Effect of roadway turnings on gas explosion propagation characteristics in coal mines

Mining Science and Technology( China)21(2011)365-369Contents lists available at ScienceDi氵Mining Science and technology(China)ELSEVIERjournalhomepagewww.elsevier.com/locate/mstcEffect of roadway turnings on gas explosion propagation characteristiin coal minesZhu Chuanjie a, b, * Lin Baiquan. b, Ye Qing Zhai Cheng,bState Key Laboratory of Coal Resources and Safe Mining, Xuzhou 221008, ChinaFaculty of Safery Engineering China University of Mining 6 Technology, Xuzhou 221116,ChinaFaculty of Energy and Safety Engineering. Hunan University of Science and Technology, Xiangtan 411201, ChinaARTICLE IN FOA BSTRACTArticle history:In order to reveal the effect of turnings on explosion propagation, experiments were performed in threeReceived 13 October 2010different pipes(single bend, U-shaped pipe and Z-shaped pipe). Flame and pressure transducers wereceived in revised form 10 November 2010 used to track the velocity at the explosion front when the pipes were filled with methane. the explosionccepted 7 December 2010Available online 12 June 2011strength was significantly enhanced due to the turbulence induced by increasing the number of turnings,while the flame speed(Sy)and peak overpressure(APmax)increased dramatically. In addition, the strengthKeywordsof the explosion increased in violence as a function of the number of turnings. However when the bendwas without methane, the turnings weakened the strength of the explosion compared with the ordinarypipe, shown by the decrease in the values of APmax and SA In addition, the propagation characteristics inPeak overpressureThe results show that the explosion propagation characteristics largely deped nd s, were also closes distribution inthe pipes and the number of tunings. the different directions of the turnings had no effecte 2011 Published by Elsevier B V. on behalf of China Universiry of Mining Technology1 Introductionflame speeds and explosion pressure [7, 8). The presence ofobstacles can generate strong turbulence and hence causes aGas explosion accidents in coal mines have occurred fre- high flame speed and explosion pressure(9-14]. The variationquently since the industrial revolution. Experimental and theo. in flame duration and thickness during flame propagation wasretical studies have been undertaken not only to understand reported by Lin and Chen [15 Lin and Gui [16]and Wangthe phenomena, but also to identify the ignition and transmis- and He 17, 18 They show that the flame duration dropped withsion laws in the flammable hydrocarbon gas with oxygen or the increase of flame speed. It was reported that the flame accel-air and explosion characteristics of industrial dust (including erated faster in pipes with larger diameters [2]. Lin et al. havecoal dust). Gas explosions occurring in underground coal mines studied the propagations of flame and explosion involving differare really complex phenomena involving many factors, spanning ent factors 119-231from pure methane and mixed compositions to geometric charThe effect of simple bifurcation and turning of a pipe onacteristics such as ignition strength and location, wall roughness, explosion propagation was, at a preliminary stage, studied bythe presence of obstacles. bifurcation and the turning of road- some investigators (24, 25 ] Savenco performed experiments inways. Over the last decade, most investigators have studied pipes with diameters of 125 mm and 300 mm in order to obtainthose factors that affect the propagation of flame and explosion the attenuation coefficient of a blast wave traveling through thepressure experimentally and by numerical simulation [1-5.bifurcation and turning of the roadway [26 The effect of geoThe results of a study on ignition strength and location metrical shape and size on explosion was investigated by Bartshowed that the flame speed and explosion pressure with igni- knecht [27]. The flame acceleration caused by turbulence wastion at the closed end of a pipe was much larger than that at observed when the flame propagates through the turning ofthe vent area [6]. The flow velocity is generated ahead of the the roadway [28, 29). As well, the variation in cross-section areaflame which generates turbulence by interaction with rough can also generate additional turbulence and hencehighwalls and hence supports a rapid burning rate and causes high flame speeds and shock wave. However the structure of tunnelsystem is so complex that there are many turnings orientedtowards different directions. We mainly focus on investigating4 Corresponding author Tel +86 15950674572the effect of roadway turning involving different numbers andE-mail addressdirections of gas中国煤化工1674-5264/s-see front matter o 2011 Published by Elsevier B.V. on behalf of China University of Mining&CNMHGdoi:0.1016mstc2011.05.006C Zhu et al/ Mining Science and Technology( China)21(2011)365-3692. Experimental equipmentment at the previous point (Sy"245. 