouple and an ion current probe were used to detect the characteristics of flame reaction zone (Chen et al., 2007, 2017; Gao et al., 2014). The

4

ACCEPTED MANUSCRIPT two sensors are positioned at 360 mm above the bottom of the duct. The thermocouple is comprised of 25 µm diameter Pt-Pt/Rh 13% wires. The ion current probe is made of 100 µm diameter Pt wires with high temperature resistivity,

RI PT

anti-oxygenic property and good conductivity. Two parallel wires of the ion current probe were set with approximately 1.0 mm apart. The composition of ion current probe is illustrated in Fig.2.

SC

In the tests, programmable logic controller was used for timing control and on-off

M AN U

control. Gas supply and powder spraying system, ignition system, high-speed photography camera and data recorder were connected together. Before the experiments, the corn starch dust weighed by an electronic scale was evenly placed on the sample dish at the bottom of the duct. The dust particles were dispersed into the

TE D

duct with help of high-pressure air. In order to weaken the influence of the airflow on dust cloud and to achieve the cloud uniform-mixed particles, the ignition time was delayed with the appropriate time. After ignition, high-speed photography camera and

EP

data recorder started to record related data. The detailed experimental conditions were

AC C

listed as follows: pressure for powder injection, 0.2 MPa; injection time, 0.2 s; ignition voltage, 15 kV; ignition delay time, 0.1 s; framing rate of high-speed photography camera, 1000 frames/s.

2.2 Experimental materials The dust samples used in the experiments was the corn starch of food grade produced by Qinhuangdao Li Hua Starch Co., Ltd (batch number: 201702044042).

5

ACCEPTED MANUSCRIPT Proximate analysis and ultimate analysis for the corn starch dust are shown in Table 1. To avoid the influence of moisture on corn starch dust, the samples were dried in a vacuum drying oven at 50 °C for 12 h before the experiments.

RI PT

Three samples with different particle size distributions were chosen for experiments. In order to obtain the samples in the same range of particle sizes, the corn starch dust were sifted in a vibration sieve with 200 mesh, 300 mesh and 400 mesh. The particle

SC

size distributions are shown in Fig.3. The median particle size D50 of the three

M AN U

samples are 65µm, 45µm and 30µm, which are considered as the mean particle diameter of the three samples in the experiments.

Thermal decomposition behavior and spontaneous combustion mechanism of combustible dust are the intrinsic factors of dust explosion (Li et al., 2017; Yang et al.,

TE D

2017). Therefore, it is necessary to study the thermal behavior of dust samples. In order to obtain the thermal behaviors of corn starch dust, thermo-gravitational analysis (TGA) was carried out using a simultaneous PerkinElmer STA-6000

EP

equipment. Three samples with different particle size distributions were subjected to a

AC C

temperature scanning programmer, which was heated from 50 to 800 °C at a heating rate of 20 °C /min. Air-flow with a rate of 20 cm3/min was used as the combustion environment for TGA analysis to make sure the experiment condition as same as the dust explosions experiments.

3. Experimental results and discussions 3.1 Thermal behaviors of corn starch dust

6

ACCEPTED MANUSCRIPT Thermogravimetric (TG) and derivative thermogravimetric (DTG) results for corn starch dust at three different particle sizes (30, 45 and 65 µm) are shown in Fig. 4. It is clear that the initial thermal decomposition temperature becomes higher with the

RI PT

increase of particle size. For all samples, the decomposition rate goes up with the decrease of particle size and the temperature of maximum weight loss rate is in advance, which occurred at 306.97 °C for 30 µm corn starch particles. Although the

SC

samples with different particle size show the same thermal behavior, the thermal

samples.

3.2 Flame propagation Behavior

M AN U

weight loss completion temperature of 65 µm sample is much higher than the other

TE D

Fig.5 shows the series of direct high-speed photographs of flame propagation process of corn starch dust clouds and the dust mass concentration in these experiments is controlled at 400 g/m3. The flame structures and position information

EP

are captured at different times. Thus, the flame surface can be approximately defined

AC C

by the illumination information on the flame from the high-speed camera images. The process from ignition to flame forming of dust clouds lasts a long time which is different from that of the premixed gas flame (Chen et al., 2012). Unlike the premixed fuel gas, the dust particles are firstly pyrolyzed before ignition and then burnt in gaseous phase, which is a slow process at lower ignition energy. The time of flame formation is about 50 ms and then the flame begins to spread around rapidly. When the flame is restricted by duct wall, it begins to accelerate to the open end of duct and

7

ACCEPTED MANUSCRIPT the flame acceleration state comes into being. As shown in these images, a large number of folds and vortices in the flame can be clearly observed and the combustion process is very uneven, which is disturbed by the spraying airflow. Therefore, the

RI PT

combustion of flame is in turbulent flow in the duct. There are many dark dots appeared at the bottom of the flame and the phenomenon can be explained as follows: the formation of agglomeration is mainly due to the small particle size of starch dust.

