- Email: [email protected]
Effects of reduced oxygen levels on flame propagation behaviors of starch dust deflagration
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
Contents lists available at ScienceDirect
Journal of Loss Prevention in the Process Industries journal homepage: www.elsevier.com/locate/jlp
Effects of reduced oxygen levels on flame propagation behaviors of starch dust deflagration
T
Hongming Zhanga, Xianfeng Chena,∗, Tian Xiea, Bihe Yuana, Huaming Daia, Song Hea, Xuanya Liub a b
School of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, China Key Laboratory of Building Fire Protection Engineering and Technology of MPS, Tianjin 300381, China
A R T I C LE I N FO
A B S T R A C T
Keywords: Dust explosion Deflagration flame Flame propagation behavior Oxygen concentration
An experimental investigation of reduced oxygen levels on flame behaviors of starch dust deflagration is studied in a half-closed dust explosion tube. Six kinds of nitrogen/air ratio mixtures are prepared as the experimental atmosphere for flame propagation. A high-speed photography is used to record flame propagation behaviors and luminous features. The characteristics of flame temperature and deflagration pressure are measured by fine thermocouple and pressure sensor, respectively. The influence of reduced oxygen concentration on flame propagation characteristics are discussed in different conditions. The results show that reducing the oxygen content can effectively retard the flame propagation process and weaken the dust combustion reaction. Compared with air atmosphere, flame acceleration process under reduced oxygen atmosphere is suppressed and the luminous intensity of flame post-combustion zone is greatly diminished. The characteristic parameters of flame velocity, flame temperature and deflagration pressure are all influenced by the reduction in oxygen concentration. Furthermore, effective partial inerting suppression can be achieved for starch deflagration flame, when the oxygen concentration is lower than 18.90%. This study is helpful for dust explosion prevention and mitigation. Additionally, the dust cloud concentration changes the flame velocity characteristic at low oxygen levels and higher dust concentration is more limited the flame acceleration.
1. Introduction Starch has a very wide range of applications in lots of processing industries, such as food processing, chemical synthesis and pharmaceutical manufacturing. However, starch dust diffusion in the processing industries cannot be effectively controlled, usually leading to increase the possibility of starch dust explosions. Starch production and processing require a high degree of environmental cleanliness, and the inerting method to reduce the ambient oxygen concentration is a very suitable application for the prevention of starch dust explosions. Studies on inerting technology have been extensively conducted, and most of the studies have focused on determination of limiting oxygen concentration (LOC) for the dust explosion (Nomura et al., 1984; Dastidar et al., 1999; Dastidar and Amyotte, 2002; Cashdollar, 2000; Going et al., 2000; Eckhoff, 2003). The LOC is the maximum oxygen concentration of a dust/air/inert gas mixture by which dust explosions cannot occur. Both EN 14034-4 (2005) and ASTM E2931 (2013) describe the standard test method for LOC of combustible dust clouds in the 1 m3 vessel or 20 L sphere apparatus and point out that the
∗
measurement of LOC is the basis for inerting and explosion prevention. Mittal (2013) experimentally measured the LOC data for two types of coals with varying volatile matter and evaluated the effect of particle size and moisture on LOC. Particle size has a comparatively small influence on the LOC of coal dust while the effect of moisture on LOC is opposite. Furthermore, the LOC measurement cannot be directly used to represent the inerting level and the processing conditions need consideration. Wilen et al. (1998) found that the LOC values of biomass dust clouds increase with increasing initial pressure in the range 0.5–1.8 MPa. This result was different from coal dust and others. Hassan et al. (2014) proposed a predictive model to assess the probability of a dust explosion occurrence under a given environment. LOC was considered as a key parameter in the model and the probabilistic model of dust explosion with respect to the change of oxygen concentration was determined. Meanwhile, reducing the oxygen content above the LOC in the atmosphere by adding nitrogen is another core mean of preventing and mitigating dust explosions (Eckhoff, 2005) and partial inerting is called for more extensive use in industrial dust explosion protection (Eckhoff, 2004). The European standard CEN/TR
Corresponding author. E-mail address: [email protected] (X. Chen).
https://doi.org/10.1016/j.jlp.2018.03.011 Received 16 November 2017; Received in revised form 16 March 2018; Accepted 16 March 2018 Available online 17 March 2018 0950-4230/ © 2018 Elsevier Ltd. All rights reserved.
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
thermocouple, a pressure sensor and a programmable logic controller. The size of square dust combustion tube was 1000 mm × 80 mm × 80 mm , which was connected by two half-meter long tubes. The front and rear sides of tube were set as optical windows, which were convenient to observe the flame propagation process. A dust dispersing device was installed at the bottom of the tube. A square vent with the size of 20 mm × 20 mm was placed at the top of the tube. The vent was opened to connect the external space for the nitrogen/air mixed gas flow through the tube to form an inert atmosphere. A pair of ignition electrodes were positioned at 50 mm above the bottom of the tube and linked with a high voltage ignition unit, which could generate ignition spark. The gas mixing system was used to form air/nitrogen mixtures with different concentrations. Furthermore, the mixed gas was filled into the tube to create an inerting atmosphere and to act as a driving gas for dust dispersion. In this study, the flame propagation behaviors were captured by a high speed photography camera (Photron FASTCAM SA1.1) with a normal lens of Nikon AF Nikkor 50 mm f/1.4. A fine thermocouple and a pressure sensor were mounted at 120 mm below the top of the tube. The thermocouple was comprised of 25 μm diameter Pt-Pt/Rh 13% wires. The flame deflagration pressure was measured by a pressure transducer (PCB Piezotronics, 113B21), which has a measuring range of 0–1379 kPa and a sensitivity of 3.703 mV/kPa. Data recorder was used to record the temperature and pressure data. Synchronous trigger switches of each system were controlled by the programmable logic controller.
15281:2006 (CEN, 2006) has systematically introduced the influence of oxygen concentration on explosion atmospheres and the methods of inerting. The supply of inert gas is confirmed and the empirical equations is provided for inerting design. Li et al. (2009) investigated the inerting effect of N2, CO2 and Ar on the magnesium dust explosion. Comparing the parameters of explosion characteristics, the inerting effect of N2 is better than Ar, and N2 has a better economic efficiency for dust explosion prevention as an inert gas. Chaudhari and Mashuga (2017) utilized a modified standard minimum ignition energy (MIE) device to measure the MIE value of partial inerting dust clouds. The partial inerting MIE was influenced by purge-induced turbulence and a higher MIE value was obtained due to the turbulence in the ignition zone. As mentioned above, studies on the dust explosion sensitivity under inerting conditions have been very extensive. However, the flame propagation process is also an essential stage for the transformation from deflagration to explosion, which is always unsteady and controlled by numerous factors (Gao et al., 2013; Chen et al., 2017; Li et al., 2017; Yang et al., 2017; Zhang et al., 2017). While, few studies are available on flame propagation characteristics of dust deflagration with the reduction of oxygen concentrations. In the present study, nitrogen was chosen as an inert gas, and the effect of reduced oxygen levels with different nitrogen/air ratios on flame propagation behaviors of starch dust deflagration was studied in a semi-closed tube. The flame propagation processes with different oxygen concentrations was recorded by high speed photography system. Flame temperature and deflagration pressure were detected by fine thermocouple and pressure sensor, respectively. Furthermore, the variation of flame propagation dynamics was explored in depth.
