Effect of Aluminum Silicate Wool on the Flame Speed and Explosion ...

0 downloads 0 Views 479KB Size Report
Abstract: Problems of decreasing the flame speed resulting from pre-mixed gas explosions and attenuating explosion overpressures are discussed. A cylindrical ...
Combustion, Explosion, and Shock Waves, Vol. 49, No. 2, pp. 153–158, 2013. c X. Q. Yan, J. L. Yu. Original Russian Text 

Effect of Aluminum Silicate Wool on the Flame Speed and Explosion Overpressure in a Pipeline X. Q. Yana and J. L. Yua

UDC 536.46

Translated from Fizika Goreniya i Vzryva, Vol. 49, No. 2, pp. 34–39, March–April, 2013. Original article submitted January 13, 2012.

Abstract: Problems of decreasing the flame speed resulting from pre-mixed gas explosions and attenuating explosion overpressures are discussed. A cylindrical test pipeline with an 89 × 4.5 mm cross section is used to study flame propagation characteristics of an acetylene–air mixture both in the empty pipeline and in the presence of aluminum silicate wool attached to the internal wall of the pipeline. Experimental results show that aluminum silicate wool, which is a kind of a fibroid porous material with a high specific surface area, decreases the increment of the outlet flame speed and attenuates drastically the explosion overpressure if the length of the porous insert exceeds the critical length. Keywords: aluminum silicate wool, flame speed, explosion overpressure, explosion suppression. DOI: 10.1134/S0010508213020044

INTRODUCTION Flame and overpressure caused by gas explosions pose a huge threat to humans and environment. Hence, the measures how to reliably and efficiently decrease flame speeds and attenuate explosion overpressures have been the focus of numerous investigations owing to wide industrial applications [1–6]. The most common examples are the commercial flame arrestors employing wire gauzes and sintered metals to quench the flame. Polymer foams [5] and foam ceramics [7, 8] have also been studied recently. Porous materials used to suppress flame propagation have attracted the attention of scientists and engineers. It is believed that the suppression effect of porous materials can be ascribed to voids (pores) in the material. The wide range of sizes and highly developed specific interfacial areas of the pores lead to efficient heat transfer [9] and nonequilibrium reactions [10]. Products made of aluminum have been used for explosion suppression for a few decades [11]. A typical kind of a porous material and a compound of aluminum a

School of Chemical Machinery, Dalian University of Technology, Dalian 116024, China; [email protected].

Fig. 1. Experimental gas explosion suppression system: (1) acetylene gas cylinder; (2) compressed air gas cylinder; (3) acetylene–air mixing tank; (4) combined pressure and vacuum gauge; (5) vacuum air pump; (6) vacuum gauge; (7) ignition source; (8) explosion control box; (9) computer and data acquisition system; (10) pipeline; (11) flame and pressure sensors; (12) aluminum silicate wool.

is aluminum silicate wool usually used as an insulation material. At the same time, it is also a novel material that can be used for explosion suppression because of its porosity and large specific area. However, there are no suppression data available for aluminum silicate wool in the literature. The suppression effect on flame speeds and explosion overpressures is experimentally studied in this work.

c 2013 by Pleiades Publishing, Ltd. 0010-5082/13/4902-0153 

153

154

Yan, Yu

Fig. 2. Pressure and flame sensors in each pipe.

EXPERIMENTAL Experimental Apparatus

Fig. 3. Voltage signal versus the vertical (1) and horizontal displacements (2).

The explosion suppression experiments were performed in a facility shown in Fig. 1, which includes a pipeline, gas mixing system, gas distribution system, ignition system, measurement system, and data acquisition system. The pipeline consists of seven pipes connected by flanges. Each of the pipes has a length of 0.3 m with an 89 × 4.5 mm cross section. The pipeline is fixed on a support located 1 m above the ground. The total length of the pipeline is 2.1 m. An acetylene–air mixture in a stoichiometric concentration was prepared by the partial pressure method in an evacuated mixing tank. The gaseous mixture was fed into the evacuated pipeline and ignited by a spark plug. To obtain a well-proportioned mixture, multipoint injection was adopted. In this experiment, four injection points were chosen. The positions of these points from the ignition source were at 0.15, 0.75, 1.35, and 1.95 m.