4 m/s). Obviously, the state offlow before the flame played a leading role in flame propagation. Itpipe, an ignition system, a circulating pump, a vacuum pump, a flow field. The value of Sy(331.6 m/s)recorded from the first pointdata acquisition system and a gas preparation system, as shown after the turning, was larger than that measured in the same posischematically in Fig. 1. The circulating system was used to make tion of the ordinary pipe. This can be explained as follows: the insure the mixture in the pipe was maintained at a uniform concen- crease of flame surface induced by turning resulted in an increasetration Ignition was actuated in the left end of the pipe using a in the heat released by the reaction which is important for thestandard 2) combustion engine spark plugspeed-up of the flame [7]. when the main pipe was filled withExperiments were carried out in three different pipes. Pipe A methane and the secondary pipe without methane, the values ofwas used to study the interaction between a single roadway turn- Sy before and outside the turning was also close to that of the ordin-ing and explosion propagation. the pipe consisted of two different ary pipe. However, when the flame propagated away from the cyl-parts: the main pipe was 80 mm x 80 mm in cross section and inder the flame speed decreased quickly after increasing over a4.5 m long and the secondary pipe was 2.0 m long with the same short distance because of the removal of the methane.Fig. 4 illustrates the peak overpressure(APmax)evolution alongvariation in explosion parameters during the explosion propaga- the experimental pipe. It shows that the three different blast wavestion in a U-shaped pipe It was a square pipe with the same cross had similar propagation characteristics before the turning. But thesection as Pipe a and including three different parts: the main pipe value of APmax outside the turning was much lower than that in-4.5 m long, the middle pipe.8 m long and the secondary pipe side when the blast wave propagated to the turning. In addition,1.5 m long. The ignition system was located on the left end of the when the secondary pipe was filled with methane, the peak over-main pipe and the end of the secondary pipe was open The inves- pressure was APmax"0.047 MPa; in contrast, the value of APmaxtigation of the explosion propagation characteristics under a was only 0.041 MPa which was lower than that measured in theZ-type"ventilation system, conducted in Pipe C, was similar to same position of the ordinary pipe(0 112 MPa)when the second-Pipe B except for the direction of the secondary pipe. A schematic ary pipe was without methane. However, when the secondary pipediagram of the three pipes is shown in Fig. 2.was filled with methane, the explosion strength became more vioMeasurements of flame speed were recorded by flame trans- lent as a result of flame speed-up with the blast wave travelingducers positioned along the pipe(see Fig. 2). The local flame dis- through the turning. The value of APmax taken from the first pointplacement speed, S, was calculated along the flame front by after the turning was 0.152>0. 124 MPa measured at the samedividing the distance(Ax) along the normal line at each point by point in an ordinary pipe. As well, it should be pointed out thatthe arrival time between two consecutive flame transducers (At). when there was no methane in the secondary pipe, the value ofPressure within the pipe was monitored using an array of piezore- APmax was consistently lower and showed a downward trend. Thissistineure transducers [3]indicated that the turning weakened the blast wave propagationThe preparation of the mixture(CHa/air)in the air pocket used a under this condition. From this exposition, we can see that the ef-partial pressure method. The pipe was, at first, pumped to a vac- fect of turning on the explosion propagation largely depends on theuum pressure of 0.06 MPa. Then the prepared mixture flowed into gas distribution: the violence of the explosion was strengthenedthe experimental pipe under the pressure difference. The concen- when the secondary pipe was filled with methane; under contrasttration of the mixture in the experimental pipe was C2=0.6C with ing conditions, the violence weakenedC the concentration of mixture in the air pocket.3. 