Home

Influence of Coal Dust on Premixed Turbulent Methane

1. 98 4 3 Correlation of turbulent burning 100 BEJETENCES at a DE RA AKA GOO Re ie Ge oe GO ee 103 5 Conclusions and receommendatlons de beds Se ese RR aula Ae 105 Appendix DE on 111 Recommended Personal Protective Equipment PPE rea 113 Din on A EE GE A 114 Calibration ER 119 Mines TCC C RE BaB R R lon l aba SRN 121 Starine RUDNIN JESC aaa ATA OR Gda ON 123 Collecting cas analysis dala ES iGo R ab aad 125 Turning off Expe mne t iese s R osse Se ee AAA 127 MATLAB Seripis e a Bad d u oe dase eden een ene 128 Using Hotwire Anemometer EE EE GED SE SG A RA BOGS 129 Appendix 3 Matlab scripts used in data analysis 132 Contents EE RE EE N 132 A3 1 Edge selection script ss se ee aaa aaa RA a R KKK ee 133 A3 2 Edge data arialsiS SCript is is az GEE Se ke gese aa in A eg be 136 A3 3 Plotting figure 4 9 KR KP ee ee Ee ee 153 A3A PlottingtigufeA LJ Ee ee es ee NE Se es 157 A3 5 Plotting figure 4 la de RE A A RD aaa 160 A3 0 Plotting figure MD sc EERS SR EG Ee be DS oe PP eves GEN Es Ge Ee ai 165 A3 7 Creating arrays of test data as a function of dust concentration
2. plot l O burn vel 08 106 00 orig ks LineWidth plotLineWidth plot l O burn vel 08 106 25 orig ro LineWidth plotLineWidth plot l O burn vel 08 106 50 orig sv Line Width plotLineWidth plot l O 1 length burn vel 08 106 75 orig burn vel 08 106 75 orig bh LineWidth plotLineWidth plot l O burn vel 10 75 00 orig ks LineWidth plotLineWidth plot l O burn vel 10 75 25 orig ro LineWidth plotLineWidth plot l O burn vel 10 75 50 orig gv Line Width plotLineWidth plot l O burn vel 10 75 75 orig bh LineWidth plotLineWidth plot l 00 burn vel 10 106 00 orig ks LineWidth plotLineWidth plot l 00 burn vel 10 106 25 orig ro LineWidth plotLineWidth plot l 00 burn vel 10 106 50 orig gv LineWidth plotLineWidth plot l 00 1 length burn vel 10 106 75 orig burn vel 10 106 75 orig bh LineWidth plotLineWidth plot xfit data yfit 10 75 00 k LineWidth plotLineWidth plot xfit data yfit 10 75 25 yfit 10 75 00 r LineWidth plotLineWidth 162 plot xfit_data yfit_10_75_50 yfit_10_75_00 g Line Width plotLine Width plot xfit data yfit 10 75 75 yfit 10 75 00 b LineWidth plotLineWidth plot xfit data yfit 10 106 25 yfit 10 75 00 r LineWidth plotLineWidth plot xfit data yfit 10 106 50 yfit 10 75 00 9g LineWidth plotLineWidth plot xfit data yfit 10 106 75 yfit 10 75 00 b LineWidth plo
3. az mm 193 11111 1114114111 111111 11111 1111111 11117111 4111111 ds 75 90 um U pms 0 185 m s 0 8 JB a 8 Ast 0 g m 196 ds 75 90 um U 0 185 m s 0 0 8 25 g m 197 ds 75 90 um U rms 0 185 m s 0 0 8 50 g m 198 ds 75 90 um U ims 1 0 185 m s 0 0 8 75 g m 199 ds 75 90 um U 0 185 m s 0 1 0 Ast 0 g m 200 ds 75 90 um U 0 185 m s 0 1 0 25 g m 201 ds 75 90 um U 0 185 m s 0 1 2 Ast 0 g m BET 204 75 90 um Um m 0 185 m s 0 1 2 25 g m 205 dst 75 90 um U 0 185 m s 0 1 2 50 g m 206 dst 75 90 um U gt 0 185 m s 0 1 2 75 g m 207 ds 75 90 um U 0 335 m s 0 0 8 Ast 0 g m NENA 68 82 1 208 ds 75 90 um U rms 0 335 m s 0 0 8 25 g m 209 ds 75 90 um U 0 335 m s 0 0 8 50 g m 210 ds 75 90 um U 0 335 m s 0 0 8 75 g m 211 dst 75 90 um U gt 0 335 m s 0 1 0 As 0 g m 212 dst 75 90 um U gt 0 335 m s 0 1 0 25 g m 213 ds 75 90 um U 0 335 m s 0 1 0 50 g m 214
4. u u 3 2 57 The velocity measurements in the experiments described below are done in cold flow without a flame similar to Kobayashi et al 5 Pope 9 discussed how the flame could have an effect on the turbulent velocity field due to the large temperature rise of the flame but with few exceptions these effects have not been studied However Chomiak 10 found that a wrinkled continuous laminar flame does not generate additional turbulence and actually reduces the overall intensity of the turbulent velocity fluctuations Many studies in the literature have used the cold flow measurement of turbulence to characterize the turbulence experienced by a flame This procedure is followed in this work as well The impact of turbulent intensity and length scale on premixed combustion has been studied by Borghi 11 Turbulence and the decay in turbulent intensity in experimental setups are discussed by Liu 12 and Roach 13 The use of perforated plates as discussed below has been shown to be a reliable way to produce predictable turbulence intensities In the HFA the turbulence intensity is controlled with the flow rate through the burner and the distance of the perforated plate from the burner exit Combined air methane flow rates of up to 4 m s are used to generate a range of turbulent intensities up to 0 532 m s Figure 3 10 shows the turbulent intensity generated as a function of perforated plate location and flow
5. 00 06 o olo olo EIS 21 o o to R O 00 N w o o R N 00 o N a S M w w o A a N R N M OO Ojo pio O olo k o N to UI o 0 196 0 014 o to RK o N o R w UI o o w 00 EE 05 65 00 w R ojo 2 00 R Ni 05 N o N to N sl to w o 65 w N 183 Appendix 5 Flame Images d 75 90 um U ims 0 024 m s 0 8 Ai 0 g m R u J 184 ds 75 90 um U rms 0 024 m s 0 8 25 g m kkkkkkk 185 ds 75 90 um Wis 0 024 m s 0 8 50 g m 11111 11141111 11141411 1141111 11111 1111111 1111111 111111 dst 75 90 um U gt 0 024 m s 1 0 0 g m b k b 7 4 a gt x 188 mm a i a Z N m 11111 1111111 11111711 1171111 11111 TTTTTT1 1111411711 111111 11111 1111111 4111111 dst 75 90 um U gt 0 024 m s 1 2 0 g m b m 3 OT s Z b b b d 5 d 192 dst 75 90 um U gt 0 024 m s 1 2 25 g m ELE 11111
6. 233 dst 75 90 um U gt 0 024 m s 0 1 0 25 g m 234 dst 75 90 um U gt 0 024 m s 0 1 0 50 g m 235 dst 75 90 um U gt 0 024 m s 0 1 0 75 g m 236 dst 75 90 um U gt 0 024 m s 1 2 25 g m 237 dst 75 90 um U rms 0 024 m s 1 2 50 g m 238 11114 1111111 dst 75 90 um U gt 0 185 m s 0 0 8 SE 25 g m 240 dst 75 90 um U gt 0 185 m s 0 0 8 50 g m 241 dst 75 90 um U gt 0 185 m s 0 0 8 75 g m 242 dst 75 90 um U gt 0 185 m s 0 1 0 25 g m 243 dst 75 90 um U gt 0 185 m s 0 1 0 50 g m 244 dst 75 90 um U gt 0 185 m s 0 1 0 75 g m RE Ne 245 dst 75 90 um U gt 0 185 m s 0 1 2 25 g m 246 dst 75 90 um U 0 185 m s 0 1 2 50 g m 247 dst 75 90 um U gt 0 185 m s 0 1 2 75 g m 248 dst 75 90 um U gt 0 335 m s 0 8 25 g m eso AE 8 0 1 249 dy 75 90 um U ms z 0 335 m s 0 0 8 50 g m 250 dst 75 90 um U gt k 0 335 m s 0 0 8 75 g
7. 154 plot l 0 10 75 1 length burn vel 10 106 75 burn vel 10 106 75 orig lam data end burn vel 10 75 10 dst _sze_origFit 4 bh Line Width plotLineWidth ylim y_axisMin y axisMaxl xlim 10 4 1 42 hold off subplot subplot 3 2 5 Parent figurel YTick 2 2 4 2 8 3 2 3 6 4 LineVVidth 2 FontWeight bold FontSize 14 FontName Times New Roman subplot 3 2 5 hold on plot l 0 12 00 burn vel 12 75 00 orig lam data end burn vel 12 75 10 dst sze origFit 1 ks LineWidth pl otLineWidth plot l 0 12 25 burn vel 12 75 25 orig lam data end burn vel 12 75 10 dst sze origFit 2 ro LineWidth pl otLineWidth plot l_0_12_50 burn_vel_12_75_50_orig lam_data end burn_vel_12_75_10_dst_sze_origFit 3 gv LineWidth pl otLineWidth plot l 0 12 75 burn vel 12 75 75 orig lam data end burn vel 12 75 10 dst sze origFit 4 bh LineWidth pl otLineWidth ylim y_axisMin y_axisMax hold off subplot subplot 3 2 6 Parent figurel YTick 2 2 4 2 8 3 2 3 6 4 LineVVidth 2 FontWeight bold FontSize 14 FontName Times New Roman subplot 3 2 6 hold on 1010 0 12 00 burn vel 12 106 00 orig lam data end burn vel 12 75 10 dst sze origFit 1 ks LineWidth p lotLineW idth plot l 0 12 25 burn vel 12 106 25 orig lam data end burn vel 12 75 10 dst sze origFit 2 ro LineWidth p lotLineW idth plot l 0 12 50 burn vel 12 1
8. 70 are required While some work has been done to make 3D images of turbulent flames using high speed cameras 39 this was not possible due to the need to use the shadowgraph with the addition of the dust particles a b Figure 3 18 Profiles of theoretical turbulent flame a side view b top view at a specific height To determine the number of images required for determining the burning velocity a parametric study was done as shown in Fig 3 19 comparing the calculated burning velocity versus the number of images used It is shown that the asymptotic velocity calculation is reached at 10 15 images To add a factor of safety 25 images are sampled for each test to determine the average turbulent burning velocity 5 Calculated Burning Velocity m s 3 10 15 20 Number of Images Figure 3 19 Comparison of calculated burning velocity versus number of images 25 used 71 3 10 Uncertainty Each individual component adds a certain amount of uncertainty to the work The mass flow controllers have an uncertainty of 1 of full scale 0 5 Ipm for air and 0 05 lpm for methane The dust feeder adds an amount of uncertainty to the flow The instantaneous fluctuations in the feed rate were not able to be quantified during the current work This could lead to uncertainty in the measurement However due to the long duration of sampling time the potential effect is minimalized An uneven distribution of dust inside
9. The dust does not There is continuous completely vaporize in burning of the particles the preheat zone behind the reaction zone The condense phase and gas phase have lean conditions There is continued burning of the dust particles behind the reaction Zone The condense phase has a The dust does not rich condition but the gas completely vaporize in phase has a lean condition the preheat zone The condense phase has a rich condition and the gas phase in the preheat zone is rich The dust does not There is continued completely vaporize in vaporization behind the the preheat zone flame The condense phase has a The dust completely rich condition and the 885 vaporizes in the preheat phase has a rich condition zone There is excess fuel behind the flame References Proust C Flame Propagation and Combustion in some dust air mixures Journal of Loss Prevention in the Process Industries 2006 19 p 89 100 Eckhoff R K Dust Explosions in the Process Industries Third Edition Third ed 2003 Boston Gulf Professional Publishing Huang Y G A Risha V Yang and R A Yetter Combustion of Bimodal Nano Micro Sized Aluminum Particle Dust in Air Proceedings of the Combustion Institute 2007 31 p 2001 2009 Sun J R Dobashi and T Hirano Structure of Flames Propagating through Aluminum Particles Cloud and Combustion Process of Particles Journal of Loss Prevention in the Pr
10. In the corrugated flame regime the flame front will be pushed around and folded by the largest eddies The smallest eddies which are just capable of affecting the flame are those with a rotational velocity assumed to be the turbulent intensity egual to the laminar burning velocity In the distributed reaction zone regime the macro eddies fold the flame front to form bulges of a size in the order of the integral length scale If these bulges extend into the unburned mixture then the local laminar burning velocity becomes less than that of an unstretched flame At the small bulges the radius of curvature is so small that the effect of quenching due to curvature is large enough to cause local extinction The flame is cut into pieces by the small eddies and these 91 pieces are scattered across the flame zone by the larger eddies As a consequence there is no well defined flame structure and the flame front consists of a collection of pockets of unburnt and burnt mixture Therefore the results of using the shadowgraph to examine the flame edge is less reliable in this range It should be understood that the discussion so far pertains to turbulent gas flames alone Additional parameters will arise for turbulent dust flames owing to the coupling between the condensed phase and gas phase Micron sized particles influence the turbulent flow structure by Crowe et al 12 1 displacement of the flow field by flow around a dispersed phase element 2
11. standard deviation BV a end if dust conc a 25 burn vel 10 75 25 1 lt Burning velocity a stanDev BV 10 75 25 1 standard deviation BV a 143 end if dust_conc a 50 burn_vel_10_75_50 1 Burning_velocity a stanDev_BV_10_75_50 1 standard deviation BV a end if dust conc a 75 burn vel 10 75 75 1 lt Burning_velocity a stanDev BV 10 75 75 1 standard deviation BV a end end if particleSize 75 amp amp flowRate a 30 if dust conc a burn vel 10 75 00 2 lt Burning_velocity a stanDev BV 10 75 00 2 standard deviation BV a end if dust conc a 25 burn vel 10 75 25 2 lt Burning velocity a stanDev BV 10 75 25 2 standard deviation BV a end if dust conc a 50 burn vel 10 75 50 2 Burning_velocity a stanDev BV 10 75 50 2 standard deviation BV a end if dust conc a 75 burn vel 10 75 75 2 Burning_velocity a stanDev BV 10 75 75 2 standard deviation BV a end end if particleSize 75 amp amp flowRate a 35 if dust_conc a burn_vel_10_75_00 3 Burning_velocity a stanDev BV 10 75 00 3 standard deviation BV a end if dust conc a lt lt 25 burn vel 10 75 25 3 lt Burning velocity a stanDev BV 10 75 25 3 standard deviation BV a end if dust_conc a 50 burn_vel_10_75_50 3 Burning_velocity a stanDev_BV_10_75_50 3 standard deviation BV a end if dust conc a 75 burn vel 10 75 75 3 lt Burning
12. 34 35 36 37 38 39 40 41 42 43 Abdel Gayed R G D Bradley and M Lavves Turbulent Burning Velocities A General Correlation in Terms of Straining Rates Proceedings of the Royal Society of London Series A Mathmatical and Physical Sciences 1987 414 1847 p 389 413 Ballal D R and A H Lefrebvre The Structure and Propagation of Turbulent Flames Proc R Soc Lond A 1975 334 p 217 234 Ballal D R The influence of laminar burning velocity on the structure and propagation of turbulent flames Proc R Soc Lond A 1979 367 p 485 502 Karlovitz B D W Denniston and F E Wells Investigation of turbulent flames Journal of Chemical Physics 1951 19 5 p 541 547 Karlovitz B 1954 AGARD London Butterworths p 247 262 Richmond J K J M Singer E B Cook J R Grumer and D S J Burgess Proc Combust Inst 1957 6 p 301 311 Damkohler G 1940 Petrov E A and A V Talantov 1959 Williams C G H C Hottel and A C Seurlock Proc Combust Inst 1949 3 p 21 40 Zotin V K and A V Talantov Izv vyssh ucheb Zaved Aviat Teknol 1966a 1 p 115 122 Zotin V K and A V Talantov Izv vyssh ucheb Zaved Aviat Teknol 1966b 3 p 98 103 Bollinger L M and D T Williams 1949 Grover J H E n Fales and A C Scurlock Proc Combust Inst 1963 9 p 21 35 Ishino Y K Takeuchi s Shiga and N Ohiwa Measurement of Instantaneous 3D Distribution of
13. Contents A1 1 Edge selection script A1 2 Edge data analysis script A1 3 Laminar data plotting script for figure 4 1 A1 4 Data validation script for figure 4 3 4 5 and 4 10 A1 5 Borghi diagram calculations for figure 4 6 A1 6 Plotting all data script for figure 4 9 A1 7 Plotting turbulent burning velocity vs turbulent intensity for Fig 4 11 A1 8 Plotting Normalized burning velocity vs turbulent intensity for Fig 4 12 A1 9 Plotting burning velocity vs dust concentration for Fig 4 14 A1 10 Fitting theory script A1 11 Creating arrays of test data as a function of dust concentration A1 12 Turbulent intensity calculation A1 13 Gas analysis data retrieval A1 14 plotAverage_noplot 132 A3 1 Edge selection script clear all close all cle dname save WOmm pos0 phi 0 8 dst 000 000g m3 V 10lpm directory EAHFA test data HFA Test Data particleSize 106 125 micron coal experimentSpecs 0mm pos0 V 010lpm dataFileName Omm pos0 phi 0 8 dst 000 000g m3 V 10lpm dname save directory particleSize experimentSpecs dataFileName dname dname save Original 1 76 Default Directory To be Opened matFileName particleSize dataFileName image start 1 image end 25 number of pictures to operate on Operating on individual images top_file dname NV Set up main database to open and look inside 15 top file z Is top file List
14. end end if particleSize 106 amp amp flowRate a 35 if dust conc a burn vel 08 106 00 3 Burning_velocity a stanDev BV 08 106 00 3 standard deviation BV a end if dust conc a lt lt 25 burn vel 08 106 25 3 Burning_velocity a stanDev BV 08 106 25 3 standard deviation BV a end if dust_conc a 50 burn_vel_08_106_50 3 Burning_velocity a stanDev BV 08 106 50 3 standard deviation BV a end if dust conc a lt lt 75 burn vel 08 106 75 3 Burning_velocity a stanDev BV 08 106 75 3 standard deviation BV a end end if particleSize 106 amp amp flowRate a 40 if dust conc a burn vel 08 106 00 4 Burning_velocity a stanDev BV 08 106 00 4 standard deviation BV a end if dust conc a lt lt 25 burn vel 08 106 25 4 Burning_velocity a stanDev BV 08 106 25 4 standard deviation BV a 140 end if dust_conc a 50 burn_vel_08_106_50 4 Burning_velocity a stanDev BV 08 106 50 4 standard deviation BV a end if dust conc a 75 burn vel 08 106 75 4 Burning_velocity a stanDev BV 08 106 75 4 standard deviation BV a end end if particleSize 75 amp amp flowRate a 10 if dust_conc a burn_vel_08_75_00 1 Burning_velocity a stanDev BV 08 75 00 1 standard deviation BV a end if dust conc a 25 burn vel 08 75 25 1 Burning_velocity a stanDev BV 08 75 25 1 standard deviation BV a end if dust conc a 50
15. 10 106 35 dst sze lam data dst sze end l OO am data dst sze end burn vel 10 106 35 dst sze orig burn vel 10 75 10 dst sze origFit sv MarkerSize plotMarkerSize Line Width plotLineWidth plot dust conc 10 106 40 dst sze lam data dst 526 6 0 1 00 lam data dst sze end 1 burn vel 10 106 40 dst sze orig burn vel 10 75 10 dst sze origFit lam data dst sze end 1 bh MarkerSize plotMarkerSize Line Width plotLineWidth ylim y_axisMin y axisMax hold off subplot subplot 3 2 5 Parent figure3 Y Tick 2 2 4 2 8 3 2 3 6 4 158 LineVVidth 2 FontWeight bold FontSize 14 FontName Times New Roman subplot 3 2 5 hold on plot dust 12 75 10 dst sze burn vel 12 75 10 dst sze ks MarkerSize plotMarkerSize LineWidth plotLi ne Width plot dust 12 75 30 dst sze lam data dst sze end OO am data dst sze end burn vel 12 75 30 dst sz e orig burn vel 12 75 10 dst sze origFit ro MarkerSize plotMarkerSize LineWidth plotLineWidth plot dust 12 75 35 dst sze lam data dst sze end 1 OO am data dst sze end burn vel 12 75 35 dst sz e orig burn vel 12 75 10 dst sze origFit sv MarkerSize plotMarkerSize LineWidth plotLineWidth plot dust 12 75 40 dst sze lam data dst sze end OO am data dst sze end burn vel 12 75 40 dst sz e orig burn vel 12 75 10 dst sze origFit bh MarkerSize plotMarkerSize LineWidth p
16. 14 phi 0 8 figurel figure Name NDim turbulent velocity axesl axes Parent figurel LineWidth 2 FontWeight bold FontSize 22 FontName Times New Roman hold on dummy plots to get the legend to have data markers and fitted curve lines if legend_plot 1 plot 1 1 rs MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 gs MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 bs MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 rv MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 gv MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 bv MarkerSize plotMarkerSize LineWidth plotLineWidth legend d 75 d_ st 25 d 75 d_ st 50 d 75 d_ st 75 d 106 d_ st 25 d 106 d_ st 50 d 106 d_ st 75 Location eastoutside end hold on plot u prime all burn vel 08 75 00 orig 1 burn vel 08 75 00 orig burn vel 08 75 10 dst sze origFit 1 ks LineW idth plotLineWidth plot u prime all burn vel 08 75 25 orig 1 burn vel 08 75 25 orig burn vel 08 75 10 dst sze origFit 2 Ts Line Width plotLineWidth plot u_prime_all burn_vel_08_75_50_orig 1 burn_vel_08_75_50_orig burn_vel_08_75_10_dst_sze_origFit 3 gs Line Width plotLine Width plot u_prime_all 1 length burn_vel_08_75_75_orig burn_vel_08_75_75_orig 1 burn_vel_08_75_75_orig burn_ vel 08
17. 40 dust conc 08 75 40 dst sze ct 10 dust_conc a burn vel 08 75 40 dst sze ct 10 Burning_velocity a ct 10 lt ct 10 1 end end if phi a 1 0 if particleSize 106 amp amp flowRate a 10 dust_conc_10_106_10_dst_sze ct_11 dust_conc a burn_vel_10_106_10_dst_sze ct_11 Burning_velocity a ct ll ct 11 1 170 end if particleSize 106 amp amp flowRate a 30 dust_conc_10_106_30_dst_sze ct_12 dust_conc a bum vel 10 106 30 dst sze ct 12 Burning velocity a ct 12 ct 12 41 end if particleSize 106 amp amp flowRate a 35 dust conc 10 106 35 dst sze ct 13 dust_conc a burn vel 10 106 35 dst sze ct 13 lt Burning velocity a ct 13 lt ct 13 1 end if particleSize 106 amp amp flowRate a 40 dust conc 10 106 40 dst sze ct 14 dust_conc a bum vel 10 106 40 dst sze ct 14 Burning velocity a ct 14 ct 14 1 end if particleSize 75 amp amp flovvRate a 10 dust conc 10 75 10 dst sze ct 15 dust conc a burn vel 10 75 10 dst sze ct 15 Burning velocity a ct 15 lt ct 15 1 end if particleSize 75 amp amp flowRate a 30 dust conc 10 75 30 dst sze ct 16 dust_conc a burn vel 10 75 30 dst sze ct 16 Burning_velocity a ct 16 lt ct 16 1 end if particleSize 75 amp amp flowRate a 35 dust_conc_10_75_35_dst_sze ct_17 dust_conc a burn_vel_10_75_35_dst_sze ct_17 Burning_
18. 6 Occupational Safety and health Administration OSHA Combustible dust expert forum meeting summary report 2011 7 Galfetti L L T De Luca F Severini L Meda G Marra M Marchetti M Regi and S Bellucci Nanoparticles for Solid Rocket Propulsion Journal of Physics Condensed Matter 2006 18 p 33 8 Buhre B J P L K Elliott C D Sheng R P Gupta and T F Wall Oxi fuel Combustion Technology for Coal Fired Power Generation Progress in Energy and Combustion Science 2005 31 4 p 283 307 9 Palmer K N Dust Explosions and Fires 1973 London United Kingdom Chapman and Hall Ltd 10 C B Parnell J R O McGee F J Vanderlick and A Contreras A Critical Evaluation of Combustible Dust Test Methods in Beyond Regulatory Compliance Making Safety Second Nature 2011 Texas A amp M University College Station Texas 11 NFPA 68 Standard on Explosion Protection by Deflagration Venting 2007 National Fire Protection Association 12 ASTM E2019 Standard Test Method for Minimum Ignition Energy of Dust Cloud in Air 2010 35 13 ASTM E1515 Standard Test Method for Minimum Explosible Concentration of Combustible Dusts 2010 14 ASTM E1226 Standard Test Method for pressure and rate of pressure rise for combustible dusts 2010 American Society for Testing and Materials ASTM 15 OSHA Hazard Communication Guidance for Combustible Dusts 2009 16 ASTM E 1354 E 1354 Standard Test Method for Heat and Visib
19. Inst 1996 The Combustion Institute Kobayashi H T Tamura K Maruta T Niioka and F A Williams Burning Velocity of Turbulent Premixed flames in a high Pressure Environment Proc Combust Inst 1996 26 p 389 396 Khramtsov V A Investigation of pressure effect on the parameters of turbulence and on turbulent burning Proc Combust Inst 1959 7 p 609 620 Liu Y and B Lenze The influence of turbulence on the burning velocity of premixed CH4 H2 flames with different laminar burning velocities Proc Combust Inst 1988 22 p 747 754 Turns S R An Introduction to Combustion Concepts and Applications 2000 New York McGraw Hill Pope S B Turbulent Premixed Flames Ann Rev Fluid Mech 1987 19 p 237 270 Chomiak J Basic Considerations in the Turbulent Flame Propagation in Premixed Gases Prog Energy Combustion Science 1979 5 p 207 221 Borghi R and D Dutoya On the Scales of the Fluctuations in Turbulent Combustion 17th Symp Int on Combustion 1979 1 p 235 244 Liu R D S K Ting and G W Rankin On the Generation of Turbulence with a Perforated Plate Experimental Thermal and Fluid Science 2004 28 p 307 316 Roach P E The Generation of Nearly Isotropic Turbulence by Means of Grids Heat and Fluid Flow 1987 8 2 p 82 92 Bruun H H Hot Wire Anemometry Principles and Signal Analysis 1995 p pg 64 Hattori H Flame Propagation in pulverized coal air mixtures in Proc Combus
20. The diffuser provided expansion and laminarization of the dust flow which is initially turbulent in the dust disperser Han et al 53 54 published results from a combustion system for laminar flame propagation in dust air mixtures The main part of the system consisted of a vertical duct 1800 mm height with 150x150 mm square cross section a shutter an ignition device a dust cloud generator and an airflow feeder with pressure controller The dimensions of the combustion duct were chosen to reduce the amount of lateral heat losses from the flame to the duct walls similar to Proust et al 51 Dust suspensions were generated through elutriation of dust particles above a fluidized bed Aspects of flame propagation were observed through a glass 1800 mm high in the front of the vertical duct Using the slide type windows of quartz glass on the side of the duct it was possible to make a laser light sheet from the side wall of the duct and change the observation area of flame propagation A pair of electrodes for spark ignitions was placed 150 mm above the lower end of the duct The dust particles were layered on a fine porous plate at the bottom of the duct Air at appropriate rates was introduced through the porous plate which acted as a flow rectifier to disperse the dust particles when the upper end of the duct was open When the duct was entirely filled with a dust cloud a time controlling system interrupted the air flow and removed the fluidized bed
21. errorbar l_0 1 length burn_vel_08_106_75 burn_vel_08_106_75 burn_vel_08_75_00 1 length burn_vel_08_75_ 75 stanDev_BV_08_106_75 burn_vel_08_75_00 1 length burn_vel_08_106_75 bv LineWidth plotLineWidth plot l O burn vel 08 106 00 orig ks LineWidth plotLineWidth plot l O burn vel 08 106 25 orig ro LineWidth plotLineWidth plot l O burn vel 08 106 50 orig sv Line Width plotLineWidth plot O 1 length burn vel 08 106 75 orig burn vel 08 106 75 orig bh LineWidth plotLineW idth plot l O burn vel 10 75 00 orig ks LineWidth plotLineWidth plot O burn vel 10 75 25 orig ro LineWidth plotLineW idth plot O burn vel 10 75 50 orig gv Line Width plotLineWidth plot O burn vel 10 75 75 orig bh LineWidth plotLineWiidth plot l 00 burn vel 10 106 00 orig ks LineWidth plotLineWidth plot l 00 burn vel 10 106 25 orig ro LineW idth plotLineWidth plot l 00 burn vel 10 106 50 orig ev LineWidth plotLineWidth plot l 00 1 length burn vel 10 106 75 orig burn vel 10 106 75 orig bh LineWidth plotLineWidth plot xfit data yfit 10 75 00 k LineWidth plotLineWidth plot xfit data yfit 10 75 25 yfit 10 75 00 r LineWidth plotLineWidth plot xfit data yfit 10 75 50 yfit 10 75 00 g LineWidth plotLineWidth plot xfit data yfit 10 75 75 yfit 10 75 00 b LineWidth plotLineWidth plot xfi
22. if dust conc a lt lt 50 burn vel 12 75 50 3 lt Burning velocity a stanDev BV 12 75 50 3 standard deviation BV a end if dust conc a lt lt 75 burn vel 12 75 75 3 lt Burning velocity a stanDev BV 12 75 75 3 standard deviation BV a end end if particleSize 75 amp amp flowRate a 40 if dust conc a 0 burn_vel_12_75_00 4 Burning_velocity a stanDev BV 12 75 00 4 standard deviation BV a end if dust conc a lt lt 25 burn vel 12 75 25 4 lt Burning velocity a stanDev BV 12 75 25 4 standard deviation BV a end if dust_conc a 50 burn_vel_12_75_50 4 Burning_velocity a stanDev BV 12 75 50 4 standard deviation BV a end if dust conc a 75 bum vel 12 75 75 4 Burning_velocity a stanDev BV 12 75 75 4 standard deviation BV a end end a zatl 1 0 08 u prime all 2 end burn vel 08 75 00 1 1 0 10 u prime all 2 end burn vel 10 75 00 1 10 12 u prime all 2 end burn vel 12 75 00 1 plotMarkerSize 10 147 plotLineWidth 3 textSize 16 figure2 figure testSize2 14 axes2 axes Parent figure2 YMinorTick on XMinorTick on FontSize testSize2 hold on dummy plots to get the legend to have data markers and fitted curve lines plot 1 1 ks MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 rs MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 gs M
23. za T 4 11 and therefore 1 12 v 1 q 4 12 E r For strong turbulence the integral on the right side of the Eg 4 10 assumes a definitive value which is egual to the time scale of the turbulence therefore Am 4 13 and 1 12 ST x 25 T L S 4 14 Combining this with Eq 4 6 and dividing by Sr yields 85 4 15 In the case of intermediate turbulence the root mean square displacement depends on the shape of the correlation function If the shape of the correlation function is approximated by a parabola es 4 16 The integral on the right hand side of Eg 4 10 may be solved for the variance of the displacement conseguently SE 4 17 12 rms mu 4 18 8 12 S and ta Small scale turbulence is believed to contain insufficient kinetic energy to distort a laminar flame but in view of Egs 4 6 and 4 7 it is reasonable to expect that the turbulent burning velocity behaves in accordance with Eg 4 4 According to Karlovitz s analysis the turbulent burning velocity should at first increase linearly but then more slowly as turbulence intensity increases this is similar to what is observed experimentally in Fig 4 3 In order to find a correlation for the effect of any particular type of turbulent motion on the turbulent burning velocity researchers have adopted a generalization of equations 4 4 4 5 4 15 and 4 18 86 4 19 where n
24. 0 2 data id avgData 3 nLines end end PLOT AVERAGE open a global figure if necessary otherwise start the plotting loop if opt plot2NevvFigure 2 outFh figure name collected averages outAh axes nextPlot add end avgH zeros nData 1 opt addErrorBars for id 1 nData find out where to plot switch opt plot2NewFigure case 0 181 data id ahOut data id ahIn set data id ahOut NextPlot add case 1 outFh figure data id ahOut axes case 2 data id ahOut outAh otherwise check whether an axes handle has been supplied if ishandle opt plot2NewFigure 8282 strcmp get opt plot2NevvFigure type axes data id ahOut opt plot2NewFigure else error unsupported option for plot2nevvFigure end end plot if opt addErrorBars err data id avgData 2 if opt plotSEM err err sgrt data id avgData 3 end end if opt horzAvg avgH id 1 plot data id ahOut data id avgData I data id avgPoints k Line Width 2 Tag avg if opt addErrorBars errH myErrorbar data id ahOut data id avgData 1 data id avgPoints err NaN length data id avgPoints 1 delete errH 1 avgH id 2 errH 2 end else avgH id 1 plot data id ahOut data id avgPoints data id avgData 1 k LineWidth 2 Tag avg if opt addErrorBars avgH id 2 myErrorbar data id ahOut data id avgPoints data id avgData 1 err end end set le
25. 1 2 there is no significant effect of on the burning velocity as the dust concentration is increased for the low turbulent intensity but a slight increase for the high turbulent intensity 4 3 Correlation of turbulent burning velocity Figure 4 14 shows the turbulent burning velocity versus the turbulent intensity Using Eq 4 19 discussed in section 4 2 1 100 de CEJ 4 23 two sets of C and n parameters are found based on fuel lean or rich conditions For gas phase equivalence ratios less than one the best fit is observed for C 2 2 n 0 2 and equivalence ratios greater than one C 1 7 n 0 2 Two values of C are used because the volatiles release by the dust in the lean phase boost the burning velocity more than with higher equivalence ratios 0 8 oe o lt 1 75 901m ma u 1 0 n 25 o 50 u 4 75 1 0 1 0 9 25 50 75 1 2 0 25 50 75 b b b b 1 rms S L st Figure 4 14 Correlation for turbulent burning velocity of hybrid flames A similar plot can also be generated for the larger particle size range tested d lt 106 12574n and is shown in Fig 4 15 In this case C 2 0 fuel lean and C 1 65 fuel rich while the exponent n remains the same n 0 2 Thus when particle size range increases a similar trend is observed although the percentage change in the value of C between fuel lean and fuel rich conditions is smaller when c