2.2. Experimental materials 2. Experimental
Wheat starch was utilized for experiments. Before experiments, dust samples were sifted in a vibration sieve with 300 mesh to ensure the particle size distribution in the same range. Then, the samples were dried and dewatered in a vacuum drying oven at 50 °C for 12 h. The median particle size (D50) of the samples was 45 μm. High purity nitrogen (purity ≥ 99.9996 vol.%) and compressed air (oxygen concentration is 21.00 vol.%) were used for preparing different gas mixtures. The dust clouds was ignited by the high-voltage transformer with an output of 15 kV. The ignition duration was 100 ms and the nominal ignition energy was approximately 10 J.
2.1. Experimental apparatus A schematic diagram of experimental apparatus was illustrated in Fig. 1, composed of nine parts: a small-scale square combustion tube, a dust dispersing device, a gas mixing system, a high voltage ignition unit, a high speed photography system, a data recorder, a
2.3. Experimental conditions Experiments were carried out in the different oxygen concentration atmosphere of 6 nitrogen/air ratios. The oxygen volume concentration for the mixed gases were 12.60%, 14.70%, 16.80%, 18.90%, 19.95% and 21.00%. The detailed conditions of the gas mixture were shown in Table 1. In order to ensure the accuracy of the oxygen concentration in the experimental atmosphere, the flow through technique relying on the purge mixed gas was used to remove the original gas in tube. The gas flow in the experiments was 5 L/min and the purge time of the mixed gas flow through the tube was calculated from the following equation (CEN, 2006): Table 1 Experimental gas atmospheres for starch dust explosions.
Fig. 1. Sketch of experimental apparatus.
147
No.
Nitrogen (%)
Air (%)
Oxygen concentration (%)
Gas composition
1 2 3 4 5 6
0 5 10 20 30 40
100 95 90 80 70 60
21.00 19.95 18.90 16.80 14.70 12.60
Air Air-Nitrogen Air-Nitrogen Air-Nitrogen Air-Nitrogen Air-Nitrogen
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
t =F
V C − Co ⎞ ln ⎛⎜ i ⎟ Q ⎝ Ci − Cf ⎠
complete inerting in dust explosion protection because of the sealed test device. In this study, the measured value of 12.60% is the maximum oxygen concentration to prevent the starch deflagration flame propagation. The reduction of oxygen content also prolongs the deflagration flame spread time in the tube. It takes about only 180 ms for the flame spreading from the bottom of the tube to the top in the air atmosphere. The flame propagation time is about 500 ms, nearly 2.8 times of which in air atmosphere in the case of 14.70% oxygen concentration. Du et al. (2014) found that the devolatilization process of starch was not affected by the changes of oxygen concentration. Therefore, the time of starch ignition and burning is increased correspondingly, and the combustion reaction rate is decreased due to the reduced oxygen content.
(1)
where, t was the time required for purging (h), F was the safety factor for purging, V was the system volume (m3), Q was the inert gas flowrate (m3/h), Cf was the oxygen content after flow purging (vol.%), Co was the initial oxygen content (vol.%), Ci was the oxygen content of inert gas (vol.%). In this study, the maximum time required for purging was at least 117 s, thus the purge time of mixed gas flow through was set as 2 min. Before experiments, a certain of dust sample weighted by an electronic balance was evenly placed on the sample dish at the bottom of the tube. The nitrogen/air mixed gas was filled into the tube and dust particles were dispersed into the tube by the air flow. Delay time for ignition was set to minimize the influence from the airflow. After the ignition, high speed photography camera and data recorder were turned on. The detailed experimental parameters were listed as follows: pressure for powder injection, 0.2 MPa; injection time, 0.2 s; ignition delay time, 0.1 s; framing rate of high speed photography camera, 1000 frames/s.
3.2. Influence of reduced oxygen concentration on flame propagation velocity The flame propagation velocities in different oxygen concentration are shown in Fig. 3. As presented in Fig. 3, the flame propagation velocities show different behaviors in the reduced oxygen content. When the oxygen concentration is between 19.95% and 21.00%, the flame spreads in an oxygen-enriched atmosphere and it is less affected by oxygen content. The high acceleration in the initial stage of flame propagation is a significant feature. Maximum flame propagation velocity appears in the middle stage of flame propagation and the values are all above 6 m/s. Due to the cooling effect of tube walls, the flame velocities decline after reaching the maximum flame propagation velocity. It takes less than 250 ms for the whole flame propagation process. However, when the oxygen concentration is below 18.90%, the flame speed begins to slow down with no obvious acceleration or deceleration due to the oxygen-poor atmosphere. There is a long period of low-speed propagation in the initial stage and the whole propagation process lasts more than 400 ms. The flame propagation in the lower oxygen content is limited. Therefore, the oxygen concentration of 18.90% is an inflection point of the reduced oxygen content for flame velocities suppression in this study. The increase of dust concentration exacerbates the restrictions on flame propagation, especially in the oxygen-poor atmosphere. When the dust concentration is less than 300 g/m3, flame propagation velocity in the oxygen-poor atmosphere shows the same growth trend as the oxygen-enriched atmosphere. Due to the low concentration of dust cloud, the oxygen content required for the combustion of dust clouds per unit volume has not reduced drastically and the process of flame acceleration is slightly suppressed. When the dust cloud concentration is above 400 g/m3, the oxygen content in the tube is consumed quickly because of the increasing dust cloud concentration and significant suppressed flame velocities. Fig. 4 shows the relationship between maximum flame propagation velocity and arrival time, when the dust concentration is 400 g/m3. There are two stages for the influence of oxygen concentration on maximum flame propagation velocity. The first stage is oxygen concentration between 18.90% and 21.00%, and the reduced oxygen content has an obvious suppression effect on maximum flame velocity in this stage. The maximum flame velocity presents linear decline with the decrease of oxygen concentration. It takes less than 180 ms for all the flame propagation in this stage to arrive the peak value. Moreover, the second stage is in the oxygen concentration range of 14.70%–18.90%. In this stage, with reducing the oxygen concentration, the maximum flame velocity is stably suppressed and the value is maintained at about 3 m/s without significant fluctuations. The time reaching maximum flame velocity is greatly prolonged, and a linear increase with decreasing oxygen concentration is obtained.