SC

As the increasing of the particle size, the particles are influenced by gravity thus the

M AN U

suspending particles will settle down. What’s more, the phenomenon that particles gathered at the bottom of duct causes partial high-intensity combustion. Comparing the luminous zone of flame, the red flame with uniform brightness is presented during the propagation process, when the average particle size of the dust cloud is 30 µm.

bottom of duct.

TE D

With increasing of the particle size, more brightness yellow flame is observed at the

Fig. 6 shows the variation of dust flame propagation velocity with different particle

EP

size distributions. The flame propagation velocity is less than 0.5 m/s within 0-50 ms.

AC C

After 50 ms, the flame propagation velocity increases rapidly. According to the flame images, the flame is accelerated out of the open end of duct, indicating that the flame propagation of dust explosion in the duct is an accelerated process. As the particle size decreases, the flame propagation velocity increases and this experimental results agree well with studies of Bidabadi et al (2010).

3.3 Reaction zone thickness

8

ACCEPTED MANUSCRIPT The combustion process of starch dust is similar to that of hydrocarbon fuel and a -

large number of intermediate ion is produced, such as H+ , OH and CH3 . The region with high ion density appears in the combustion reaction zone and is inclined to the

RI PT

side of combustion products. Therefore, the fluctuation of ion current can be taken as an important parameter to characterize the structure of combustion reaction zone (Han et al., 2000). Based on Ju’s research, the combustion reaction zone thickness of visible

SC

flame is represented by the half value of peak width of ion current (Ju et al., 1998).

following formula:

M AN U

The combustion reaction zone thickness of visible flame δ is defined as the

δ=vf ×∆t

(1)

where vf is the mean propagation velocity of flame, and ∆t is the time for half peak

TE D

width of ion current curve.

Fig. 7 shows the ion current signal waveform of the dust flame with the average particle size of 30 µm and the mass concentration is 600 g/m3. A single ion current

EP

peak appears at about 91 ms after ignition, and then the intensity decreases. At about

AC C

176 ms, the ion current turns to be almost zero. The curve is smooth without fluctuation, indicating that under this case, the flame combustion is controlled by the premixed gas explosion, where a homogeneous chemical reaction occurs (Gao et al., 2015). The combustion reaction zone thickness is obtained as 42.5 mm which is much thicker than the premixed gas flame. The relationship between flame combustion reaction zone thickness and particle sizes can be obtained in Fig. 8. The particle size of dust greatly affects the reaction

9

ACCEPTED MANUSCRIPT zone structure of flame, and its thickness increases with the rise of particle size. Combined with TG-DTG curves, it can be seen that under the same heating rate, the decomposition of smaller particle size happens earlier than that of larger particle size.

RI PT

At the same time, the pyrolysis rate of smaller particles is higher than that of larger ones. Therefore, the dust clouds with smaller particle size can be devolatillzed in a short period in the process of flame propagation. The flame propagation type is

SC

similar to that of premixed gas explosion, and the chemical reaction is controlled by

M AN U

the kinetics in homogeneous phase. Thus the flame combustion reaction zone thickness is relatively thin. With increasing of particle size, the whole combustion process tends to be controlled by the pyrolysis and gasification rate of the particles. The reaction time and zone of flame combustion get increased, resulting in the

TE D

enlarged thickness of combustion reaction zone. The thickness of flame reaction zone of corn starch dust clouds ranges from 24 mm to 60 mm, which is several dozen times

2015).