of voltage ΔU versus the candle displacements. It can be found that, when the horizontal displacement Δx increases from 0 to 5 mm, the voltage decreases from ≈2 V nearly to zero rapidly, indicating that the flame sensor can only detect the flame whose light exactly passes through the diode. The voltage decreases more gently with the increase in the vertical displacement. However, the vertical displacement of 15 mm adopted in these experiments is appropriate for the photoconductive diode to detect the light signal. The pressure at the point O (see Fig. 2) was measured by the pressure sensor. The flame speed at the point O approximately equals the average speed on the segment AB, which is calculated according to the distance and duration of the flame. In these experiments, a distance of 50 mm was chosen to improve the accuracy. The ignition system is a spark plug with 48 J energy released each time, controlled by a computer and an explosion control box. A data acquisition card with 48 single channels is used to record all the pressure and flame signals and to control the ignition system.

Instrumentation

Porous Aluminum Silicate Wool

The pressure sensor is an MD-HF piezoelectric sensor with a maximum range of 2 MPa and a sampling frequency of 200 kHz. Its outer signal voltage is 0–5 V. The flame sensor consists of a photoconductive diode in a detector. The detector is a hollow steel tube with a 7 × 1.5 mm cross section, designed to protect the diode and to reduce the spot size of the flame light passing through the diode. Figure 2 illustrates the arrangement of sensors in each pipe. Prior to conducting the experiments, a lighted candle was moved first horizontally, and then vertically away from the tip of the sensor to verify the feasibility of the flame sensors. Figure 3 shows the changes

Aluminum silicate wool is made of aluminum oxide (mass fraction 51%), silicon dioxide (46%), and other impurities. The sales company declared that it has the following physical properties: bulk density ρ = 240 kg/m3 , specific heat cp = 0.255 J/(kg · K), thermal conductivity coefficient λ = 0.106 W/(m · K), porosity ε = 0.8, specific surface area A = 10 m2 /g, average pore size d = 20 μm, and fiber diameter d0 =2–4 μm. Figure 4 demonstrates aluminum silicate wool with 10 mm thickness used in experiments. To fix the material on the internal wall of the pipeline, a ferrous wire was coiled into a cylinder as a framework. Aluminum silicate wool was wrapped along the framework and tied

Effect of Aluminum Silicate Wool on the Flame Speed

Fig. 4. Aluminum silicate wool and framework.

155

Fig. 6. Two scenarios selected in the experiments.

outlet flame speed v0 , and outlet explosion overpressure p0 were measured at the same sensing points in the empty pipeline.

EXPERIMENTAL RESULTS Flame Propagating in the Empty Pipeline Fig. 5. Aluminum silicate wool placing into the pipeline.

by strings before being fixed in the pipeline, as is shown in Fig. 5. Aluminum silicate wool used in the experiment can resist high temperatures up to 1273 K. Experimental Scheme The experiment was started when the acetylene–air mixture at a stoichiometric concentration of 7.7 vol.% was ignited at the ignition source end; the other end was open. Two scenarios were studied for comparison: (1) flame propagating in the empty pipeline; (2) flame propagating in the pipeline with aluminum silicate wool. The experiment arrangement and the layout of aluminum silicate wool are schematically illustrated in Fig. 6. The position of the porous material Lx and the length of the porous material l were varied in the experiments (see Fig. 6): Lx = 0.4, 0.7, 1.0, and 1.3 m; l = 0.2, 0.3, 0.4, 0.5, and 0.6 m. The inlet flame speed vi , inlet explosion overpressure pi , outlet flame speed v0 , and outlet explosion overpressure p0 were measured in experiments with aluminum silicate wool. In order to compare the effect of aluminum silicate wool on suppression characteristics, the inlet flame speed vi , inlet explosion overpressure pi ,

At first, the flame propagating in the empty pipeline was investigated after the acetylene–air mixture was ignited. Figure 7 illustrates the distribution of the flame speed and explosion overpressure along the empty pipeline. The minimum and maximum standard deviations for the measured flame speeds are 5 and 52 m/s, corresponding to the positions Lx = 0.4 and 1.3 m. Moreover, the minimum and maximum standard deviations for the explosion overpressures are 29 and 54 kPa, at the positions Lx =1.3 and 0.7 m, respectively. It can be seen that the flame speed increases linearly with the increase in the distance from the ignition source. The acceleration of the flame propagating in the pipeline is obvious. However, the pressure first increases and then decreases with increasing distance, with the inflection at Lx = 1.6 m. The decrease in pressure can be ascribed to the venting effect at the pipeline outlet. Influence of Aluminum Silicate Wool on the Flame Speed After the acetylene–air mixture was ignited, the flame propagating in the pipeline with aluminum silicate wool was investigated. Figure 8 shows the changes in the outlet flame speed at different porous material lengths at the position Lx = 0.4 m. Results measured at same testing points in the empty pipeline are also given with a dashed line for comparison.