3. Results and discussions 3.1. Influence of reduced oxygen concentration on flame propagation process Fig. 2 shows a series of direct high-speed photographs of flame propagation process of starch dust deflagrations with different oxygen concentrations and the dust mass concentration in these experiments is controlled at 400 g/m3. Suspended dust clouds were ignited by ignition electrodes and the flame began to propagate from the bottom to vent. In the upward propagation process, the flame structures and position information were recorded. Fig. 2 (a)-(e) are the flame propagation processes at the oxygen concentration of 21.00%, 19.95%, 18.90%, 16.80% and 14.70%. It can be seen from the high-speed photography images that the luminous area behind the front of the flame is filled in the entire tube. Unlike the premixed gas flame propagation process, the combustion of starch particles cannot be completed instantaneously at the flame front and this area is the post-combustion zone of the deflagration flame propagation. Unburned dust particles conduct devolatilization and combustion in the post-combustion zone, and further provide heat to promote flame spread upward. The flame luminous intensity in the post-combustion zone is the external manifestation of dust combustion reaction. In the air atmosphere with 21.00% oxygen concentration, yellow light-emitting area is the dominant and the overexposed white light appears locally, due to the sufficient chemical reaction under oxygen-rich conditions. The direct effects of reduced oxygen content on flame propagation is the diminished luminous intensity of the post-combustion zone. When the oxygen concentration is reduced to 18.90%, the flame emission brightness is dominated by light red, and the range of yellow luminous area in the flame is gradually significantly. When the oxygen concentration is 14.70%, the dust clouds can be ignited and propagate in the tube, but the luminous brightness of the flame is almost invisible. What's more, the flame is extinguished automatically when it spread to the 850 mm from the bottom of the tube. Because of the limitation of the oxygen content, the deflagration flame takes away the oxygen during the propagation process, making the unburned dust particles unable to support combustion and spontaneous extinguishment. When the oxygen concentration is 12.60%, the starch cloud cannot be ignited after several experiments. Furthermore, starch clouds with other dust cloud concentrations also cannot be ignited on this oxygen concentration. Result indicates that the limiting oxygen concentration in these experimental conditions is approximately 12.60%. However, this value is a little higher than the LOC of 11% measured under standard test conditions (NFPA, 2008). The results of the standard tests are relatively more accurate for the
3.3. Influence of reduced oxygen concentration on flame temperature In order to monitor the temperature behavior of flame combustion 148
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
Fig. 2. Flame propagation at different oxygen concentrations: (a) 21.00%; (b) 19.95%; (c) 18.90%; (d) 16.80%; (e) 14.70% (dust concentration: 400 g/m3).
Fig. 5 shows the temperature curves of flame combustion zone measured by fine thermocouple when the dust concentration is 400 g/ m3. Reducing the oxygen concentration delays the detecting time of the thermocouple and prolongs the flame temperature curve rise time. In the air atmosphere, the maximum flame temperature is about 800 °C, and the values decreases with the oxygen content reduction. When the oxygen concentration decreases to 14.70%, the maximum flame temperature is only about 70% of the value for air atmosphere. The combustion mechanism of starch usually includes such steps as particle heating, devolatilization, volatile-air mixture and homogeneous combustion. (Du et al., 2014; Zhang et al., 2017). The decrease in the flame temperature indicates that the combustion of dust particles is incomplete and the heat of combustion is reduced at a lower oxygen
zone in the tube, the flame temperature was measured by the fine thermocouple installed in the middle of the tube. The data measured by the thermocouple needs to be compensated due to the thermal inertia. It is assumed that the convective heat transfer process between thermocouple wires is mainly from thermal radiation and heat conduction, thus the thermocouple temperature correction formula can be expressed as follows (Ballantyne and Moss, 1977; Gao et al., 2014):
T = Tm + τ
dTm dt
(2)
Where Tm is the temperature measured by the thermocouple, and τ is the response time of thermocouple. In our experiments, the value of τ is approximately 6.7 × 10−3 s. 149
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
Fig. 3. Flame propagation velocity curves at different oxygen concentration. Dust concentrations: (a) 200 g/m3, (b) 300 g/m3, (c) 400 g/m3, (d) 500 g/m3, (e) 600 g/m3, (f) 700 g/m3.
temperature at each dust cloud concentration decreases greatly. The reduction of oxygen content greatly enhanced the resistance of the combustion reaction. Decreasing the oxygen concentration can cause partial inerting in the tube, thus this is an effective but underutilized mitigation technique for dust explosion mitigation (Eckhoff, 2004; Hoppe and Jaeger, 2005; Chaudhari and Mashuga, 2017). Thus, the influence of flame temperature verifies again that oxygen concentrations below 18.90% are effective for partial inerting on the deflagration flame.
concentration. Thus, the reduced emission rate of volatile surrounding the unburned particles and the prolonged devolatilization time delays the flame propagation. The combustion of volatile further consumes the oxygen in the environment and the combustion of dust particles is further inhibited, thereby the temperature of flame combustion zone is reduced with the decreasing of oxygen concentration. The curves of maximum flame temperature of dust cloud at different concentrations are shown in Fig. 6. As similar to the law of maximum flame propagation velocity, in a higher oxygen concentration atmosphere (18.90%–21.00%), the maximum flame temperature increases from lower dust concentration until the dangerous mass concentration of 500 g/m3 (Radandt et al., 2001), and then which tends to be almost constant. The result shows that the effect of high dust concentration on combustion kinetics of starch is more pronounced on higher oxygen concentration. The endothermic effect of cold source of the unburnt particles results in a slower temperature rise of the particles and the release rate of combustion heat is inhibited in the end. When the oxygen concentration is lower than 18.90%, the maximum flame
3.4. Influence of reduced oxygen concentration on flame deflagration pressure The flame deflagration pressure is the overpressure on the confined space during flame propagation. A pressure sensor is installed on the side of the tube to measure the deflagration pressure and the pressure curves at different oxygen concentrations are shown in Fig. 7. The maximum pressure appears in the air atmosphere is about 37 kPa, and 150
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
Fig. 4. Relationship between maximum flame propagation velocity and time (dust concentration: 400 g/m3). Fig. 7. Pressure curves at different oxygen concentrations (dust concentration: 400 g/ m3).
the peak pressures of starch deflagration flame on other oxygen contents decrease. The threat of deflagration is diminished by the reduced oxygen concentration. When the oxygen concentration is less than 18.90%, the maximum deflagration pressure is significantly reduced and pressure curve is decayed slowly at the pressure peak, which may be caused by the slow propagation of the combustion wave at low oxygen concentration and the pressure is maintained for a long time. Fig. 8 shows the variation of the maximum deflagration pressure with oxygen concentration. Due to the relatively adequate oxygen supply during the flame propagation process, the combustion process is less affected by oxygen concentration and the explosion pressure is not significantly suppressed. When the oxygen concentration is in the lower content (less than 18.90%), the deflagration pressures of each dust concentration is reduced obviously. For the lower oxygen content, most of the dust particles cannot be burned completely and the unburned particles will take away the combustion heat. The energy release from flame wave is absorbed and the pressure decreases rapidly until it cannot propagate. The inerting effect on the deflagration flame process is fully functioned by reducing the oxygen content.