26. 23 nozzleArea pi 4 nozzleDiameter 2 vel flow flowRate 60000 nozzleArea m s velocity based on flow rate in tube vel flow save a 1 vel flow fileToReadl dname V fileName newDatal importdata fileToRead1 vars fieldnames newDatal for i 1 length vars assignin base vars i newDatal varsfi end calculate rms value num_samples 100000 173 time data 1 num_samples 1 s time stamp E all data 1 num_samples 2 volts voltage from anemometer E E all E_bar mean E volts average voltage E_bar_save a 1 E_bar cal_factor vel_flow E_bar m s volts u 0 000215 exp 7 918735 E m s u_bar mean u u prime u u bar u_prime_max max abs u_prime RMS TIMEssgrt sum T Time T Time length T Time u prime rms2 lt sgrt sum u prime u prime length u prime u prime rms sgrt mean u prime 2 u prime rms save a u prime rms ACF acf u prime 1000 1 0 u_bar sum ACF auto_corr xcorr u_prime u_prime 10000 1 0 u_bar sum auto_corr length auto_corr 2 end 1 100000 Tl u T2 T_1 offset 0 T 1 sin 0 0 01 480 C First temperature profile left T_2 offset 1 length T_1 T_1 1 dength T_1 offset maxlag 1000 maximum size of sampling lag window_size length T_1 2 maxlag 1 length of data profile to use T_ls T_I maxlag 1 maxlag window_size create Ist correlated profile
27. Also if you have added error bars with errorbar instead of myErrorbar the error bars are included in the averaging and you will get unexpected results created with MATLAB ver 7 10 0 59 R2010a on Mac OS X Version 10 6 2 Build 10C540 created by jonas DATE 26 Jan 2010 Go Go Go Ho Ho Go Yo Fo Go Go Fo o Vo Fo Yo Ho Go Yo Fo Fo Go Fo Fo Go Go Fo Go Go Ho Go Fo Ho Ho Go Yo Ho Fo Go Yo Fo Fo Yo Fo Go Go Yo Go Go Jo Po Go Yo Ho Go Yo Vo Fo Go Go Fo Fo Go Fo Po Go Jo TEST INPUT set defaults opt struct addErrorBars true horzAvg false interpMethod linear plot2NewFigure 0 useRobustMean true plotSEM true find all axes handles plot data to new figure if necessary if nargin lt 1 l isempty handleOrData handleOrData gcf end class cell is data class double is handle if isa handleOrData cell if isEven length handleOrData error Data needs to be supplied in x y pairs e g xl y1 x2 y2 end plot a new figure figure plot handleOrData handleOrData gca end loop through handles to get list of axes handles Skip improper handles ahList for ih 1 length handleOrData if ishandle handleOrData ih Assume it s a 3D plot if the view is not standard 2D if stremp get handleOrData ih type axes amp amp all get handleOrData ih View 0 90 ahList ahList handleOrData ih elseif
28. Analyzer Fuel control system for Hybrid Flame Analyzer Water cooling system for Hybrid Flame Analyzer Building Annular Ring Pilot Flame for Turbulent Burner Hybrid Flame Analyzer Simple shadowgraph design description Hybrid Flame Analyzer Gas analysis for combustion system Hybrid Flame Analyzer How to use mass flow controllers Hybrid Flame Analyzer Changing perforated plate in Hybrid Flame Analyzer Calibrating volumetric dust feeder Hybrid Flame Analyzer Setting up hot wire anemometer for Hybrid Flame Analyzer Checking hotwire anemometer voltage for Hybrid Flame Analyzer 112 Recommended Personal Protective Equipment PPE Gloves Safety Glasses Lab Coat Respirator dust mask Steel toed boots 113 Turning on HFA Start water cooling 10 Iph to much more will rupture water cooling fittings The water is controlled using the sink taps Rotate the knob toward the sink to turn the water on and away from you to turn the water off It does not need to be turned very far lt 1 8 turn to get the recommended flow The flowmeter clamped to the sink will show the flow going through the tubing the stainless steel valve can be used to control the flow but it is recommended to use the sink knob itself to avoid building up pressure in the line between the sink and the flowmeter inlet There is a clear plastic water flow indicator as shown below When the red ball is turning it is easy to see the water is flowing without ha
29. Burn vel dst part szevOl ME Ee Ge GE 170 A3 8 Turbulent intensity 190 3 9 Gas analysis data retrieval RKK a 193 A3 10 plotAverage noplot ee ee ee ee 194 Appendix 4 Error Bar values standard deviation of velocity calculation 200 Appendix 5 Flame Images sca ESSEN a Gn k elvei 201 List of figures 1 1 Diagram of explosion sphere 1 2 Pressure vs time curve and change in pressure vs time curve from an explosion sphere 1 3 Diagram of explosion sphere with increasing turbulence as the flame propagates 2 1 Schematic illustration of the structure of a premixed dust air flame 2 2 Schematic of flame structure in a dust air flame 3 1 Hybrid Flame Analyzer HFA combustion chamber 3 2 HFA exhaust system diagram 3 3 Diagram of experimental section of Hybrid Flame Analyzer HFA 3 4 Images of burner nozzles 3 5 Diagram of turbulent burner nozzle 3 6 Turbulent burner parts 3 7 Image of premixed methane oxygen pilot flame 3 8 Images of perforated plates 3 9 Calibration curve for hot wire anemometer 3 10 Turbulent intensity versus flow rate 3 11 Comparison of calculated turbulent intensity vs number of samples 3 12 Diagram of dust feeder bloc
30. Combust Inst 1957 36 35 Hattori H Flame Propagation in pulverized coal air mixtures in Proc Combust Inst 1957 36 Burgoyne J H and V D Long Some Measurements of the Burning Velocity of Coal in air Suspensions 1n Conference on Science in the use of coal 1958 37 Palmer H B D J Seery and W F Marshall A study of the burning velocity of laminar coal dust flames 1962 38 William F Marshall J The Effect of Concentration and Particle Size on the Burning Velocity of Laminar Coal Dust Flames in Department of Fuel Technology 1964 Pennylvania State University 39 Mason W E and M J G Wilson Laminar Flames of Lycopodium Dust in Air Combustion and Flame 1967 11 3 p 195 200 40 Bryant J T The Combustion of Premixed Laminar Graphite Dust Flames at Atmospheric Pressure Combustion Science and Technology 1971 2 p 389 399 41 Strehlow R A L D Savage and S C Sorenson Coal Dust Combustion and Suppression in AIAA SAE 10th Propulsion Conference 1974 San Diego CA 42 Milne T A and J E Beachey The Microstructure of Pulverized Coal Air Flames I Stabilization of Small Bunsen Burner and direct sampling techniques Combustion Science and Technology 1977 16 p 123 138 43 Bradley D Z Chen S El Sherif S E D Habik and G John Structure of Laminar Premixed Carbon Methane Air Flames and Ultrafine Coal Combustion Combustion and Flame 1994 96 p 80 96 44 Goroshin S I Fomenko an
31. Files inside main folder z cellstr ls top file Turn cells from Is function into strings cc c 3 length c Set up a matrix without the and produces by the s function S z size cc 9 Find the size of matrix containing names of files inside of main database a image start This counter is set to 3 to account for the and at the beggining of each matrix created by Is image image start ref width check 0 while image lt image end close all file char cellstr top file char cc image File to be operated on file_name char cc image display file being operated on in command window Fo Yo Go Fo Fo To Vo Go Go Vo Vo Fo Yo Fo Ho Yo Go Go Fo Yo Yo Fo Fo Fo Yo Yo Yo Yo Go Go Fo Go Yo Yo Go Go Go Po Po Put code to operate on each file in a folder here fileToRead2 file imemp lt imread fileToRead2 colour map of imported image imemp imrotate imcmp 90 Rotate image image_magnification 63 crop 2 if crop imcmp imcrop imemp 500 1300 500 8501 elseif crop imcmp imcrop imemp 500 1100 750 1050 end imemp imemp 2 change to blue channel only 133 imemp imadjust imcmp size_y size_x spare size imcmp dname_x_pix_save dname_save x_pix_save txt dname y pix save dname_save Ny pix save txt if ref width check x pix save zeros 4 1 y pix save zeros 4 1 save dname x pix save x pix Save ascii double
32. J T Verheijen S M Lemkowitz and B Scarlett 1996 Dust Explosions in Spherical Vessels Prediction of the Pressure Evolution and Determination of the Burning Velocity and Flame Thickness Volume 26 Eckhoff R K Dust Explosions in the Process Industries 2003 Boston Gulf Professional Publishing 27 Dahoe A E R S Cant and B scarlett On the Decay of Turbulence in the 20 Liter Explosion Sphere Flow Turbulence and Combustion 2001 67 p 159 184 28 Cashdollar K L Overview of Dust Explosibility Characteristics Journal of Loss Prevention in the Process Industries 2000 13 p 183 199 29 Skjold T Review of the DESC project Journal of the Loss Prevention in the Process Industries 2007 20 p 291 302 30 Arntzen B J H C Salvesen H F Nordhaug LE Storvik and O R Hansen CFD Modelling of Oil Mist and Dust Explosion Experiments 31 Robinson G F Pollutant Formation in Turbulent Flames in Field of Mechanical Engineering dt Astronautical Sciences 1974 Northwestern University Evanston Illinois 32 Joshi N D Gravitational Effects on Particle Cloud Flames in Mechanical Engineering 1984 State University of New York at Stony Brook New York 33 Cassel H M A K D Gupta and S Guruswamy Factors Affecting Flame Propagation Through Dust Clouds Third Symposium on Combustion Flame and Explosion Phenomena 1949 p 185 190 34 Ghosh B D Basu and N K Roy Studies of Pulverized Coal Flames in Proc
33. Local Burning Velocity on a Trubulent Premixed Flame by Non Scanning 3D CT Reconstruction in Proc Combust Inst 2009 Fells I and H G Rutherford Burning velocity of methane air flames Combustion and Flame 1969 13 p 130 Halpern C 1958 Res Natl Bur Std p 535 Lee J Burning velocity measurements in aluminum air suspensions using bunsen type dust flames 2001 Rallis C J and A M Garforth The Determination of Laminar Burning Velocity Prog Energy Combustion Science 1980 6 p 303 329 75 4 Results and Analysis Chapter 4 discusses the experimental results for laminar and turbulent flames using the hybrid flame analyzer HFA discussed in Chapter 3 4 1 Laminar flames Figure 4 1 a f shows the laminar burning velocity Sus as a function of dust concentration for the three gas phase equivalence ratios d 0 8 1 0 1 2 and two dust particle sizes 75 90 um 106 125 um The subscript L denotes laminar while st denotes the presence of dust st represents staub in the German language for dust The naming convention is consistent with that adapted by NFPA 68 and 69 For the 75 90 um particle range and equivalence ratio 0 8 and 1 2 Figs 4 1a and 4 le it is observed that the burning velocity is reduced when compared to the gas only value as the dust concentration is increased from 0 to 75 g m The effect is small less than 10 but more than the experimental uncertainty A
34. T_ls T 15 mean T_1 Normalizing the temperature profile sigma_13 std T_1 std T_2 Calculating standard deviation 1 zeros maxlag 1 1 9 create initial CC coefficient matrix Cross Correlation claculations for i maxlag T_2s T_2 i maxlag 1 i maxlag window_size T_2s T_2s mean T_2 CCC1 i maxlag 1 T 2s T 1s length T 2s sigma 13 Cross correlation 1st side CCC1 i 1 T 25 1s mean T 1 2 Cross correlation 1st side end lag spacing 1 2 maxlag 1 maxlag 1 Create matrix is lag spacings Plot Comparison close all figure plot CCC1 ylabel Correlation Coefficient 174 pause 0 2 1 02 8 u bar sum CCC1 1 100000 1 0 1 02 a if position 1 amp amp height 0 Turb int posl htO ct 01 1 lt u prime rms u prime max posl htO ct 01 1 u prime max 1 0 posl htO ct 01 1 1 0 ct 01 ct 01 1 elseif position 3 amp amp height Turb int pos3 htO ct 02 1 u prime rms u prime max pos3 htO ct 02 1 u prime max 1 0 80583 htO ct 02 1 0 ct_02 ct 02 1 elseif position 6 amp amp height Turb int pos6 htO ct 03 1 u_prime_rms u prime max pos6 ht ct 03 1 u prime max 1 8086 0 03 1 0 ct_03 ct 03 1 elseif position 1 amp amp height 3 Turb int posl ht3 ct 04 1 u_prime_rms u prime max posl ht3 ct 04 1 u prime max 1 0 posl ht3 ct 04 1 1 0 ct 04 ct 04 1 elseif position 3
35. YTick 2 2 4 2 8 3 2 3 6 4 LineVVidth 2 FontWeight bold FontSize 14 FontName Times New Roman hold on plot l 0 08 00 burn vel 08 75 00 orig lam data end burn vel 08 75 10 dst sze origFit 1 ks LineWidth pl otLineVVidth plot l 0 08 25 burn vel 08 75 25 orig lam data end burn vel 08 75 10 dst sze origFit 2 To LineWidth pl otLineWidth plot l 08 50 burn vel 08 75 50 orig lam data end burn vel 08 75 10 dst sze origFit 3 sv LineWidth pl otLineWidth 153 plot l 0 08 75 1 length burn vel 08 75 75 burn vel 08 75 75 orig lam data end burn vel 08 75 10 dst s ze_origFit 4 bh LineW idth plotLineWidth ylim y axisMin y axisMax hold off axis tight subplot subplot 3 2 2 Parent figurel YTick 2 2 4 2 8 3 2 3 6 4 LineVVidth 2 FontWeight bold FontSize 14 FontName Times New Roman subplot 3 2 2 hold on plot l 0 08 00 burn vel 08 106 00 orig lam data end burn vel 08 75 10 dst sze origFit 1 ks LineWidth p lotLineW idth plot l 0 08 25 burn vel 08 106 25 orig lam data end burn vel 08 75 10 dst sze origFit 2 ro LineWidth p lotLineW idth plot l 0 08 50 burn vel 08 106 50 orig lam data end burn vel 08 75 10 dst sze origFit 3 sv LineWidth plotLineW idth plot l 0 08 75 1 length burn vel 08 106 75 burn vel 08 106 75 orig lam data end burn vel 08 75 10 dst _sze_origFit 4 bh Line W
36. and F A Williams Burning Velocity of Turbulent Premixed flames in a high Pressure Environment Proc Combust Inst 1996 26 p 389 396 3 Grover J H E n Fales and A C Scurlock Proc Combust Inst 1963 9 p 21 35 4 Dahoe A Dust Explosions a Study of Flame Propagation in Applied Sciences 2000 Delft University of Technology p 298 5 Damkohler G NACA Tech Memo 1112 1947 National Advisory Committee for Aeronautics Washington 6 Schelkin K I On combustion in a turbulent flow NACA Tech Memo 1110 1947 National Advisory Committee for Aeronautics Washington 7 Karlovitz B D W Denniston and F E Wells Investigation of turbulent flames Journal of Chemical Physics 1951 19 5 p 541 547 8 Taylor G I Diffusion by continuous movements Proceedings of the Royal Society of London Series A Mathmatical and Physical Sciences 1921 20 p 196 212 9 Glassman I Combustion 1996 Academic Press San Diego Calif 10 Turns S R An Introduction to Combustion Concepts and Applications 2000 New York McGraw Hill 11 Peters N Laminar flamelet concepts in turbulent combustion Proc Combust Inst 1986 21 p 1231 1250 12 Crowe C M Sommerfeld and Y Tsuji Multiphase Flows with Droplets and Particles 1998 Boston CRC Press 13 Gore R A and C T Crowe The effect of particle size on modulating turbulent intensity Intl J Multiphase Flow 1989 15 p 279 14 Crowe C T On mode
37. at a temperature higher than ambient Thus possibility of autoignition is higher 44 1 Operating pressure bar Certain facilities can operate at pressures other than atmospheric Studies have shovvn that thermodynamic and thermo kinetic properties vary vvith pressure 45 Relative humidity Major of quantity of water vapor in ambient air 46 Confinement Dimensions of the enclosure which is considered to be at constant temperature and pressure and surrounds given test apparatus or control volume under consideration 47 Turbulence Flow instability represented by chaotic state of fluid Reynolds number motion with dissipative structure 48 Detonability limit Condition outside which self sustained propagation of detonation wave cannot be realized Test methods starting with ASTM and IEC are standard test methods Some standard test methods are not designed for dust per se but can be easily modified to include dust samples LASTM B761 06 Standard Test Method for Particle Size Distribution of Metal Powders and Related Compounds by X ray Monitoring of Gravity Sedimentation ii ASTM E 1226 05 Standard Test Method for Pressure and Rate of Pressure Rise for Combustible Dust iii ASTM E 1491 06 Standard Test Method for Minimum Autoignition Temperature of Dust Clouds iv ASTM 1515 07 Standard Test Method for Minimum Explosible Concentration of Combustible Dusts v ASTM E 1945 02 2008 Standard test Me
38. burn vel 08 75 50 1 Burning_velocity a stanDev BV 08 75 50 1 standard deviation BV a end if dust conc a lt lt 75 burn vel 08 75 75 1 lt Burning velocity a stanDev BV 08 75 75 1 standard deviation BV a end end if particleSize 75 amp amp flowRate a 30 if dust conc a burn vel 08 75 00 2 lt Burning velocity a stanDev BV 08 75 00 2 standard deviation BV a end if dust conc a lt lt 25 burn vel 08 75 25 2 lt Burning velocity a stanDev BV 08 75 25 2 standard deviation BV a end if dust_conc a 50 burn_vel_08_75_50 2 Burning_velocity a stanDev BV 08 75 50 2 standard deviation BV a end if dust conc a lt lt 75 burn vel 08 75 75 2 lt Burning velocity a stanDev BV 08 75 75 2 standard deviation BV a end end if particleSize 75 amp amp flowRate a 35 if dust conc a burn vel 08 75 00 3 lt Burning velocity a stanDev BV 08 75 00 3 standard deviation BV a end if dust conc a lt lt 25 burn vel 08 75 25 3 lt Burning velocity a stanDev BV 08 75 25 3 standard deviation BV a 141 end if dust_conc a 50 burn_vel_08_75_50 3 Burning_velocity a stanDev BV 08 75 50 3 standard deviation BV a end if dust conc a lt lt 75 burn vel 08 75 75 3 Burning velocity a stanDev BV 08 75 75 3 standard deviation BV a end end if particleSize 75 amp amp flowRate a 40 if dust conc a 0 burn_vel_08_75_
39. burn vel 08 75 10 dst sze orig 4 burn vel 106 10 dst sze burn vel 08 106 10 dst sze 1 end burn vel 08 106 10 dst sze orig 1 burn vel 106 30 dst 526 burn vel 106 30 dst sze lam data dst sze end burn vel 08 106 10 dst sze orig 2 burn vel 106 35 dst sze burn vel 106 35 dst sze lam data dst sze end burn vel 08 106 10 dst sze orig 3 burn vel 106 40 dst SZe burn vel 106_40_dst sze lam_data_dst_sze end burn_vel_08_106_10_dst_sze_orig 4 burn vel 75_10_dst sze burn vel 10 75 10 dst sze 1 end burn vel 10 75 10 dst sze orig 1 150 burn vel 10_75_30_dst burn vel 10_75_30_dst sze sze lam data dst sze end burn vel 10 75 10 dst sze orig 2 burn vel 10 75 35 dst burn vel 10 75 35 dst sze sze lam data dst sze end burn vel 10 75 10 dst sze orig 3 burn vel 10 75 40 dst burn vel 10 75 40 dst sze sze lam data dst sze end burn vel 10 75 10 dst sze orig 4 burn vel 10 106 10 dst sze burn vel 10 106 10 dst sze 1 end burn vel 10 106 10 dst sze orig 1 burn vel 10 106 30 dst sze burn vel 10_106_30_dst sze lam_data_dst_sze end burn_vel_10_106_10_dst_sze_orig 2 burn vel 10_106_35_dst sze burn vel 10_106_35_dst sze lam_data_dst_sze end
40. burn_vel_10_106_10_dst_sze_orig 3 burn vel 10_106_40_dst sze burn vel 10_106_40_dst sze lam_data_dst_sze end burn_vel_10_106_10_dst_sze_orig 4 burn vel 12_75_10_dst burn vel 12_75_30_dst burn vel 12_75_30_dst sze burn_vel_12_75_10_dst_sze 1 end burn_vel_12_75_10_dst_sze_orig 1 526 sze lam data dst sze end burn vel 12 75 10 dst sze orig 2 burn vel 12 75 35 dst burn vel 12 75 35 dst sze sze lam data dst sze end burn vel 12 75 10 dst sze orig 3 burn vel 12 75 40 dst burn vel 12 75 40 dst sze sze lam data dst sze end burn vel 12 75 10 dst sze orig 4 burn vel 12 106 10 dst sze burn vel 12 106 10 dst sze 1 end burn vel 12 106 10 dst sze orig 1 burn vel 12 106 30 dst SZe burn vel 12_106_30_dst sze lam_data_dst_sze end burn_vel_12_106_10_dst_sze_orig 2 burn vel 12_106_35_dst SZe burn vel 12_106_35_dst sze lam_data_dst_sze end burn_vel_12_106_10_dst_sze_orig 3 burn vel 12_106_40_dst SZe burn vel 12_106_40_dst sze lam_data_dst_sze end burn_vel_12_106_10_dst_sze_orig 4 Standard deviation stanDev_BV_08_75_00_orig stanDev_BV_08_75_00 stanDev_BV_08_75_25_orig stanDev_BV_08_75_25 stanDev BV 08 75 50 orig stanDev BV 08 75 50 stanDev BV 08 75 75 orig stanDev BV
41. chamber This door is composed of 0 3175cm 1 8 thick aluminum frame with a 25 4 10 by 20 32 cm 8 by 0 238 cm 3 32 plate glass allowing to see inside the combustion chamber The door was attached to the main aluminum frame using a one piece door hinge EPDM rubber weather sealing 0 794 cm 5 16 and 1 51 cm 19 32 wide is used to seal the door Pressure 50 clamps not shown are used to hold the door closed during testing The top of the combustion chamber contains a fume hood m to remove combustion products Figure 3 1a Picture of combustion chamber 51 Figure 3 1b Hybrid Flame Analyzer HFA combustion chamber exploded view 3 3 Exhaust system The HFA s exhaust system is shown in Fig 3 2 Combustion products are removed from the combustion chamber through a water cooled fume hood The water cooled tubing not shown consists of 0 635 cm 44 OD copper tubing wrapped around the aluminum hood Water flows at arate of 10 lpm The combustion products are pulled into exhaust tubing by a centrifugal pump To help cool the exhaust products and prevent any pressure differential in the combustion chamber excess air is pulled into the exhaust ducting through a makeup air system The makeup 52 air ducting contains an s bend to prevent hot combustion products from escaping into the laboratory The cooled combustion products are exhausted out of the lab through more 10 16 cm 4 diameter tubing
42. combustible This property accounts for this variable 21 Particle shape Quantitatively shape factors and coefficients are used as Pattern parameters in equations governed by particle shape recognition techniques 22 Particle size m Characteristic dimension of irregularly shaped particle Image Analysis representing the diameter of equivalent sphere with Microscope 23 Particle size distribution Statistical term that quantifies fluctuations in size and ASTM B761 06 shape of particles of given dust sample 24 Bulk density g cm VVeight of dust per unit volume 25 Porosity Measure of difference in densities of dust bulk and dust particle because of void spaces between particles in the bulk 26 Degree of compaction of Ratio of volume under specified pressure to volume under powder ambient pressure for a given mass of dust and configuration of dust pile 27 Moisture content in dust Weight percentage of water content in given dust sample 28 Layer thickness mm Minimum thickness of dust layer of a give particle size needed to cause a deflagration 29 Surface area volume ratio Ratio of surface area to volume of given dust particles can of dust 1 m be used to relate the arbitrary particle shapes to standard shapes like cube sphere cylinder etc 30 Suspension Ease with which particles can be suspended in air 31 Dispersibility Degree of dispersion in a dust cloud depends on ASTM E 1945 cohesiveness of particles settling velocity mo
43. dummy plots to get the legend to have data markers and fitted curve lines plot 1 1 ks MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 kv MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 gs MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 gv MarkerSize plotMarkerSize LineWidth plotLineWidth 10 1 1 5 MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 gv MarkerSize plotMarkerSize LineWidth plotLineWidth legend Nphi 0 8 106 Aphi 0 8 75 phi 1 0 106 phi 1 0 75 phi 1 2 106 phi 1 2 75 Location northwest Orientation horizontal end subplot subplot 3 2 1 Parent figure3 Y Tick 2 2 4 2 8 3 2 3 6 4 LineWidth 2 FontWeight bold FontSize 14 FontName Times New Roman subplot 3 2 1 hold on plot dust 08 75 10 dst sze burn vel 08 75 10 dst sze ks MarkerSize plotMarkerSize LineWidth plotLi ne Width plot dust conc 08 75 30 dst sze lam data dst 522 4 OO am data dst sze end burn vel 08 75 30 dst sz e orig burn vel 08 75 10 dst sze origFit ro MarkerSize plotMarkerSize LineWidth plotLineWidth plot dust 08 75 35 dst sze lam data dst sze end OO am data dst sze end burn vel 08 75 35 dst sz e orig burn vel 08 75 10 dst sze origFit sv MarkerSize plotMarkerSize LineWidth plotLineWidth plot dust conc 08 75 40 dst sze lam data ds
44. et al 76 both studied hybrid mixtures of coal dust methane air hybrid flames Liu showed that the hybrid mixture had a lower flammability limit than coal dust flames meaning that hybrid mixtures were more hazardous than a gas or dust alone Benedetto et al 75 showed that the turbulence generated by the expanding products of combustion needs to be quantified in order to determine the correct turbulent burning velocity These two studies injected coal 32 dust methane clouds into a combustion chamber ignited the clouds using electronic igniters and analyzed the clouds using either visual or Schlieren measurements recorded on a high speed camera While they noted the importance of turbulent intensity Benedetto et al 75 and Liu et al 76 were not able to quantify it Laminar hybrid flames of methane coal and air were successfully studied by Xie et al 77 78 using a Bunsen burner style burner nozzle similar to the one used for laminar flames in this study 1 4 4 Modeling of dust flames The earliest work on dust air premixed flames was reported by Nusselt 79 in 1924 who investigated coal mine explosions and focused on incorporating particle radiation in the classical gaseous premixed flame model developed by Mallard and Le Chattelier 80 in 1883 Effects of conduction devolatilization gas phase reaction and diffusion were subsequently added by several researchers with a comprehensive review by Eckhoff 26 in 2003 Noteworthy in t
45. flame interacts with an eddy S These assumptions lead to 4 5 If u ms gt gt Syz the first term under the root sign is made negligible by the second term and the turbulent burning velocity becomes independent of the laminar burning velocity This is in accordance with Damkohler s hypothesis Karlovitz et al 7 derived several expressions for the effect of large scale turbulence gt 6 on the turbulent burning velocity from the assumption that an additional velocity produced by the turbulent diffusion S has to be added to the laminar burning velocity S S S 4 6 The additional velocity was taken into account by dividing the root mean square displacement due to the turbulence by the average time interval during which a flame element interacts with an eddy be b s 4 7 r 4 7 If the turbulent flow field is characterized by the root mean sguare of the instantaneous velocity fluctuations and the autocorrelation coefficient 84 A 1 HELP sa EC 2 4 8 17 u t the length scale time scale and variance of the displacement are related through the following relationships 8 l U ms T ele 4 9 2 plzkiz 4 10 In the case of weak turbulence that is u s lt lt S the root mean square displacement within the interaction time between a flame element and a turbulent eddy becomes by integrating Eq 4 10 x
46. for the optimum concentration of the dust cloud 13 Deflagration index Ks Rate of pressure rise at maximum pressure during a dust ASTM E1226 bar m s deflagration normalized to unit volume 14 Minimum explosion Minimum concentration of a combustible dust cloud ASTM E 1515 concentration MEC sufficient to increase the pressure by 1 atmosphere 14 7 g m psi or 1 01bar due to deflagration Dust assumed to be well dispersed in air 15 Minimum ignition energy Minimum energy sufficient to ignite most easily ignitable ASTM E 2019 MIE mJ concentration of fuel in air 16 Autoignition temperature of Lowest set temperature of the surface at which dust layer ASTM E 2021 layer C on it will ignite spontaneously 17 Autoignition temperature of Minimum temperature at which a dust cloud will self ASTM E 1491 06 cloud C ignite Godbert Greenwald Furnace Test 18 Limiting oxygen Minimum oxygen concentration at the limit of ASTM E 2079 108 concentration LOC flammability for the worst case most flammable fuel concentration Physical Parameters 19 Thermal conductivity of Amount of heat transmitted through a unit thickness in a dust W m C direction normal to a surface of unit area caused due to a unit temperature gradient 20 Mass of combustible Typically a fugitive dust layer can contain inerts which are particulate solid g not
47. from the bottom of the duct To reduce the influence of the initial turbulence of flow in the duct the ignition time was delayed by 0 3 0 5 s before ignition The flame will propagate upwards in a quasi quiescent medium Dust concentration was determined by measuring the decrease of mass of dust in the movable system fluidized bed The process of 29 flame propagation was recorded by several video cameras An ion probe and thermocouple with schlieren optical system were used to examine the structure of the combustion zone and the temperature distribution simultaneously Dobashi et al 55 published results from an experiment to burn stearic acid particles in 2006 Stearic acid was heated to become liquid and sprayed through a two phase nozzle The sprays of liquefied stearic acid quickly solidified into suspended combustible particles The concentration and particle size distribution were controlled by supply pressures of liquid and air to the nozzle Ignition was started after some duration from the end of the spraying in order to sufficiently suppress the flow turbulence induced by spraying After ignition by an electric spark a flame propagated outward from the ignition point In this set up the flame propagation in an open field could be observed The propagating flame was recorded by a CCD video camera 1 4 2 Turbulent dust flame experiments 1 4 2 1 Stationary flames Turbulent gas flames have been reviewed by Bradley et al 56 Wi
48. from the rate of strain and 3 the Kolmogorov microscale which typifies the smallest dissipative eddies These length scales and the intensity can be combined to form 3 Reynolds numbers R u v Ry u 2 v and Ra u rms o li Iv with an inter relationship that can be derived 9 as R R R Similarly the length scale associated with laminar flame structures in reacting flows is the characteristic thickness of a premixed flame calculated here using 6 2a S where is the thermal diffusivity of air 10 estimated at 650 K Comparison of an appropriate chemical length with a fluid dynamic length provides a nondimensional parameter that has a bearing on the relative rate of reaction Nondimensional numbers of this type are called Damkohler numbers and given the symbol Da For large Da the chemistry is fast reaction time is short and reaction sheets of various wrinkled types may occur For small Da the chemistry is slow compared to the fluid mechanics and well stirred flames may occur Diagrams defining the regimes of premixed turbulent combustion in terms of the nondimensional groups discussed above have been proposed by several researchers cf Peters 11 Turns 10 To determine the regime in which the current experiments exist two such diagrams are examined One adapted from Turns 10 is a plot of the Da number 2 St Vs the L U ms turbulent Reynolds number ms bo based on the integral length
49. generation of wakes behind particles 3 dissipation of turbulence transfer of turbulence energy to the motion of the dispersed phase 4 modification of velocity gradients in the carrier flow field and corresponding change in turbulence generation 5 introduction of additional length scales which may influence the turbulence dissipation and 6 disturbance of flow due to particle particle interaction Considering fluid mechanics alone no combustion additional length scales may also need to be considered Some obvious examples are the diameter of the particles and the average inter particle spacing The wakes produced by particles yield a length scale on the order of the particle size If the particle size is smaller than the Kolmogorov scale the particle diameter is probably not a significant length scale affecting the dissipation If the concentration particles introduced into a flow yields an average interparticle spacing smaller than the inherent dissipation length scale the particles may interfere with existing eddies breaking them up so that the new dissipation length scale is proportional to the average interparticle spacing rather than the geometry such as the size of the perforated hole size 92 Gore and Crowe 13 have shown that a critical parameter that offers a demarcation of particle size which causes the turbulent intensity to either decrease or increase with the addition of particles in the flow is given by the ratio d l