EP

larger than that of premixed gas flame (Lafay et al., 2008; Li et al., 2014; Gao et al.,

AC C

Under same particle size conditions, flame reaction zone thickness increases firstly and then remains almost constant with the increase of the dust cloud concentration. This explanation for this phenomenon is related to the oxygen content in the duct. When the concentration of dust cloud is relatively low, the oxygen content in the duct is sufficient, and the volatile products of dust cloud can be completely burned, and the flame thickness is relatively thin. With increasing of the dust cloud concentration, the pyrolysis product of dust cloud increases continuously, and the content of oxygen is

10

ACCEPTED MANUSCRIPT insufficient to support the complete combustion of dust cloud. The heterogeneous combustion of flame in dust cloud is more obvious, leading to the increase of flame thickness. When the flame reaction zone thickness reaches to a certain value, the dust

RI PT

cloud concentration increases and the thickness of flame reaction zone remains almost unchanged due to the limitation of oxygen in the duct. According to the fitting result of the flame combustion zone thickness, the R2 is above 0.98, thus the change rule of

SC

different particle sizes has a high regularity.

M AN U

3.4 Preheat zone thickness

According to thermodynamics and combustion theory on flame propagation, the combustion temperature of dust cloud plays an important role in flame propagation and structure changes owing to that temperature rise can accelerate the chemical

TE D

reaction process. In this study, a fine thermocouple together with ion current probe are applied to detect the structural of flame preheat zone. During the experiment, the fine thermocouple and ion probe were installed on the same plane of the duct, thus data

EP

concerning to the flame is detected at the same position.

AC C

The response time of fine thermocouple is 6.7 ms. In order to eliminate the influence of thermal inertia, the thermocouple must be revised and validated before each test. It is assumed that the convective heat transfer process is much larger than that of heat conduction and radiation. The temperature correction formula is shown as follows ( Ballantyne and Moss, 1977; Gao et al., 2012): T = Tm + τ

dTm dt

（2）

11

ACCEPTED MANUSCRIPT Where, Tm is the temperature measured by the thermocouple, and τ is the constant of time. In our experiments, the value of τ is approximately 6.7×10-3 s. Fig. 9 shows the temperature curves and ion current curves of corn starch dust

RI PT

flame with various particle sizes. In addition, the temperature data in the figures is revised by Eq. (2). As can be seen from the diagram, the thermocouple responses more rapidly than that of ion current, which indicates that the temperature rises earlier

SC

than the formation of ion current. t1 is the onset time of temperature, t2 is the onset

M AN U

time of ion current. With the increasing of ion current, the temperature rises rapidly. According to Gao’s research (Gao et al., 2014), during the period of t1 to t2 , the temperature of the flame increases without any chemical reaction. The flame temperature rising is mainly caused by the heat transfer and heat radiation from the

TE D

chemical reaction zone. The flame does not spread to the area yet, only the dust particles are heated to pyrolysis, then burn does not occur. After t2 , with ion current value increasing, the chemical reaction increases obviously. It shows that the dust

EP

cloud is ignited and the flame spreads into the combustion zone. As a result, the area

AC C

from t1 to t2 corresponds to the flame preheat zone. The temperature value at t2 can be regarded as the minimum ignition temperature of dust clouds. With the increasing of particle size, the minimum ignition temperature increases. An ideal dust flame with the thicknesses defined is schematically shown in Fig.10. The preheat zone is in the front of the reaction zone and the dust particles are pyrolyzed in this area. Dust volatile gas is ignited at the flame front and the flame simultaneously ignites the surrounding dust particles that are being pyrolyzed. The flame fronts are continuously

12

ACCEPTED MANUSCRIPT spreading toward the preheat zone. Generally, the reaction zone thickness of visible flame can be characterized by the ion current wave. Using this characteristic, the flame preheat zone structures of flame can be obtained.

RI PT

Fig.11 shows the thickness variation of flame preheat zone in corn starch dust cloud. It can be seen from Fig.11 that the thickness of the corn starch cloud flame preheat zone is between 10-38 mm. The thickness of flame preheat zone becomes thinner with

SC

the particle size decreasing, which means the dust cloud with smaller particle size

M AN U

shows higher pyrolysis rate and the formation process of combustible gas phase is much faster. In smaller particle size dust, the flame tends to show homogeneous combustion, thus the flame preheat zone becomes thinner. Under the premise of same particle size, flame preheat zone thickness increases with the rise of dust cloud

TE D

concentration. The thickness shows a quadratic curve variation, which deceases firstly and then increasing and the preheat zone thickness reaches to the minimum value when the concentration of dust cloud is about 500 g/m3. The reason is related to the

EP

explosion characteristics of corn starch (Radandt et al., 2001). When the concentration

AC C

of dust cloud is about 500 g/m3, an optimum concentration causes the maximum explosion pressure (Pmax) and maximum pressure rise rate (Kst) to the maximum value. According to the combustion theory, the fuel-air equivalence ratio of this dust cloud concentration in this study is very close to 1. Thus, the dust clouds combustion is the most sufficient and the flame behaviors are almost the same as gaseous flame. Meanwhile, the flame propagation velocity reaches to the maximum value, and the thinnest flame preheat zone is also obtained.