Fig. 5. Flame temperature at different oxygen concentrations (dust concentration: 400 g/ m3).
Fig. 6. Maximum flame temperature of different dust cloud concentrations.
Fig. 8. Maximum explosion pressure at different dust cloud concentrations.
151
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
4. Conclusions
Chaudhari, P., Mashuga, C.V., 2017. Partial inerting of dust clouds using a modified standard minimum ignition energy device. J. Loss Prev. Process. Ind. 48, 145–150. Chen, X.F., Zhang, H.M., Chen, X., Liu, X.Y., Niu, Y., Zhang, Y., Yuan, B.H., 2017. Effect of dust explosion suppression by sodium bicarbonate with different granulometric distribution. J. Loss Prev. Process. Ind. 49, 905–911. Dastidar, A.G., Amyotte, P.R., 2002. Determination of minimum inerting concentrations for combustible dusts in a laboratory-scale chamber. Process Saf. Environ. Protect. 80 (6), 287–297. Dastidar, A.G., Amyotte, P.R., Going, J., Chatrathi, K., 1999. Flammability limits of dusts: minimum inerting concentrations. Process Saf. Prog. 18 (1), 56–63. Du, B., Huang, W.X., Kuai, N.S., Yuan, J.J., Liu, L., Ren, Y.D., 2014. Comparative study of explosion processes controlled by homogeneous and heterogeneous combustion mechanisms. J. Loss Prev. Process. Ind. 30, 155–163. Eckhoff, R.K., 2003. Dust Explosions in the Process Industries, third ed. Gulf Professional Publishing, Amsterdam. Eckhoff, R.K., 2004. Partial inerting - an additional degree of freedom in dust explosion protection. J. Loss Prev. Process. Ind. 17, 187–193. Eckhoff, R.K., 2005. Current status and expected future trends in dust explosion research. J. Loss Prev. Process. Ind. 18, 225–237. Gao, W., Mogi, T., Sun, J., Dobashi, R., 2013. Effects of particle thermal characteristics on flame structures during dust explosions of three long-chain monobasic alcohols in an open-space chamber. Fuel 113, 86–96. Gao, W., Yu, J.L., Mogi, T., Zhang, X.Y., Sun, J.H., Dobashi, R., 2014. Effects of particle thermal characteristics on flame microstructures during dust explosions of three longchain monobasic alcohols in a half-closed chamber. J. Loss Prev. Process. Ind. 32, 127–134. Going, J.E., Chatrathi, K., Cashdollar, K.L., 2000. Flammability limit measurements for dusts in 20-L and 1-m3 vessels. J. Loss Prev. Process. Ind. 13, 209–219. Hassan, J., Khan, F., Amyotte, P.R., Ferdous, R., 2014. A model to assess dust explosion occurrence probability. J. Hazard Mater. 268, 140–149. Hoppe, T., Jaeger, N., 2005. Reliable and effective inerting methods to prevent explosions. Process Saf. Prog. 24 (4), 266–272. Li, G., Yuan, C.M., Fu, Y., Zhong, Y.P., Chen, B.Z., 2009. Inerting of magnesium dust cloud with Ar, N2 and CO2. J. Hazard Mater. 170 (1), 180–183. 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 powder based on thermal analysis. Powder Technol. 309, 68–73. Mittal, M., 2013. Limiting oxygen concentration for coal dusts for explosion hazard analysis and safety. J. Loss Prev. Process. Ind. 26, 1106–1112. NFPA, 2008. Standard on Explosion Prevention Systems. NFPA 69. National Fire Protection Association, Quincy, MA. Nomura, S., Torlmoto, M., Tanaka, T., 1984. Theoretical upper limit of dust explosion in relation to oxygen concentration. Ind. Eng. Chem. Process Des. Dev. 23, 420–423. Radandt, S., Shi, J.Y., Vogl, A., Deng, X.F., Zhong, S.J., 2001. Cornstarch explosion experiments and modeling in vessels ranged by height/diameter ratios. J. Loss Prev. Process. Ind. 14 (6), 495–502. Wilen, C., Rautalin, A., Garcia-Torrent, J., Conde-Lazaro, E., 1998. Lnerting biomass dust explosions under hyperbaric working conditions. Fuel 77 (9–10), 1089–1092. Yang, M., Chen, X.F., Wang, Y.J., Yuan, B.H., Niu, Y., Zhang, Y., Liao, R.Y., Zhang, Z.M., 2017. Comparative evaluation of thermal decomposition behavior and thermal stability of powdered ammonium nitrate under different atmosphere conditions. J. Hazard Mater. 337, 10–19. Zhang, H.M., Chen, X.F., Zhang, Y., Niu, Y., Yuan, B.H., Dai, H.M., He, S., 2017. Effects of particle size on flame structures through corn starch dust explosions. J. Loss Prev. Process. Ind. 50, 7–14.
In this work, flame propagation behaviors of starch dust explosion under reduced oxygen content were experimentally investigated and the changing rules of flame characteristic parameters were determined. For the reduction of oxygen content, flame propagation time in the tube was greatly prolonged and the luminous intensity of the flame postcombustion zone was diminished. When the oxygen concentration was lower than 12.60%, the completely inerting effect on flame propagation process was obtained for all the dust concentrations. Flame acceleration was limited by the oxygen levels and it was also related to the dust concentrations. When the dust concentration was more than 400 g/m3, the flame propagation velocity was more greatly suppressed by the reduced oxygen content. The values of maximum flame velocity, flame temperature and deflagration pressure were all significantly decreased on the condition of the oxygen concentration below 18.90%. With further reduction of oxygen content, these flame characteristic parameters always maintained at a low level and the propagation process decelerated to extinguish. By which, the starting point of availably partial inerting on flame propagation behaviors was the oxygen concentration of 18.90%. Acknowledgments The authors gratefully acknowledged the financial supports from the National Key Research and Development Program of China (Grant No. 2017YFC0804705), the National Natural Science Foundation of China (Grant Nos. 51374164 and 51774221) and the Key Laboratory of Building Fire Protection Engineering and Technology of MPS (KFKT2014ZD03). References ASTM Standard E2931, 2013. Standard Test Method for Limiting Oxygen (Oxidant) Concentration of Combustible Dust Clouds. ASTM International, West Conshohocken, PA. Ballantyne, A., Moss, J.B., 1977. Fine wire thermocouple measurements of fluctuating temperature. Combust. Sci. Technol. 17 (1–2), 63–72. BS EN 14034–4, 2005. Determination of Explosion Characteristics of Dust Clouds- Part 4: Determination of the Limiting Oxygen Concentration LOC of Dust Clouds. British Standards Institution. Cashdollar, K.L., 2000. Overview of dust explosibility characteristics. J. Loss Prev. Process. Ind. 13, 183–199. CEN, 2006. Guidance on Inerting for the Prevention of Explosions. CEN/TR15281, May 2006. CEN Management Centre, rue de Stassart, 36, B-1050. Brussels.