50. had to be determined in separate tests at nominally identical dust cloud generation conditions i e rotating speed of the feeding screw conveyor and vibration mode of the dust disperser The dust concentration was measured gravimetrically A manually operated sliding tray was inserted into the tube like a gate valve By simultaneously closing the tube at the top by a conventional sliding gate valve the volume of dust cloud between the top valve and the tray was trapped Immediately before performing an 27 explosion test the dust feed was stopped and the bottom end of the tube closed by a gate valve located just below the ignition zone The ignition source was a propane flame generated by injecting a small pocket of propane air mixture into the bottom region of the explosion tube and igniting by means of an electric spark located at the tube axis By means of this apparatus a flame traveled vertically upwards away from the ignition source and could be determined as a function of the average dust concentration Proust 4 50 used a vertical square tube 10 by 10 cm with a length of 1 5 m Another apparatus with larger dimensions was also used by Proust et al 51 3 m long tube the cross section of the duct was square 0 2 m x 0 2m and over 2 m The tube was made of glass in order to obtain good conditions for visualization The suspension was generated through the elutriation of dust particles above a fluidized bed Ignition of the mixture w
51. in the burner which as discussed play a major role in the turbulent burning velocity 34 1 6 Organization of the thesis The thesis is organized into five chapters as follows Chapter 1 provides a broader background to the topic of dust deflagrations as well as a literature review related to the topic Chapter 2 analyzes the structure of a dust flame This chapter forms the body of a paper that is currently under review in the Fire Safety Journal submitted in Dec 2011 Chapter 3 is a detailed discussion of the experimental set up capable of analyzing a turbulent dust flame Chapter 4 is a discussion of the experimentally observed behavior of a turbulent hybrid flame Chapter 5 summarizes the conclusions of this study A total of five appendices Al to A5 are provided at the end of the document as supplementary material References 1 Hoekstra G Knowledge of wood dust explosions not widespread in B C industry 2012 Vancouver Sun 2 Frank W L and M L Holcomb Housekeeping Solutions 3 Eckhoff R K Dust Explosions in the Process Industries Third Edition Third ed 2003 Boston Gulf Professional Publishing 4 Proust C A Few Fundamental Aspects About Ignition and Flame Propagation in Dust Clouds Journal of Loss Prevention in the Process Industries 2006 19 p 104 120 5 Abbasi T and S A Abbasi Dust explosions cases causes consequences and control Journal of Hazardous Materials 2007 140 p 44 77
52. initial pressure Similar expressions are also used in numerical CFD codes that are used to model dust explosions 29 Modeling codes like FLACS 30 and similar modeling programs use an empirical correlation S7 F Sz U rms 1o 20 30 which correlates the turbulent burning velocity Sr as a function of the laminar burning velocity Sr turbulent intensity u ms and integral length scale Jo It should be noted that none of these are measured for dust air mixtures Additional parameters are needed to create a similar relationship for combustion including particles The effect of dust particles on the turbulent burning velocity has not been thoroughly analyzed in the literature and is the focus of this work 17 1 4 Prior related work To understand the work already published on dust deflagrations a literature review was conducted to find relevant information on dust flames turbulent gas flames experimental methods for studying burning velocity and hybrid flame experiments Excellent reviews of publications involving dust flames have been published by Robinson 31 Joshi 32 and Eckhoff 3 Much of the work reported here is gathered using their reviews as a starting point 1 4 1 Laminar dust flame experiments 1 4 1 1 Stationary flames Cassel et al 33 1949 were one of the first to publish results using an experimental burner capable of making dust air mixtures The procedure consisted of blowing gas jets onto a layer of
53. m 251 dst 75 90 um U gt 0 335 m s 0 1 0 25 g m 252 dst 75 90 um U gt 0 335 m s 0 1 0 50 g m 253 dst 75 90 um U gt 0 335 m s 0 1 0 75 g m 254 dst 75 90 um U gt 0 335 m s 0 1 2 25 g m 55 5 8 5 8 38 8 5 4 5 8 6 6 68 04 Rt St hn MA 0 es MH MI xn BR Ag 255 dst 75 90 um U gt 0 335 m s 0 1 2 50 g m 256 dst 75 90 um U 0 335 m s 0 1 2 75 g m 257 dst 75 90 um U gt k 0 532 m s 0 0 8 25 g m 258 dg 75 90 um U gt 0 532 m s 0 8 50 g m Ni 382 9 6 5 95 8 8 6 18 22 09 dd 259 dst 75 90 um U gt 0 532 m s 0 1 0 25 g m 260 dst 75 90 um U gt 0 532 m s 0 1 0 50 g m 261 dy 75 90 um U ms 0 532 m s 1 2 25 g m Sl mt re 9 4 262 dg 75 90 um U gt 0 532 m s 1 2 50 g m 263
54. methane air flames seeded with graphite The stainless steel tube of the matrix burner was of 76 mm diameter while the matrix at the end of it comprised of a disk of graphite impregnated copper This facilitated the drilling of approximately 2500 holes of Imm diameter in it with a distance between hole centers of 1 5mm The burner tube moved vertically within a copper frame tube of 254 mm diameter A smaller window enabled the flame to be observed and photographed Methane air mixtures were ignited by a retractable igniter and burned gasses were exhausted along the water cooled flame tube to atmosphere by a water cooled rotary exhauster A gate valve located in the exhaust line acted as a critical flow control and prevented back transmission of pressure pulses to the flame A large gate valve at the exhauster provided coarse control of the pressure The dried air and methane mass flow rates were metered separately by orifices and mixed in a mixing chamber Thereafter the mixture divided into two streams one of which passed through a fluidized bed to entrain the graphite An important difference from previous work arose from the necessity to operate with a higher overall mass fraction of graphite in the final mixture The gas and entrained particles passed into a top conical section and along a tube of 6 mm diameter to join the gas flow that had bypassed the bed before entering the burner tube The mass of graphite entrained was found by weighing at know
55. of the dust tube would also lead to uncertainty in the measurement Qualitative measurements of the uniformity of the dust concentration were not able to be carried out Makeup air was controlled using a rotometer with an uncertainty of 5 of full scale This could lead to slight variations in the ambient flow speed inside the combustion chamber The high speed shadowgraph images were taken with a shutter speed of 125 microseconds The flame can move some amount during this exposure time depending on the level of turbulent intensity The flame edges were selected manually using a MATLAB program The uncertainty associated with selecting the images by hand is not quantified in the current work Since the uncertainty of the experiment could not be quantitatively measured error bars are calculated as the standard deviation of the burning velocity as calculated from the maximum heights of the individual flame edges This provides a region of 95 confidence in the calculated result as shown in Chapter 4 72 3 11 Experimental matrix Table 3 3 Experimental matrix total flow rate pm gt ses es wm ERR 75 95 um Coal r 4 r 10 30 35 40 Ulm m s 0 0 185 0 335 0 532 Dust Conc C La pjsjs s 025 hiz sie 66 lsi lsi ame 1915 oa NES EDEN EE EES DEER GEEIS OE EDS sose o so 987 1 os EENDE go 29215285 oem sralvas 51 526 0218 EDS EENDE 670 982 ose t bilini ESE EES DE DER OER OE 92 865 5532 4
56. scale commonly referred to as V 88 the Borghi diagram and another adapted from Peters 11 is a plot of VS commonly L L referred to as the modified Borghi diagram Figure 4 5 a shows the functional relationship between Da and Rer and Fig 4 6 b shows the relationship between and characterizing L L the fluid mechanics of the current experimental setup 1 8 Figure 4 5 a Parameters Da vs Rer b Parameters for Borghi diagram Figures 4 6 a and b depict the characteristic parametric relationships of premixed turbulent combustion The Da number for the experiments used in this work ranged from 1 1 to 8 5 while the turbulence Reynolds number Rer ranged from 1 6 to 3 1 This range is shown in Fig 4 6 b as a red rectangle For the current set of experiments this range is hard to analyze due to the large ranges on the X and Y axis The regime of the current experiments is once again within the rectangular region shaded red and shown in Fig 4 6 b This range includes the distributed corrugated and wrinkled reaction zone 89 turbulence Weak d Reaction 10 sheets Well stirred reactor 10 Da Distributed U ns reaction zone Flamelets in S eddies Corrugated flame s 2 zone 18 NC Eg i Wrinkled flame Ss Distributed Laminar reactions flame zone zone 10 7 1 105 10 TT T T TI T 1 10 Re T ly 6 a b Figure 4
57. script 1 pixel 15 88mm 368pix 0 043 152mm pix clear all close all cle Script to operate on all files in a folder dname C Users Public Documents WPI research Hybrid Flame Analyzer HFA Matlab Codes Edge Analysis EdgeDataAll_35lpm only Default Directory To be Opened dname C Users Public Documents WPI research Hybrid Flame Analyzer HFA Matlab Codes Edge Analysis EdgeDataAll_copyNoDustFiles Default Directory To be Opened pix_to_m 15 88 368 1000 mm pix u prime all 0 0241 0 1854 0 3352 0 5323 turbulent intensity for 10 30 35 40 Ipm u_bar 1 001 3 03 3 53 4 04 flow velocity u_prime_williams 0 0 0993 0 1995 0 3289 0 4593 burning velocity williams 0 3394 0 4931 0 7844 1 096 1 2013 1 00 0 0027 0 0016 0 0014 0 0011 D_per_plate 0 001 m diameter of holes in perforate plate 10 u_prime_all u_bar 1_00 D_per_plate solution method 1 average height of flame edge 2 using plot average function 3 fitting average line to shape of cone sol_method 2 plotAll 1 value 1 will plot figures of all tests 90 Set up basic file name path to read top_file dname VI Set up main database to open and look inside s top file Is top file List Files inside main folder z 61158 15 top file Turn cells from Is function into strings cc c 3 length c Set up a matrix without the and produces by the s function S size
58. similar experimental trend was also observed recently by Xie et al 1 The effects of the dust particles on the burning velocity are likely to originate from three competing sources the energy absorption by the dust particles a decreasing effect the increase in local equivalence ratio due to fuel vapor released from the particles increasing in fuel lean but decreasing effect in fuel rich and the effect of the particles on the local fluid mechanics of the flame sheet c f Fig 4 2 which could increase or decrease Sts In the current case the laminar burning velocity shows a general decreasing trend mainly due to the heat absorption by the particles which overcomes the effect of change of effective equivalence ratio due to pyrolysis of the coal particles In the stoichiometric case 1 0 there is almost no effect as the dust particle concentration is increased the maximum variation is only 6 5 This is because the effective equivalence ratios in these cases reach slightly rich limit where the burning velocity 76 becomes maximum and this compensates for the decrease flame temperature due to heat absorption by the dust particles At higher particle size range of 106 125 microns the laminar burning velocity is almost constant with a variation of approximately 4 5 around the mean value over all the dust concentrations For the lean case 0 8 at a particle loading of 25 g m the measured burning velocity is increase
59. tabs save dname y pix save pix save ascii double tabs 70 dname S T save dname save S_T_save txt S T save l 0 save dname S T save S T save ascii double tabs end ref width check 1 ifref width check 0 scale coef load dname scale coef save end load dname x pix save load dname y pix save count 01 max 1000 count 01 lt 1 figure axis on imshovv imcmp InitialMagnification image magnification Border tight axis on while count 01 lt count 01 max if count 0151 hold on plot x pix y pix end x_pix count_01 y_pix count_01 ginput 1 Grab x and y hold on plot x_pix y_pix if count 01 gt 1 if x pix count 01 1 x pix count 01 kk pix count 01 1 y_pix count_01 break end end count 01 count_01 1 end dname image save dname save edit file name jpg saveas gcf dname image Save close all 134 save text data x pix x_pix 1 end 1 y_pix y_pix 1 end 1 if length x_pix gt 1 dist_tot a 1 x_check x_pix y_check y_pix X_pix_save 1 length x_pix a x_pix y_pix_save 1 length y_pix a y_pix save dname_x_pix_save x pix Save ascii double tabs save dname_y_pix_save y pix save ascii double tabs clear x pix y pix imcmp save matFileName mat a zatl end image image 1 end 135 A3 2 Edge data analysis
60. the current 5 33 1 6 Organization of the thesis 35 References LS UA NO ak 35 2 Structure of a Dust Flame Ee z eo ee EG Ee eG eae 40 2 1 Premixed or Non Premixed DEE Ee Ge a seve ek 40 22 klame SU E S aaa ba OG eee s 41 3 Experimental Apparatus Construction and Procedure se ee ee ee ee ee ed ee 49 3 SUM MAT AN ANE Ge N NA ANE Ge A 49 3 2 Combustion oka DEK RE 49 3 3 Exhaust syste Meee Se DR A sean ee oak ka head 52 3 4 Bufrner test Se HOR EE se EE gees Ese ee cole OR Eed ts 53 3 5 Burner nozzle design 55 3 6 Fuel control Systems satin en ced puede ETA 62 5 7 Optical EG oL EE Ge 65 3 8 Directions for using HFA ccccccccssssssssnsccecesecsesssstscsececeesesssesanseeeceeeesesssanaesececessessassaneeees 67 3 9 HEA data analysis EE RR AE HO R N 67 3 11 Experimenta lit atrix iss din ET EG OW GE ints EE 73 Refere A a 73 4 Results atid Analysis sesse es see 76 AA Laminar tla MESO nena EE GOO al aa de 76 4 2 Turbulentflames sies ree i ee ed 80 42 1 Gas flames Validation Stud dr OE W OG NE Eg ER GE 80 4 2 2 Turbulent combustion regimes 87 4 2 4 Effect of dust concentration on burning
61. the pulverized material This material was continuously agitated by magnetically vibrating an iron diaphragm which forms the bottom of the container The particles were carried away by the gas current into a vertical pipe whose upper end is connected to a vertical glass tube which serves as the burner tube The dust receptacle was a brass cylinder 15 24 cm diameter and 10 16 cm high The pipe extends into the container to a distance of 226 06 cm from the diaphragm Two gas jet orifices on opposite sides entering the receptacle 2 54 cm above the bottom were directed tangentially and turned downward at an angle of 45 deg To obtain variations of the dust concentration at constant rate of flow a valve controlled bypass was provided between the top of the container and the outlet of the pipe so that the gas entraining the dust could be diminished while the rising cloud was diluted with practically dust free gas The apparatus could run tests over a period of 10 minutes without refueling To ensure fully developed flow at the burner port a length of 3 feet was used for the 2 54 cm glass tubes The feeding mechanism was calibrated by weighing filtered samples from a constant volume of dust laden gas aspirated from the emerging cloud 18 Ghosh et al 34 published results from two experiments for studying dust flames in 1957 one inside a furnace and one in open air The apparatus used for studying pulverized coal flames inside a furnace consisted of a
62. to propose a theory that covered a range of wrinkled and severely wrinkled flames which is discussed next It should be noted that although a laminar burning velocity Sr is a physiochemical and chemical kinetic property of the unburned mixture a turbulent burning velocity 57 is in reality a mass consumption rate per unit area divided by the unburned gas mixture density Thus S7 must depend on the properties of the turbulent field in which it exists a b c d U ms 0 024 m s Uims70 185 m s U ms 0 335 m s U ms 0 532 m s Figure 4 4 Flame images at various turbulent intensities CH air gas only 0 1 To further analyze the problem a theoretical treatment similar to Dahoe 4 is utilized on the current experimental data In the case of large scale low intensity turbulence the instantaneous flame front will be wrinkled while the transport properties remain the same The wrinkles increase the flame front area per unit cross section of the turbulent flame brush which results in a higher propagation velocity without a change in the instantaneous local flame structure itself The instantaneous flame surfaces in such a turbulent flame are known as laminar flamelets With this picture in mind Damkohler 5 and Schelkin 6 derived the earliest models for the turbulent 82 burning velocity Both researchers equated the mass flux m through the cross sectional area of the flame brush Ar to the mass flow of the unburnt mixtu
63. u primeDivS L An plot u primeDivS L S T eq 148 k LineWidth plotLine Width C 22 n 20 u_primeDivS_L u_prime_smooth burn_vel_08_75_50_orig 1 S T eg 148 1 C u primeDivS L 4n plot u primeDivS L S T eq 148 k LineWidth plotLineWidth axis 0 15 0 55 0 65 131 legend 08 75 00 08 75 25 08 75 50 08 75 75 08 106 00 08 106 25 08 106 50 08 106 75 40 75 00 10 75 25 10 73 30 10 73 79 2 10 106 00 10 106 25 10 106 50 10 106 75 2 75 00 12 75 25 1257350 12 73 Ve 12 106 00 12 106 25 12 106 50 12 106 75 legend Wambda_fst 0 Wambda_fst 25 Wambda_fst 50 Wambda_fst 75 Mambda_fst 0 Uambda_fst 25 Wambda_fst 50 Mambda st lt 75 lambda st z0 Mambda stj lt 25 Mambda stJ 50 Vambda st lt 75 Mambda_fst 0 UWambda_fst 25 Wambda_fst 50 Uambda_fst 75 Mambda_fst 0 Wambda_fst 25 Wambda_fst 50 Mambda st lt 75 lambda stj lt 0 Mambda st lt 25 Mambda stj lt 50 Mambda st lt 75 legend Wambda 0 Wambda 25 Wambda 50 Wambda 75 Mambda 0 Wambda 25 Wambda 50 Wambda 75 Mambda 0 Wambda 25 WUambda 50 Wambda 75 Mambda 0 Wambda 25 Wambda 50 Wambda 75 Mambda 0 Wambda 25 Mlambdas50 Wambda 75 Mambda 0 Wambda 25 Uambda 50 Wambda 75 p 1 1 1000 450 1600 270 1000 6501 p 1 1 1000 450 10 270 1000 6501 set g
64. velocity Positions 1 3 and 6 as shown in Fig 3 6a as the notches in the side of the slit the first notch is covered by water cooling tubing and is not used are located 10 15 and 30 mm below the nozzle exit respectively The red dots in Fig 3 10 indicate the perforated plate location and flow velocity range used in the current study This set of conditions was chosen because it matched the turbulent intensities used by Kobayashi et al 5 and further when the 1mm perforated plate was raised to position 3 the flame flashed back inside of the burner Due to time constraints all of 58 the possible perforated plate and flow rate combinations were not tried with a flame to determine which combinations had a stable condition Figure 3 5 Diagram of turbulent burner nozzle 39 Figure 3 6 Turbulent burner parts a side view of turbulent burner without pilot gas fitting b pilot flame gas fitting c pilot flame spacing insert Figure 3 7 Image of premixed methane oxygen pilot flame Figure 3 8 Images of perforated plates 60 5 4 5 4 235 y 00021567 98735 E 3 5 2 5 S 3 1 5 A 1 0 5 1 05 1 1 1 15 1 2 1 25 1 3 Voltage V Figure 3 9 Calibration curve for hot wire anemometer 1 60 1mm pos 1 1 40 1mm pos3 1mm pos 6 2mm pos 1 1 20 2mmpos3 2mm pos 6 2 5099 6053 1 00 Bam poes z
65. 0 4n_12_00 plot 1 12 S TL 12 00 k LineVVidth plotLineVVidth hold off subplot 3 2 6 hold on 1010 0 12 burn vel 12 106 00 ks LineWidth plotLineWidth plot l 0 12 burn vel 12 106 25 ro LineWidth plotLineWidth plot l 0 12 burn vel 12 106 50 sv LineWidth plotLineWidth plot l 0 12 1 length burn vel 12 106 75 burn vel 12 106 75 bh LineWidth plotLineWidth hold off p get 0 monitorpositions p 1 1 1000 450 1600 270 1000 950 set gcf position p 2 163 plot xfit_data yfit_12_75_00 k Line Width plotLineWidth plot xfit data yfit 12 75 25 yfit 12 75 00 r LineWidth plotLineWidth plot xfit data yfit 12 75 50 yfit 12 75 00 g LineVVidth plotLineVVidth plot xfit data yfit 12 75 75 yfit 12 75 00 b LineWidth plotLineWidth plot xfit data yfit 12 106 25 yfit 12 75 00 r LineWidth plotLineWidth plot xfit data yfit 12 106 50 yfit 12 75 00 g LineWidth plotLineWidth plot xfit data yfit 12 106 75 yfit 12 75 00 b LineWidth plotLineWidth hold off xlabel u u bar O D pp FontSize textSize ylabel S 5 L S T S L gas only phi 1 2 FontSize textSize axis 0 09 0 15 1 1 251 164 A3 6 Plotting figure 4 15 plot ND SLdivSLgasOnlyv01 plot ND SLdivSLgasOnlyv02 6fig plot ND SLv03 6fig legend plot 0 plotMarkerSize 10 plotLineWidth 3 testSize2 14 phi
66. 0 75 35 dst sze burn vel 75 40 dst sze orig burn vel 10 75 40 dst sze burn vel 106 10 dst sze orig burn vel 10 106 10 dst sze burn vel 106 30 dst sze orig burn vel 10 106 30 dst sze burn vel 106 35 dst sze orig burn vel 10 106 35 dst sze burn vel 106 40 dst 526 orig burn vel 10 106 40 dst sze burn vel 75 10 dst sze orig burn vel 12 75 10 dst sze burn vel 75 30 dst sze orig burn vel 12 75 30 dst sze burn vel 75 35 dst sze orig burn vel 12 75 35 dst sze burn vel 75 40 dst sze orig burn vel 12 75 40 45 sze burn vel 106 10 dst sze orig burn vel 12 106 10 dst sze burn vel 106 30 dst sze orig burn vel 12 106 30 dst sze burn vel 106 35 dst sze orig lt burn vel 12 106 35 dst sze burn vel 106 40 dst sze orig burn vel 12 106 40 dst sze burn vel 75 10 dst burn vel 75 30 dst burn vel 75 30 dst sze burn vel 08 75 10 dst sze l end burn vel 08 75 10 dst sze orig 1 sze sze lam data dst sze end burn vel 08 75 10 dst sze orig 2 burn vel 75 35 dst burn vel 75 35 dst 526 sze lam data dst sze end burn vel 08 75 10 dst sze orig 3 burn vel 75 40 dst burn vel 75 40 dst sze sze lam data dst sze end
67. 0 8 figurel figure Name NDim turbulent velocity axesl axes Parent figure1 LineWidth 2 FontWeight bold FontSize 22 FontName Times New Roman hold on dummy plots to get the legend to have data markers and fitted curve lines if legend_plot 1 plot 1 1 rs MarkerSize plotMarkerSize Line Width plotLineWidth 10 1 1 55 Marker Size plotMarkerSize Line Width plotLineWidth plot 1 1 bs Marker Size plotMarkerSize Line Width plotLineWidth plot 1 1 rv MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 gv MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 bv MarkerSize plotMarkerSize LineW idth plotLineWidth legend d 75 d_ st 25 d 75 d_ st 50 d 75 d_ st 75 d 106 d_ st 25 d 106 d_ st 50 d 106 d_ st 75 Location eastoutside end hold on plot u prime all burn vel 08 75 00 orig 1 burn vel 08 75 00 orig burn vel 08 75 00 orig 1 ks LineWidt h plotLineWidth plot u prime all burn vel 08 75 25 orig 1 burn vel 08 75 25 orig burn vel 08 75 25 orig 1 rs LineWidt h plotLineWidth plot u prime all burn vel 08 75 50 orig 1 burn vel 08 75 50 orig burn vel 08 75 50 orig 1 gs LineWidt h plotLineWidth plot u prime all 1 length burn vel 08 75 75 orig burn vel 08 75 75 orig 1 burn vel 08 75 75 orig burn vel 08 75 75 orig 1 bs LineWidth plotLineW idth plot
68. 00 4 Burning_velocity a stanDev BV 08 75 00 4 standard deviation BV a end if dust conc a lt lt 25 burn vel 08 75 25 4 lt Burning velocity a stanDev BV 08 75 25 4 standard deviation BV a end if dust conc a lt lt 50 burn vel 08 75 50 4 lt Burning velocity a stanDev BV 08 75 50 4 standard deviation BV a end if dust conc a lt lt 75 burn vel 08 75 75 4 lt Burning velocity a stanDev BV 08 75 75 4 standard deviation BV a end end end if phi a lt lt 1 0 if particleSize 106 amp amp flowRate a 10 if dust conc a 0 burn_vel_10_106_00 1 Burning_velocity a stanDev_BV_10_106_00 1 standard deviation BV a end if dust conc a 25 burn vel 10 106 25 1 Burning_velocity a stanDev BV 10 106 25 1 standard deviation BV a end if dust conc a 50 burn vel 10 106 50 1 Burning_velocity a stanDev BV 10 106 50 1 standard deviation BV a end if dust conc a 75 burn vel 10 106 75 1 Burning_velocity a stanDev BV 10 106 75 1 standard deviation BV a end end if particleSize 106 amp amp flowRate a 30 if dust_conc a burn_vel_10_106_00 2 Burning_velocity a stanDev BV 10 106 00 2 standard deviation BV a end if dust conc a lt lt 25 burn vel 10 106 25 2 Burning_velocity a 142 stanDev BV 10 106 25 2 standard deviation BV a end if dust conc a lt lt 50 burn vel 10 106 50 2 Burning_velocity a s
69. 06 50 orig lam data end burn vel 12 75 10 dst sze origFit 3 sv LineWidth plotLineW idth plot l 0 12 75 1 length burn vel 12 106 75 burn vel 12 106 75 orig lam data end burn vel 12 75 10 dst _sze_origFit 4 bh Line Width plotLineWidth ylim y_axisMin y_axisMax hold off p 1 1 1000 450 100 120 850 2750 set gcf position p 2 dummy plots to get the legend to have data markers and fitted curve lines if legend_plot plot 1 1 rs MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 gs MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 bs MarkerSize plotMarkerSize Line Width plotLineVVidth plot 1 1 rv MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 gv MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 bv MarkerSize plotMarkerSize LineWidth plotLineWidth legend d 75 d_ st 25 d 75 d_ st 50 d 75 d_ st 75 155 d 106 d_ st 25 d 106 d_ st 50 d 106 d_ st 75 Location castoutside end 156 A3 4 Plotting figure 4 13 plot_legend 0 plotMarkerSize 10 plotLineWidth 3 testSize2 18 y_axisMin 1 9 y_axisMax 4 particleSize 75 106 figure3 figure Name Dust concentration plotting laminar data axes2 axes Parent figure3 YMinorTick on XTick 0 25 50 75 XMinorTick on FontSize testSize2 hold on if plot_legend
70. 08 75 75 stanDev stanDev stanDev stanDev stanDev BV_08_106_00 orig stanDev_BV_08_106_00 BV_08_106_25 orig stanDev BV 08 106 25 BV 08 106 50 orig lt stanDev BV 08 106 50 BV 08 106 75 orig stanDev BV 08 106 75 BV 10 75 00 orig stanDev BV 10 75 00 stanDev BV 10 75 25 orig stanDev BV 10 75 25 stanDev BV 10 75 50 orig stanDev BV 10 75 50 stanDev BV 10 75 75 orig stanDev BV 10 75 75 stanDev stanDev stanDev stanDev stanDev BV 10 106 00 orig stanDev BV 10 106 00 BV 10 106 25 orig stanDev BV 10 106 25 BV 10 106 50 BV 10 106 75 orig stanDev BV 10 106 50 orig stanDev BV 10 106 75 BV 12 75 00 orig stanDev BV 12 75 00 stanDev BV 12 75 25 orig stanDev BV 12 75 25 stanDev BV 12 75 50 orig stanDev BV 12 75 50 stanDev BV 12 75 75 orig stanDev BV 12 75 75 151 stanDev BV 12 106 00 orig stanDev BV 12 106 00 stanDev BV 12 106 25 orig stanDev BV 12 106 25 stanDev BV 12 106 50 orig stanDev BV 12 106 50 stanDev BV 12 106 75 orig stanDev BV 12 106 75 close all plotting functions plot ND SLv01 plot ND SL dst sev0l plot ND SLdivSLgasOnlyv01 plot ND SL dst sev02 lamOnly plot williams data v0l plot SL 02 lamOnly plot ND SL dst sev02 lamOnly plot ND 51 03 6 plot ND SLdivSLgasOnlyv04 6fig plot ND SL dst sev03 turbOnly 152 A3 3 Plot
71. 1 x_curve end num_pt_ave end 138 y_curveSmooth 1 num_pt_ave y_curve 1 num_pt_ave y_curveSmooth num_pt_ave 1 length CCC_N1 CCC_N2 num_pt_ave 1 end y_curveSmooth length CCC_N1 1 length CCC_N1 num_pt_ave 1 y_curve end num_pt_ave end plot x curveSmooth y curveSmooth LineWidth 4 axis 0 1000 0 10001 hold off if a pause 0 5 end end height smooth a max y curveSmooth min y curveSmooth pix to m height ave clicked lines a mean height dist tot median a median dist tot nozzleDiameter 0 0145 m nozzleArea pi 4 nozzleDiameter 2 vel flow a flowRate a 60000 nozzleArea m s velocity based on flow rate in tube volFlow a flowRate a 60000 flame_area a dist tot median a 2 pi 2 nozzleDiameter Burning velocity ave length a volFlow a flame area a height ave a height smooth a Burning velocity ave height a vel_flow a sin atan 0 5 nozzleDiameter height_ave a Burning velocity all height vel_flow a sin atan 0 5 nozzleDiameter height standard deviation BV a std Burning velocity all height Burning velocity a Burning velocity ave height a 7000 Vo Go Go Fo Yo Yo Vo Fo Fo Yo Yo Go Go Go Go Fo Fo Yo Go Go Fo Fo o Yo Go Go Fo Fo Yo Go Go Go Po o particleSizeSave a lt particleSize clear x curveSmooth y curveSmooth CCC NI CCC N2 CCC1 CCC2 x y x plot nz y plot nz clear avgH avgData x curve y curve x plot nz y plot nz clear x
72. 3mm pos 6 EP ad Amm pos 3 0 80 4mm pos 3 imm pos6 A 0 60 0 40 0 20 7 0 00 10 15 20 25 30 35 40 45 Flow Rate LPM Figure 3 10 Turbulent intensity versus flow rate The integral length scale of the turbulence can be calculated using 14 i 3 3 where iz is the average flow velocity and p 7 is the autocorrelation of the velocity fluctuation 61 Table 3 1 shows the flow rate flow velocity turbulent intensity integral length scale and Re using the nozzle diameter as the characteristic diameter values for tests performed in this work using a 1 mm perforated plate for the turbulent flow regime Table 3 1 Integral length scale calculations Flow Flow rate Flow velocity Wms VOU rms lo Re regime pm m s m s mm Turbulent 30 3 03 0 185 6 1 1 6 2802 35 3 53 0 335 9 5 1 4 3264 40 4 04 0 532 13 2 1 1 3736 Laminar 10 1 00 0 024 0 024 2 07 926 Figure 3 11 shows a calculation of the turbulent intensity versus the number of samples used Due to the consistent nature of the result 100 000 samples were used in the results shown Turbulent intensity m s o o o o o o N w o 0 50000 100000 150000 200000 Number of samples 250000 300000 Figure 3 11 Comparison of calculated turbulent intensity versus number of samples used in the calculation 3 6 Fuel control system The g
73. 5 Startup New Database put in name 129 When asked to reate new project click yes gt put in name When asked to configure system now click yes gt click 1 D probe gt pick type of probe 55P11 gt 1Dsupporters gt 1D straight short 5SH20 gt cable 4 meter gt click save When asked would you like to setup hardware click yes I ve been leaving the default I haven t been doing a calibration but maybe I should To collect anemometer data using the MiniCTA At top of the window click Run run default setup type in name click on setup click on A D box gt under measurement set Sampling frequency and Number of Samples 100000 kHz 300000 samples Click ok for A D setup Click ok for define default setup Click run to start collecting data Data file will show up in Database window To view data Double click file name in database window In raw data selection click load gt close window Data will load in new window To export data load data Click file gt export gt set name amp location gt set type to tab delimited I think gt click save you have to do it for each file individually as far as I can tell 130 The hot wire is mounted above the burner exit using a ring stand C clamped to the experiment frame see video if I forget to include a picture here Turbulent intensity calculations are done using a the matlab code below 131 Appendix 3 Matlab scripts used in data analysis
74. 6 Characteristic parametric relationships of premixed turbulent combustion a diagram recreated from Turns 10 b modified Borghi diagram recreated from Dahoe 4 also shown in Peters 11 Region of testing in this work is shown by area shaded in red An illustrative sketch of the turbulent flame structure in these regimes is shown in Fig 4 7 a c 4 For the relatively low levels of turbulence created in this work the testing mostly existed in the laminar flamelet regime where the macro structure is not rapid enough to destroy the laminar flame structure to such a degree that the laminar burning velocity becomes an irrelevant parameter and the chemistry is so fast that every change in the flame shape due to the large eddies is being reflected in the turbulent burning rate as the flame propagates normal to itself This flamelet regime is divided into the wrinkled and corrugated sub sections Fig 4 8 a and b If the turbulent intensity is less than the laminar burning velocity and assuming that the turbulent intensity is the rotation speed of the largest eddies then the eddies cannot fold the 90 flame The turbulence only wrinkles the flame front and the turbulent burning velocity is largely determined by the laminar flame propagation b Corrugated Flame 1 2 3 c Distributed reaction zone d Well stirred reactor Figure 4 7 Diagrams of turbulent flame structure 1 burned mixture 2 reaction zone 3 unburned mixture 4