13

ACCEPTED MANUSCRIPT 4. Conclusions The thickness of flame combustion reaction zone and preheat zone through corn starch dust clouds with three different particle size distributions are characterized and

RI PT

the changing rules of flame structures are determined. For the three different particle size distributions, the thickness range of flame reaction zone is 24-60 mm and the minimum thickness corresponds to the average particle size of 30 µm. The flame

SC

reaction zone thickness increases constantly with the increase of particle size.

M AN U

Additionally, the reaction zone structures are also influenced by the dust concentration. There is a positive correlation between the particle size and preheat zone thickness and the thickness range in this research is 10-38 mm. The preheat zone thickness shows a quadratic curve relationship with dust concentration. The thermal behavior

TE D

analysis shows that burning rate and burning time of corn starch dust are also affected by particle size distributions, which is reflected in the flame propagation behaviors

AC C

EP

and flame structures.

Acknowledgments

The authors gratefully acknowledge the financial supports from the National Key

Research and Development Program of China (Grant Nos. 2016YFC0802801 and 2017YFC0804705) and the National Natural Science Foundation of China (Grant Nos. 51374164 and 51774221).

14

ACCEPTED MANUSCRIPT

RI PT

References [1] Zhong Shengjun, Deng Xufan, 2000. Modeling of maize starch explosions in a 12 m3 silo. Journal of Loss Prevention in the Process Industries. 13(3), 299-309.

SC

[2] Chao Jenny, Dorofeev Sergey B., 2015. Evaluating the overall efficiency of a

Process Industries. 36, 63-71.

M AN U

flameless venting device for dust explosions. Journal of Loss Prevention in the

[3] Radandt S., Shi Jianye, Vogl A., Deng X. F., Zhong S. J., 2001. Cornstarch explosion experiments and modeling in vessels ranged by height/diameter ratios.

TE D

Journal of Loss Prevention in the Process Industries. 14(6), 495-502. [4] Ramírez Á., Garcíatorrent J., Aguado P. J., 2009. Determination of parameters used to prevent ignition of stored materials and to protect against explosions in food

EP

industries. Journal of Hazardous Materials. 168(1), 115-120.

AC C

[5] Eckhoff Rolf K.,2003. Dust Explosions in the Process Industries, third ed. Gulf Professional Publishing, Amsterdam. [6] Amyotte Paul R., Pegg Michael J., Khan Faisal I., 2009. Application of inherent safety principles to dust explosion prevention and mitigation. Process Safety and Environmental Protection. 87(1), 35-39. [7] Abbasi T., Abbasi S. A., 2007. Dust explosions-cases, causes, consequences, and control. Journal of Hazardous Materials. 140(1-2), 7-44.

15

ACCEPTED MANUSCRIPT [8] Amyotte Paul R., Eckhoff Rolf K., 2010. Dust explosion causation, prevention and mitigation: An overview. Journal of Chemical Health and Safety. 17(1), 15-28. [9] Yuan Chunmiao, Amyotte Paul R., Hossain Md Nur, Li Chang, 2014. Minimum

TiO2 powder. Journal of Hazardous Materials. 275, 1-9.

RI PT

ignition temperature of nano and micro Ti powder clouds in the presence of inert nano

[10] Amyotte Paul R., Khan Faisal I., Basu Atreyee, Dastidar Ashok G., Dumeah

SC

Robert K., Erving William L., 2006. Explosibility parameters for mixtures of

M AN U

pulverized fuel and ash. Journal of Loss Prevention in the Process Industries. 19(2-3), 142-148.

[11] Cloney Chris T., Ripley Robert C., Amyotte Paul R., Khan Faisal I., 2013. Quantifying the effect of strong ignition sources on particle preconditioning and

26(6), 1574-1582.