2. Explosion propagation in U-shaped pipe3. Results and discussionFig 5 provides the flame speed, S, at different measurementpoints in the U-shaped pipe (i.e. Pipe B). When the entire pipe3.1. Effect of single tuming on explosion propagationwas filled with methane, the value of Sy measured before the firstturning increased as a function of the distance along the pipeFig 3 shows the flame speed at different measurement points and its rate of growth was also close, which shows that the beharalong Pipe A with a single turning. The behavior of the flame prop- ior of the explosion was similar to that in the ordinary pipe beforeagation before the turning was similar to that in the ordinary pipe the first turning. As the flame traveled to the first turning, theie, a straight pipe with the same length as Pipe a but without flame speed, taken from the point outside the turning, was onlyturning). When the pipe was filled with methane S. measured out- 24.6 m/s which was an order of magnitude lower than that mea-side the turning, was 26.6 m/s which was lower than the measure- sured at the two points just before and after this point( the valueswere 263 4 m/s and 324.9 m/s). The mechanism was equivalent tothat in the pipe with a single turning mentioned earlier. when theflame propagated away from the first turning, the value of Sy wasPressuresignificantly higher than that measured in the ordinary pipe asreducing valvethe flame accelerated continuously. The flame speed, S, obtaineduisition systemfrom the point outside the second turning was 35. 8 m/s whichwas much lower than the values recorded from the two adjacentpoints. The flame continued to accelerate as it left the second turn-Flameing The value of the flame speed observed from the first point justExperiment channelafter the second turning was 527 1 m/ s which was higher than thevalue(385 1 m/s)observed in the ordinary pipe. It was also higherCHa Airthan the value (465.0 m/s)observed in Pipe B In the case that thebend (i.e theCirculating pumpane, the val中国煤化工 arning was5 wer thanthose at theCNMHGthan those measuredFig. 1 Schematic of experimental apparatusin Pipe a anducpaPa, Litany, a, the number of turningsC Zu et al/ Mining Science and Technology( China)21(2011)365-369Fl8FI7 F16F15 FAPI甚-PP4F1le F3F4 F5B6 F7BF3F4FF6F7月8HFIF2 F3F4 F5F6 F9- E PSCFig. 2. Scheme of experimental channel and position of transducers(Ff: flame transducers; Pi: pressure transducers)Ordinary pipeOrdinary pipe70Pipe A(the bend filled with methane)U-shaped pipe(the bend filled with methaneU-shaped pipe(the bendfilled without methane0060LDFig. 5. Flame speed for U-shaped pipe.Fig 3. Flame speed for a single roadway turning (LD: ratio of length to diameter).increased the flame propagated faster when the bend was filledwith methane, while the flame speed showed a downward trendOrdinary pippewhen the bend was not filled with methanePipe A( the band filled with methaneFig. 6 illustrates the peak overpressure(APmax)evolution in the025Pipe A(the bend filledU-shaped pipe It is clearly seen from this figure that the evolutionwithout mahan)and value of APmax along the pipe was similar to that of Pipe a be-fore the first turning, whether the bend was filled with methane ornot In the case that the bend was filled with methane, the value ofg0.5APmax outside the turning was also much lower than that insidewhen the blast wave propagated to the second turning. In addition,the value measured inside the turning was 0.22 MPa which washigher than that measured in the ordinary pipe(0. 136 MPa)andPipe A, due to the turbulence induced by the two turnings. whenthe bendxplosion pressuredeclined signifi中国煤化工 of the two tum-CNMHGng was 0.097 MPaFlg 4. Comparison of APmux in ordinary pipe and Pipe Awhich was lower utne oruimdly pipe(0. 136 MPa). InC Zhu et aL/ Mining Science and Technology(China )21(2011)365-369Ordinary pipesecond turning were similar. However, the flame speed dropped0.25U-shaped pipe(the bend filled with methanequickly when there was no methane in the bend. Fig. 8 showsU-shaped pipethe peak overpressure APmax evolution along the experimental0.20filled without methanepipe. A comparison of Figs. 6 and 8 show that the evolution ofblast wave was similar and the values were close. In additionwhen the pipe was filled with methane, the values of the explo-g0.10sion pressure were higher due to the flame acceleration inducedby turbulence, but it declined due to the weakening of turningsfor cases with no methane in the bend. these results show thatthe direction of turnings had no effect on explosions4. ConclusionsFIg 6. Comparisons of APmax in ordinary pipe and U-shaped pipe(1 when the pipe was filled with methane, the violence of theexplosion was strengthened with an increase in the numberof turnings, which can be explained by the effect of turbu-Ordinary pipelence induced by turnings. This can be seen from the700Z-shaped pipe(the bend filled with methane)increase in the values of the flame speed (S) and explosion分600pressure(APmax) The turning played a role in strengtheningexplosion propagation under this condition.(2)The values of Sy and APmax dropped significantly when com-pared with the ordinary pipe when the bend was not filledwith methane. The turning played a role in weakening theexplosion propagation under this condition.