75. 6 DE EE 106 125 um Coal ua af MENEER r r 5 ka an 50 75 123 208 293 27 68 27 15 26 64 TADA 2 325 0 293 548 803 32 29 2 851 0 293 548 803 31 67 3 357 0 293 548 803 31 08 2 712 0 336 633 930 3 326 01336 633 930 3917 0 336 633 930 2 009 99879 solan risi DIEN Table 3 3 shows the test matrix of experiments conducted in this study A total of 92 tests were performed The numbers highlighted in grey are the dust feeder settings for the prescribed dust concentration based on the feeder calibration curve and represent individual tests size of the coal was determined using a sieve shaker bituminous with approximately 30 volatiles References The particle The fuel dust is Pittsburgh seam coal 1 Bradley D Z Chen S El Sherif S E D Habik and G John Structure of Laminar Premixed Carbon Methane Air Flames and Ultrafine Coal Combustion Combustion and Flame 1994 96 p 80 96 Andrews G E D Bradley and S B Lwakabamba Turbulence and Turbulent Flame Propagation A Critical Appraisal Combust Flame 1975 24 p 285 304 3 Lewis B and G V Elbe Stability and Structure of Burner Flames Journal of Chemical Physics 1943 11 p 75 97 73 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Goroshin S I Fomenko and J H S Lee Burning Velocities in Fuel Rich Aluminum Dust Clouds in Proc Combust
76. 75 10 dst sze origFit 4 bs LineWidth plotLineWidth plot u prime all burn vel 10 75 00 orig 1 burn vel 10 75 00 orig burn d LineWidth plotLineWidth plot u_prime_all burn_vel_10_75_25_orig 1 burn_vel_10_75_25_orig burn LineW idth plotLineWidth plot u prime all burn vel 10 75 50 orig 1 burn vel 10 75 50 orig bum d LineW idth plotLineWidth plot u prime all burn vel 10 75 75 orig l burn vel 10 75 75 orig burn d LineW idth plotLineWidth plot u prime all burn vel 12 75 00 orig 1 burn vel 12 75 00 orig burn A LineWidth plotLineWidth 160 vel_10_75_10_dst sze origFit 1 k vel 10 75 10 dst sze origFit 2 rd vel_10_75_10_dst sze origFit 3 x vel 10 75 10 dst sze origFit 4 b vel_12_75_10_dst sze origFit 1 k plot u prime all burn vel 12 75 25 orig 1 burn vel 12 75 25 orig burn vel 12 75 10 dst sze origFit 2 1 LineW idth plotLineWidth plot u prime all burn vel 12 75 50 orig 1 bum vel 12 75 50 orig burn vel 12 75 10 dst sze origFit 3 s LineWidth plotLineWidth plot u_prime_all burn_vel_12_75_75_orig 1 burn_vel_12_75_75_orig burn_vel_12_75_10_dst_sze_origFit 4 b LineWidth plotLineVVidth ylim 1 41 C 1 70 n 20 u prime smooth 0 max u prime all min u prime all 200 max u prime all u primeDivS Lu prime smooth burn vel 12 75 75 orig 1 S T eq 148 1 C
77. 8 conditions For higher dust concentrations the shadovvgraph intensity is significantly reduced as shown in Fig 4 12 see image on extreme right This is mainly because of the increased brightness of the flame due to coal dust particles burning beyond the combustion zone This reduces the contrast of the shadowgraph image and makes the selection of the flamed edge more difficult at higher dust concentrations 98 75 90um st 4 W lt o r 3 2 Z Z v Sus 2 4 2 a 0 2 4 6 8 x 10 1 0 i 3 6 Sp 32 k A x 2 6 g 5 2 4 2 0 2 4 6 8 1 2 S 4 u 0 185 mis m x 7 0 335 mis I Um Us 0 532 m s Figure 4 13 Nondimensionalized burning velocity as a function of dust concentration Figure 4 13 shows the burning velocity with respect to dust concentration while holding turbulent intensity constant For the smaller particle size d 75 9074n the increase in dust concentration tends to have a varying effect on the burning velocity of the mixture In the lean case 9 0 8 the addition of coal dust has a tendency to decrease the burning velocity slightly This may be due to an increase in the local equivalence ratio from a fuel lean to a fuel rich 99 d 106 125 7 4 3 6 3 2 go 9 9 24 2 0 2 4 6 8 x10 4 3 6 3 2 2 2 4 Z b 4 2 0 2 4 6 8 x10 4 3 6 3 2 v 25 o 8 2 0 2 4 6 8 condition at higher
78. Data d yy data id xData d else Xx data id xData d yy data id yData d end if stremp opt interpMethod hist associate points in xx with averagePoints avgPoints data id avgPoints 180 meanDelta mean diff avgPoints avgPoints avgPoints meanDelta 2 avgPoints end meanDelta 2 n binIdx histc xx avgPoints goodldx binIdx gt 0 use accumarray to get average yes if opt useRobustMean avgTmp d accumarray binIdx goodIdx yy goodIdx robustMean stdTmp d accumarray binldx goodldx yy goodldx robustS td else avgTmp d accumarray binIdx goodIdx yy goodIdx mean stdTmp d accumarray binIdx goodIdx yy goodIdx std end else make unigue xx uidx unique xx yy yy uidx remove NaNs anyNaN isnan xx isnan yy xx anyNaN yy anyNaN interpolate if length xx gt 3 amp amp length yy gt 3 avgTmp d interpl xx yy data id avgPoints opt interpMethod end end end if nLines lt lt 1 amp amp stremp opt interpMethod hist data id avgData 1 avgTmp data id avgData 2 stdTmp data id avgData 3 1 elseif opt useRobustMean amp amp nLines gt 4 data id avgData 1 data id avgData 2 iid robustMean avgTmp 2 ctMat zeros size avgTmp ctMat iid 1 data id avgData 3 sum ctMat 2 else data id avgData 1 nanmean avgTmp 2 data id avgData 2 nanstd avgTmp
79. FontSize 14 FontName Times New Roman subplot 3 2 3 hold on plot dust 10 75 10 dst sze burn vel 10 75 10 dst sze ks MarkerSize plotMarkerSize LineWidth plotLi ne Width plot dust 10 75 30 dst sze lam data dst sze end 1 OO am data dst sze end burn vel 10 75 30 dst sz e orig burn vel 10 75 10 dst sze origFit ro MarkerSize plotMarkerSize LineWidth plotLineWidth plot dust conc 10 75 35 dst sze lam data dst sze end 1 OO am data dst sze end burn vel 10 75 35 dst sz e orig burn vel 10 75 10 dst sze origFit sv MarkerSize plotMarkerSize LineW idth plotLineW idth plot dust conc 10 75 40 dst sze lam data dst sze end 1 OO am data dst sze end burn vel 10 75 40 dst sz e orig burn vel 10 75 10 dst sze origFit bh MarkerSize plotMarkerSize LineW idth plotLineW idth ylim y axisMin y axisMax hold off subplot subplot 3 2 4 Parent figure3 YTick 2 2 4 2 8 3 2 3 6 4 LineVVidth 2 FontWeight bold FontSize 14 FontName Times New Roman subplot 3 2 4 hold on plot dust 10 106 10 dst sze burn vel 10 106 10 dst sze kv MarkerSize plotMarkerSize LineWidth plot LineWidth plot dust 10 106 30 dst sze lam data dst sze end l OO am data dst sze end burn vel 10 106 30 dst sze orig burn vel 10 75 10 dst sze origFit ro MarkerSize plotMarkerSize Line Width plotLineWidth plot dust
80. Influence of Coal Dust on Premixed Turbulent Methane Air Flames Scott R Rockwell A Dissertation Submitted to the Faculty of VVorcester Polytechnic Institute In partial fulfillment of the requirements for the Degree of Doctor of Philosophy in Fire Protection Engineering July 2012 APPROVED Professor Ali S Rangwala Advisor Professor Kathy A Notarianni Head of Department Professor Simon W Evans Professor Sanjeeva Balasuriya Dr Alfonso F Ibarreta Professor Forman A Williams Table of Contents Table of Contenis ia 2 Erst of SUES boo K gustan tat EA EEN RR EE EER 5 Liso table RE ME OE RO EE EK EE ad n n ae 6 NomenclaluTe si iji RA EE EE OE lt 7 Acknowledgments RE RE RAA GE Ee ee 8 Aba bab ob aa o s 9 iese dee EE OE EE A EN A DA 10 141 General OV ENVIS Wi er es ou Ge GR a 10 1 2 Hazard assessment used in dust industry 11 1 3 Thesexplosiofspheres EE EE GEE EE EE GR EE adi dua Ge ee EE ee ee 12 1 4 Prior related work a a a O AO AR AOR O AAC 18 1 4 1 Laminar dust flame 15 18 1 4 2 Turbulent dust flame experiments ee Ee Ge Ee AE ee ee Ge ee seen AE Ge 30 1 4 3 Hybrid flame experiments 32 1 4 A Modeling of dust flames ss B AO A CA 33 1 5 Goals and objectives of
81. Line Width plotLineWidth plot 1 1 bh MarkerSize plotMarkerSize LineWidth plotLineWidth legend phi 0 8 d 75 d_st 00 phi 0 8 d 75 d_st 25 phi 0 8 d 75 d_st 50 phi 0 8 d 75 d_st 75 phi 1 0 d 75 d_st 00 phi 1 0 d 75 d_st 25 phi 1 0 d 75 d_st 50 phi 1 0 d 75 d_st 75 pici 2 de75 d 117 01 plisd 2 d 75d st 25 phiel 2 d 75 d_st 50 hied 2 d 75 0_ ste phi 0 8 d 106 d_st 00 phi 0 8 d 106 d_st 25 phi 0 8 d 106 d_st 50 phi 0 8 d 106 d_st 75 phi 1 0 d 106 d_st 00 phi 1 0 d 106 d_st 25 phi 1 0 d 106 d_st 50 phi 1 0 d 106 d_st 75 phi 1 2 d 106 d_st 00 phi 1 2 d 106 d_st 25 phi 1 2 d 106 d_st 50 phi 1 2 d 106 d_st 75 Location eastoutside Linear fit for laminar data dust_concFit 0 25 50 75 lam fit O8 polyfit dust concFit burn vel 08 75 10 dst sze orig 1 lam fit 10 polyfit dust concFit burn vel 10 75 10 dst sze orig 1 lam fit 12 polyfit dust concFit burn vel 12 75 10 dst sze orig 1 burn vel 08 75 10 dst sze origFit dust concFit lam fit 08 1 Ham fit 08 2 burn vel 10 75 10 dst sze origFit dust concFit lam fit 10 1 lam fit 102 burn vel 12 75 10 dst sze origFit dust concFit lam fit 12 1 Ham fit 12 2 normalizing data burn vel 08 75 00 orig burn vel 08 75 00 148 burn_vel_08_75_25_orig burn_vel_08_75_25 burn vel 08 75 50 orig burn vel 08 75 50 burn vel 08 75 75 orig lt burn vel 08 75 75 burn vel 08 106 00
82. Sugar Port Manufacturer Sugar dust Wentworth GA b Coal mine 1701000006 A 20061 5 dust Hu Plastic dust North Carolina USA 2003 8 Manufacturer Rubber recycling Rouse Polymerics International plant Rubber dust Inc Vicksburg MS Ford Motor Company Rouge Povverhouse Coal dust Complex Dearborn MI 1999 6 Shell mold Phenol We manufacturing formaldehyde resin 10 Over the last 20 years advances in expanding chemical metallurgical and pharmaceutical industries have given birth to a steadily increasing number of new finely divided combustible materials 3 4 In a review by Abbasi and Abbasi 5 dust deflagrations caused a total of 125 casualties and 398 injuries between 1980 and 2003 These explosions were caused by a wide range of dust particles including grain aluminum coal textile rubber tantalum resin and others A recent report from the Occupational Safety amp Health Administration OSHA 6 further investigating accidents involved with dust related deflagrations has shown that the problem is still significant From a fundamental viewpoint dust combustion is studied for three main reasons the risk of explosions and fire often caused by fugitive dust produced by industrial processes 3 propulsion such as when aluminum dust is used as a stabilizer in rocket motors 7 and energy production such as in oxy coal combustors 8 This work focuses on analyzing the risk of explosions c
83. T Hirano Mechanisms of flame propagation through combustible particle clouds Journal of Loss Prevention in the Process Industries 1996 9 3 p 225 229 70 Chen D L J H Sun O S Wang and Y Liu Combustion Behaviors and Flame Structure of Methane Coal Dust Hybrid in a Vertical Rectangle Chamber Combust Sci and Tech 2008 180 p 1518 1528 71 Amyotte P R KJ Mintz MJ Pegg Y H Sun and K I Wilkie Laboratory Investigation of the Dust Explosibility Characteristics of Three Nova Scotia Coals Journal of Loss Prevention in the Process Industries 1991 4 2 p 102 109 38 72 Bradley D G Dixon Lewis and S E D Habik Lean Flammability Limits and Laminar Burning Velocities of CH4 Air Graphite Mixtures and Fine Coal Dusts Combustion and Flame 1989 77 p 41 50 73 Ju Y and C K Law Dynamics and Extinction of Non Adiabatic Particle Laden Premixed Flames Proc Combust Inst 2000 28 p 2913 2920 74 Andac M G F N Egolfopoulos and C S Campbell Effects of Combustible Dust Clouds on the Extinction Behavior of Strained Laminar Premixed Flames in Normal Gravity Proc Combust Inst 2002 29 p 1487 1493 75 Benedetto A D A Garcia Agreda O Dufaud I Khalili R sanchirico N Cuervo L Perrin and P Russo Flame Propagation of Dust and Gas Air Mixtures in a Tube in MCS 7 The Comb Institute it 2011 Chia Laguna Cagliari Sardinia Italy 76 Liu Y J Sun and D Chen Flame Propagation in Hy
84. T Ogawa Behavior of flames propagating through lycopodium dust clouds in a vertical duct Journal of Loss Prevention in the Process Industries 2000 13 p 449 457 54 Han O S M Yashima T Matsuda H Matsui A Miyake and T Ogawa A study of flame propagation mechanisms in lycopodium dust clouds based on dust particles behavior Journal of Loss Prevention in the Process Industries 2001 14 3 p 153 160 37 55 Dobashi R and K Senda Detailed analysis of flame propagation during dust explosinos by UV band observations Journal of Loss Prevention in the Process Industries 2006 19 p 149 153 56 Bradley D M Z Hag R A Hicks T Kitagawa M Lawes C G W Sheppard and R Woolley Turbulent Burning Velocity Burning Gas Distribution and Associated Flame Surface Definition Combustion and Flame 2003 133 p 415 430 57 Williams F A An Approach to Turbulent Flame Theory Journal of Fluid Mechanics 1970 40 2 p 401 421 58 Pope S B Monte Carlo Calculations of Premixed Turbulent Flames The Combustion Institute 1981 18 p 1001 1010 59 Borghi R and D Dutoya On the Scales of the Fluctuations in Turbulent Combustion 17th Symp Int on Combustion 1979 1 p 235 244 60 Chomiak J Basic Considerations in the Turbulent Flame Propagation in Premixed Gases Prog Energy Combustion Science 1979 5 p 207 221 61 Ballal D R and A H Lefrebvre The Structure and Propagation of Turbulent Flames Pro
85. The flow through the exhaust system is 0 0178 m s 1068 Ipm Exhaust L 90 cm 4 10 cm gt Water cooled exhaust hood t Combustion chamber Figure 3 2 HFA exhaust system diagram 3 4 Burner test section To determine the best way to study hybrid flames a literature search for published methods of experimental burning velocity measurements of flames was conducted Based on this study full details are given in the literature review in Cha 1 and the critical reviews by Andrews et al 2 and Lewis and Elbe 3 the anchored Bunsen burner experimental design used in this work was chosen This style of experiment is the simplest to use and analyze and allows a turbulent flame which can be studied for an extended period of time facilitating easier instrumentation and measurement accuracy This is important because turbulent flames are inherently not steady state therefore average quantities determined about the flame should come from many 53 measurements taken over time This requires the flame to be anchored at the burner exit for several minutes Figure 3 3 shows a diagram of the hybrid flame analyzer s test section The side view and top view of the combustion chamber are shown the outline of combustion chamber a the point source of light b uses a bulb from a projector 480 watt A steel plate with a pin hole in the center is used to create the point source This point source of light is placed at the fo
86. _velocity a stanDev BV 10 75 75 3 standard deviation BV a end end if particleSize 75 amp amp flowRate a 40 if dust_conc a burn_vel_10_75_00 4 Burning_velocity a stanDev BV 10 75 00 4 standard deviation BV a end if dust conc a 25 burn vel 10 75 25 4 lt Burning velocity a stanDev BV 10 75 25 4 standard deviation BV a end 144 if dust_conc a 50 burn_vel_10_75_50 4 Burning_velocity a stanDev BV 10 75 50 4 standard deviation BV a end if dust conc a lt lt 75 burn vel 10 75 75 4 lt Burning velocity a stanDev BV 10 75 75 4 standard deviation BV a end end end if phi a lt lt 1 2 if particleSize 106 amp amp flowRate a 10 if dust conc a burn vel 12 106 00 1 Burning_velocity a stanDev BV 12 106 00 1 standard deviation BV a end if dust conc a lt lt 25 burn vel 12 106 25 1 Burning_velocity a stanDev BV 12 106 25 1 standard deviation BV a end if dust conc a 50 burn vel 12 106 50 1 Burning_velocity a stanDev BV 12 106 50 1 standard deviation BV a end if dust conc a lt lt 75 burn vel 12 106 75 1 Burning_velocity a stanDev BV 12 106 75 1 standard deviation BV a end end if particleSize 106 amp amp flowRate a 30 if dust conc a burn vel 12 106 00 2 Burning_velocity a stanDev BV 12 106 00 2 standard deviation BV a end if dust conc a lt lt 25 burn vel 12 106 25 2 Bur
87. a flame containing condense phase fuel Not only can the fuel change but for a given fuel the flame can behave differently given the amount of vaporization which takes place These figures show the increased complexity of the dust air problem and give a clear impression on the need to study this behavior 44 Representation of heat and mass transfer processes in the preheat zone Burning Vaporizing particles inthe Particles in the Vaporizing particles in preheat zone Fuel particles in the Ambient Preheat Reaction Convection zone zone zone ambient zone Figure 2 2 Schematic illustration of the structure of a premixed dust air flame 45 Reaction Zone Preheat Zone Convection Zone T lt T A m o lt 1 ae TT B 42 Nozna I de lt 1 Re TT C o gt 1 g T ET o gt 1 Y EEEE B L 1 Figure 2 3 Schematic of flame structure in a dust air flame To study the dust flame problem a hybrid flame is the optimum tool because it allows the creation of all 5 scenarios by varying the gas phase and condense phase equivalence ratios These scenarios are succinctly described in table 2 1 46 Table 2 1 Fuel concentration scenarios in hybrid flames Ambient zone Preheat zone Convection zone The dust completely vaporizes in the preheat There is excess oxygen zone The condense phase and gas phase have a lean condition
88. ame is thus coupled to a Nusselt flame 2 2 Flame structure Fundamentally flame propagation in dust flames requires three sequential processes heating and devolatilization of the particles mixing of the volatiles and ultimately combustion of the mixture 9 The last step can involve gas phase combustion of the volatiles released by the condensed fuel or surface reactions or a combination of both and is the most complicated The three processes are illustrated in Fig 2 2 where five potential scenarios based on equivalence ratios and are presented The variable represents the equivalence ratio based on the total condense phase fuel in the ambient zone whereas represents the equivalence ratio based on the volatized gas vapor evolved at the end of the preheat zone Equivalence ratio can be calculated using 10 41 N fuel MW ae n MW 2 1 N fuel MW ar MW ACIE stoichiomeric air air air where n is the number of moles and MW is the molecular weight There are five scenarios because scenario B can have 6 greater or less than 1 In Fig 2 2 label A denotes the condition where lt 1 and lt 1 Label B denotes the condition where 4 can be greater or less than 1 but is less than one Label C denotes the condition where both d and are greater than 1 In conditions C the particles completely vaporize in the preheat zone while in condition Cy the particles do not comple
89. amp amp height 3 Turb int 053 ht3 ct 05 1 u prime rms u prime max pos3 ht3 ct 05 1 u prime max 1 0 pos3 ht3 ct 05 1 0 ct 05 ct 05 1 elseif position 6 amp amp height 3 Turb_int_pos6_ht3 ct_06 1 u_prime_rms u_prime_max_pos6_ht3 ct_06 1 u_prime_max 1 8086 ht3 ct 06 1 0 ct_06 ct 06 1 end o Fo Go Fo Fo Fo Yo Vo Fo Fo Yo Yo Yo Yo Go Go Go Go Fo Yo Yo Go Go Go Fo Yo Yo Yo Go Fo Fo Go Yo Go To a zatrl end 175 A3 9 Gas analysis data retrieval Pulls in max of 3600 header lines used to get data shouldn t been collecting for more than an hour anyway clear all close all cle fileName 120411 phi 1 test matlab capture TXT filePath J VTerribite Drive Documents My Documents Folder Copied 2 WPI research Turbulent flame Hybrid Flame Analyzer HFA Gas Analyser Hyperterminal fileToRead1 filePath V fileName DELIMITER X HEADERLINES 3600 Import the file rawDatal importdata fileToRead1 DELIMITER HEADERLINES rawDatal cell cellstr rawDatal for ct 1 l length rawDatal cell dataOneLine cell2mat rawDatal cell ct 1 oxygenPercent ct 1 1 str2num dataOneLine 46 50 end 176 A3 10 plotAverage_noplot Note This is a modified code from the matlab central exchange and was not origionally written by the author function avgH avgData plotAverage handleOrData avgPoin
90. ance of the net heat transfer divided by heat of gasification These additional parameters influence the burning dynamics of particle air flames as discussed further in Fig 2 2 which shows a sketch of the flame structure for the five types of eguivalence ratios combinations considered The profiles of mass fraction of condense phase fuel Y mass fraction of vaporized fuel Yrc sm the vaporization rate the reaction rate and the temperature T across ambient preheat reaction and convection zones are shown in Fig 2 3 Case A represents the conditions where lt land all of the condense phase fuel is vaporized as shown in Fig 2 1 A When P lt 1 fuel is the limiting reactant and is completely consumed in the reaction zone Vaporization predominantly takes place in the preheat zone with the mass fraction of the condense phase particles Y dropping to zero and the mass fraction of the fuel vapor reaching a maximum in the preheat zone The temperature increases through the preheat zone attains the maximum value in the reaction zone and remains constant in the convection zone where losses can be neglected Case B represents the conditions of lt land particles continue to burn even in the convection 8 zone resulting in 7 gt T The inset labeled in Fig 2 2 shows the convection zone in case B where the fuel particles continue to burn in the presence of excess o
91. and propane went directly into the base of the burner The powder and gases were mixed in the burner and were expelled through the burner screen Typical operating procedures consisted of establishing a propane oxygen nitrogen flame introducing the powder by activating the vibrator and finally reducing the propane nitrogen and oxygen flows until the desired flame conditions were obtained Strehlow et al 41 published results from a steady state burner in 1974 The basic objective of the burner design was to obtain two relatively large area coaxial streams with flat laminar 22 velocity profiles such that the central stream could be completely surrounded by hot products from the combustion of a gaseous fuel in the outer stream The two innermost regions of the burner were fed by combustible streams The inner rectangular test stream region can be fed a mixture of fuel air and suppressant consisting of up to five different gasses and two different solid powders all independently metered The coaxial annular region directly outside of the inner stream could be fed by a fuel air mixture This outer region provided an atmosphere containing products of combustion of a non suppressed premixed laminar flame and therefore represented a continuous strong igniter for the inner test flow region The flow area outside the ignition flame could contain only air and served to shield the outer edge of the flame from external disturbances The burner was enclos
92. ant parameters quantifying the hazard associated with a particular type of dust their classification and test methods is provided in Appendix 1 Of these typically three quantities 3 the minimum ignition energy MIE 12 the minimum explosible concentration MEC 13 and the deflagration index Kg 14 are mainly used and incorporated in industrial standards For example dust hazards are ranked by the Occupational Safety and Health Administration OSHA using the dust deflagration index Kg 15 based on ASTM E 1226 16 The deflagration index is related to thermokinetic parameters governing both the flame propagation as well as pressure build up in deflagration and is measured using the explosion sphere apparatus The MIE is the minimum spark energy required to ignite a fuel mixture It is found experimentally using the Modified Hartmann Tube apparatus by creating a cloud of premixed fuel and sending a spark of known energy 1 kJ through the mixture The MEC represents the minimum amount of dust in terms of g m that can be ignited using an explosion sphere ASTM E 1226 14 EN 13673 17 1 3 The explosion sphere The explosion sphere shown in Fig 1 1 is an experimental device for measuring the deflagration index Kg discussed earlier It is based on the early experimental work by Andrews 12 et al 18 Abdel Gayed and Bradley 19 and many subsequent publications by the Leeds group cf Bradley 20 where burning velocit
93. arkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 bs MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 kv MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 rv MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 gv MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 bv MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 k4 MarkerSize plotMarkerSize LineWidth plotLineWidth 10 1 1 MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 g MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 b MarkerSize plotMarkerSize Line Width plotLine Width 10 1 1 ko MarkerSize plotMarkerSize LineW idth plotLineWidth plot 1 1 ro MarkerSize plotMarkerSize LineVVidth plotLineVVidth plot 1 1 g0 MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 bo MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 kx MarkerSize plotMarkerSize LineW idth plotLineWidth plot 1 1 rx MarkerSize plotMarkerSize LineW idth plotLineWidth plot 1 1 gx MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 bx MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 kh MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 rh MarkerSize plotMarkerSize LineWidth plotLineWidth plot 1 1 sh MarkerSize plotMarkerSize
94. as achieved using an electrically heated tungsten wire The flame propagates from the open end of the tube at the bottom up to the closed end at the top Ionization probes were used to determine the flame location thermocouples were used to measure the maximum flame temperature and collimated photodiodes to record the light emitted by the flame front Two kinds of photographic records were performed self emitted light and laser tomographic records Dust concentrations were determined by measuring the decrease of mass of the elutriator and by metering the air flow rate The laser tomographic system was used to control the homogeneity of the suspension Goroshin et al 52 published results from an experiment which consisted of two parts a dust feeder and a disperser The dust was fed via a syringe type device which had an internal diameter of 2 5 cm and a maximum piston stroke of 20 cm The rate at which the dust was supplied to the flow and hence the dust concentration in the suspension was controlled by varying the piston speed with the help of a special electromechanical system The range of the piston speed was 28 0 5 3 cm min The dust was dispersed at the base of a conical chamber through the impact of a high velocity cylindrical jet issuing from an adjustable circular slot A Pyrex flame tube 5 cm i d and 120 cm length where combustion experiments were performed was connected to the dispersion chamber through an 8 conical diffuser
95. as phase fuel equivalence ratio and flow rate is controlled using a pair of mass flow controllers 50 Ipm full scale for air and 10 Ipm full scale for methane with uncertainties of 1 of full scale The gas phase equivalence ratio ranged from 0 8 1 2 Dust particle feed rate is controlled using a volumetric screw feeder which is calibrated for different dusts and particle 62 sizes similar to the setup used by Hattori et al 15 The dust is mixed with the CH4 air in the bottom of the 50 cm vertical feeder tube ID lt 14 5 mm Figure 3 12 shows a diagram of the dust injector block used to add coal dust Pittsburgh seam c f table 3 2 for property data into the premixed fuel mixture before it travels to the nozzle exit Dust is held in an agitated hopper a and fed into the burner feed system using a 0 635 cm 4 helix b housed inside of a stainless steel tube c The dust e is fed into a wooden block h with a thin slit 0 159 cm 1 16 wide i The methane air mixture d coming up through 1 27 cm 2 copper tubing g creates a shear flow in the thin slit i entraining dust similar to the experiment used by Kolbe et al 16 This prevents clumping of the dust and helps provide continuous injecting of the dust into the fuel stream The combined dust gas f mixture exits the block through a 1 27 cm 72 copper tube g Side View Figure 3 12 Diagram of dust feeder block The dust feeder is calibrated by collectin
96. ating arrays of test data as a function of dust concentration Burn_vel_func_dst_part_szev01 Burn vel func dst szev0l if phi a 0 8 if particleSize 106 amp amp flowRate a 10 dust conc 08 106 10 dst sze ct 3 dust conc a bum vel 08 106 10 dst sze ct 3 Burning velocity a ct3 ct3 1 end if particleSize 106 amp amp flowRate a 30 dust conc 08 106 30 dst sze ct 4 dust conc a bum vel 08 106 30 dst sze ct 4 Burning velocity a ct 4 ct 4 1 end if particleSize 106 amp amp flowRate a 35 dust conc 08 106 35 dst sze ct 5 dust conc a bum vel 08 106 35 51 sze ct 5 Burning_velocity a ct 5 ct 5 1 end if particleSize 106 amp amp flowRate a 40 dust conc 08 106 40 dst sze ct 6 dust conc a burn vel 08 106 40 dst sze ct 6 Burning_velocity a ct 6 ct 6 1 end if particleSize 75 amp amp flowRate a 10 dust conc 08 75 10 dst sze ct 7 dust conc a bum vel 08 75 10 dst sze ct 7 Burning velocity a ct 7 ct 7 1 end if particleSize 75 amp amp flowRate a 30 dust conc 08 75 30 dst sze ct 8 dust conc a burn vel 08 75 30 dst sze ct 8 lt Burning velocity a ct 8 lt ct 8 11 end if particleSize 75 amp amp flowRate a 35 dust conc 08 75 35 dst sze ct 9 lt dust conc a bum vel 08 75 35 dst sze ct 9 Burning velocity a ct 9 ct 9 1 end if particleSize 75 amp amp flowRate a
97. aused by mixtures of combustible gasses and dusts hybrid fuels This type of explosion often occurs in coal mines which start with a methane air explosion and entrain coal dust as the flame propagates down the mine gallery see table 1 1 for recent explosions involving this type of dust 1 2 Hazard assessment used in dust industry Palmer 9 describes a dust explosion deflagration in a facility as a series of explosions The first primary explosion is relatively small It ruptures the initial containment releasing a pressure wave followed by a relatively slow fire front All subsequent explosions following the primary are referred to as secondary explosions Secondary explosions can result in much higher pressures resulting in significantly greater damage to both personnel and property Dust mine explosions often have a primary explosion fueled by a methane air cloud and secondary 11 explosions fueled by mixtures of methane and coal dust As discussed by Parnell 10 after the dust explosions in 2009 OSHA revised its Combustible Dust National Emphasis NEP program The NEP looked into 64 industries with over 1000 inspections and found more than 4000 combustible dust related violations This exhaustive report has led OSHA to consider new rules for facilities handling combustible dust A dust is any finely divided solid with a mean diameter less than 420 um National Fire Protection Association NFPA 68 standard 11 A list of all relev
98. bda 0 Wambda 25 WUambda 50 Wambda 75 Mambda 0 Wambda 25 Uambda 50 Wambda 75 Mambda 0 Wambda 25 Wambda 50 Wambda 75 axis 0 2 1 4 p 1 1 1000 450 1600 270 1000 650 set gcf position p 2 error autobreak plot l_00 burn_vel_08_75_00_orig ks Line Width plotLineWidth plot l_00 burn_vel_08_75_25_orig ro Line Width plotLine Width plot l_00 burn_vel_08_75_50_orig sv Line Width plotLineWidth plot l_00 1 length burn_vel_08_75_75_orig burn_vel_08_75_75_orig bh LineW idth plotLineWidth xlabel u u_fbar 1_0 D_fpp FontSize textSize ylabel S 5 L S T S L gas only 5 1 0 8 FontSize textSize errorbar 1 0 burn vel 08 75 25 burn vel 08 75 00 stanDev BV 08 75 25 burn vel 08 75 00 rs LineWidth plotLineW idth 167 1 O burn vel 08 75 50 burn vel 08 75 00 stanDev BV 08 75 50 burn vel 08 75 00 gs LineWidth plotLineWidth errorbar l_0 1 length burn_vel_08_75_75 burn_vel_08_75_75 burn_vel_08_75_00 1 length burn_vel_08_75_75 stanDev_BV_08_75_75 burn_vel_08_75_00 1 length burn_vel_08_75_75 bs LineWidth plotLineWidth errorbar 1_0 burn_vel_08_106_25 burn_vel_08_75_00 stanDev_BV_08_106_25 burn_vel_08_75_00 rv LineWid th plotLineW idth errorbar 1 O burn vel 08 106 50 burn vel 08 75 00 stanDev BV 08 106 50 burn vel 08 75 00 gv LineVVi dth plotLineWidth