TE D

distribution in the 20-L chamber. Journal of Loss Prevention in the Process Industries.

[12] Eckhoff R. K., 2002. Minimum ignition energy (MIE): a basic ignition sensitivity

EP

parameter in design of intrinsically safe electrical apparatus for explosive dust clouds.

AC C

Journal of Loss Prevention in the Process Industries. 15(4), 305-310. [13] Gao Wei, Yu Jianliang, Mogi Toshio, Zhang Xinyan, Sun Jinhua, Dobashi Ritsu, 2014. Effects of particle thermal characteristics on flame microstructures during dust explosions of three long-chain monobasic alcohols in a half-closed chamber. Journal of Loss Prevention in the Process Industries. 32, 127-134.

16

ACCEPTED MANUSCRIPT [14] Gao Wei, Mogi Toshio, Sun Jinhua, Yu Jianliang, Dobashi Ritsu, 2013. Effects of particle size distributions on flame propagation mechanism during octadecanol dust explosions. Powder Technology. 249(249), 168-174.

RI PT

[15] Gao Wei, Mogi Toshio, Yu Jianliang, Yan Xingqing, Sun Jinhua, Dobashi Ritsu, 2015. Flame propagation mechanisms in dust explosions. Journal of Loss Prevention in the Process Industries. 36, 186-194.

SC

[16] Han O. S., Yashima M., Matsuda T., Matsui H., Miyake A., 2000. Behavior of

M AN U

flames propagating through lycopodium dust clouds in a vertical duct. Journal of Loss Prevention in the Process Industries. 13(6), 449-457.

[17] Bidabadi Mehdi, Haghiri Ali, Rahbari Alireza, 2010. The effect of Lewis and Damköhler numbers on the flame propagation through micro-organic dust particles.

TE D

International Journal of Thermal Sciences. 49(3), 534-542.

[18] Proust Christophe, 2006. Flame propagation and combustion in some dust-air mixtures. Journal of Loss Prevention in the Process Industries. 19(1), 89-100.

EP

[19] Proust Christophe, 2006. A few fundamental aspects about ignition and flame

AC C

propagation in dust clouds. Journal of Loss Prevention in the Process Industries. 19(2-3), 104-120.

[20] Di Benedetto A., Russo P., 2007. Thermo-kinetic modelling of dust explosions. Journal of Loss Prevention in the Process Industries. 20(4-6), 303-309. [21] Chen Xianfeng, Sun Jinhua, Liu Yi, Liu Xuanya, Chen Sining, Lu Shouxiang, 2007. Microstructure of premixed propane/air flame in the transition from laminar to turbulent combustion. Chinese Science Bulletin. 52(5), 685-691.

17

ACCEPTED MANUSCRIPT [22] Chen Xianfeng, Zhang Hongming, Chen Xi, Liu Xuanya, Niu Yi, Zhang Ying, Yuan Bihe, 2017. Effect of dust explosion suppression by sodium bicarbonate with different granulometric distribution. Journal of Loss Prevention in the Process

RI PT

Industries, https://doi.org/10.1016/j.jlp.2017.02.012. [23] Li X., Shang Y. J., Chen Z. L., Chen X. F., Niu Y., Yang M., Zhang Y., 2017. Study of spontaneous combustion mechanism and heat stability of sulfide minerals

SC

powder based on thermal analysis. Powder Technology. 309, 68-73.

M AN U

[24] Yang Man, Chen Xianfeng, Wang Yujie, Yuan Bihe, Niu Yi, Zhang Ying, Liao Ruoyu, Zhang Zumin, 2017. Comparative evaluation of thermal decomposition behavior and thermal stability of powdered ammonium nitrate under different atmosphere conditions. Journal of Hazardous Materials.337,10-19.

TE D

[25] Chen Xianfeng, Zhang Yin, Zhang Ying, 2012. Effect of CH4–Air Ratios on Gas Explosion Flame Microstructure and Propagation Behaviors. Energies. 5(12), 4132-4146.

EP

[26] Ju Wen-Jun, Dobashi Ritsu, Hirano Toshisuke, 1998. Reaction zone structures

AC C

and propagation mechanisms of flames in stearic acid particle clouds. Journal of Loss Prevention in the Process Industries. 11(6), 423-430. [27] Gao Wei, Mogi Toshio, Rong Jianzhong, Yu Jianliang, Yan Xingqing, Dobashi Ritsu, 2015. Motion behaviors of the unburned particles ahead of flame front in hexadecanol dust explosion. Powder Technology. 271, 125-133.