152
Contents lists available at ScienceDirect
Journal of Loss Prevention in the Process Industries journal homepage: www.elsevier.com/locate/jlp
Effects of reduced oxygen levels on flame propagation behaviors of starch dust deflagration
T
Hongming Zhanga, Xianfeng Chena,∗, Tian Xiea, Bihe Yuana, Huaming Daia, Song Hea, Xuanya Liub a b
School of Resources and Environmental Engineering, Wuhan University of Technology, Wuhan 430070, China Key Laboratory of Building Fire Protection Engineering and Technology of MPS, Tianjin 300381, China
A R T I C LE I N FO
A B S T R A C T
Keywords: Dust explosion Deflagration flame Flame propagation behavior Oxygen concentration
An experimental investigation of reduced oxygen levels on flame behaviors of starch dust deflagration is studied in a half-closed dust explosion tube. Six kinds of nitrogen/air ratio mixtures are prepared as the experimental atmosphere for flame propagation. A high-speed photography is used to record flame propagation behaviors and luminous features. The characteristics of flame temperature and deflagration pressure are measured by fine thermocouple and pressure sensor, respectively. The influence of reduced oxygen concentration on flame propagation characteristics are discussed in different conditions. The results show that reducing the oxygen content can effectively retard the flame propagation process and weaken the dust combustion reaction. Compared with air atmosphere, flame acceleration process under reduced oxygen atmosphere is suppressed and the luminous intensity of flame post-combustion zone is greatly diminished. The characteristic parameters of flame velocity, flame temperature and deflagration pressure are all influenced by the reduction in oxygen concentration. Furthermore, effective partial inerting suppression can be achieved for starch deflagration flame, when the oxygen concentration is lower than 18.90%. This study is helpful for dust explosion prevention and mitigation. Additionally, the dust cloud concentration changes the flame velocity characteristic at low oxygen levels and higher dust concentration is more limited the flame acceleration.
1. Introduction Starch has a very wide range of applications in lots of processing industries, such as food processing, chemical synthesis and pharmaceutical manufacturing. However, starch dust diffusion in the processing industries cannot be effectively controlled, usually leading to increase the possibility of starch dust explosions. Starch production and processing require a high degree of environmental cleanliness, and the inerting method to reduce the ambient oxygen concentration is a very suitable application for the prevention of starch dust explosions. Studies on inerting technology have been extensively conducted, and most of the studies have focused on determination of limiting oxygen concentration (LOC) for the dust explosion (Nomura et al., 1984; Dastidar et al., 1999; Dastidar and Amyotte, 2002; Cashdollar, 2000; Going et al., 2000; Eckhoff, 2003). The LOC is the maximum oxygen concentration of a dust/air/inert gas mixture by which dust explosions cannot occur. Both EN 14034-4 (2005) and ASTM E2931 (2013) describe the standard test method for LOC of combustible dust clouds in the 1 m3 vessel or 20 L sphere apparatus and point out that the
∗
measurement of LOC is the basis for inerting and explosion prevention. Mittal (2013) experimentally measured the LOC data for two types of coals with varying volatile matter and evaluated the effect of particle size and moisture on LOC. Particle size has a comparatively small influence on the LOC of coal dust while the effect of moisture on LOC is opposite. Furthermore, the LOC measurement cannot be directly used to represent the inerting level and the processing conditions need consideration. Wilen et al. (1998) found that the LOC values of biomass dust clouds increase with increasing initial pressure in the range 0.5–1.8 MPa. This result was different from coal dust and others. Hassan et al. (2014) proposed a predictive model to assess the probability of a dust explosion occurrence under a given environment. LOC was considered as a key parameter in the model and the probabilistic model of dust explosion with respect to the change of oxygen concentration was determined. Meanwhile, reducing the oxygen content above the LOC in the atmosphere by adding nitrogen is another core mean of preventing and mitigating dust explosions (Eckhoff, 2005) and partial inerting is called for more extensive use in industrial dust explosion protection (Eckhoff, 2004). The European standard CEN/TR
Corresponding author. E-mail address: [email protected] (X. Chen).
https://doi.org/10.1016/j.jlp.2018.03.011 Received 16 November 2017; Received in revised form 16 March 2018; Accepted 16 March 2018 Available online 17 March 2018 0950-4230/ © 2018 Elsevier Ltd. All rights reserved.
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
thermocouple, a pressure sensor and a programmable logic controller. The size of square dust combustion tube was 1000 mm × 80 mm × 80 mm , which was connected by two half-meter long tubes. The front and rear sides of tube were set as optical windows, which were convenient to observe the flame propagation process. A dust dispersing device was installed at the bottom of the tube. A square vent with the size of 20 mm × 20 mm was placed at the top of the tube. The vent was opened to connect the external space for the nitrogen/air mixed gas flow through the tube to form an inert atmosphere. A pair of ignition electrodes were positioned at 50 mm above the bottom of the tube and linked with a high voltage ignition unit, which could generate ignition spark. The gas mixing system was used to form air/nitrogen mixtures with different concentrations. Furthermore, the mixed gas was filled into the tube to create an inerting atmosphere and to act as a driving gas for dust dispersion. In this study, the flame propagation behaviors were captured by a high speed photography camera (Photron FASTCAM SA1.1) with a normal lens of Nikon AF Nikkor 50 mm f/1.4. A fine thermocouple and a pressure sensor were mounted at 120 mm below the top of the tube. The thermocouple was comprised of 25 μm diameter Pt-Pt/Rh 13% wires. The flame deflagration pressure was measured by a pressure transducer (PCB Piezotronics, 113B21), which has a measuring range of 0–1379 kPa and a sensitivity of 3.703 mV/kPa. Data recorder was used to record the temperature and pressure data. Synchronous trigger switches of each system were controlled by the programmable logic controller.
15281:2006 (CEN, 2006) has systematically introduced the influence of oxygen concentration on explosion atmospheres and the methods of inerting. The supply of inert gas is confirmed and the empirical equations is provided for inerting design. Li et al. (2009) investigated the inerting effect of N2, CO2 and Ar on the magnesium dust explosion. Comparing the parameters of explosion characteristics, the inerting effect of N2 is better than Ar, and N2 has a better economic efficiency for dust explosion prevention as an inert gas. Chaudhari and Mashuga (2017) utilized a modified standard minimum ignition energy (MIE) device to measure the MIE value of partial inerting dust clouds. The partial inerting MIE was influenced by purge-induced turbulence and a higher MIE value was obtained due to the turbulence in the ignition zone. As mentioned above, studies on the dust explosion sensitivity under inerting conditions have been very extensive. However, the flame propagation process is also an essential stage for the transformation from deflagration to explosion, which is always unsteady and controlled by numerous factors (Gao et al., 2013; Chen et al., 2017; Li et al., 2017; Yang et al., 2017; Zhang et al., 2017). While, few studies are available on flame propagation characteristics of dust deflagration with the reduction of oxygen concentrations. In the present study, nitrogen was chosen as an inert gas, and the effect of reduced oxygen levels with different nitrogen/air ratios on flame propagation behaviors of starch dust deflagration was studied in a semi-closed tube. The flame propagation processes with different oxygen concentrations was recorded by high speed photography system. Flame temperature and deflagration pressure were detected by fine thermocouple and pressure sensor, respectively. Furthermore, the variation of flame propagation dynamics was explored in depth.