99. brid Mixture of Coal Dust and Methane Journal of Loss Prevention in the Process Industries 2007 20 p 691 697 71 Xie Y Study of Interaction of Entrained Coal Dust Particles in Lean Methane Air Premixed Flames in Fire Protection Engineering 2011 Worcester Polytechnic Institute Worcester MA 78 Xie Y V Raghavan and A S Rangwala Study of interaction of entrained coal dust particles in lean methane air premixed flames Combust Flame 2012 159 p 2449 2456 79 Nusselt W Die Verbrennung und die Vergasung der Kohle auf dem Rost 1924 68 p 124 80 Mallard E and H L l Chatelier Recherches Experimentales et Theoretiques sur la Combustion des Melanges Gazeux Explosifs Annals des Mines 1883 4 81 Seshadri K A L Berlad and V Tangirla The Structure of Premixed Particle Cloud Flames Combustion and Flame 1992 89 p 333 342 82 Bidabadi M and A Rahbari Modeling Combustion of Lycopodium Particles by Considering the Temperature Difference between the Gas and the Particles Combustion Explosion and Shock Waves 2009 45 3 p 278 285 83 Smoot D and M D Horton Propagation of Laminar Pulverized Coal Air Flames Prog Energy Combustion Science 1977 3 p 235 258 84 Krazinski J L Buckius R O and K H Coal Dust Flames A Review and Development of a Model for Flame Propagation Progress in energy and Combustion Science 1979 5 p 31 71 85 Slezak S E R O Buckius and H Krier A modle o
100. burn vel 12 75 50 orig lam data end burn vel 12 75 00 orig lam data end burn vel 12 75 75 lt burn vel 12 75 75 orig lam data end burn vel 12 75 00 orig lam data end burn vel 12 106 00 lt burn vel 12 106 00 orig lam data end burn vel 12 106 00 orig lam data end burn vel 12 106 25 lt burn vel 12 106 25 orig lam data end burn vel 12 106 00 orig lam data end burn vel 12 106 50 lt burn vel 12 106 50 orig lam data end burn vel 12 106 00 orig lam data end 149 burn vel 12 106 75 burn vel 12 106 75 orig lam data end burn vel 12 106 00 orig lam data end 1 001 0 81 1 0 10 1 0 1 O lam data end burn vel 75 10 dst sze orig burn vel 08 75 10 dst sze burn vel 75 30 dst sze orig bum vel 08 75 30 dst sze burn vel 75 35 dst sze orig lt burn vel 08 75 35 dst sze burn vel 75 40 dst sze orig lt burn vel 08 75 40 dst sze burn vel 106 10 dst sze orig burn vel 08 106 10 dst sze burn vel 106 30 dst sze orig burn vel 08 106 30 dst sze burn vel 106 35 dst sze orig lt burn vel 08 106 35 dst sze burn vel 106 40 dst sze orig burn vel 08 106 40 dst sze burn vel 75 10 dst sze orig lt burn vel 10 75 10 dst sze burn vel 75 30 dst sze orig bum vel 10 75 30 dst sze burn vel 75 35 dst sze orig lt burn vel 1
101. c R Soc Lond A 1975 334 p 217 234 62 Kobayashi H T Tamura K Maruta T Niioka and F A Williams Burning Velocity of Turbulent Premixed flames in a high Pressure Environment Proc Combust Inst 1996 26 p 389 396 63 Smallwood GJ O L Gulder D R Snelling B M Deschamps and I Gokalp Characterization of Flame Front Surfaces in Turbulent Premixed Methane Air Combustion Combustion and Flame 1995 101 p 461 470 64 Filatyev S A J F Driscoll C D Carter and J m Donbar The Study of the Turbulent Burning Velocity by Imaging the Wrinkled Flame Surface in 40th Aerospace Sciences Meeting amp Exhibit 2002 Reno NV 65 Hertzberg M K L Cashdollar and C P Lazzara The Limits of Flammability of Pulverized Coals and other dusts in Proc combust Inst 1981 The combustion Institute 66 Li Y C W Kauffman and M Sichel An Experimental Study of Deflagration to Detonation Transition Supported by Dust Layers Combustion and Flame 1995 100 p 505 515 67 Sun J R Dobashi and T Hirano Temperature profile across the combustion zone propagating through an iron particle cloud Journal of Loss Prevention in the Process Industries 2001 14 p 463 467 68 Ju W R Dobashi and T Hirano Dependence of flammability limits of a combustible particle cloud on particle diameter distribution Journal of Loss Prevention in the Process Industries 1998 11 p 177 185 69 Chen J L R Dobashi and
102. cal explosion vessels Components Operating Conditions Standard 20 L vessel 36 L vessel Reservoir volume 0 6 1 LL Initial pressure in vessel 0 6 barg 0 3 barg Fast acting valve time 45 ms 50 ms Pressure at time of ignition 0 barg 0 barg Ignition delay time 60 ms 75 ms Viewing window Pressure transducer Electro chemical ignition Dispersion nozzle Dust reservoir Dispersion air Time Figure 1 2 Pressure versus time curve and change in pressure versus time curve from an explosion sphere 13 14 The deflagration index is determined by an explosion sphere apparatus based on the maximum rate of pressure rise and the volume of the explosion sphere 25 given by the cube root law PH v K Z v 1 1 where Vo is the volume of the explosion sphere and dP dt is the change in pressure over time Tt has been shown Eckhoff 26 that the deflagration index changes as the size of the explosion sphere changes This makes the investigation of a dust flame rather difficult and also complicates the hazard classification as the quantity used to characterize the hazard is now dependent on the experimental apparatus The problem arises mainly due to the increase in turbulent intensity caused by the expanding combustion products in a constant volume vessel Figure 1 3 shows an illustrative sketch of an expanding flame front at four different times inside of a typical explosion spher
103. cal point of a bi convex lens b with a 100 mm diameter and a 200 mm focal length This creates a 100 mm diameter test section of parallel light d inside the combustion chamber The parallel light passes through the flame 1 and through a second identical bi convex lens which reduces the diameter of the image This reduction makes the image small enough to fit on the sensor of a digital single reflective lens camera with a 1 1 macro lens f with the focus set to infinity To reduce the intensity of the coal dust emissions a short pass filter e with a cutoff of 550 nm is placed in front of the camera lens similar to the experiment by Goroshin et al 4 The flame 1 is fueled from a methane source h an air source i and a dust hopper j The dust is injected into the fuel air mixture using the injector block k as described in detail in Fig 3 11 Known turbulent intensities are created using a set of perforated plates as described below Makeup air is injected into the combustion chamber through the 4 fitting o and distributed in the subsection of the combustion chamber g Combustion products are removed from the water cooled n 12 cm diameter exhaust duct A removable spark igniter s is used to ignite the pilot flame similar to the experiment used by Bradley et al 1 54 Top view Figure 3 3 Diagram of experimental section of Hybrid Flame Analyzer HFA 3 5 Burner nozzle design At the top of the vertical feede
104. cc 9 Find the size of matrix containing names of files inside of main database a 1 This counter is set to 3 to account for the and at the beggining of each matrix created by Is ct_3 1 ct_4 1 5 1 ct_6 1 ct_7 1 8 1 ct_9 1 ct_10 1 ct_11 1 ct_12 ct 13 1 ct_14 1 ct_15 l ct 16 l ct 17 1 ct_18 1 ct_19 1 ct_20 ct 21 1 ct_22 l ct 23 1 1 24 1 25 1 ct_26 I while a lt S 1 close all file char cellstr top file char cc a File to be operated on data_n char cc a file_name char cc a if str2num file name 1 2 75 flowRate a str2num file name 56 57 dust conc a str2num file name 45 47 phi a str2num file_name 32 34 particleSize 75 end if str2num file_name 1 3 106 136 flowRate a str2num file name 58 59 dust_conc a str2num file_name 47 49 phi a str2num file_name 34 36 particleSize 106 end if flowRate a 10 u_prime a u_prime_all 1 elseif flowRate a 30 u_prime a u prime all 2 elseif flowRate a 35 u_prime a u_prime_all 3 elseif flowRate a 40 u_prime a u prime all 4 end fileNameLoad top_file file name load fileNameLoad mat x_pix_save load fileNameLoad mat pix save l X_non_zero nonzeros x_pix_save x plot x pix save median x_non_zero y_plot max max y pix save y pix save SS size x_plot if
105. cf position p 2 error autobreak 161 plot l 00 burn vel 08 75 00 orig ks LineWidth plotLineWidth plot l 00 burn vel 08 75 25 orig ro Line Width plotLineVVidth plot l 00 burn vel 08 75 50 orig sv Line Width plotLineWidth plot l 00 1 length burn vel 08 75 75 orig burn vel 08 75 75 orig bh LineWidth plotLineWidth xlabel u u bar O D pp FontSize textSize ylabel S 5 LYS T S L gas only phi 0 8 FontSize textSize errorbar 1 O burn vel 08 75 25 burn vel 08 75 00 stanDev BV 08 75 25 burn vel 08 75 00 rs LineWidth plotLineW idth errorbar 1 O burn vel 08 75 50 burn vel 08 75 00 stanDev BV 08 75 50 burn vel 08 75 00 gs LineVVidth plotLineWidth errorbar l_0 1 length burn_vel_08_75_75 burn_vel_08_75_75 burn_vel_08_75_00 1 length burn_vel_08_75_75 stanDev_BV_08_75_75 burn_vel_08_75_00 1 length burn_vel_08_75_75 bs LineWidth plotLineWidth errorbar 1_0 burn_vel_08_106_25 burn_vel_08_75_00 stanDev_BV_08_106_25 burn_vel_08_75_00 rv LineWid th plotLine Width errorbar 1_0 burn_vel_08_106_50 burn_vel_08_75_00 stanDev_BV_08_106_50 burn_vel_08_75_00 gv LineWi dth plotLineVVidth 1 O I length burn vel 08 106 75 burn vel 08 106 75 burn vel 08 75 00 I length burn vel 08 75 75 stanDev BV 08 106 75 burn vel 08 75 00 1 1ength burn vel 08 106 75 bv LineWidth plotLineWidth gt
106. chamber Just after the movable tube had moved down to its bottom position the suspended iron dust was ignited by an electric spark A flame then started to propagate throughout the iron particle cloud Ju et al and Chen et al 68 69 published results from a constant pressure flash fire burner The system consists of an atomizing nozzle cylindrical ducts and electric heaters To minimize the influence of air flow on cloud behavior a piece of aluminum plate was placed closely around the nozzle In the experiments the fuel in a reservoir was heated to become liquid just above its melting point and sprayed by the nozzle The liquid droplets turned into solid particles during their rise along to the test section The distribution of particles diameters was controlled by changing the pressure of the feeding air and fuel To avoid influences of turbulence caused by fuel spraying on the combustion phenomena the ignition time was delayed by 0 5 s after the end of fuel spraying The particle cloud was ignited at its centre by an electric spark Just before the particle cloud ignited the middle part of the duct was moved down Thus the combustion of the particle cloud could be kept free from the influence of the wall 1 4 3 Hybrid flame experiments Hybrid flames have been studied by a number of researchers Chen et al 70 Amyotte et al 71 Bradley et al 72 Ju et al 73 Andac et al 74 Relevant to this work Benedetto et al 75 and Liu
107. circulatory system for producing a coal dust suspension and an electrically heated furnace within which combustion took place with the formation of a flame The circulatory system consisted of a blower the inlet and outlet of which were connected by a loop Coal dust and air were circulated through this loop and the suspension produced was fed into the burner through an outlet tube attached vertically axially to the elbow of the descending limb of the loop A vibrator was placed against the descending limb of the apparatus in order to minimize settling of coal dust on the tube walls The burner tube as a vertical water cooled copper tube 5mm ID connected to an outlet tube of the circulatory systems by means of a short piece of rubber tubing The tip of the burner tube projected one inch from the cooling jacket When placed in the operating position the tip was flushed with the ceiling of the furnace cavity A mirror allowed observation of the flame from the bottom The circulatory system was air tight therefore the rate of flow suspension was obtained from the rate at which air was introduced into the system The rate of coal flow was determined by removing the burner tube collecting the coal flowing through the coal outlet of the loop for one minute and weighing The apparatus used for studying flames in open air consisted of a blower for producing coal dust suspensions Pulverized coal was kept in an inclined conical flask and was introduced into t
108. city significantly compared to larger particle sizes and lower concentration ranges The experimental data is used to develop a correlation similar to turbulent gas flames to facilitate modeling of the complex behavior 1 Introduction 1 1 General overview The hazards of dust combustion are often overlooked in industrial safety In industries that manufacture transport process or use combustible dusts accidental dust deflagrations represent a real hazard to both personnel and equipment Dust explosions cause injuries fatalities and significant financial cost The insurance company FM Global reported that between 1983 and 2006 the cost of 166 manufacturing plant dust explosions were 284 million 1 The Ford motor Company Power house explosion in 1999 caused over 1 billion in damage 2 Table 1 1 lists a few of the most recent industrial explosions caused by dust and hybrid fuels Table 1 1 Recent incidents of industrial dust or hybrid flame explosions 2 Industry Type Fuel Location Date Fatalities Saw mill Wood dust O ALA 2012 2 Prince George Canada Babine Forest Products in Saw mill Wood dust Burns Lake Canada 2012 2 5 andes Hoeganaes Corporation TN 2011 3 Manufacturer USA se incidents Methane amp coal Upper Branch mine West Coal mine dust Virginia USA 2010 29 Coal mine po Pike River Nevv Zealand 2010 29 Sugar Imperial
109. clature A pre exponential factor Ar cross sectional area of flame brush AL vvrinkled laminar flame area d diameter of burner ds dust diameter Da Damkohler number E Activation energy h height of flame cone k thermal conductivity K deflagration index lo integral length scale m mass flux MW molecular weight n number of moles Eq 1 3 n number of samples Eq 3 2 P pressure Q Heat of combustion R Radius of cone base Rer Turbulent Reynolds number SL laminar burning velocity St turbulent burning velocity t time u velocity u velocity fluctuation u average velocity U ims turbulent intensity Vo volume X displacement Greek half angle of flame cone Eq 3 5 thermal diffusivity Sec 4 2 2 L laminar flame thickness density 2 r autocorrelation of velocity Ast dust concentration equivalence ratio b time flame element interacts vvith eddy Subscripts 8 gas max rms st maximum value ambient parameter root mean square dust Acknowledgments The author would like to thank the National Science Foundation Graduate Research Fellowship program and the Koerner Fellowship program for funding this work The author would also like to thank Dr K A Joshi for his counsel the faculty of the WPI FPE program for their guidance and his parents for their support over the years Abstract The hazard associated with dust deflagrations has increased over the last decade indust
110. concentrations and turbulent intensities In fact it can be noted that at the highest dust concentration As 75 g m and turbulent intensities the burning velocity for all three equivalence ratios is approximately constant 3 2 A similar trend is also observed with larger particle size range used ds 106 120 um This result shows that it may be possible that at sufficiently high turbulent intensities the burning behavior becomes independent of the chemistry of the gas flame but is controlled only by the size and concentration of the dust particles in the flame The smaller the particles the higher the burning velocity For the stoichiometric case 1 0 the increase in dust concentration shows a minimal effect on the burning velocity except for the high turbulent intensity which is slightly increased at the high dust concentration For the rich case 1 2 the increase in particle concentration also shows minimal effects except for the highest turbulent intensity which is slightly increases For the large particle size d 106 125um the effect of the increase in particle loading is more distinct In the lean case A 0 8 the increase in concentration causes a distinct decrease in the burning velocity For the stoichiometric case 1 0 the burning velocity also causes a decrease in the burning velocity as the concentration is increased but to a lesser extent than for the lean case For the rich case
111. d researchers to make a distinction between two types of dust flames 7 the Nusselt flame and the volatile flame In the Nusselt flame strictly heterogeneous combustion occurs at the surface of the particles sustained by the diffusion of oxygen towards the particles surface Therefore a Nusselt flame which on a macroscopic scale may seem like premixed combustion consists of an ensemble of local diffusion flames as shown in Fig 2 la In the case of the volatile flame vapors volatiles and or pyrolysis gases are produced by the particles prior to or during 40 combustion When mixed with air these gases and vapors burn as a premixed gas Depending on the nature of the solid three distinct mechanisms have been proposed for the combustions of particles in volatile flames 8 1 Devolatilization and burning of volatiles followed by combustion of a solid residue as shown in Fig 2 1b 2 Melting followed by evaporation and subsequently vapor phase burning as shown in Fig 2 1c e g sulphur plastics 3 Evaporation through a solid oxide shell followed by combustion of the vapor outside the shell e g metals like magnesium and aluminum 2 1 a b or c When a flame propagates through clouds of coal dust and many organic powders additional complexities arise These occur because after the homogeneous combustion of the liberated volatiles has occurred combustion of the remaining solid char may take place in the tail of the flame The volatile fl
112. d J H S Lee Burning Velocities in Fuel Rich Aluminum Dust Clouds in Proc Combust Inst 1996 The Combustion Institute 45 Lee J Burning velocity measurements in aluminum air suspensions using bunsen type dust flames 2001 46 Andac M G F N Egolfopoulos C S Cambell and R Lauvergne Effects of inert dust clouds on the extinction of strained laminar flames at normal and micro gravity Proc Combust Inst 2000 28 p 2921 2929 47 Kolbe M Laminar Burning Velocity Measurements of Stabilized Aluminum Dust Flames in Mechanical Engineering 2001 Concordia University Montreal Quebec Canada 48 Gonzalez O J F Richards and J D D Rivera Measurement of Flame Speed in Copper Concentrate Clouds Journal of the Chilean Chemical Society 2006 51 2 p 869 874 49 Palmer K N and P S Tonkin Coal Dust Explosions in a Large Scale Vertical Tube Apparatus Combustion and Flame 1971 17 p 159 170 50 Proust C Flame Propagation and Combustion in some dust air mixures Journal of Loss Prevention in the Process Industries 2006 19 p 89 100 51 Proust C and B Veyssiere Fundamental Properties of Flames Propagating in Starch Dust Air Mixtures Combustion Science and Technology 1988 62 4 p 149 172 52 Goroshin S M Bidabadi and J H S Lee Quenching Distance of Laminar Flame in Aluminum Dust Clouds Combustion and Flame 1996 105 p 147 160 53 Han O S M Yashima T Matsuda H Matsui A Miyake and
113. d but at the two higher particle loadings 50 and 75 g m the burning velocity is decreased This also occurs in the fuel rich case 0 1 2 and in the stoichiometric case 9 1 0 the 50 g m case is slightly increased but the 25 and 75 g m cases are slightly decreased These fluctuations are within the uncertainty of the measurement The trend for the larger sized particles may be due to the combined effect of heat absorption by the particles compensated by the increase in the effective equivalence ratio due to increased pyrolysis resulting from increased surface area of the particles at a similar gas velocity 77 d 75 90um d 106 125um 0 8 0 45 0 4 S st 0 35 m s 0 3 a u u 0 25 a 0 20 40 60 80 0 45 9 1 0 v 3 0 4 v v Ta 0 35 m s 0 3 0 25 d 0 20 40 60 80 1 2 0 45 0 4 S 0 35 0 35 m s v 0 0 v v 0 25 f 0 20 40 60 80 8 A 3 Figure 4 1 Laminar flame as a function of dust concentration Figure 4 2 shows a comparison of a sample set of laminar flames 1 2 a gas only b A 50g m c A 100 g m 0 200 g m These shadowgraph images are from video recordings using a Nikon d90 fitted with a macro lens and can be used for a qualitative understanding of the influence of particle concentration on the nature of the flame sheet The lines in the shadowgraph represent the premixed gas phase reaction zone The influence of the particles on the smoothness of the flame sheet is evi
114. dent in comparing Fig 4 2 a and b 78 representing a gas only and a gas flame with dust particle ds 75 90 um A 50 g m As shown in Fig 4 2 the particles cause slight bending of the flame sheet but do not create turbulent structures as observed with turbulent flames discussed later In Fig 4 2 c small sections of the premixed gas flame closer to the apex begins to extinguish In Fig 4 2 d the entire top of the premixed flame is extinguished This extinguishment phenomenon may occur because the dust particles are absorbing energy as they heat up Also as the dust concentration increases the production of volatile gases in the preheat zone will increases which may raise the local equivalence ratio above the upper flammability limit for the gas mixture This phenomenon is exacerbated by the lengthening of the residence time of a particle in the preheat zone as the burning velocity is reduced The observed reduction in contrast of the flame tip observed in Fig 4 2 may also be due to light saturation in the optical set up or decrease in relative temperature change As shown in Fig 4 2 the contrast of the flame in the shadowgraph is reduced as the dust concentration is increased The increase in dust causes an increase in the light emissions from the coal dust passing through the optics and collected on the camera sensor thereby saturating it in regions with high emissivity In the gas only flame the ambient gas temperature outside th
115. ds 75 90 um U rms 0 335 m s 0 1 0 75 g m 1 215 dst 75 90 um U 0 335 m s 0 1 2 As 0 g m 216 ds 75 90 um U 0 335 m s 0 1 2 25 g m 217 dst 75 90 um U gt 0 335 m s 0 1 2 50 g m 218 ds 75 90 um U rms i 0 335 m s 1 2 75 g m 219 dst 75 90 um U gt K 0 532 m s 0 8 Ast 0 g m 18 878 8 220 dst 75 90 um U gt 0 532 m s 0 8 25 g m 221 dg 75 90 um U gt 0 532 m s 0 8 50 g m 222 dst 75 90 um U gt 0 532 m s 1 0 Ast 0 g m 223 dst 75 90 um U gt 0 532 m s 0 1 0 Ast 25 g m 224 dst 75 90 um U gt 0 532 m s 0 1 0 50 g m 225 dst 75 90 um U 0 532 m s 0 1 2 Ast 0 g m 87 6 0 7 8 86 5 8 8 6 03 6 8 4 p 3 8 8 8 99 px MEE RUE 4 N ARR s R 0 2 V diz Z a iz k sa 227 dst 75 90 um U gt 0 532 m s 0 1 2 25 g m 228 ds 75 90 um U 0 024 m s 0 8 25 g m 231 dst 75 90 um U 0 024 m s 0 8 50 g m 232 dst 75 90 um U gt 0 024 m s 0 8 75 g m
116. e The graph shown in Fig 1 3 shows the turbulent intensity at different time instants Since the flame front is nonstationary and accelerates as the flame grows in size the turbulent intensity will also correspondingly increases as shown in Fig 1 3 Location t indicates when the flame is initiating usually using a chemical igniter or a spark f indicates the laminar flame propagating outward away from the ignition point tz denotes onset of turbulence which wrinkles the flame At the turbulent intensity u further increases as shown in the inset to Fig 1 3 If it is assumed that the dust particles are sufficiently small and well mixed to behave as premixed flames and this can be argued as discussed in Chapter 2 the laminar burning velocity SL or the velocity at which the flame front propagates normal to itself and relative to the flow into the unburnt mixture is very important For a turbulent flow the turbulent burning velocity Sr is equal to the mean normal velocity and depends on the turbulent intensity u and the integral length scale Note that the turbulent burning velocity becomes an averaged quantity as per its 15 definition Further turbulence increases the mass consumption rate of the reactants or reactant mixture to values much greater than those that can be obtained with laminar flames A greater mass consumption rate increases the chemical energy release rate and hence the power generated from a cer
117. e l along with the particle size ds and concentration As of condense phase fuel to provide a measure of the burning velocity of gas dust or hybrid flames The HFA is divided into several sections combustion chamber exhaust system burner nozzles dust feeder and optical setup which are explained in this chapter 3 2 Combustion chamber The HFA s combustion chamber is used to contain the dust and ash generated and minimizes ambient air disturbances Figures 3 la b shows the details of the combustion chamber The external frame a is made out of 3 81 cm 1 5 angle aluminum 0 3175 cm 1 89 thick The external dimensions of the frame are 44 cm 17 25 tall and 17 8 cm 12 wide The frame is held together using 8 32 bolts with Loctite to prevent the nuts from coming loose over time The edges of the frame were sealed using a high temperature RTV gasket maker Permatex Two of the walls were made of plate glass b 25 4 cm 10 tall 20 32 cm 8 wide and 0 238 cm 3 32 thick Rubber gasket 0 159 cm 1 16 thick is used between the glass and the aluminum to prevent leaks and help prevent the glass from cracking The glass was held onto the aluminum frame by eight tabs two on each side 49 The combustion chamber is divided into two sections the main section with the burner and a small section at the bottom c 7 62 cm 3 high where the makeup air is injected and allowed to disperse The air enters the combust
118. e Analyzer Fuel control system for Hybrid Flame Analyzer Water cooling system for Hybrid Flame Analyzer Building Annular Ring Pilot Flame for Turbulent Burner Hybrid Flame Analyzer 9 Simple shadowgraph design description Hybrid Flame Analyzer 10 Gas analysis for combustion system Hybrid Flame Analyzer 11 How to use mass flow controllers Hybrid Flame Analyzer 12 Changing perforated plate in Hybrid Flame Analyzer 13 Calibrating volumetric dust feeder Hybrid Flame Analyzer 14 Setting up hot wire anemometer for Hybrid Flame Analyzer 15 Checking hotwire anemometer voltage for Hybrid Flame Analyzer HEMP HE The MATLAB scripts used in the analysis of images and data acguisition are shown in Appendix 3 3 9 HFA data analysis Abdel Gayed 26 provides an excellent review of methods that have been used to measure turbulent burning velocities There are three main methods used to determine the turbulent 67 burning velocity with a stabilized vertical Bunsen burner type of flame The first is by determining an average flame angle used by Ballal et al 27 28 Karlovitz et al 29 30 Richmond et al 31 and Kobayashi et al 5 among others The second method is called the inner area method and was used by Damkohler 32 Khramtsov 6 Petrov et al 33 Williams et al 34 and Zotin et al 35 36 The total area method which involves using the area of a right angle cone fitted to the flame shape is us
119. e a day to once a week See the gas analyzer users manual for instructions on how to do the calibration The calibration gasses should be tied into the sample line before the heat sink so that the calibration gas goes through the same processing as the sample gas Dust Feeder A filter topped hopper was built to catch the dust out of the burner after a 1 minute run Using flow rate of 10 lpm and running the hopper for 1 minute at every 100 counts a curve fit is made of the dust feeder output To do the calibration the burner nozzles need to be replaced with the copper tube shown below one end of it has been sanded down so that it easily fits into the filter assembly shown below Run the test for 1 minute take the dust catch housing off and weight it Repeat for the full range of the feeder Two calibration curves are shown below 119 3 gt 75 90 um Coal s 2 5 106 125 um Coal 2 m E o 2 y 0 0032x 0 1436 o 5 m 91 5 o w a 2 1 g y 0 0029x 0 1037 e o 0 5 5 0 200 400 600 800 1000 Dust feeder calibration 120 Things to check Gas bottle levels When running a large number of tests it is more efficient to keep more than 1 bottle of air in the lab at one time Filter levels Acid filter should be changed periodically it does not have an indicator desiccant should be changed when it turns purple The absorbed water can be removed using an oven there is one i
120. e a point source of light This is placed at the focal length of the biconvex lens mounted in the side of the combustion chamber This creates a column of parallel light which passes through the flame and into an identical bi convex lens This second lens condenses the image so that the test section can be captured on a camera lens The image is captured on a Canon EOS 5D digital camera attached with a macro lens with the focus set to infinity Images are recorded at an average rate of 3 frames per second using a 66 shutter speed of 1 8000 seconds fstop of 2 8 reported and ISO of 800 A remote shutter release is used to prevent any camera movement due to handling of the camera A short wave optical filter is used on the front of the macro lens with a cutoff of 550 nm to reduce the effect of the bright orange yellow emissions of the burning dust particles The camera is mounted on a 20 kg block of concrete to prevent movement during testing 3 8 Directions for using HFA A user s manual for operating the HFA is shown in App 2 Instructional videos were also created on how to run tests using the hybrid flame analyzer and using individual components which can be found at www firesciencetools com in the Hybrid Flame Analyzer section These videos include Hybrid Flame Analyzer startup sequence Hybrid Flame Analyzer shutdown sequence Running laminar flame tests Electrical system for Hybrid Flame Analyzer Exhaust system for Hybrid Flam
121. e premixed flame is significantly lower than the flame temperature In a hybrid mixture the preheated coal dust continues to burn after leaving the premixed gas flame front causing relative difference between the premixed flame temperature and the surrounding gas to be lower This effect reduces the intensity of the shadow created by the region on either side of the reaction zone 79 Figure 4 2 Comparison of laminar flames 9 1 2 dx 75 90 um a gas only b 2 50 g m 6 4 100 g m 4 2 200 g m Original videos are available for viewing at www firesciencetools com 4 2 Turbulent flames 4 2 1 Gas flames validation study To validate the experimental apparatus and procedure the turbulent methane air flame data is compared with data from published work by Kobayashi et al 2 as shown in Fig 4 3 Fig 4 3 shows the turbulent burning velocity Sr of a methane air 6 1 0 flame as a function of turbulent intensity u ms Turbulence is generated using a perforated plate Imm hole diameter blockage ratio of 50 placed 30 mm below the exit of the nozzle similar to that used by Kobayashi et al 2 Error bars representing the uncertainty in the measurement are one standard deviation of the burning velocity calculated from the individual flame heights from 25 images These bars represent a 95 confidence level that the burning velocity exists within the range in the laminar case the error bars do not exceed the size of