156

Fig. 7. Distribution of the flame speeds and explosion overpressures along the empty pipeline.

Yan, Yu

Fig. 9. Outlet flame speed versus the porous material length (1–3) at Lx = 0.7 (1 and 4), 1.0 (2 and 5), and 1.3 m (3 and 6) (points 4–6 show the results obtained in the empty pipeline).

Fig. 8. Outlet flame speed of the porous material versus the porous material length (1) and in the empty pipeline (2). Fig. 10. Inlet flame speed and inlet overpressure versus the critical length of the porous material.

It can be seen from Fig. 8 that the outlet flame speed first increases and then decreases with the increase in the porous material length, with the inflection at Lx = 0.2 m. Compared to the speeds obtained in the empty pipeline, it is found that aluminum silicate wool has two effects on the flame speed: acceleration and deceleration. The acceleration effect results from both the decrease in the cross-sectional area of the pipeline and from the disturbances generated by the combustion reaction. The deceleration effect is due to flame cooling and quenching, because the large amount of pores divide the flame into small flame fragments. If the porous material length is smaller than 0.4 m, the acceleration effect dominates. At l > 0.4 m, the deceleration effect is gradually preponderant. If the deceleration effect is exactly counteracted by the acceleration effect, the critical length is reached, ≈0.4 m, at an inlet flame speed vi = 143 m/s. At Lx = 0.4 m, the lowest outlet speed is about v0 = 280 m/s at l = 0.6 m, which is greater than the

inlet flame speed vi = 143 m/s, but smaller than the outlet flame speed v0 = 346 m/s in the empty pipeline. This means that aluminum silicate wool is incapable of decreasing the inlet flame speed, but can decrease the increment of the outlet flame speed if the porous material length is greater than the critical length in this experiment. For a changed porous material position, a series of outlet flame speeds versus the porous material length is shown in Fig. 9. It is seen that the acceleration and deceleration effects still exist at different positions. However, the critical length varies with the position. As the inlet flame speed is affected by the position, it is concluded that the critical length for speed has a close relationship with the inlet flame speed. For example, the critical length is about lcr = 0.41 m for the inlet flame speed vi = 238 m/s, lcr = 0.44 m for vi = 337 m/s, and lcr = 0.46 m for vi = 448 m/s. Figure 10 illustrates this relationship.

Effect of Aluminum Silicate Wool on the Flame Speed

Fig. 11. Outlet explosion overpressure versus the porous material length (curve 1) (Lx = 0.4 m) and in the empty pipeline (curve 2).

Influence of Aluminum Silicate Wool on the Explosion Overpressure Figure 11 shows the outlet explosion overpressure in the empty pipeline and in the presence of aluminum silicate wool at the position Lx = 0.4 m. It can be found that, with aluminum silicate wool in the pipeline, the outlet overpressure increases and then decreases with the increase in the porous material length. It is obvious that, in the case of overpressures, two effects of aluminum silicate wool also exist, like those for the flame speed. First, aluminum silicate wool can increase the explosion overpressure because of the contractible crosssectional area of the pipeline and the induced turbulence of the combustion reaction. Then, aluminum silicate wool can also partially absorb the pressure wave by its porosity and large specific area. At the position Lx = 0.4 m, the critical length for overpressure is about 0.3 m. Dislike the flame speed, the outlet overpressure is reduced to a lower value than the inlet explosion overpressure at the length Lx = 0.6 m. It can be concluded that, compared to decreasing flame speed, aluminum silicate wool is more efficient on attenuating the pressure wave. Figure 12 shows the outlet explosion overpressure versus the porous material length at the positions Lx = 0.7, 1.0, and 1.3 m. At the positions Lx = 0.7 and 1 m, the pressure first increases and then decreases. The difference of the case at Lx = 1.3 m can be ascribed to the venting effect. As is shown in Fig. 7, venting at the end of the pipeline can cause the overpressure declining from Lx = 1.3 m in the pipeline. However, the outlet overpressures are lower than the inlet values with the porous material length of 0.6 m at any position, demonstrating the function of aluminum silicate wool in attenuating the overpressure. It is found that the critical length for overpressure is related to the inlet explosion overpressure p0 . The critical length for overpressure is ≈0.3 m for inlet over-

157

Fig. 12. Outlet explosion overpressure versus the porous material length (1–3) at the points Lx = 0.7 (3 and 6), 1.0 (2 and 4), and 1.3 m (1 and 5) (curves 4–6 show the results for the empty pipeline).

pressure p0 = 200 kPa, 0.32 m for 305 kPa, and 0.35 m for 357 kPa, as is shown in Fig. 10.