18

ACCEPTED MANUSCRIPT [28] Lafay Y., Renou B., Cabot G., Boukhalfa M., 2008. Experimental and numerical investigation of the effect of H2 enrichment on laminar methane–air flame thickness. Combustion and Flame. 153(4), 540-561.

RI PT

[29] Li D., Chen X., Zhang Y., Rangwala A., 2014. Premixed CH4-Air Flame Structure Characteristic and Flow Behavior Induced by Obstacle in an Open Duct. Advances in Mechanical Engineering. 7, 248279-248279.

SC

[30] Gao Wei, Dobashi Ritsu, Mogi Toshio, Sun Jinhua, Shen Xiaobo, 2012. Effects

M AN U

of particle characteristics on flame propagation behavior during organic dust explosions in a half-closed chamber. Journal of Loss Prevention in the Process Industries. 25(6), 993-999.

[31] Ballantyne A., Moss J. B., 1977. Fine Wire Thermocouple Measurements of

AC C

EP

TE D

Fluctuating Temperature. Combustion Science and Technology. 17:(1-2), 63-72.

19

ACCEPTED MANUSCRIPT

RI PT

Table captions

Table 1

Proximate Analysis (wt %) Fixed Vdaf

Ash

Carbon 0.8

C

H

O

N

S

39.27

6.56

53.03

0.04

0.00

TE D

85.6

Ultimate Analysis (wt %)

AC C

EP

13.6

M AN U

Proximate and ultimate analysis of corn starch dust

SC

Table 1 Proximate and ultimate analysis of corn starch dust

20

ACCEPTED MANUSCRIPT

RI PT

Figure captions Fig. 1 Schematic diagram of experimental apparatus.

Fig. 3 Particle size distribution of corn starch dust.

SC

Fig. 2 Sketch of ion current probe.

M AN U

Fig. 4 TG-DTG curves of the corn starch dust: (a) TG; (b) DTG.

Fig. 5 High-speed direct images of flames with different particle size distributions: (a)

30 µm; (b) 45 µm; (c) 65 µm (dust concentration: 400 g/m3).

g/m3).

TE D

Fig. 6 Variation of flame propagation velocity with time (dust concentration: 400

Fig. 7 Ion current fluctuation curve (average particle size: 30 µm, dust concentration:

EP

600g/m3).

Fig. 8 Thickness of flame combustion zone.

AC C

Fig. 9 Relationships between flame temperature and ion current with different particle

sizes : (a) 30 µm; (b) 45 µm; (c) 65 µm (dust concentration: 400 g/m3). Fig. 10 Sketch of an idealized dust flame thickness Fig. 11 Thickness of flame preheat zone.

21

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

Fig. 1. Schematic diagram of experimental apparatus.

Fig. 2. Sketch of ion current probe.

22

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

Fig. 3. Particle size distribution of corn starch dust

23

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Fig. 4. TG-DTG curves of the corn starch dust: (a) TG; (b) DTG.

24

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

TE D

Fig. 5. High-speed direct images of flames with different particle size distributions: (a)

AC C

EP

30 µm; (b) 45 µm; (c) 65 µm (dust concentration: 400 g/m3).

25

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Fig. 6. Variation of flame propagation velocity with time (dust concentration: 400

AC C

EP

TE D

g/m3).

Fig. 7. Ion current fluctuation curve (average particle size: 30 µm, concentration: 600g/m3).

26

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

Fig. 8. Thickness of flame combustion zone.

27

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Fig. 9. Relationships between flame temperature and ion current with different

AC C

particle sizes: (a) 30 µm; (b) 45 µm; (c) 65 µm (dust concentration: 400 g/m3).

28

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

Fig. 10. Sketch of an idealized dust flame thickness

Fig. 11. Thickness of flame preheat zone.

29

ACCEPTED MANUSCRIPT Dust particle size plays a key role in both flame reaction zone and preheat zone. The thickness ranges of flame reaction zone and preheat zone are determined. Preheat zone thickness has a quadratic curve relationship with dust concentration.

AC C

EP

TE D

M AN U

SC

RI PT

Flame structures are influenced by the thermal behavior of dust.