122. e radius of the burner and maintains the constant spacing of the third tube q attached to the pilot fuel housing This insert 0 is 0 071 cm 0 028 thick and has 8 vertical slots cut into the inside p 0 127 cm 0 0507 deep The top of the insert was machined on a lathe to create a double notch at the top The deeper notch n allows gas to distribute around the circumference of the tube evenly and is approximately 0 254 cm 0 1007 deep The second notch m creates an anchored methane oxygen flame and is approximately 0 127 cm 0 050 deep Figure 3 6 shows pictures of the individual components in the turbulent burner nozzle a shows a side view of the main burner tube without the pilot flame assembly or water cooling b shows the fitting which allows the pilot fuel gas to be added and evenly distributed around the circumference c shows the spacing insert which keeps the spacing of the pilot gap constant around the circumference and increases the velocity of the oxygen methane mixture Figure 3 8 shows an image of the perforated plates used in this work Five perforated plates were created having a variety of hole diameters 4mm a 3mm b 2mm Imm 0 and 0 6mm e These round perforated plates are similar to the ones used by Khramtsov 6 The perforated plates are 56 mounted in identical nylon tubes f which have a threaded hole for the adjusting pin The 1 mm perforated plate has a blockage ratio area of holes tota
123. ed by Bollinger et al 37 and Grover et al 38 The calculation of the turbulent burning velocity in this work is similar to Grover et al 38 who averaged the measured flame height for 22 images to determine the burning velocity of a turbulent flame The area method uses S usina 3 5 to calculate the burning velocity where iz is the mean flow velocity and oz is the half angle of the right cone with a height equal to the mean flame height This method uses a number of simplifying assumptions as discussed by Lewis et al 3 1 The burning velocity is constant over the whole cone surface 2 The boundary between unburned and burned gases approximates a mathematical surface with the temperature changing abruptly from the initial to the final on passing through it 3 The flow lines retain their direction and velocity from the orifice right up to the cone surface Figure 3 17 shows a diagram of the process used to get the average flame height in this work a shows a sample shadowgraph image collected 0 8 u lt 0 185 2 lt 0 Using a MATLAB program shown in App the image is cropped the blue channel is extracted the intensity of the image is increased and the edge of the flame is selected by manually clicking along the edge shown as a blue line in b The pixel locations are converted to a distance with 1 68 pixel being equal to 0 04315 mm The location of the selected points is stored as part of a matrix c Th
124. ed on all sides by 10 inch high windows to provide shielding from room disturbances and free edge mixing Coal dust and one suppressant powder could be fed to the central stream by means of the feed disk scraper blade The dust was kept suspended in the feed tube in the diffuser and as it passed through the bore tubes by means of a white noise acoustic energy field in the burner tube Coal dust concentrations were determined at the burner head using a total capture technique by weighing the coal dust captured over a specified time Milne et al 42 published results from a new experiment in 1977 Dry air from a cylinder entered at the bottom of a glass storage section and passed through a sintered metal porous disk This fluidizing air passed up through the column of coal dust One portion flowed out through the exit tube and the rest exited through a filter and a flowing meter at the top of the apparatus An additional flow of gas to assist transport of the coal air along the tube and into the burner was provided near the entrance of the intake tube Best results were obtained when a stirrer was 23 added to continuously agitate the fluid bed of coal when the outtake tube was periodically reamed out and when the pressure in the fluidizer was held constant with a pressure controller Bradley et al 43 published results from an experiment in 1994 which used a graphite entraining fluidized bed and burner to provide flat laminar adiabatic
125. eing walked on 125 ASTM standards E2058 for FPA and E1354 for Cone Calorimeter describe the use of gas analyzers and the equations involved It is to be noted that the gas analyzer measures a percentage of the selected gas in the exhaust duct The flow rate in the duct needs to be known to be able to use a percentage This could be accomplished by adding a bi directional probe inside the duct The vane anemometer and hot wire anemometer will be compromised by the as from the coal particles The T connection after the pump releases excess pressure from the pump if this pressure is not released the gas reading takes about 10 minutes to change Currently the gas analyzer is the only part of the analyzer set up for use due to the high uncertainty in the CO and CO2 The lab does not currently have gas tanks to calibrate these 126 Turning off Experiment Make sure all flames are extinguished Change set point of mass flow controllers to zero Open combustion chamber Vacuum combustion chamber clean Wipe of lens and glass of shadowgraph with lens cleaning wipes Bleed gas lines Turn off gas bottle top valves Bleed out gas lines o Methane o Air o Oxygen do not bleed methane and oxygen into combustion chamber together Close pressure regulator valve Close shutoff valve Close valves between Unplug electronics Main power strip on instruments Power supply for mass flow controllers Check and empty moisture trap i
126. ency shutoff switch which will turn off the electronics and the gas flow in the event of a situation in which the user needs to leave the area rapidly The methane being used is lab grade 99 9 pure and therefore has no odorant therefor a methane detector should be added to the laboratory area so that any potential leaks can be detected A nitrogen purge should be added to the exhaust system due to the possibility of a fire caused by a build of coal particles And lastly an automated image analysis program would significantly speed up the data analysis process 1 Skjold T Review of the DESC project Journal of the Loss Prevention in the Process Industries 2007 20 p 291 302 2 ISO 6184 1 Explosion Protection Systems Part 1 Determination of explosion indices of combustible dusts in air International Organization for Standardization ISO 1985 107 Appendix 1 Parameters quantifying the hazard associated with a dust Name of Parameter Description Established Test symbol units Methods or Apparatus Thermodynamic Parameters 1 Heat of combustion J g Amount of energy released per unit mass undergoing a Bomb calorimeter combustion reaction 2 1 Combustion efficiency Fraction of energy that is utilized in pressure build up Law of Conservation of Energy 3 Radiant heat fraction Fraction of total heat released that is transferred via Radiant flux radiatio
127. end end if particleSize 75 amp amp flovvRate a 10 if dust_conc a burn_vel_12_75_00 1 Burning_velocity a stanDev BV 12 75 00 1 standard deviation BV a end if dust conc a lt lt 25 burn vel 12 75 25 1 lt Burning velocity a stanDev BV 12 75 25 1 standard deviation BV a end if dust_conc a 50 burn_vel_12_75_50 1 Burning_velocity a stanDev BV 12 75 50 1 standard deviation BV a end if dust conc a 75 burn_vel_12_75_75 1 Burning_velocity a stanDev BV 12 75 75 1 standard deviation BV a end end if particleSize 75 amp amp flowRate a 30 if dust_conc a burn_vel_12_75_00 2 Burning_velocity a stanDev BV 12 75 00 2 standard deviation BV a end if dust conc a 25 burn vel 12 75 25 2 Burning_velocity a stanDev BV 12 75 25 2 standard deviation BV a 146 E U 5 end if dust_conc a 50 burn_vel_12_75_50 2 Burning_velocity a stanDev BV 12 75 50 2 standard deviation BV a end if dust conc a 75 burn vel 12 75 75 2 Burning_velocity a stanDev BV 12 75 75 2 standard deviation BV a end end if particleSize 75 amp amp flowRate a 35 if dust conc a 0 burn_vel_12_75_00 3 Burning_velocity a stanDev BV 12 75 00 3 standard deviation BV a end if dust conc a 25 burn vel 12 75 25 3 lt Burning velocity a stanDev BV 12 75 25 3 standard deviation BV a end
128. erage line point x matrix Left x start left x aveDataRaw indicies 1 y matrix Left y start left y aveDataRaw indicies 1 x int left x start left x aveDataRaw indicies 1 x start left num pt ave 1 x aveDataRaw indicies 1 y int left interpl x matrix Left y matrix Left x int left x matrix Right lt x start right x aveDataRaw indicies end y matrix Right y start right y aveDataRaw indicies end x int right x aveDataRaw indicies end x start right x aveDataRaw indicies end num pt ave 1 x start right y int right interpl x matrix Right y matrix Right x int right x curve 1 num pt ave lt x int left x curve num pt ave l num pt ave length indicies x aveDataRaw indicies x curve num pt ave length indicies 1 2 num pt ave length indicies x int right curve l num pt ave y int left curve num pt ave l num pt ave length indicies y aveDataRaw indicies y_curve num_pt_ave length indicies 1 2 num_pt_ave length indicies y int right Simple Low Pass Filter num pt ave num pt ave 2 CCCI x curve CCC2 curve for ii num pt avetl length CCC1 num pt ave l CCC N141 mean CCCl ii num pt ave ii num pt ave CCC N2Gi mean CCC2 ii num pt ave iitnum_pt_ave end x curveSmooth 1 num pt ave x curve l num pt ave x curveSmooth num pt ave l length CCC N1 CCC NI num pt ave l end x_curveSmooth length CCC_N1 1 length CCC_N1 num_pt_ave
129. ermine x or y points for calculating the average for id 1 nData if isempty data id avgPoints if opt horzAvg collect y pts cat 2 data id yData else collect x pts cat 2 data id xData end cluster keep multiples for averaging pts unique pts d pdist pts follow TMW notation Z linkage d cutoff is half the average step size Of course this could theoretically lead to too wide spacing Hoewever if there are many points that overlap REALLY well robustMean gives a cutoff that is way too low cutoff mean diff unique pts 2 clust cluster Z cutoff cutoff criterion distance for every cluster calculate mean tmp NaN max clust 1 for c 1 max clust tmp c robustMean pts clust c end remove NaN sort data id avgPoints sort tmp isfinite tmp elseif isscalar data id avgPoints if opt horzAvg collect y pts cat 2 data id yData else collect x pts cat 2 data id xData end linearly space N points data id avgPoints linspace min pts max pts data id avgPoints 2 data id avgPoints 1 end end now that we know the location get the value of the average nLines length data id xData avgTmp NaN length data id avgPoints nLines stdTmp avgTmp for d 1 nLines if there are multiple abscissa points with the same value interpolation fails Thus pick the first point if necessary if opt horzAvg xx data id y
130. f Flame Propagation in Righ Mixtures of Coal Dust in Air Combustion and Flame 1985 59 p 251 265 39 2 Structure of a Dust Flame 2 1 Premixed or Non Premixed flame zone dust particles 2 6 with individual flames dust particles flame zone KU 5 oa with individual o o 00 9 o pyrolyzing flames 0070 pyrolyzing dust particles o 20 dust particles 50 90 o 9 99 9 oo a b c Figure 2 1 Types of dust flames a non continuous flames around individual particles b continuous gas flame with individual particles continuing to burn after flame zone c continuous gas flame front Unlike a premixed gas flame a mixture of dust and oxidizer involves a multiphase flow which causes difficulty in both experiments and modeling 1 5 Since gas combustion is a process involving only one phase homogeneous combustion the reactants are represented by their smallest entities i e molecules 6 When the fuel and oxidant are thoroughly mixed they are separated only by molecular distances Premixed combustion is therefore guaranteed down to very small scales By contrast dust explosions and dust flames involve the combustion of a dust air suspension A dust cloud which is uniform when viewed at a macro scale e g cloud radius may not be considered premixed at a small scale e g inter particle distance This cause
131. f gas analysis was done Turn off water cooling Check to make sure there is no combustion in the exhaust hood 127 MATLAB Scripts There are several MATLAB scripts created for use with the HFA Image Analysis pixel point only Area of flame based on pixel data Turbulent Intensity Gas Analyzer data from Hyperterminal These are shown in Appendix 3 128 Using Hotwire Anemometer For the setup of the hotwire anemometer a video was made Naming convention has two parts In the MiniCTA software when data is collected experiments are named KKH Where from left to right the numbers correspond to the perforated plate hole size in mm the flow rate in Ipm the perforate plate position 1 6 one on top and the height above the burner in cm When the data is exported the naming convention is YYMMDD mm lpm pos tem KHz sec S Where from left to right the numbers correspond to the date hole diameter flow rate perforate plate position anemometer location above the burner in cm sampling rate number of samples of data collection and the stands for the smaller anemometer which was the only one not broken at the time this was written The probe used is a 55P11 on the bottle the following data was on the bottle R20 3 80 ohms R2 0 5 ohms Alpha20 0 36 degC RR tot alpha_20 R20 T_sensor T 0 To use the CTA software Install CTA programs and Drivers on the two provided CD s Run MiniCTA v4 0
132. g the dust coming out of the nozzle over the range of settings for each dust used For each calibration point the dust feeder is operated for one minute 63 The dust is collected in a filtered dust hopper which allows gas to escape but collects the dust particles as shown in Fig 3 13 similar to Cassel et al 17 and Gosh et al 18 the dust air mixture flows up through c the dust is collected in the open area d and the air passes through a dust filter e The filter is held on by a rubber gasket g and 4 bolts f The output is weighed on a scale producing a linear line which is fitted to an equation as shown in Fig 3 14 These regression equations are used to provide the settings required for the desired dust concentration in terms of g m for each test The coal dust is sieved to different sizes using Retsch AS300 Sieve Shaker It is important to maintain the purity while sieving different materials therefore whenever a different material is sieved all sieving steel pans are cleaned by a Retsch 12 ultrasonic cleaner Table 3 2 Pittsburgh seam coal properties 19 20 E 65 4 kJ mol A 6 6x10 Us k 70 1 WK p 492 kg m Q 3 04x10 J kg a b Figure 3 13 Diagram of dust hopper used to calibrate dust feeder a side view of dust catch b top view of dust catch 64 75 90 um Coal 22 2 5 106 125 um Coal a 2 a E o 2 y 0 0032x 0 1436 ri ped o 2 d 91 5
133. gend name set avgH id 1 DisplayName Average line end loop data to plot CLEANUP warning oldWarn if nargout 0 clear avgH end if nargout gt 1 for id nData 1 1 avgData id data id avgPoints data id avgData end end 182 Appendix 4 Error Bar values standard deviation of velocity calculation Velocity m s Standard Deviation Turbulent Intensity Turbulent Intensity 75 90 75 90 NN m s 0 821 90 0 008 0 08 0 427 0 084 0 461 0 016 7 w gt to olo o CO 00 NO N oo o ojoje 51 4 se AR DIS o to w R gt w js EB OD w 65 gt w ojo NIN 05 2 RK O 00 dd N U1 B SOJO 1 BRIN 1 00 75 90 75 90 BIT o 09 to o SE N w S R Nu olo co m a to SIE 1 00 O P 2581 to co 4 5 olo olo o o SIR o to co 2 AK N o N R m o N o o P ta o UI R o o N o UI oo U O WIN olo 2 55 olo 05 UT Ojo olo olo 00 N alg Oje NIN O WIN Q co OJo oo o o olo o o N o ko o R N N vo pa to to Q R R 00 o N 00 o o N N o Ww e o 00 N EE NIN 00 O U U ojo 03 BIS 66 Ojo js 00
134. he blower by gently vibrating the flask The coal concentration in the suspension could be varied by varying the speed of the blower and the rate of vibrating the flask The coal concentration in the suspension was measured by aspirating a known volume of the suspension through a dust filter and weighing the coal collected 19 Hattori 35 published results from a steady state experimental method in 1957 Pulverized coal air mixtures were discharged from a burner into free air and ignited by an ignition source placed in the center of the burner Flame propagated into the mixture and an inverted cone flame front was formed Pulverized coal stored in hopper was fed continuously by a screw feeder driven by variable speed D C motor Coal and air from the screw feeder were uniformly mixed in a cyclone mixer and led into the burner tube The mixture passed through an annular space formed by the burner tube and ignition gas tube and was discharged upward to open atmosphere Acetylene as ignition gas was also discharged upward to open air through the ignition gas tube Electromagnetic vibrators were attached to the hopper and the mixer lest the pulverized coal should adhere to the walls When the acetylene was ignited a steady inverted cone flame was formed Burgoyne et al 36 published results from a downward pointing steady state burner in 1958 The suspension for combustion was formed by jet impaction of a regular supply of solid by the carrier a
135. he addition of dust particles in the turbulent gas phase lean case 96 unburned flame burned mixture 5 mixture 1 A pC S lt Ja lu ins en bo E 0 x Figure 4 10 The effect of turbulence on the temperature profile in the preheat zone In the case of higher particle sizes d 106 12Sum at a given intensity the injection of the particles either enhances or dissipates the turbulence level This is also dependent on the size distribution of the particles mean value of the particle diameter in the injected lot Therefore the combined effects of an increase or decrease in the turbulent intensity due to particle injection and the particle size distribution cause a nearly random variation in these cases However in this case also as the intensity is increased the ratio of turbulent to laminar burning velocity also increases An increasing trend with concentration of the dust is also observed however only at the higher equivalence ratio of 1 2 The influence of concentration on the burning velocity is further analyzed in the next section 97 4 2 4 Effect of dust concentration on burning velocity Dust 0 g m Dust 25 g m Dust 50 g m Dust 75 g m Figure 4 12 Images of turbulent flames at various dust concentrations 0 8 u ms 0 185 m s and d 75 90 um Figure 4 12 shows sample images over the range of dust concentration used for a lean 0
136. his group of literature is the work by Seshadri et al 81 as it is the first study that incorporated both gas and condensed phase kinetics and thereby systematically analyzed the influence of volatization on dust flame dynamics Recent work by Bidabadi and Rahbari 82 extended the theory to include the effects of inter particle conduction as well A detailed literature is available in Smoot and Horton 83 Krazinski et al 84 and Slezak et al 85 1 5 Goals and objectives of the current study Initiation and propagation of dust deflagrations are extremely complex phenomena due to the interaction between solid particles and the gaseous flame front In comparison with premixed gas deflagration a dust oxidizer deflagration depends on the rate of evolution of volatiles the mixing of these volatiles with the oxidizer surrounding the particles coupling of the particles and gas phase oxidation as well as radiative energy exchange between the flame and its surroundings 33 Though engineering tools such as the DESC code produced by Gexcon have been created due to the complications discussed above a comprehensive mathematical theory to predict deflagration mechanisms of dust clouds is at present beyond reach Although vast amount of testing both small scale 20 liter explosion vessel and large scale tests have been done over the last 50 years most theories that connect the data to models are heavily empirical and the problem has never bee
137. idth plotLineWidth ylim y_axisMin y_axisMax hold off subplot subplot 3 2 3 Parent figurel YTick 2 2 4 2 8 3 2 3 6 4 LineVVidth 2 FontWeight bold FontSize 14 FontName Times New Roman subplot 3 2 3 hold on plot l 0 10 burn vel 10 75 00 orig lam data end burn vel 10 75 10 dst sze origFit 1 ks LineWidth plotLi ne Width plot l 0 10 25 burn vel 10 75 25 orig lam data end burn vel 10 75 10 dst sze origFit 2 ro LineWidth pl otLineW idth plot 10 50 burn vel 10 75 50 orig lam data end burn vel 10 75 10 dst sze origFit 3 sv LineWidth pl otLineW idth plot l 0 10 75 burn vel 10 75 75 orig lam data end burn vel 10 75 10 dst sze origFit 4 bh LineWidth pl otLineW idth ylim y axisMin y axisMax xlim 0 4 1 42 subplot subplot 3 2 4 Parent figurel XTick 0 4 0 6 0 8 1 1 2 1 4 YTick 2 2 4 2 8 3 2 3 6 4 LineVVidth 2 FontWeight bold FontSize 14 FontName Times New Roman subplot 3 2 4 hold on plot l 0 10 00 burn vel 10 106 00 orig lam data end burn vel 10 75 10 dst sze origFit 1 ks LineWidth p lotLine Width 10 0 0 10 25 burn vel 10 106 25 orig lam data end burn vel 10 75 10 dst sze origFit 2 ro LineWidth p lotLine Width plot l_0_10_50 burn_vel_10_106_50_orig lam_data end burn_vel_10_75_10_dst_sze_origFit 3 gv LineWidth plotLineWidth
138. ies of gas flames were measured by initiating a flame from a central spark and recording the spherical flame propagation in a closed vessel optically via a quartz window or by recording the pressure time trace by placing a pressure transducer at the vessel walls as shown in Fig 1 1 This set up was adopted for dust air pre mixtures as discussed below a detailed history is given in Eckhoff 21 The standard dust explosion vessel is equipped with a vacuum dust dispersion ignition and pressure sensor systems The standard procedure begins by placing a measured quantity of a dust sample in a reservoir as shown in Fig 1 1 Prior to ignition the dust air mixture is discharged into the vessel through a fast acting valve and a rebound nozzle The dispersed dust cloud is ignited after a specified ignition delay time The ignition source is typically two chemicals igniters 5 kJ each positioned near the center of the vessel The main operating conditions for two typical explosion vessels are shown in Table 1 2 Measures for the energy content and the reactivity of the dust air suspension are derived from the pressure time history as shown in Fig 1 2 Both the maximum pressure and deflagration index are determined from the same type of experiments in constant volume explosion vessels Further details are described in standards ISO 6184 1 22 EN 14034 1 23 EN 14034 2 24 and ASTM E 1226 14 13 Table 1 2 Main operating conditions for two typi
139. igures Sq s is used to normalize the turbulent velocity data to prevent the increase in uncertainty caused by the division of two experimentally measured points the laminar data was fit to linear lines and these functions were used to calculate the value used for the normalization 94 d 75 90um d 106 125um g 08 4 4 3 6 3 6 i 3 2 ki 3 2 7 x 2 8 2 4 2 4 2 a 2 b 0 5 1 1 5 2 0 5 1 1 5 2 d 1 0 4 4 3 6 3 6 r 32 3 2 S q 28 v 2 8 d 2 2 04 06 0 8 1 12 1 44 04 06 0 8 1 12 1 44 9 1 2 4 4 036 3 6 3 2 3 2 v i a TE y 2 4 e 24 f 2 2 O 000 7 0 5 1 15 2 0 5 1 1 5 2 o Agt 25 gim Um Was v 50 gim Sig 52 YE 7 75 gim Figure 4 9 Turbulent burning velocity vs turbulent intensity In general Fig 4 9 shows that the turbulent burning velocity is more than two times larger than the laminar counter part for each and every case studied The turbulent to laminar burning velocity ratio increases as the turbulent intensity is increased for all cases More interestingly in most of the cases where smaller particle range is used d 75 90um as the dust concentration is increased to 75 g m at a given intensity the ratio of turbulent to laminar flame velocity is seen to increase significantly This is primarily due to the effect of an increase in the 95 turbulence level due to the interaction of smaller sized particles which also increases with increased nu
140. ing Repeat as needed Start flowing air through burner and pilot gasses ignite pilot gasses with spark adjust pilot gasses as desired Start methane flowing at desired rate using mass flow controller Take the desired number of pictures Run long enough to get gas analysis data Set dust feeder settings Start dust feeder Take desired number of pictures 124 Collecting gas analysis data Turn on colt trap orange covered switch Turn on Hyperterminal software Turn off driver for mouse computer thinks 232 usb adapter is a mouse Windows button drivers amp printers gt gigaware USB to serial cable com5 right click gt properties gt hardware tab disable mouse or driver if I remember right Com5 Baud rate 9600 no bias continuous output also settings for servomex 4000 You want to collect hyperterminal data as the experiment occurs use Transfer gt capture text gt name file gt stop when done with test There is approximately a 30 second delay in the gas analyzer measurement Data is recorded at 1 Hz Cold Trap particle Acid Filter Filter amp moisture Gas Sample trap Exhaust duct Gas Particle E 7 Analyzer filter 27 0 25 Ipm Pump pressure release The gas analyzer is connected to the computer using a RS232 extension cable RS232 to usb Desiccant adapter and USB adapter extinction cable This cable is run up and over along the drop ceiling to prevent it from b
141. ing the exhaust flow A flame trap was 21 fitted at the bottom of the burner tube In the exhaust system the surplus air was filtered and measured and then passed to a valve system The pump was arranged so that all of the suspension formed in the fluidized bed could be exhausted or any desired fraction could be made to flow up the burner tube The concentration of the suspension emerging from the burner could be measured by attaching to it a reducing nozzle with plastic tube leading to a weighed filter followed by a pump control valves and bubble meter The sampling time was at least one minute Bryant 40 built a steady state burner apparatus in 1971 The various gases and solids were introduced into the narrow channel at the base of the burner where they were mixed and their flow streamlined as they passed upward through the conical volume and the coarse screen off which the flame was stabilized The screen was required to prevent flashback In some experiments the flame was surrounded by a blanket of oxygen The powder dispersing device was a modified S S White Model F abrasive cutting unit The modifications consisted of the removal of the powder container and vibrating table from the original cabinet to a position immediately adjacent to the burner manufacture of a gas tight cap for the container and the installation of remote controls for the vibrating table and carrier gas The rates were determined by collecting and weighing The oxygen
142. ion chamber through hundreds of 0 159 cm 1 16 holes drilled into a 28 guage steel plate which separates the section d similar to the experiment used in Bradley et al 1 The makeup air is controlled using a flowmeter During tests 30 Ipm of air is injected into the combustion chamber by a 0 635 cm Swagelok female tube adapter e The air is distributed through a 1 27 cm 72 copper tube not shown with 1 inch wide slits cut into the side to help distribute the air in the lower section Water for cooling the burner is injected and removed through two 0 635 cm 4 Swagelok female tube adapters f The fuel for the burner pilot flame is injected into the combustion chamber through another 0 635 cm 47 Swagelok female tube adapter g The third side of the combustion chamber is made of a plate of 28 guage galvanized steel plate 1 The water cooling fittings f pilot fuel gas i a biconvex lens h and the spark igniter j are connected through this steel plate and sealed with high temperature RTV gasket maker The optics system uses two bi convex lenses h which are attached to the combustion chamber The spark igniter j is mounted on a 30 48 cm 12 aluminum rod surrounded by a rubber housing This housing allows the igniter to be moved inside of the combustion chamber allowing it to ignite the pilot and then be moved out of the way The 4 side of the combustion chamber is a door 1 to access the inside of the combustion
143. ir and was burnt on a downward pointing water cooled burner The tendency of buoyancy to distort the flame was countered by an extraction system mounted below the burner Two types of burner nozzles were used a convergent nozzle with a throat bore ratio of approximately W and a short tube 2 1 5 in long Flames propagated only if a form of energy addition were maintained and a convenient source was found to be an annular premixed coal gas air flame formed at the periphery of the coal air stream The concentration of the cloud was determined either before ignition or after extinction of the flame by collecting the issuing coal dust on a filter Palmer et al and William et al 37 38 published results from a steady state laminar dust flame burner in 1962 Dispersion of the dust was achieved my means at a unit at the bottom of the 20 burner In the dust dispersion unit was a hypodermic needle through which the input gas stream passed at a pressure drop of about 20 psig The exiting high velocity gases impinged on the dust bed thus generating the dust dispersion The dust reached the hypodermic needle through an opening cut in the base of the burner tube A constant supply of dust was kept moving in to the dispersion chamber by the rotation of a brass container which enclosed the entire dispersion unit In addition to the gas flowing through the hypodermic needle another stream of gas flowed through a central dilution tube This stream was used t
144. is known as the bending exponent and C is a parameter that contains the influence of the scale of turbulence This work Eg 4 4 Eq 4 5 Eg 4 15 Eq 4 19 0 1 0 2 0 3 0 4 0 5 U m s Figure 4 5 Correlation of experimental results 1 Equations 4 4 4 5 4 15 and 4 19 are shown along with the current experimental data in Fig 4 5 It is observed that Damkohler s Eq 4 4 Schlekin s Eq 4 5 and Eq 4 15 under predict the experimental data The best fit is obtained by using Eq 4 19 using values of C 1 6 and n 0 3 similar to those used by Dahoe 4 for propane air flames 4 2 2 Turbulent combustion regimes Figure 4 4 showing the flame images epitomizes how the reaction zone of a flame can be affected by a turbulent field To understand the effect a suitable starting point is the consideration of the quantities that determine the fluid characteristics of the system The structure of the 87 turbulent velocity field may be presented in terms of two parameters the scale and the intensity of the turbulence The intensity is the square root of the turbulent kinetic energy which essentially gives a root mean square velocity fluctuation v Based on the three length scales used in turbulence 9 1 the integral length scale which characterizes the large eddies or the length beyond which various fluid mechanical quantities become essentially uncorrelated 2 the Taylor microscale 2 which is obtained
145. is process is repeated 25 times and using another MATLAB program shown in App the location of the flame edges are combined as shown in d and averaged as shown in e The average shown in e is done by averaging the height location of the curves moving across the horizontal axis between the two average anchoring points at the edges of the burner nozzle A linear line is shown connecting the cutoff point to the base location but is not employed in the analysis method used The resulting curve is smoothed using a point averaging method and the maximum height of the fitted curve is used to calculate the half angle as afosa a tan 3 6 where d is the internal diameter of the nozzle exit and h is the mean flame height Using this procedure the calculated burning velocities for turbulent methane air flames match reasonably well with published data as shown in the next section 69 d e Figure 3 17 Analysis method of turbulent images The method shown above was used due to the difficulty in determining a quantitative total flame surface area as opposed to the total area method Figure 3 18 shows top and side profiles for a theoretical turbulent flame As shown to completely quantify the surface area of a turbulent flame instantaneous measurements of the side view a and top view profiles b at each height along the flame this could require 100 s or 1000 s of slices similar to the way an MRI is done