MECHANISM OF ALUMINUM SILICATE WOOL SUPPRESSING FLAME SPEED AND EXPLOSION OVERPRESSURE As was discussed above, the high porosity and specific surface area of aluminum silicate wool provide a large space for the combustion reaction. When the reaction occurs, aluminum silicate wool divides the flame into numerous fragments and slows down the combustion reaction, hence, suppressing the acceleration of the flame speed. It is mentioned in the literature [8] that porous materials can suppress transverse waves generated from gas-explosion shock waves, which is the reason why detonation waves can propagate in a self-sustained mode. A large amount of pores exist in aluminum silicate wool. When the explosion shock waves expand and propagate in the porous material, the energy of the shock waves is mostly consumed by the elastic and plastic deformation of aluminum silicate wool. Hence, the explosion overpressure is effectively attenuated.

CONCLUSIONS An experimental study of flame propagation in a pipeline with aluminum silicate wool for a stoichiometric acetylene–air mixture was conducted. Data were collected for various porous material positions and lengths, and the dependences of the explosion overpressure and flame speed on these characteristics were discussed. The results indicate that, due to specific features of the high specific area structures, aluminum silicate wool

158

Yan, Yu

can decrease the increment of the outlet flame speed, compared to the experiments in the empty pipeline, and remarkably attenuate the explosion overpressure if the porous material length is greater than the critical length. Moreover, the critical length for flame speed is related to the inlet flame speed, whereas the critical length for explosion overpressure is also influenced by the inlet explosion overpressure. The authors would like to thank the National Natural Science Foundation Project of China (Project No. 50974027) and the Liaoning Province science and technology plan project.

REFERENCES 1. V. S. Babkin, A. A. Korzhavin, and V. A. Bunev, “Propagation of Premixed Gaseous Explosion Flames in Porous Media,” Combust. Flame 87 (2), 182–190 (1991). 2. K. L. Edwards and M. J. Norris, “Materials and Constructions Used in Devices to Prevent the Spread of Flames in Pipelines and Vessel,” Mater. Design 20 (5), 245–252 (1999). 3. L. D. Mare et al., “Experimental and Numerical Study of Flammability Limits of Gaseous Mixtures in Porous Media,” Experim. Therm. Fluid. Sci. 21 (3), 117–123 (2000).

4. V. Bubnovich and M. Toledo, “Analytical Modeling of Filtration Combustion in Inert Porous Media,” Appl. Therm. Eng. 27 (7) 1144–1149 (2007). 5. R. Zalosh, “Deflagration Suppression Using Expanded Metal Mesh and Polymer Foams,” J. Loss. Prevent. Proc. 20 (4–6), 659–663 (2007). 6. C. Hohansen and G. Ciccarelli, “Combustion in a Horizontal Channel Partially Filled with a Porous Media,” Shock Waves 18 (2) 97–106 (2008). 7. H. I. Joo and G. Ciccarelli, “Flame Quenching Performance of Ceramic Foam,” Combust. Sci. Technol. 178 (10–11) 1755–1769 (2006). 8. B. Nie, X. He, R. Zhang, et al., “The Roles of Foam Ceramics in Suppression of Gas Explosion Overpressure Quenching of Flame Propagation,” J. Hazard. Mater. 192 (2), 741–747 (2011). 9. M. A. Mujeebu, M. Z. Abdullah, M. Z. A. Bakar, et al., “Applications of Porous Media Combustion Technology—A review,” Appl. Energ. 86 (9), 1365–1375 (2009). 10. A. A. M. Oliveira and M. Kaviany, “Nonequilibrium in the Transport of Heat and Reactants in Combustion in Porous Media,” Prog. Energ. Combust. 27 (5), 523–546 (2001). 11. A. M. Birk, “Review of Expanded Aluminum Products for Explosion Suppression in Containers Holding Flammable Liquids and Gases,” J. Loss. Prevent. Proc. 21 (5), 493–505 (2008).