2.2. Experimental materials 2. Experimental
Wheat starch was utilized for experiments. Before experiments, dust samples were sifted in a vibration sieve with 300 mesh to ensure the particle size distribution in the same range. Then, the samples were dried and dewatered in a vacuum drying oven at 50 °C for 12 h. The median particle size (D50) of the samples was 45 μm. High purity nitrogen (purity ≥ 99.9996 vol.%) and compressed air (oxygen concentration is 21.00 vol.%) were used for preparing different gas mixtures. The dust clouds was ignited by the high-voltage transformer with an output of 15 kV. The ignition duration was 100 ms and the nominal ignition energy was approximately 10 J.
2.1. Experimental apparatus A schematic diagram of experimental apparatus was illustrated in Fig. 1, composed of nine parts: a small-scale square combustion tube, a dust dispersing device, a gas mixing system, a high voltage ignition unit, a high speed photography system, a data recorder, a
2.3. Experimental conditions Experiments were carried out in the different oxygen concentration atmosphere of 6 nitrogen/air ratios. The oxygen volume concentration for the mixed gases were 12.60%, 14.70%, 16.80%, 18.90%, 19.95% and 21.00%. The detailed conditions of the gas mixture were shown in Table 1. In order to ensure the accuracy of the oxygen concentration in the experimental atmosphere, the flow through technique relying on the purge mixed gas was used to remove the original gas in tube. The gas flow in the experiments was 5 L/min and the purge time of the mixed gas flow through the tube was calculated from the following equation (CEN, 2006): Table 1 Experimental gas atmospheres for starch dust explosions.
Fig. 1. Sketch of experimental apparatus.
147
No.
Nitrogen (%)
Air (%)
Oxygen concentration (%)
Gas composition
1 2 3 4 5 6
0 5 10 20 30 40
100 95 90 80 70 60
21.00 19.95 18.90 16.80 14.70 12.60
Air Air-Nitrogen Air-Nitrogen Air-Nitrogen Air-Nitrogen Air-Nitrogen
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
t =F
V C − Co ⎞ ln ⎛⎜ i ⎟ Q ⎝ Ci − Cf ⎠
complete inerting in dust explosion protection because of the sealed test device. In this study, the measured value of 12.60% is the maximum oxygen concentration to prevent the starch deflagration flame propagation. The reduction of oxygen content also prolongs the deflagration flame spread time in the tube. It takes about only 180 ms for the flame spreading from the bottom of the tube to the top in the air atmosphere. The flame propagation time is about 500 ms, nearly 2.8 times of which in air atmosphere in the case of 14.70% oxygen concentration. Du et al. (2014) found that the devolatilization process of starch was not affected by the changes of oxygen concentration. Therefore, the time of starch ignition and burning is increased correspondingly, and the combustion reaction rate is decreased due to the reduced oxygen content.
(1)
where, t was the time required for purging (h), F was the safety factor for purging, V was the system volume (m3), Q was the inert gas flowrate (m3/h), Cf was the oxygen content after flow purging (vol.%), Co was the initial oxygen content (vol.%), Ci was the oxygen content of inert gas (vol.%). In this study, the maximum time required for purging was at least 117 s, thus the purge time of mixed gas flow through was set as 2 min. Before experiments, a certain of dust sample weighted by an electronic balance was evenly placed on the sample dish at the bottom of the tube. The nitrogen/air mixed gas was filled into the tube and dust particles were dispersed into the tube by the air flow. Delay time for ignition was set to minimize the influence from the airflow. After the ignition, high speed photography camera and data recorder were turned on. The detailed experimental parameters were listed as follows: pressure for powder injection, 0.2 MPa; injection time, 0.2 s; ignition delay time, 0.1 s; framing rate of high speed photography camera, 1000 frames/s.
3.2. Influence of reduced oxygen concentration on flame propagation velocity The flame propagation velocities in different oxygen concentration are shown in Fig. 3. As presented in Fig. 3, the flame propagation velocities show different behaviors in the reduced oxygen content. When the oxygen concentration is between 19.95% and 21.00%, the flame spreads in an oxygen-enriched atmosphere and it is less affected by oxygen content. The high acceleration in the initial stage of flame propagation is a significant feature. Maximum flame propagation velocity appears in the middle stage of flame propagation and the values are all above 6 m/s. Due to the cooling effect of tube walls, the flame velocities decline after reaching the maximum flame propagation velocity. It takes less than 250 ms for the whole flame propagation process. However, when the oxygen concentration is below 18.90%, the flame speed begins to slow down with no obvious acceleration or deceleration due to the oxygen-poor atmosphere. There is a long period of low-speed propagation in the initial stage and the whole propagation process lasts more than 400 ms. The flame propagation in the lower oxygen content is limited. Therefore, the oxygen concentration of 18.90% is an inflection point of the reduced oxygen content for flame velocities suppression in this study. The increase of dust concentration exacerbates the restrictions on flame propagation, especially in the oxygen-poor atmosphere. When the dust concentration is less than 300 g/m3, flame propagation velocity in the oxygen-poor atmosphere shows the same growth trend as the oxygen-enriched atmosphere. Due to the low concentration of dust cloud, the oxygen content required for the combustion of dust clouds per unit volume has not reduced drastically and the process of flame acceleration is slightly suppressed. When the dust cloud concentration is above 400 g/m3, the oxygen content in the tube is consumed quickly because of the increasing dust cloud concentration and significant suppressed flame velocities. Fig. 4 shows the relationship between maximum flame propagation velocity and arrival time, when the dust concentration is 400 g/m3. There are two stages for the influence of oxygen concentration on maximum flame propagation velocity. The first stage is oxygen concentration between 18.90% and 21.00%, and the reduced oxygen content has an obvious suppression effect on maximum flame velocity in this stage. The maximum flame velocity presents linear decline with the decrease of oxygen concentration. It takes less than 180 ms for all the flame propagation in this stage to arrive the peak value. Moreover, the second stage is in the oxygen concentration range of 14.70%–18.90%. In this stage, with reducing the oxygen concentration, the maximum flame velocity is stably suppressed and the value is maintained at about 3 m/s without significant fluctuations. The time reaching maximum flame velocity is greatly prolonged, and a linear increase with decreasing oxygen concentration is obtained.