146. isture content 32 Agglomeration A mass conserving number reducing process that shifts the particle size distribution towards larger sizes 33 Terminal settling velocity Velocity of a particle when the drag force and buoyancy of dust particle m s force balance equal the gravitational pull 34 Speed of sound in dust Plays an important role in all compressible flow cloud m s phenomena Chemical Parameters 35 Chemical composition Molecular formula of the sample gives important information like Molecular Weight acidic or basic nature special affinity for other chemicals 36 Reactivity with water Electrical Parameters 37 Volume resistivity Measure of electrostatic ignition hazard of the dust TEC 60093 38 Charge relaxation time Time duration of charge retention in a dust TEC 61340 2 1 2000 39 Chargeability Propensity of dust particles to become charged when TEC 61340 2 flowing or air bourn 1 2000 External Parameters facility related 40 Size of partial volume This factor will depend on construction type volume of explosion that can be handled by the construction initial cloud that can be formed number of vents installed and nature of dust 109 41 Type of construction Based on NFPA 220 standard on types of building construction 42 Room volume m Total volume of room enclosure where fugitive dust accumulation is possible 43 1 Operating temperature C 1 Certain facilities could operate
147. k 3 13 Diagram of dust hopper used to calibrate dust feeder 3 14 Dust feeder calibration curves 3 15 Visual images of burner flames 3 16 Shadowgraph images of burner flames 3 17 Analysis method of turbulent images 3 18 Profiles of theoretical turbulent flame 3 19 Comparison of calculated burning velocity versus number of images used 4 1 Laminar Flame as a function of dust concentration 4 2 Comparison of laminar flames 4 3 Turbulent burning velocity of a methane air flame vs turbulent intensity 4 4 Flame images at various turbulent intensities 4 5 Comparison of this work with published data 4 6 Borghi diagram parameters 4 7 Characteristic parametric relationships of premixed turbulent combustion 4 8 Diagrams of turbulent flame structure 4 9 All data as a function of turbulent intensity 4 10 Error bars on data 4 11 Turbulent burning velocity vs turbulent intensity 4 12 Influence of dust on the burning velocity of a gas flame 4 13 Images of turbulent flames at various dust concentrations 4 14 Nondimensionalized burning velocity as a function of dust concentration 4 15 Combined fitted curves of test data List of tables 1 1 Recent incidents of industrial dust or hybrid flame explosions 1 2 Main components for two typical explosion vessels 2 1 Fuel concentration scenarios in hybrid flames 3 1 Integral length scale calculations 3 2 Pittsburgh seam coal properties 3 3 Experimental matrix 4 1 Curve fitting parameters Nomen
148. l area of 50 The perforated plate design is similar to work by Kobyashi et al 5 and Liu et al 7 The annular pilot shown in Fig 3 7 is similar to the one used by Kobayashi et al 5 It is necessary to hold the flame due to the high flow rates used to generate turbulent intensity and is fueled by methane and oxygen mixture 1 This mixture was used because of the higher burning velocity which compared to air prevents the turbulence in the main burner flow from disturbing the pilot Both burner nozzles have water cooling 10 liters per hour controlled by a flowmeter made out of copper tubing wound around the burner diameter with thermal grease Arctic Silver Ceramique Thermal Compound to increase conductive heat transfer The main burner flow is measured using a hot wire anemometer Dantec Dynamic 9055P011 sampling at a rate of 100 kHz The platinum plated tungsten wire sensor has a diameter of 5 microns and is 1 25 mm long The hot wire anemometer was calibrated using the average bulk flow velocity through the burner based on the mass flow controller The calibration curve follows a power law relationship as shown in Fig 3 9 Turbulent flow can be described using 8 u u u 3 1 where u is the flow velocity u is the average flow velocity and u is the fluctuating component of the flow velocity The turbulent intensity is defined as the root mean square RMS of the turbulent fluctuation in the u 8 and can be calculated using
149. le Smoke Release Rates for Materials and Products Using an Oxygen Comsumption Calorimeter 2009 17 EN 13673 1 Determination of the maximum explosion pressure and the maximum rate of pressure rise of gases and vapours Determination of the maximum explosion pressure 2003 18 Andrews G E D Bradley and S B Lwakabamba Turbulence and Turbulent Flame Propagation A Critical Appraisal Combust Flame 1975 24 p 285 304 19 Abdel Gayed R G and D Bradley Dependence of turbulent burning velocity on turbulent reynolds number and ratio of laminar burning velocity to R M S turbulent velocity Proc Combust Inst 1977 16 p 1725 1735 20 Bradley B How Fast Can We Burn Proc Combust Inst 1992 24 p 247 262 21 Eckhoff R K Explosion Hazards in the Process Industries 2005 Houston TX 22 ISO 6184 1 Explosion Protection Systems Part 1 Determination of explosion indices of combustible dusts in air nternational Organization for Standardization 180 1985 23 EN 14034 I Determination of explosion characteristics of dust clouds Part I Determination of the maximum explosion pressure Pmax of dust clouds 2004 European Committee for Standardization CEN Brussels 24 EN 14034 2 Determination of explosion characteristics of dust clouds part 2 Determination of the maximum rate of explosion pressure rise of dust clouds 2006 European Committee for Standardization CEN Brussels 25 Dahoe A E J F Zevenbergen P
150. length scale and particle size This relatively crude model is used for the current data to provide a mathematical representation of the trends Once more testing is done as discussed below a more precise model can be developed 105 The author would like to provide a number of recommendations for the continued use of the HFA First off it is important to show that the instrument can produce reproducible measurements To do this and produce more data tests could be done at more frequent dust concentrations with the same flow conditions used in this work The data should fall on the same line The overall purpose of this work is to help provide industry with a new tool for designing protection systems therefore results from these tests should be compared with large scale explosion experiments to show that the data from the lab scale experiments can be correlated to the large scale explosions Once the lab scale is shown to match with the full scale the HFA experimental results should be coupled to industry in two ways First the empirical models provided through the experimentation should be used in future modeling programs such as DESC 1 Second the turbulent burning velocity should be tied into structure vent design as it is tied to Ky currently in the design codes and standards 2 After confidence has been established in this new apparatus and repeatability shown a variety of dust types should be tested such as steel and other metal dusts along
151. lliams 57 Pope 58 Borghi et al 59 Chomiak 60 and Ballal 61 Some turbulent burner design ideas from Kobayashi et al 62 Smallwood et al 63 and Filatyev 64 have been incorporated in the current design discussed in section 3 1 4 2 2 Non stationary flames Hertzberg et al 65 published results from a 7 8 liter flammability chamber a modified and larger version of the standard 1 2 liter Hartmann apparatus This instrument included a dust probe pressure transducer oxygen sensor dust cup and ignition point The top plate of the chamber was fitted with a sapphire window assembly through which the infrared radiance of the explosion could be measured The normal procedure was to spread a measured mass of dust 30 uniformly around the disperser cone The top plate was then bolted and the chamber partially evacuated to about 0 2 atm The air dispersion tank was pressurized to 5 atm This 0 2 sec air impulse dispersed the dust mixed with it and raised the chamber pressure to 1 0 atm After another 0 1 sec delay to allow for more uniform dispersion the ignition source was energized If the mixture was flammable the developing pressure and infrared spectral radiance were monitored When flame propagation was complete and after the combustion products cooled the residual oxygen content was measured and dust or gas samples could be taken for analysis Li et al 66 published results from a long tube to study dust combusti
152. lotLineWidth 1 65 n 20 u_prime_smooth 0 max u_prime_all min u_prime_all 200 max u_prime_all u primeDivS L u prime smooth burn vel 12 75 75 orig 1 S T eg 148 1 C u primeDivS L n 166 plot u primeDivS L S T eg 148 k LineWidth plotLineWidth C 2 00 n 20 u_primeDivS_L u_prime_smooth burn_vel_08_75_50_orig 1 S T eg 148 1 C u primeDivS L 4n plot u primeDivS L S T eg 148 k LineWidth plotLineWidth axis 0 15 0 55 0 65 131 legend 08 75 00 08 75 25 08 75 50 08 75 75 08 106 00 08 106 25 08 106 50 08 106 75 410 75 005510 75 25 10 75 50 10 75 75 10 106 00 10 106 25 10 106 50 10 106 75 12 75 00 12 75 2512 7350 12 73 75 12 106 00 12 106 25 12 106 50 12 106 75 legend Wambda_fst 0 Wambda_fst 25 Wambda_fst 50 Wambda_fst 75 Mambda_fst 0 UWambda_fst 25 Wambda_fst 50 Mambda st lt 75 Mambda_fst 0 UWambda_fst 25 Wambda_fst 50 Mambda st lt 75 Mambda_fst 0 UWambda_fst 25 Wambda_fst 50 Mambda st lt 75 lambda fstJ 0 Vambda stj lt 25 lambda stJ 50 Vambda st lt 75 lambda stj lt 0 Mambda st lt 25 Mambda stj lt 50 Mambda st lt 75 legend Wambda 0 Nambda 25 Nambda 50 Mambda 75 Mambda 0 Wambda 25 Uambda 50 Vambda 75 Mambda 0 Wambda 25 WUambda 50 Wambda 75 Mam
153. lotLineWidth ylim y_axisMin y axisMax hold off subplot subplot 3 2 6 Parent figure3 Y Tick 2 2 4 2 8 3 2 3 6 4 LineVVidth 2 FontWeight bold FontSize 14 FontName Times New Roman subplot 3 2 6 hold on plot dust_conc_12_106_10_dst_sze burn_vel_12_106_10_dst_sze kv MarkerSize plotMarkerSize LineWidth plot Line Width plot dust 12 106 30 dst sze lam data dst sze end l 00 lam data dst sze end burn vel 12 106 30 dst 526 orig burn vel 12 75 10 dst sze origFit ro MarkerSize plotMarkerSize Line Width plotLineW idth plot dust conc 12 106 35 dst sze lam data dst sze end l 00 1am data dst sze end burn vel 12 106 35 dst sze orig burn vel 12 75 10 dst sze origFit sv MarkerSize plotMarkerSize Line Width plotLineWidth plot dust 12 106 40 dst sze lam data dst 526 6 0 1 00 lam data dst sze end 1 burn vel 12 106 40 dst sze orig burn vel 12 75 10 dst sze origFit lam data dst sze end 1 bh MarkerSize plotMarkerSize LineWidth plotLineWidth ylim y axisMin y axisMax hold off p get 0 monitorpositions p l1 1 1000 450 100 270 850 2750 set gcf position p 2 159 A3 5 Plotting figure 4 14 plot_ND_SLdivSLgasOnlyv01 plot_ND_SLdivSLgasOnlyv02_6fig plot_ND_SLv03_6fig legend_plot 0 plotMarkerSize 10 plotLineWidth 3 testSize2
154. ls for turbulence modulation in fluid particle flows International Journal of Multiphase Flow 2000 26 p 719 727 15 Rockwell S R and A S Rangwala Effect of Coal Particles on Turbulent Burning Velocity of Methane Air Premixed Flames in Technical Meeting of the Eastern States Section of the Combustion Institute 2011 Storrs CT 16 Bradley B How Fast Can We Burn Proc Combust Inst 1992 24 p 247 262 17 Arntzen B J Modelling of turbulence and combustion for simulation of gas explosions in complex geometries in Applied Mechanics Thermodynamics and Fluid Dynamics 1998 Norwegian University of Science and Technology 18 Wingerden V B J Arntzen and P Kosinski Modelling of dust explosions VDI Berichte 2001 1601 p 411 19 Agreda A G Study of Hybrid Mixture Explosions in Chemical Engineering 2003 Degli Studi di Napoli Federico II 20 Liu Y J Sun and D Chen Flame Propagation in Hybrid Mixture of Coal Dust and Methane Journal of Loss Prevention in the Process Industries 2007 20 p 691 697 103 21 Chen D L J H Sun Q S Wang and Y Liu Combustion Behaviors and Flame Structure of Methane Coal Dust Hybrid in a Vertical Rectangle Chamber Combust Sci and Tech 2008 180 p 1518 1528 104 5 Conclusions and recommendations Using a combination of experimental methods found in the literature a new apparatus called the Hybrid Flame Analyzer HFA is designed constructed and used
155. mber of particles present at higher concentrations 14 This enhances the overall heat and mass transfer in the small sized particles and as a result the burning velocity increases While the increasing trend is observed for all three equivalence ratios tested it is highest for the fuel lean cases since there is also an increase in the local equivalence ratio as discussed below For a fixed planar flame sustained by an isotropic turbulent flow of a combustible mixture with a constant cp the Favre averaged one dimensional energy equation is given by 4 PesS dio 27 dx Lu rms dx 0 4 20 Where the turbulent thermal conductivity is expressed as the product of the turbulent length scale lo and root mean square of the turbulent velocity fluctuations w Equation 4 20 rms be solved with the boundary conditions 0 T T 4 21 xo to obtain T T pc S 0 VATE 4 22 T ds L ms This solution is plotted in Fig 4 10 to illustrate how the width of the preheat Zone depends on the turbulent diffusion of heat into the unburnt mixture ahead of the flame When the turbulence is intensified S u decreases and the width of the preheat zone increases As the preheat zone extends further into the unburnt mixture the fuel particles are exposed to a higher temperature longer and consequently release more volatiles This provides an explanation for why the burning velocity increases with t
156. n analyzed from a fundamental viewpoint Identification of the controlling parameters of dust deflagration mechanisms is crucial to our understanding of the problem As a first step a scientific experimental platform is needed to understand the physical and chemical processes that control the behavior of dust flames in both laminar and turbulent flow fields The objective of this study is to develop such an experimental platform capable of measuring the laminar and turbulent burning velocity of a dust air premixed flame as a function of properties specific to the reactants such as dust particle size and concentration The experimental set up is then used to analyze the a particle gas air premixed system composed of micron sized coal dust particles 75 90 and 106 120 um in a premixed CHa air 0 8 1 0 and 1 2 flame This work will ultimately improve the knowledge on fundamental aspects of dust flames which is essential for the development of mathematical models This study is the first of its kind where different parameters that govern flame propagation in a spatially uniform cloud of volatile particles are systematically analyzed These parameters include initial particle radius number density or concentration turbulent intensity and length scale The major improvement of the experiment used in this work beyond the experiments described in the work above is the ability to control and quantify the turbulent intensity and integral length scale
157. n mode measurements 4 Latent heat of vaporization Amount of heat required to vaporize a unit mass of fuel Differential J g Scanning Calorimeter 5 Adiabatic flame Maximum possible temperature achieved by the Theoretical temperature C combustion reaction in a constant pressure process Calculations 6 Specific heat of dust J g Amount of energy reguired per unit mass of dust to Differential K increase the temperature of the dust by one unit Scanning Calorimeter Thermo kinetic Parameters 7 Laminar burning velocity Velocity at which unburned gases move through a None m s combustion front in the direction normal to the front surface 8 1 Propagation speed of Rate at which a exothermic oxidation reaction front moves smoldering reaction front in the direction of non reactive zone of a dust layer m s 9 Rate of reaction in the gas Rate at which the reactant gas concentration depletes phase g s 10 Rate of reaction in the solid Identifies the smoldering combustion of a dust layer phase surface chemical Smoldering layers can release combustible vapors such as reaction rate g s CO CH4 which can lead to a gas deflagration 11 Maximum closed volume Maximum pressure reached during a dust deflagration for ASTM E1226 deflagration pressure bar the optimum concentration of the dust cloud 12 Maximum closed volume Rate of pressure rise at maximum pressure reached during ASTM E1226 rate of pressure rise bar s a dust deflagration
158. n the fire lab Dust hopper level make sure there is enough dust in the feeder to run tests Camera memory level empty before test 121 The exhaust duct should be vacuumed out periodically to prevent the buildup of dust particles and ash Shadowgraph alignment Water catch shown above make sure it is empty 122 Starting Running test Camera settings Shutter speed 1 8000 Fstop 2 8 ISO 800 Align shadowgraph Install perforate plate and set in desired position Set mass flow controllers to desired flow rate Turn on Combustion chamber makeup air 30 Ipm Turn on methane for pilot 200 cc min reading on flow meter Ignite with spark Turn on oxygen for pilot 700 cc min reading on flow meter Allow pilot to reach steady state there will be a distinct high pitch sound Turn on central burner air Turn on central burner methane Turn on Dust Turn on shadowgraph light Take 100 pictures 35 seconds using the remote and holding down the large button Turn off Dust Turn off Shadowgraph light Turn off Central burner methane Turn off Central burner air Turn off pilot oxygen and wait for diffusion flame to form Turn off pilot methane 123 Briefly turn on main burner air to blow out pilot flame Wait for picture to copy to compact flash disk in camera Create new folder on computer with test details Copy images to folder Delete images on compact flash disk Put disk back in camera Change dust sett
159. n time intervals Linearity and constancy of calibration were maintained for up to one hour Goroshin et al and Lee 44 45 published results using an experimental setup comprised of a water cooled laminar dust burner nozzle The dust dispersion system included a syringe type 24 dust feeder and a circular annular high velocity gas jet sheet The system had an ability to produce a uniform dust flow for a wide range of dust concentrations for duration of up to six minutes A long stainless steel tube of 70 cm length and inner diameter 25 mm was connected to the dispersion chamber through a small angle conical diffuser This provided laminarization of the initially turbulent dust flow as it exited the dust disperser The dust flow exited the combustion tube through a small angle conical nozzle A water cooled brass ring with a triangular cross section was used as a flame holder located 1 cm above the nozzle exit An auxiliary stream of Na concentric to the dust stream was used to maintain the cylindrical configuration of the dust cloud issuing from the burner The flame shapes were recorded with a Canon single lens reflex camera with a bellows macrophoto attachment at a scale of 3 1 A neutral filter with an optical density of about three had to be used to attenuate the flame radiation Andac et al 46 published results from a counter flow experiment to study flame extinction from inert particles The experimental configuration includes the
160. ning_velocity a stanDev BV 12 106 25 2 standard deviation BV a end if dust_conc a 50 burn_vel_12_106_50 2 Burning_velocity a stanDev BV 12 106 50 2 standard deviation BV a end if dust conc a lt lt 75 burn vel 12 106 75 2 Burning_velocity a stanDev BV 12 106 75 2 standard deviation BV a end end if particleSize 106 amp amp flowRate a 35 if dust conc a burn vel 12 106 00 3 Burning_velocity a stanDev BV 12 106 00 3 standard deviation BV a end if dust conc a lt lt 25 burn vel 12 106 25 3 Burning_velocity a stanDev BV 12 106 25 3 standard deviation BV a 145 end if dust_conc a 50 burn_vel_12_106_50 3 Burning_velocity a stanDev_BV_12_106_50 3 standard deviation BV a end if dust conc a 75 burn vel 12 106 75 3 Burning_velocity a stanDev BV 12 106 75 3 standard deviation BV a end end if particleSize 106 amp amp flowRate a 40 if dust conc a 0 burn vel 12 106 00 4 Burning_velocity a stanDev BV 12 106 00 4 standard deviation BV a end if dust conc a 25 burn vel 12 106 25 4 Burning_velocity a stanDev BV 12 106 25 4 standard deviation BV a end if dust conc a 50 burn vel 12 106 50 4 Burning_velocity a stanDev BV 12 106 50 4 standard deviation BV a end if dust conc a 75 burn vel 12 106 75 4 Burning_velocity a stanDev BV 12 106 75 4 standard deviation BV a
161. nnel that encountered the upward moving dust pile and entrained it By forcing the air jet through a thin slot a very high rate of shear was created sufficient to provide the necessary turbulence to dislodge the dust particles The mixture was laminarized by expanding the dust flow through a diffuser A brass elbow fitting had the ability to decrease or increase the dust flow without affecting the dust air dispersion or concentration The ejector connected the main burner tube to a smaller bypass side tube Following the location of the ejector were two sections of stainless steel tubing that made up the main burner tube Resting on the brass connector was a glass tube that encompassed the second upper steel tube Regular dry air was made to flow in this glass tube at relatively low flow rates to provide an enveloping blanket or protective co flow for the exiting dust air flow This co flow existed so that the dust air mixture remained in a laminar column like form once it exited from the conical nozzle and recirculation eddies forming at the nozzle exit could be prevented The dust flow finally exited the tube though a conical brass nozzles which could have varying contraction angles The flame directly stabilized on the nozzle eliminating the uncertainty in flow rate that might occur from gas entrainment into the flame from the surrounding atmosphere beneath a cooling ring 26 Gonzalez et al 48 published results from an inverted burner to stud
162. non zero x ploty plot x aveDataRaw y aveDataRaw indicies clear x matrix Lefty matrix Left x int left y int left clear x matrix Right y matrix Right x int right y int right Burn vel func dst part szev l Plot funct part size 01 if phi a lt lt 0 8 if particleSize 106 amp amp flowRate a 10 if dust conc a burn vel 08 106 00 1 Burning_velocity a stanDev BV 08 106 00 1 standard deviation BV a height ave 08 106 00 1 1 lt height ave a end if dust conc a lt lt 25 139 burn_vel_08_106_25 1 Burning_velocity a stanDev BV 08 106 25 1 standard deviation BV a end if dust conc a 50 burn vel 08 106 50 1 Burning_velocity a stanDev BV 08 106 50 1 standard deviation BV a end if dust conc a lt lt 75 burn vel 08 106 75 1 Burning_velocity a stanDev BV 08 106 75 1 standard deviation BV a end end if particleSize 106 amp amp flowRate a 30 if dust conc a burn vel 08 106 00 2 Burning_velocity a stanDev BV 08 106 00 2 standard deviation BV a end if dust conc a lt lt 25 burn vel 08 106 25 2 Burning_velocity a stanDev BV 08 106 25 2 standard deviation BV a end if dust conc a lt lt 50 burn vel 08 106 50 2 Burning_velocity a stanDev BV 08 106 50 2 standard deviation BV a end if dust conc a lt lt 75 burn vel 08 106 75 2 Burning_velocity a stanDev BV 08 106 75 2 standard deviation BV a
163. o E 2 A g y 0 0029x 0 1037 2 5 0 5 A a 0 O 0 200 400 600 800 1000 Dust feeder calibration Figure 3 14 Dust feeder calibration curves feed rate versus feeder setting from 1 1000 3 7 Optical system The HFA uses a shadowgraph to determine the flame edge of the premixed portion of the hybrid flame Shadowgraphs have been used by a number of researchers to study the burning velocity of gas flames including Sherrat et al 21 Garner et al 22 Anderson et al 23 24 and Whol et al 25 Figure 3 15 shows an example of visual images taken of a a methane air flame and b a hybrid flame including coal dust It is clear that the premixed flame edge cannot be determined from visual images Figure 3 16 shows shadowgraph images of a a methane air only flame and b a hybrid flame including coal dust The flame edges are clear in both of these cases though in b contrast is reduced due to emissions from the coal This effect is further discussed in Chapter 4 65 Figure 3 15 Visual images of burner flames a methane air only 0 8 UW yms 0 532m s b hybrid flame including coal dust Asz 50 g m ds 106 125 um Figure 3 16 Shadowgraph images of burner flame a methane air only b hybrid flame including coal dust As 50 g m d 75 90 um The shadowgraph shown in Fig 3 16 uses the fan cooled bulb 480 W from a projector covered by a steel plate with a pin hole in the center to mak
164. o decrease the proportion of the gas used to disperse the dust thus allowing the concentration to vary Regardless of the flow rate through the central tube the pressure behind the hypodermic needle was normally maintained at 20 psig The generated dust cloud rose vertically through the burner tube 2 cm ID 56 cm length the top 22 cm of which could be heated electronically which was tapped continuously by a 60 cycle electromagnet vibrator Surrounding the upper 30 cm of the burner tube was a 5 cm ID aluminum jacket through which flowed the auxiliary nitrogen stream From there it flowed unrestricted to the top where it was accelerated through a nozzle In order to maintain a stable flame consistently it was necessary to use a flame holding device a brass ring with a conical cross section where the apex of the cone was oriented downward toward the burner The brass ring was customarily heated before igniting the flames Mason et al 39 published results from a laminar steady state dust flame burner in 1967 where fluidizing air was supplied from a humidifying and metering system via a Manostat and control valve The resulting suspension flew up into the diverging section where the flow was divided part passed up the burner tube 10 9 mm bore and the surplus was exhausted The concentration of the suspension could be varied by adjusting the flow of fluidizing air and the flow velocity up the burner tube could be varied independently by alter
165. ocess Industries 2006 19 p 769 773 Horton M D F P Goodson and L D Smoot Characteristics of Flat Laminar coal Dust Flames Combust Flame 1977 28 p 187 195 Dahoe A Dust Explosions a Study of Flame Propagation in Applied Sciences 2000 Delft University of Technology p 298 Eckhoff R K Dust Explosions in the Process Industries 2003 Boston Gulf Professional Publishing Bardon M F and D E Fletcher Dust Explosions Science Progress Oxford 1983 68 p 459 473 Hertzberg M K L Cashdollar and C P Lazzara The Limits of Flammability of Pulverized Coals and other dusts in Proc combust Inst 1981 The combustion Institute 47 10 Turns S R An Introduction to Combustion Concepts and Applications 2000 New York McGraw Hill 48 3 Experimental Apparatus Construction and Procedure 3 1 Summary The primary objective of this study is to develop an experimental platform to accurately measure the turbulent burning velocity of a hybrid flame with the capability of systematic variation of the parameters which influence the problem such as particle size dust type turbulent intensity integral length scale dust concentration and gas phase equivalence ratio To accomplish these goals a new instrument called a Hybrid Flame Analyzer HFA was designed instrumented and constructed during this study This instrument can control the laminar burning velocity Sr turbulent intensity u ms and length scal
166. ompared to the smaller particle size ranges This is mainly due to the decrease in the pyrolysis rate of coal dust particles with an increase in diameter 101 0 8 4 106 125um 8 4 0 u 4 25 z 50 8 A lt 75 1 0 1 0 9 25 50 75 1 2 A 1 0 A 1 25 A 50 0 0 5 1 1 5 A 1 75 u Figure 4 15 Correlations for turbulent burning velocity of hybrid flames d 106 125um The modeling coefficients from Fig 4 5 4 14 and 4 15 are listed in table 4 1 Table 4 1 Modeling coefficients ds 0 C n Gas only 1 1 6 0 3 lt 2 2 02 75 90 21 1 7 0 2 lt 2 0 0 2 106 125 21 1 65 0 2 It is shown that the C coefficient which includes the turbulence effects is increasing with the addition of dust this change is highest in the lean cases when the local equivalence ratio is increased by the addition of fuel vapor from the dust The n coefficient known as the bending 102 coefficient is lower for the experiments involving dust This means that due to the influence of the dust particles the burning velocity is not leveling off as much as the pure gas case the turbulent intensity is increased References 1 Xie Y V Raghavan and A S Rangwala Study of interaction of entrained coal dust particles in lean methane air premixed flames Combust Flame 2012 159 p 2449 2456 2 Kobayashi H T Tamura K Maruta T Niioka
167. on called the Flame Acceleration Tube FAT which was a 70 m long 30 cm diameter tube The FAT was instrumented with static pressure transducers dynamic pressure transducers and photodiodes at eight stations along the tube A four wavelength optical pyrometer was mounted near the end of the tube The initiator consisted of two parts a 2 44 m long and 5 08 cm diameter detonation tube separated from the FAT by a Mylar diaphragm which was filled with a flammable mixture followed by a 3 m long section of the FAT in which dust was dispersed by loading it into a V channel fitted with air injection holes A specially designed cart equipped with a 6 L dust pan an auger two motors and a fan was used to travel inside the FAT to deposit a dust layer with a predetermined thickness and width on the bottom of the FAT In 2001 Sun et al 67 published results using an experimental setup in which a flame could propagate in an open field without any influence from the chamber wall This experiment was comprised of an air supplying part a controller part a combustion chamber an ignition part a laser light source a temperature measurement setup and a high speed video camera with a microscopic optical system The combustion chamber 76 mm inside diameter was provided with an air nozzle a sample dish a pair of ignition electrodes and a movable tube Before the 31 movable tube started to move down the iron dust was dispersed by air into the combustion