()The comparison of experiments in the U-shaped andZ-shaped pipes shows that their flame and blast wavepropagation was similar and the values were also close.120Therefore, the direction of turnings had no effect on theexplosion.Flg 7. Flame speed for Z-shaped pipe.AcknowledgmentsOrdinary pipeFinancial support for this work, provided by the national natu025z2由甲 ped pipe(the bend filled with methaneral Science Foundation of China(No. 50574093) the Key Programfilled withoutof the Nlatural Science Foundation of China(No50534090) the National Basic Research and Development Program0.15of China(No. 2005CB221506) the National Science Foundation forYoung Scholars of China(No. 50804048 ), the National Key Technol-Eo.10ogy R&D Program(No. 2007BAK29B01)Research Innovation Pro-gram for College Graduates of Jiangsu Province and the pen005Foundation of State Key laboratory of Explosion Science and Technology(No. KFJ10-19M), is gratefully acknowledged.100LDReferencesFig 8. Comparisons of APmux in ordinary pipe and Z-shaped pi[1] Kuchta JM. Investigation of fire and explosion accidents in the chemical,ining and fuel-related industries-a manual, United States, Department ofconclusion, the evolution of explosion pressure in a U-shaped pipe[2] Xu D, The study on gas explosion wave propagation law and affected factors inine. Beijing: China University of Mining Technology: 2003 lin Chineselxplosion pressure declined faster compared with a single turning. l4 brplosion J China Coal So and analysis of dynamics feature parameter in idepends on the gas distribution in a pipe and the number of turnings In the case where the bend was not filled with methane the14 Ibrahima SS, Masri AR. The effects of obstructions on overpressurHowever, when the bend was filled with methane. the violence of[5I Fairweather M. Hargrave GK, Ibrahima SS, Walker DG. Studiesthe explosion was strengthened by the turbulence induced by theflame propagation in explosion tubes. Combust Flame 1999: 116(4[6] Bjerketvedt D Bakke JR Van wingerden K Gas explosion handbook J Hazard997:52(1}:1-15[7I Lin BQ Jian CG, Zhou SN. Inducement of turbulence and its3.3. Explosion propagation in Z-shaped pipetransmission in gas explosion. J China Univ Mining Technol 2003: 32(2):Measurements of the flame speed Sr in the Z-shaped pipe(i.e,Pipe C)are presented in Fig. 7. The figure shows that the behav- (91 Silvestrini M Genova B,. Parisi G, Leon Trujillo F]. Flame acceleration and DDTior of the flame propagation was similar to that in the U-shapedspeed outside the first turning was lower. But it developed more (10)H/ 3a ance for smooth and obstacles filled tubes. J Loss Prev Process Indpipe. As well, when the pipe was filled with methane, the flame中国煤化工 concentration on flamequickly while propagating away from the turnings, which can beCNMHGtute; 1984 (CMi Report Noexplained by the effect of turbulence. The results observed in the8434035C Zhu et al/ Mining Science and Technology( China)21(2011)365-369[11] Xu JD, Yang GY. Numerical simulation of the barricade encouraging effect in [211 Jian CG. Study on the properties and influencing factors of gas explosionhe process of gas explosion propagation JCoal Sci Eng 2004: 29(1): 53-6[12] Yao HX. Fan BC. The simulation of turbulent acceleration flame induced by [22) Gui XH. Study on the propagation characteristics of dynamic unsteady flameobstacles. J Nanjing Univ Sci Technol 1999: 23(2): 109-12 lin Chinesel[13] Zhou KY. Li ZF. Flame front acceleration of propane-air deflagration in straightTechnology: 2001 [ in Chirtubes Explosion Shock Waves 2000: 20(2): 137-42[23] Xie B, Fan BC, Wang KQ, Xid ZZ. Experimental study of flame acceleration and[14] Lin BQ. Zhou SN. Influence of barriers on flame transmission and explosipressure induced by baffle obstacle. J China Coal Soc 2002: 27(6): 627-30ave in gas explosion. J China Univ Mining Technol 1999: 28(2): 104-7124] Zhai C Jian CG. Lin BQ. Experimental research of tube furcations influence of[15] Lin BQ. Chen BH. Analysis and lab measurement of flame thickness in gasame transmission speed in gas explosion. Jiangsu Coal 2004(1 ) 46-47, 66 finxplosion. J China Univ Mining Technol 2000: 29(1 ) 45-7 lin Chinesel[16] Lin BQ, Gui XH. Computer simulation on the temperature field of gas explosion [25] Wang L Li TC. Propagation of air shock wave in rectangular bend tunnel and itsand lab measurement for the flame thickness. J Exp Mech 2002: 17(2): 227-33ton.J Nat Disasters 2004: 13(4): 146-50 fin Chinese.[26] Savenco CK. Underground air shock wave. Beijing: Metallurgical Industry[17] Wang CY. He XQ, An experimental study of flame thickness in gas explosion.Explos Mater 2001: 30(2): 28-32 lin Chinese].118] Wang CY, He XQ, Expess and sustainingtime of gas explosion. Coal Sci Technol 2001: 29(8): 27-30 (in Chinese].[28]xia C]. Zhou KY. Shen ZW. Dong YX, Nian WM. Wang HL Experimental st[19 Lin BQ Zhou SN, Zhang RG. The inducing condition of shock waves in gasn the propagation characteristics of unsteady gaseous detonation throughexplosion. J Exp Mech 1998: 13(4): 463-8 [ in Chinese90 round bend. J Exp Mech 2002: 17(4): 438-43 (in Chinese[20 Lin BQ, Zhang RG, Lv HH. Researchmechanism and flame [291transmission in gas explosion. J China Coal Soc 1999: 24(1): 56-9 linChinese).2005 in Chinese lH中国煤化工CNMHG