3. Results and discussions 3.1. Influence of reduced oxygen concentration on flame propagation process Fig. 2 shows a series of direct high-speed photographs of flame propagation process of starch dust deflagrations with different oxygen concentrations and the dust mass concentration in these experiments is controlled at 400 g/m3. Suspended dust clouds were ignited by ignition electrodes and the flame began to propagate from the bottom to vent. In the upward propagation process, the flame structures and position information were recorded. Fig. 2 (a)-(e) are the flame propagation processes at the oxygen concentration of 21.00%, 19.95%, 18.90%, 16.80% and 14.70%. It can be seen from the high-speed photography images that the luminous area behind the front of the flame is filled in the entire tube. Unlike the premixed gas flame propagation process, the combustion of starch particles cannot be completed instantaneously at the flame front and this area is the post-combustion zone of the deflagration flame propagation. Unburned dust particles conduct devolatilization and combustion in the post-combustion zone, and further provide heat to promote flame spread upward. The flame luminous intensity in the post-combustion zone is the external manifestation of dust combustion reaction. In the air atmosphere with 21.00% oxygen concentration, yellow light-emitting area is the dominant and the overexposed white light appears locally, due to the sufficient chemical reaction under oxygen-rich conditions. The direct effects of reduced oxygen content on flame propagation is the diminished luminous intensity of the post-combustion zone. When the oxygen concentration is reduced to 18.90%, the flame emission brightness is dominated by light red, and the range of yellow luminous area in the flame is gradually significantly. When the oxygen concentration is 14.70%, the dust clouds can be ignited and propagate in the tube, but the luminous brightness of the flame is almost invisible. What's more, the flame is extinguished automatically when it spread to the 850 mm from the bottom of the tube. Because of the limitation of the oxygen content, the deflagration flame takes away the oxygen during the propagation process, making the unburned dust particles unable to support combustion and spontaneous extinguishment. When the oxygen concentration is 12.60%, the starch cloud cannot be ignited after several experiments. Furthermore, starch clouds with other dust cloud concentrations also cannot be ignited on this oxygen concentration. Result indicates that the limiting oxygen concentration in these experimental conditions is approximately 12.60%. However, this value is a little higher than the LOC of 11% measured under standard test conditions (NFPA, 2008). The results of the standard tests are relatively more accurate for the
3.3. Influence of reduced oxygen concentration on flame temperature In order to monitor the temperature behavior of flame combustion 148
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
Fig. 2. Flame propagation at different oxygen concentrations: (a) 21.00%; (b) 19.95%; (c) 18.90%; (d) 16.80%; (e) 14.70% (dust concentration: 400 g/m3).
Fig. 5 shows the temperature curves of flame combustion zone measured by fine thermocouple when the dust concentration is 400 g/ m3. Reducing the oxygen concentration delays the detecting time of the thermocouple and prolongs the flame temperature curve rise time. In the air atmosphere, the maximum flame temperature is about 800 °C, and the values decreases with the oxygen content reduction. When the oxygen concentration decreases to 14.70%, the maximum flame temperature is only about 70% of the value for air atmosphere. The combustion mechanism of starch usually includes such steps as particle heating, devolatilization, volatile-air mixture and homogeneous combustion. (Du et al., 2014; Zhang et al., 2017). The decrease in the flame temperature indicates that the combustion of dust particles is incomplete and the heat of combustion is reduced at a lower oxygen
zone in the tube, the flame temperature was measured by the fine thermocouple installed in the middle of the tube. The data measured by the thermocouple needs to be compensated due to the thermal inertia. It is assumed that the convective heat transfer process between thermocouple wires is mainly from thermal radiation and heat conduction, thus the thermocouple temperature correction formula can be expressed as follows (Ballantyne and Moss, 1977; Gao et al., 2014):
T = Tm + τ
dTm dt
(2)
Where Tm is the temperature measured by the thermocouple, and τ is the response time of thermocouple. In our experiments, the value of τ is approximately 6.7 × 10−3 s. 149
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
Fig. 3. Flame propagation velocity curves at different oxygen concentration. Dust concentrations: (a) 200 g/m3, (b) 300 g/m3, (c) 400 g/m3, (d) 500 g/m3, (e) 600 g/m3, (f) 700 g/m3.
temperature at each dust cloud concentration decreases greatly. The reduction of oxygen content greatly enhanced the resistance of the combustion reaction. Decreasing the oxygen concentration can cause partial inerting in the tube, thus this is an effective but underutilized mitigation technique for dust explosion mitigation (Eckhoff, 2004; Hoppe and Jaeger, 2005; Chaudhari and Mashuga, 2017). Thus, the influence of flame temperature verifies again that oxygen concentrations below 18.90% are effective for partial inerting on the deflagration flame.
concentration. Thus, the reduced emission rate of volatile surrounding the unburned particles and the prolonged devolatilization time delays the flame propagation. The combustion of volatile further consumes the oxygen in the environment and the combustion of dust particles is further inhibited, thereby the temperature of flame combustion zone is reduced with the decreasing of oxygen concentration. The curves of maximum flame temperature of dust cloud at different concentrations are shown in Fig. 6. As similar to the law of maximum flame propagation velocity, in a higher oxygen concentration atmosphere (18.90%–21.00%), the maximum flame temperature increases from lower dust concentration until the dangerous mass concentration of 500 g/m3 (Radandt et al., 2001), and then which tends to be almost constant. The result shows that the effect of high dust concentration on combustion kinetics of starch is more pronounced on higher oxygen concentration. The endothermic effect of cold source of the unburnt particles results in a slower temperature rise of the particles and the release rate of combustion heat is inhibited in the end. When the oxygen concentration is lower than 18.90%, the maximum flame
3.4. Influence of reduced oxygen concentration on flame deflagration pressure The flame deflagration pressure is the overpressure on the confined space during flame propagation. A pressure sensor is installed on the side of the tube to measure the deflagration pressure and the pressure curves at different oxygen concentrations are shown in Fig. 7. The maximum pressure appears in the air atmosphere is about 37 kPa, and 150
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
Fig. 4. Relationship between maximum flame propagation velocity and time (dust concentration: 400 g/m3). Fig. 7. Pressure curves at different oxygen concentrations (dust concentration: 400 g/ m3).
the peak pressures of starch deflagration flame on other oxygen contents decrease. The threat of deflagration is diminished by the reduced oxygen concentration. When the oxygen concentration is less than 18.90%, the maximum deflagration pressure is significantly reduced and pressure curve is decayed slowly at the pressure peak, which may be caused by the slow propagation of the combustion wave at low oxygen concentration and the pressure is maintained for a long time. Fig. 8 shows the variation of the maximum deflagration pressure with oxygen concentration. Due to the relatively adequate oxygen supply during the flame propagation process, the combustion process is less affected by oxygen concentration and the explosion pressure is not significantly suppressed. When the oxygen concentration is in the lower content (less than 18.90%), the deflagration pressures of each dust concentration is reduced obviously. For the lower oxygen content, most of the dust particles cannot be burned completely and the unburned particles will take away the combustion heat. The energy release from flame wave is absorbed and the pressure decreases rapidly until it cannot propagate. The inerting effect on the deflagration flame process is fully functioned by reducing the oxygen content.
Fig. 5. Flame temperature at different oxygen concentrations (dust concentration: 400 g/ m3).
Fig. 6. Maximum flame temperature of different dust cloud concentrations.
Fig. 8. Maximum explosion pressure at different dust cloud concentrations.
151
Journal of Loss Prevention in the Process Industries 54 (2018) 146–152
H. Zhang et al.
4. Conclusions
Chaudhari, P., Mashuga, C.V., 2017. Partial inerting of dust clouds using a modified standard minimum ignition energy device. J. Loss Prev. Process. Ind. 48, 145–150. Chen, X.F., Zhang, H.M., Chen, X., Liu, X.Y., Niu, Y., Zhang, Y., Yuan, B.H., 2017. Effect of dust explosion suppression by sodium bicarbonate with different granulometric distribution. J. Loss Prev. Process. Ind. 49, 905–911. Dastidar, A.G., Amyotte, P.R., 2002. Determination of minimum inerting concentrations for combustible dusts in a laboratory-scale chamber. Process Saf. Environ. Protect. 80 (6), 287–297. Dastidar, A.G., Amyotte, P.R., Going, J., Chatrathi, K., 1999. Flammability limits of dusts: minimum inerting concentrations. Process Saf. Prog. 18 (1), 56–63. Du, B., Huang, W.X., Kuai, N.S., Yuan, J.J., Liu, L., Ren, Y.D., 2014. Comparative study of explosion processes controlled by homogeneous and heterogeneous combustion mechanisms. J. Loss Prev. Process. Ind. 30, 155–163. Eckhoff, R.K., 2003. Dust Explosions in the Process Industries, third ed. Gulf Professional Publishing, Amsterdam. Eckhoff, R.K., 2004. Partial inerting - an additional degree of freedom in dust explosion protection. J. Loss Prev. Process. Ind. 17, 187–193. Eckhoff, R.K., 2005. Current status and expected future trends in dust explosion research. J. Loss Prev. Process. Ind. 18, 225–237. Gao, W., Mogi, T., Sun, J., Dobashi, R., 2013. Effects of particle thermal characteristics on flame structures during dust explosions of three long-chain monobasic alcohols in an open-space chamber. Fuel 113, 86–96. Gao, W., Yu, J.L., Mogi, T., Zhang, X.Y., Sun, J.H., Dobashi, R., 2014. Effects of particle thermal characteristics on flame microstructures during dust explosions of three longchain monobasic alcohols in a half-closed chamber. J. Loss Prev. Process. Ind. 32, 127–134. Going, J.E., Chatrathi, K., Cashdollar, K.L., 2000. Flammability limit measurements for dusts in 20-L and 1-m3 vessels. J. Loss Prev. Process. Ind. 13, 209–219. Hassan, J., Khan, F., Amyotte, P.R., Ferdous, R., 2014. A model to assess dust explosion occurrence probability. J. Hazard Mater. 268, 140–149. Hoppe, T., Jaeger, N., 2005. Reliable and effective inerting methods to prevent explosions. Process Saf. Prog. 24 (4), 266–272. Li, G., Yuan, C.M., Fu, Y., Zhong, Y.P., Chen, B.Z., 2009. Inerting of magnesium dust cloud with Ar, N2 and CO2. J. Hazard Mater. 170 (1), 180–183. 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 powder based on thermal analysis. Powder Technol. 309, 68–73. Mittal, M., 2013. Limiting oxygen concentration for coal dusts for explosion hazard analysis and safety. J. Loss Prev. Process. Ind. 26, 1106–1112. NFPA, 2008. Standard on Explosion Prevention Systems. NFPA 69. National Fire Protection Association, Quincy, MA. Nomura, S., Torlmoto, M., Tanaka, T., 1984. Theoretical upper limit of dust explosion in relation to oxygen concentration. Ind. Eng. Chem. Process Des. Dev. 23, 420–423. Radandt, S., Shi, J.Y., Vogl, A., Deng, X.F., Zhong, S.J., 2001. Cornstarch explosion experiments and modeling in vessels ranged by height/diameter ratios. J. Loss Prev. Process. Ind. 14 (6), 495–502. Wilen, C., Rautalin, A., Garcia-Torrent, J., Conde-Lazaro, E., 1998. Lnerting biomass dust explosions under hyperbaric working conditions. Fuel 77 (9–10), 1089–1092. Yang, M., Chen, X.F., Wang, Y.J., Yuan, B.H., Niu, Y., Zhang, Y., Liao, R.Y., Zhang, Z.M., 2017. Comparative evaluation of thermal decomposition behavior and thermal stability of powdered ammonium nitrate under different atmosphere conditions. J. Hazard Mater. 337, 10–19. Zhang, H.M., Chen, X.F., Zhang, Y., Niu, Y., Yuan, B.H., Dai, H.M., He, S., 2017. Effects of particle size on flame structures through corn starch dust explosions. J. Loss Prev. Process. Ind. 50, 7–14.
In this work, flame propagation behaviors of starch dust explosion under reduced oxygen content were experimentally investigated and the changing rules of flame characteristic parameters were determined. For the reduction of oxygen content, flame propagation time in the tube was greatly prolonged and the luminous intensity of the flame postcombustion zone was diminished. When the oxygen concentration was lower than 12.60%, the completely inerting effect on flame propagation process was obtained for all the dust concentrations. Flame acceleration was limited by the oxygen levels and it was also related to the dust concentrations. When the dust concentration was more than 400 g/m3, the flame propagation velocity was more greatly suppressed by the reduced oxygen content. The values of maximum flame velocity, flame temperature and deflagration pressure were all significantly decreased on the condition of the oxygen concentration below 18.90%. With further reduction of oxygen content, these flame characteristic parameters always maintained at a low level and the propagation process decelerated to extinguish. By which, the starting point of availably partial inerting on flame propagation behaviors was the oxygen concentration of 18.90%. Acknowledgments The authors gratefully acknowledged the financial supports from the National Key Research and Development Program of China (Grant No. 2017YFC0804705), the National Natural Science Foundation of China (Grant Nos. 51374164 and 51774221) and the Key Laboratory of Building Fire Protection Engineering and Technology of MPS (KFKT2014ZD03). References ASTM Standard E2931, 2013. Standard Test Method for Limiting Oxygen (Oxidant) Concentration of Combustible Dust Clouds. ASTM International, West Conshohocken, PA. Ballantyne, A., Moss, J.B., 1977. Fine wire thermocouple measurements of fluctuating temperature. Combust. Sci. Technol. 17 (1–2), 63–72. BS EN 14034–4, 2005. Determination of Explosion Characteristics of Dust Clouds- Part 4: Determination of the Limiting Oxygen Concentration LOC of Dust Clouds. British Standards Institution. Cashdollar, K.L., 2000. Overview of dust explosibility characteristics. J. Loss Prev. Process. Ind. 13, 183–199. CEN, 2006. Guidance on Inerting for the Prevention of Explosions. CEN/TR15281, May 2006. CEN Management Centre, rue de Stassart, 36, B-1050. Brussels.
152