168. orig lt burn vel 08 106 00 burn vel 08 106 25 orig lt burn vel 08 106 25 burn vel 08 106 50 orig lt burn vel 08 106 50 burn vel 08 106 75 orig lt burn vel 08 106 75 burn vel 10 75 00 orig lt burn vel 10 75 00 burn vel 10 75 25 orig lt burn vel 10 75 25 burn vel 10 75 50 orig lt burn vel 10 75 50 burn vel 10 75 75 orig lt burn vel 10 75 75 burn vel 10 106 00 orig lt burn vel 10 106 00 burn vel 10 106 25 orig lt burn vel 10 106 25 burn vel 10 106 50 orig burn vel 10 106 50 burn vel 10 106 75 orig bum vel 10 106 75 burn vel 12 75 00 orig lt burn vel 12 75 00 burn vel 12 75 25 orig lt burn vel 12 75 25 burn vel 12 75 50 orig lt burn vel 12 75 50 burn vel 12 75 75 orig lt burn vel 12 75 75 burn vel 12 106 00 orig lt burn vel 12 106 00 burn vel 12 106 25 orig lt burn vel 12 106 25 burn vel 12 106 50 orig lt burn vel 12 106 50 burn vel 12 106 75 orig lt burn vel 12 106 75 lam data 1 if 1 include laminar data if 2 exclude laminar data lam data dst 526 1 burn vel 08 75 00 lt burn vel 08 75 00 orig lam data end burn vel 08 75 00 orig lam data end burn vel 08 75 25 lt burn vel 08 75 25 orig lam data end burn vel 08 75 00 orig lam data end burn vel 08 75 50 burn vel 08 75 50 orig lam data end burn vel 08 75 00 orig lam data end burn vel 08 75 75 lt burn vel 08 75 75 orig lam data end burn vel 08 75 00 orig lam data end 1 bu
169. pairs addErrorBars if 1 error bars are added if 0 not Default 1 horzAvg if 1 average is calculated horizotally along x instead of vertically Default 0 interpMethod interpolation method for estimating data values in between support points See help interpl for supported methods Default linear Use interpMethod hist if you want to take the average of all points in the vicinity of the data good for scattered data points plot2NewFigure if 1 average is plotted in separate figure If 0 average is plotted on top of the individual data lines If 2 or an axes handle the average lines of all the plots are collected in the same figure Default 0 useRobustMean if 1 the robust mean is taken discarding outliers for the average curve If 0 the normal mean is used Default 1 plotSEM if 1 SEM if 0 the standard deviation is plotted Default 1 OUTPUT avgH handle s to average line plus errorbar handle if applicable avgData cell array with x y err n array of x values y values 177 standard deviation not std of the mean of the average line and number of inlier lines for each data containing axes Divide err by sqrt n for SEM REMARKS 1 This function only works for 2D plots it ignores axes where the View is not set to 0 90 2 Since the function looks for axes children of type line it won t work for e g bar plots
170. plotAll figurel figure axes axes Parent figurel YDir reverse hold axesl all end ct_l 1 while ct 1 lt SS 2 x_plot_nz nonzeros x_pix_save ct_1 y_plot_nz y_pix_save 1 length x_plot_nz ct_1 X x_plot_nz y y_plot_nz x_left_min ct_1 x 1 x_right_min ct_1 x end y_left_min ct_1 y 1 y_right_min ct_1 y end count 02 1 while count 02 lt length x point dist count 02 sgrt x count 02 1 x count 02 2 y count 02 1 y count 02 2 count 02 count 02 1 end dist tot ct 1 sum point dist pix to m if plotAll 1 plot x plot nz y plot nz s end height ct 1 max y plot nz min y plot nz pix to m clear x plot nz y plot nz ct 1 lt ct 1 1 end 137 if sol_method 2 lavgH avgData plotAverage pause 0 5 avgH avgData plotAverage noPlot averageData cell2mat avgData x_aveDataRaw averageData 1 y_aveDataRaw averageData 2 calculate starting position of flames on each side x_start_left mean x_left_min x start right mean x right min y start left mean y left min y start right mean y right min if y_start_left gt y start right y start right y start left end if y start right gt y start left y start left y start right end indicies find x aveDataRaw gt x start left amp x aveDataRaw lt x start right num pt ave 20 Interpolate between average starting point and first av
171. r tube two different water cooled nozzles with internal diameters of 14 5 mm are attached to the top of the feeder tube as shown in Fig 3 4 The first nozzle a is a straight tube used for creating laminar flames Laminar flames are generated using a combined air methane flow rate of 10 lpm The second nozzle b uses a set of perforated plates to generate turbulence and has a methane oxygen annular pilot to anchor the flame The details of construction of the turbulent burner nozzle are illustrated in Figs 3 5 and 3 6 The turbulent flame a fueled by the dust air mixture j is anchored to the burner nozzle using a 33 methane oxygen pilot flame b The nozzle tip is water cooled using 1 8 copper tubing c The pilot flame fuel air mixture e is injected through pilot fuel housing d with a 0 635 cm 4 Swagelok tube to MNPT fitting not shown The stainless steel housing g with an internal diameter 1 of 14 5 mm was cooled using 0 318cm 1 8 copper tubing f similar to Bradley et al 1 1994 and Kobayashi et al 5 Turbulence is generated by nylon perforated plates h mounted 10 30 mm from the nozzle exit k shown in Fig 3 8 The pilot flame housing shown in a close up view is made up of three sequential copper tubes The inner tube g has the same inner diameter as the stainless steel tube g and is 0 036 cm 0 014 thick The 2 tube 0 is an insert which both creates the uniform high speed flow around th
172. re through the wrinkled laminar flame area Ar m p A SPA 4 1 A 4 2 S A Damkohler 151 proposed that the ratio of the area of the vvrinkled laminar flame and the cross section of the turbulent flame brush could be approximated by A S U ms 1 u rms 4 3 A S S 4 4 In the limit u s gt gt Sr Eq 4 4 implies that the turbulent burning velocity becomes independent of the laminar burning velocity and the chemistry has no effect on the propagation velocity This is known as the Damkohler hypothesis Schelkin 6 proposed another approximation for the surface of the wrinkled laminar flame by reasoning that turbulence creates conical bulges in a laminar flame and that the increased flame surface is proportional to the average cone area divided by the average cone base If the radius of the cone base and the cone height are respectively denoted by R and h then the surface area 1 2 of the cone base and the cone mantle are egual to 77 and zR R N Thus when a circular 83 element of a planar laminar flame is bulged into a cone the surface area increases by a factor R h 2 R Schelkin 6 assumed that the diameter of the cone base is proportional to the L S rms 1 L average length scale of the turbulence R oc 17 27 and that the apothem scales as hocu He considered the apothem to be proportional to the average fluctuating velocity u s and the time during which an element of the
173. ries that manufacture transport process or use combustible dusts Identification of the controlling parameters of dust deflagration mechanisms is crucial to our understanding of the problem The objective of this study is to develop an experimental platform called the Hybrid Flame Analyzer HFA capable of measuring the laminar and turbulent burning velocity of gas dust and hybrid gas and dust air premixed flames as a function of properties specific to the reactants such as dust particle size and concentration In this work the HFA is used to analyze a particle gas air premixed system composed of coal dust particles 75 90 um and 106 120 um in a premixed CHa air 4 0 8 1 0 and 1 2 flame This work ultimately aims to improve the knowledge on fundamental aspects of dust flames which is essential for the development of mathematical models This study is the first of its kind where multiple different parameters that govern flame propagation initial particle radius particle concentration gas phase equivalence ratio turbulent intensity and integral length scale are systematically analyzed in a spatially uniform cloud of volatile particles forming a stationary flame The experiments show that the turbulent burning velocity is more than two times larger than the laminar counter part for each and every case studied It is observed that smaller particles and larger concentrations gt 50 g m tend to enhance the turbulent burning velo
174. rn vel 08 106 00 lt burn vel 08 106 00 orig lam data end burn vel 08 106 00 orig lam data end burn vel 08 106 25 lt burn vel 08 106 25 orig lam data end burn vel 08 106 00 orig lam data end burn vel 08 106 50 lt burn vel 08 106 50 orig lam data end burn vel 08 106 00 orig lam data end burn vel 08 106 75 lt burn vel 08 106 75 orig lam data end burn vel 08 106 00 orig lam data end 1 burn vel 10 75 00 bum vel 10 75 00 orig lam data end burn vel 10 75 00 orig lam data end burn vel 10 75 25 lt burn vel 10 75 25 orig lam data end burn vel 10 75 00 orig lam data end burn vel 10 75 50 lt burn vel 10 75 50 orig lam data end burn vel 10 75 00 orig lam data end burn vel 10 75 75 lt burn vel 10 75 75 orig lam data end burn vel 10 75 00 orig lam data end burn vel 10 106 00 lt burn vel 10 106 00 orig lam data end burn vel 10 106 00 orig lam data end burn vel 10 106 25 lt burn vel 10 106 25 orig lam data end burn vel 10 106 00 orig lam data end burn vel 10 106 50 lt burn vel 10 106 50 orig lam data end burn vel 10 106 00 orig lam data end burn vel 10 106 75 burn vel 10 106 75 orig lam data end burn vel 10 106 00 orig lam data end 1 burn vel 12 75 00 lt burn vel 12 75 00 orig lam data end burn vel 12 75 00 orig lam data end burn vel 12 75 25 lt burn vel 12 75 25 orig lam data end burn vel 12 75 00 orig lam data end burn vel 12 75 50 lt
175. st_conc_12_75_35_dst_sze ct_25 dust_conc a burn_vel_12_75_35_dst_sze ct_25 Burning_velocity a ct 25 ct 25 1 end if particleSize 75 amp amp flowRate a 40 dust conc 12 75 40 dst sze ct 26 dust_conc a bum vel 12 75 40 dst sze ct 26 Burning_velocity a ct 26 ct 26 1 end end 172 A3 8 Turbulent intensity calculation clear all close all cle format long folderName Imm perf plate test data only filePath E NHFA test data Turbulent Intensity measurements dname filePath V folderName 90 Set up basic file name path to read top_file dname VI Set up main database to open and look inside s top file Is top file 9 List Files inside main folder z cellstr ls top file Turn cells from Is function into strings cc c 3 length c Set up a matrix without the and produces by the s function S size cc 9Find the size of matrix containing names of files inside of main database a zl This counter is set to 3 to account for the and at the beggining of each matrix created by Is ct 01 1 ct 02 lt 1 ct 03 lt 1 ct 04 1 ct 05 1 ct 06 1 while a lt lt S 1 close all file char cellstr top file char cc a File to be operated on data_n char cc a fileName char cc a nozzleDiameter 0 0145 m flowRate str2num fileName 12 13 position str2num fileName 17 height str2num fileName 22
176. stremp get handleOrData ih type figure chH get handleOrData ih Children rm legends 178 legendIdx stremp legend get chH Tag ahList ahList chH legendIdx ok lt AGROW gt end end end if isempty ahList error no valid axes handles found in handleOrData or children thereof end check for other optional inputs if nargin lt 2 avgPoints end if isEven length varargin error options must be specified as parameter name parameter value pairs end for i 1 2 length varargin opt varargin i varargin i 1 end turn off robutsMean warning oldWarn warning warning off ROBUSTMEAN INSUFFICIENTDATA CALCULATE AVERAGE nAh length ahList data 1 nAh struct x Data yData avgPoints avgPoints ahln num2cell ahList ahOut avgData for ia nAh 1 1 count down in case we remove entries find data in axes chH get data ia ahIn Children remove errorBars not lines chH stremp line get chH Type ismember get chH Tag errorBar avg if isempty chH if no valid children discard axes data ia else get data if length chH data ia xData get chH XData data ia yData get chH Y Data else data ia xData get chH X Data data ia yData get chH Y Data end end end nData length data if nData lt 1 error no line plots found in the axes provided end 179 det
177. t Inst 1957 Kolbe M Laminar Burning Velocity Measurements of Stabilized Aluminum Dust Flames in Mechanical Engineering 2001 Concordia University Montreal Quebec Canada Cassel H M A K D Gupta and S Guruswamy Factors Affecting Flame Propagation Through Dust Clouds Third Symposium on Combustion Flame and Explosion Phenomena 1949 p 185 190 Ghosh B D Basu and N K Roy Studies of Pulverized Coal Flames in Proc Combust Inst 1957 Kobayashi H J B Howard and A F Sarofim Coal devolatilization at high temperatures Proc Combust Inst 1977 16 p 411 415 Reddy P P R Amyotte and M J Pegg Effect of inerts on layer ignition temperature of coal dust Combust Flame 1988 114 p 41 53 Sherrat S and J W Linnett The determination of flame speeds in gaseous mixtures Trans Faraday Soc 1948 44 p 596 608 Garner F H R Long and G R Ashforth Determination of burning velocities in benzene air mixtures Fuel 1949 28 12 p 272 276 Anderson J W and R S Fein Measurements of normal burning velocities and flame temperatures of Bunsen flames J Chem Phys 1949 17 p 1268 1273 Anderson J W and R S Fein Measurment of normal burning velocities of propant air flames from shadow photographs J Chem Phys 1950 19 p 441 443 Whol K N P Kapp and C Gazley The stability of open flames Proc Combust Inst 1949 3 p 3 20 74 26 27 28 29 30 31 32 33
178. t data yfit 10 106 25 yfit 10 75 00 r LineWidth plotLineWidth plot xfit data yfit 10 106 50 yfit 10 75 00 9g LineWidth plotLineWidth plot xfit data yfit 10 106 75 yfit 10 75 00 b LineWidth plotLineWidth hold off xlabel u u bar 0 D pp FontSize textSize ylabel S T S L S T S L gas only phi 1 0 FontSize textSize axis 10 09 0 15 1 1 25 figure2 figure axes2 axes Paren figure2 YMinorTick on XMinorTick on FontSize testSize2 168 hold on if legend_plot 1 10 1 1 ks MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 rs MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 gs MarkerSize plotMarker ize Line Width plotLineWidth plot 1 1 bs MarkerSize plotMarkerSize LineWidth plotLineVVidth plot 1 1 rv MarkerSize plotMarkerSize LineVVidth plotLineVVidth plot 1 1 gv MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 bv MarkerSize plotMarkerSize LineWidth plotLineWidth legend gas only d 75 d_ st 25 d 75 d_ st 50 d 75 d_ st 75 d 106 d_ st 25 d 106 d_ st 50 d 106 d_ st 75 Location eastoutside end subplot 3 2 5 hold on plot l 0 12 burn vel 12 75 00 ks LineWidth plotLineWidth plot l 0 12 burn vel 12 75 25 ro LineWidth plotLineWidth plot l 0 12 burn vel 12 75 50 gv LineWidth plotLineWidth plo
179. t l 0 12 burn vel 12 75 75 bh LineWidth plotLineWidth C 12 00 40 n 12 00 25 Su L 12 00 burn_vel_12_75_00_orig 1 S_TL_12_00 1 C_12_00 u_prime_all 2 end Su_L_12_00 4n_12_00 plot l 0 12 5 TL 12 00 k LineWidth plotLineWidth hold off subplot 3 2 6 hold on plot l 0 12 burn vel 12 106 00 ks Line Width plotLineWidth plot l 0 12 burn vel 12 106 25 ro LineWidth plotLineWidth plot l 0 12 burn vel 12 106 50 gsv LineWidth plotLineWidth plot l 0 12 1 1ength burn vel 12 106 75 burn vel 12 106 75 bh LineWidth plotLineWidth hold off p get 0 monitorpositions p 1 1 1000 450 1600 270 1000 950 set gcf position p 2 plot xfit data yfit 12 75 00 K LineWidth plotLineWidth plot xfit data yfit 12 75 25 yfit 12 75 00 r LineWidth plotLineWidth plot xfit data yfit 12 75 50 yfit 12 75 00 g LineVVidth plotLineVVidth plot xfit data yfit 12 75 75 yfit 12 75 00 b LineWidth plotLineWidth plot xfit data yfit 12 106 25 yfit 12 75 00 r LineWidth plotLineWidth plot xfit data yfit 12 106 50 yfit 12 75 00 g LineWidth plotLineWidth plot xfit data yfit 12 106 75 yfit 12 75 00 b LineWidth plotLineWidth hold off xlabel u u bar 0 D pp FontSize textSize ylabel S T S L S T S L gas only phi 1 2 FontSize textSize axis 0 09 0 15 1 1 25 169 A3 7 Cre
180. t sze end OO am data dst sze end 1 burn vel 08 75 40 dst sze orig burn vel 08 75 10 dst sze origFitdam data dst sze end 1 bh MarkerSize plotMarkerSize LineWidth plotLineWidth ylim y axisMin y axisMax hold off subplot subplot 3 2 2 Parent figure3 Y Tick 2 2 4 2 8 3 2 3 6 4 LineWidth 2 FontWeight bold FontSize 14 FontName Times New Roman subplot 3 2 2 157 hold on plot dust conc 08 106 10 dst sze burn vel 08 106 10 dst sze kv MarkerSize plotMarkerSize LineWidth plot LineWidth plot dust conc 08 106 30 dst sze lam data dst sze end l OO am data dst sze end burn vel 08 106 30 dst 526 orig burn vel 08 75 10 dst sze origFit ro MarkerSize plotMarkerSize Line Width plotLineWidth plot dust 08 106 35 dst sze lam data dst sze end l 00 1am data dst sze end burn vel 08 106 35 dst sze orig burn vel 08 75 10 dst sze origFit sv MarkerSize plotMarkerSize Line Width plotLineWidth plot dust conc 08 106 40 dst sze lam data dst sze end l 00 1am data dst sze end 1 burn vel 08 106 40 dst sze orig burn vel 08 75 10 dst sze origFit lam data dst sze end 1 bh MarkerSize plotMarkerSize LineWidth plotLineWidth ylim y axisMin y axisMax hold off subplot subplot 3 2 3 Parent figure3 Y Tick 2 2 4 2 8 3 2 3 6 4 LineVVidth 2 FontWeight bold
181. tLineWidth hold off xlabel u u bar O D pp FontSize textSize ylabel 5 T S L S T S L gas only phi 1 0 FontSize textSize axis 10 09 0 15 1 1 251 figure2 lt figure axes2 axes Parent figure2 YMinorTick on XMinorTick on FontSize testSize2 hold on if legend plot 1 plot 1 1 ks MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 ts MarkerSize plotMarkerSize Line Width plotLineWidth plot 1 1 gs MarkerSize plotMarkerSize LineW idth plotLineWidth plot 1 1 bs MarkerSize plotMarkerSize LineW idth plotLine Width plot 1 1 rv MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 gv MarkerSize plotMarkerSize Line Width plotLine Width plot 1 1 bv MarkerSize plotMarkerSize LineWidth plotLineWidth legend gas only d 75 d_ st 25 d 75 d_ st 50 d 75 d_ st 75 d 106 d_ st 25 d 106 d_ st 50 d 106 d_ st 75 Location eastoutside end subplot 3 2 5 hold on plotd_0_12 burn_vel_12_75_00 ks LineWidth plotLineWidth plot l 0 12 burn vel 12 75 25 ro LineWidth plotLineWidth plot l 0 12 burn vel 12 75 50 gv LineWidth plotLineWidth plot l 0 12 burn vel 12 75 75 bh LineWidth plotLineWidth C 12 00 40 12 00 25 Su 1 12 00 burn_vel_12_75_00_orig 1 S_TL_12_00 1 C_12_00 u_prime_all 2 end Su_L_12_0
182. tain deflagration Hence from a practical standpoint it is important to develop laboratory experiments which can accurately characterize and re create turbulence levels similar to those found in accidental explosions Currently there is no methodology to incorporate or measure the turbulent burning velocity or the necessary parameters to quantify the turbulence and lo using the current design of the explosion sphere It thus becomes questionable to use the explosion sphere and relate the P vs t curve to industrial explosions in a meaningful way Dahoe et al 27 show that no formal cube root law agreement could be found between a 20 L sphere and a 1000 L sphere explosion vessel though there has been some success with normalization using an estimation of flame thickness by Dahoe 25 This discrepancy limits the application of the cube root law in the design of deflagration venting and further shows the need to quantify the levels of turbulence and the corresponding effect on flame speed 16 Pressure measurement Laminar Highly Turbulent Turbulent Figure 1 3 Diagram of explosion sphere with increasing turbulence as the flame propagates The deflagration index has also been used to estimate the laminar burning velocity for a given dust concentration and particle size using expressions such as 28 K Sa E si 1 2 ssd mx Arm o where 4 84 is an empirical constant Ping is the maximum pressure and P is the
183. tanDev BV 10 106 50 2 standard deviation BV a end if dust conc a lt lt 75 burn vel 10 106 75 2 Burning_velocity a stanDev BV 10 106 75 2 standard deviation BV a end end if particleSize 106 amp amp flowRate a 35 if dust conc a burn vel 10 106 00 3 Burning_velocity a stanDev BV 10 106 00 3 standard deviation BV a end if dust conc a lt lt 25 burn vel 10 106 25 3 Burning_velocity a stanDev BV 10 106 25 3 standard deviation BV a end if dust conc a lt lt 50 burn vel 10 106 50 3 Burning_velocity a stanDev BV 10 106 50 3 standard deviation BV a end if dust conc a 75 burn vel 10 106 75 3 Burning_velocity a stanDev BV 10 106 75 3 lt standard deviation BV a end end if particleSize 106 amp amp flowRate a 40 if dust conc a burn vel 10 106 00 4 Burning_velocity a stanDev BV 10 106 00 4 standard deviation BV a end if dust conc a lt lt 25 burn vel 10 106 25 4 Burning_velocity a stanDev BV 10 106 25 4 standard deviation BV a end if dust_conc a 50 burn_vel_10_106_50 4 Burning_velocity a stanDev BV 10 106 50 4 standard deviation BV a end if dust conc a lt lt 75 burn vel 10 106 75 4 Burning_velocity a stanDev BV 10 106 75 4 standard deviation BV a end end if particleSize 75 amp amp flowRate a 10 if dust conc a burn vel 10 75 00 1 lt Burning velocity a stanDev BV 10 75 00 1
184. tely vaporize in the preheat zone The color gradients shown indicate the mass fraction of fuel vapor present with a darker color representing higher mass fraction The inset labeled D shows a close up of the ambient zone where the random distribution of both particle separation and size in a potential dust air flame is highlighted The inset labeled E shows a close up of the preheat zone During this process the differences in particle size will play a significant role as smaller particles get heated up faster and vaporize almost completely while larger particles continue to be in the condensed phase as they move into the reaction and convection zones The inset labeled F shows a close up view of a single vaporizing particle The inset labeled G shows the surface of a particle in the preheat zone where the fuel changes phase from solid to gas and premixes with the oxidizer to establish a 42 flame front At this stage it is possible that the burning is localized on the surface alone however this condition is not analyzed in the current study It should be noted that the change in phase slows down the burning velocity significantly as compared to a gas flame Further as shown in G the heat transfer fluxes related to in depth conduction 4 and radiation Grain Arad also play a significant role The inset also shows the vaporization rate m v w which is determined by an energy bal
185. the data marker 80 Reasonably good agreement is observed between the two experimental methods Similar to Kobayashi et al s data the burning velocity increases as turbulent intensity increases and eventually begins to level off when higher levels of turbulent intensity are reached It is also interesting to note that Kobayashi et al used the angle method c f Fig 3 on pg 391 of 2 to extract the turbulent burning velocity from schlieren images of turbulent flames whereas in this study an alternative method similar to Grover et al 3 is used as discussed in Chapter 3 and provides similar results O Kobayashi et al This work 0 0 1 0 2 03 04 0 5 U s Figure 4 3 Turbulent burning velocity of a methane air flame 1 0 vs turbulent intensity Figure 4 4 a d shows a comparison of the shadowgraph results for the laminar and turbulent gas only flames Figure 4 4 a shows the smooth and clearly defined edge of a laminar flame Figures 4 4 b d showing turbulent flames with increasing turbulent intensity clearly show the wrinkled combustion zone Further flame wrinkling is observed to increase as the turbulent intensity is increased The increased wrinkling causes an increase in the reaction zone area which 81 means that the flame structure can consume the fuel air mixture at a faster rate This results in an increase in the value of the burning velocity as shown in Fig 4 3 Damkohler 5 was one of the first
186. thod for Percent Dispersibility vi ASTM E 2019 03 2007 Standard Test Method for Minimum Ignition Energy of a Dust Cloud in Air vii ASTM E 2021 06 Standard Test Method for Hot Surface Ignition Temperature of Dust Layers viii ASTM E 2079 07 Standard Test Methods for Limiting Oxygen oxidant Concentration in Gases and Vapors ix IEC 60093 Methods of test for volume resistivity and surface resistivity of solid electrical insulating materials x IEC 61340 2 1 2002 06 Measurement methods Ability of materials and products to dissipate static electric charge xi IEC 61340 2 2 2000 067 Measurements methods Measurement of chargeability 110 Appendix 2 HFA User s Manual Hybrid Flame Analyzer HFA User s Manual v01 Last revised 2012 Combustion Lab Salsbury Lab 214 Worcester Polytechnic Institute Worcester MA 01609 111 Potential Dangers of this instrument Glass breaking Electrical shock Burning Respiratory Irritation Explosion Suffocation Instructional videos 15 instructional videos were made to help students learn how to use the HFA These are available at www firesciencetools com on the hybrid flame analyzer page 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Hybrid Flame Analyzer startup sequence Hybrid Flame Analyzer shutdovvn sequence Running laminar flame tests Electrical system for Hybrid Flame Analyzer Exhaust system for Hybrid Flame
187. ting figure 4 9 plot_ND_SLdivSLgasOnlyv01 plot_ND_SLdivSLgasOnlyv02_6fig plot_ND_SLv03_6fig legend_plot 0 plotMarkerSize 10 plotLineWidth 3 testSize2 14 x_axisMin 0 x_axisMax 80 y_axisMin 1 9 y_axisMax 4 08 00 u_prime_all 2 end burn_vel_08_75_10_dst_sze_origFit 1 08_25 u_prime_all 2 end burn_vel_08_75_10_dst_sze_origFit 1 08 50 u_prime_all 2 end burn_vel_08_75_10_dst_sze_origFit 1 08_75 u_prime_all 2 end burn_vel_08_75_10_dst_sze_origFit 1 S ooo 10 00 u prime all 2 end burn vel 10 75 10 dst sze origFit 1 10 25 lt u prime all 2 end burn vel 10 75 10 dst sze origFit 1 10 50 u prime all 2 end burn vel 10 75 10 dst sze origFit 1 10 75 u prime all 2 end burn vel 10 75 10 dst 526 origFit 1 S ooo 12 00 lt u prime all 2 end burn vel 12 75 10 dst sze origFit 1 12 25 u prime all 2 end burn vel 12 75 10 dst sze origFit 1 12 50 u prime all 2 end burn vel 12 75 10 dst sze origFit 1 12 75 u prime all 2 end burn vel 12 75 10 dst sze origFit 1 phi 0 8 figurel figure Name NDim turbulent velocity axes2 axes Parent figurel YMinorTick on XMinorTick on FontSize testSize2 hold on subplot subplot 3 2 1 Parent figurel
188. to study the effects of coal dust on the burning velocity of CH4 air flames as a function of particle size particle concentration turbulent intensity and gas phase equivalence ratio The burning velocity of the premixed section of the turbulent flames was calculated using the area method as found in the literature The turbulent flow was characterized and shown to fall in the laminar flamelet regime of the Borghi diagram The main conclusions based on the experiments are 1 The turbulent burning velocity is more than tvvo times larger than the laminar counter part for each and every case studied The turbulent to laminar burning velocity ratio increases as the turbulent intensity is increased for all cases The ratio of turbulent to laminar flame velocity is seen to increase significantly as particle size decreases and dust concentration increases This is primarily due to the effect of an increase in the turbulence level due to the interaction of smaller sized particles which also increases with increased number of particles present at higher concentrations While the increasing trend is observed for all three equivalence ratios tested it is highest for the fuel lean cases since there is also an increase in the local equivalence ratio n S U n ki can be used to correlate the Lost Lost An empirical correlation of the form experimental data where C amp n are functions of the gas phase equivalence ratio integral
189. ts varargin PLOTAVERAGE plots an average line into a plot and more SYNOPSIS avgH avgData plotAverage handleOrData avgPoints parameterName parameterValue INPUT handleOrData handle to figure or axes of the plot to average Can be vectors of figures or of axes handles In a figure with multiple subplots the average is calculated for each subplot individually Alternatively provide a cell array with x1 y1 x2 y2 where xi yi are vectors of different data sets With the latter form a plot is generated with figure plot x1 y1 x2 y2 Optional If empty plotAverage calls gcf to find the current figure avgPoints points on the x axis or y axis see below where the average is to be calculated If empty the points are selected by locally clustering data points and robustly averaging of the position within each cluster This works best if the data on the corresponding axes indeed cluster into more or less evenly spaced clusters If this is not the case it is probably better to input avgPoints If avgPoints is a scalar N the axis is split into N equally spaced points between the minimum and the maximum of the data excluding the minimum and maximum Note If you want to specify separate avgPoints for each of the axes handles passed to plotAverage pass avgPoints as a cell array plotAverage supports the following parameterName parameter Value
190. u prime all burn vel 08 106 00 orig 1 burn vel 08 106 00 orig burn vel 08 75 10 dst sze origFit 1 kv LineWidth plotLine Width plot u_prime_all burn_vel_08_106_25_orig 1 burn_vel_08_106_25_orig burn_vel_08_75_10_dst_sze_origFit 2 rv LineWidth plotLineWidth plot u_prime_all burn_vel_08_106_50_orig 1 burn_vel_08_106_50_orig burn_vel_08_75_10_dst_sze_origFit 3 ev LineWidth plotLine Width 165 plot u_prime_all 1 length burn_vel_08_106_75_orig burn_vel_08_106_75_orig 1 burn_vel_08_106_75_orig bu rn vel 08 75 10 dst sze origFit 4 bv Line Width plotLineWidth plot u_prime_all burn_vel_10_75_00_orig 1 burn_vel_10_75_00_orig burn_vel_10_75_00_orig 1 kd Line Widt h plotLine Width plot u_prime_all burn_vel_10_75_25_orig 1 burn_vel_10_75_25_orig burn_vel_10_75_25_orig 1 rd LineWidt h plotLine Width plot u_prime_all burn_vel_10_75_50_orig 1 burn_vel_10_75_50_orig burn_vel_10_75_50_orig 1 gd Line Widt h plotLine Width plot u prime all burn vel 10 75 75 orig 1 burn vel 10 75 75 orig burn vel 10 75 75 orig 1 bd LineWidt h plotLineWidth plot u prime all burn vel 10 106 00 orig 1 burn vel 10 106 00 orig burn vel 10 75 10 dst sze origFit 1 ko LineW idth plotLineWidth plot u prime all burn vel 10 106 25 orig 1 burn vel 10 106 25 orig burn vel 10 75 10 dst sze origFit 2 ro LineWidth plotLineWidth plot
191. u prime all burn vel 10 106 50 orig 1 burn vel 10 106 50 orig burn vel 10 75 10 dst sze origFit 3 go LineWidth plotLineWidth plot u prime all 1 length burn vel 10 106 75 orig burn vel 10 106 75 orig 1 burn vel 10 106 75 orig bu rn vel 10 75 10 dst sze origFit 4 bo LineWidth plotLineW idth plot u prime all burn vel 12 75 00 orig 1 burn vel 12 75 00 orig burn vel 12 75 00 orig 1 k LineWidt h plotLineWidth plot u prime all burn vel 12 75 25 orig 1 burn vel 12 75 25 orig burn vel 12 75 25 orig l r LineWidt h plotLineWidth plot u prime all burn vel 12 75 50 orig 1 burn vel 12 75 50 orig burn vel 12 75 50 orig 1 8 LineWidt h plotLineWidth plot u prime all burn vel 12 75 75 orig 1 burn vel 12 75 75 orig burn vel 12 75 75 orig 1 b LineWidt h plotLineWidth plot u prime all burn vel 12 106 00 orig 1 burn vel 12 106 00 orig burn vel 12 75 10 dst sze origFit 1 kh LineW idth plotLineWidth plot u prime all burn vel 12 106 25 orig 1 burn vel 12 106 25 orig burn vel 12 75 10 dst sze origFit 2 rh LineWidth plotLineWidth plot u prime all burn vel 12 106 50 orig 1 burn vel 12 106 50 orig burn vel 12 75 10 dst sze origFit 3 gh LineWidth plotLine Width plot u prime all 1 length burn vel 12 106 75 orig burn vel 12 106 75 orig 1 burn vel 12 106 75 orig bu rn vel 12 75 10 dst sze origFit 4 bh LineVVidth p
192. use of two counter flowing jets exiting from two opposing burners The particle seeder utilizes a piston which was attached beneath the bottom burner and fed the particles into the flow at a constant rate Chemically inert aluminum oxide and nickel alloy particles were used The particle mass delivery was determined by both the piston speed and the flow rate The gas flow enters the top of the piston shaft through sixteen 1 mm diameter holes equally spaced around the shaft which locally increase the gas velocity and improve the entrainment of particles into the flow This design allowed for seeding under both normal and microgravity However it should be obvious that the particle pickup was strongly affected by gravitational forces The particle seeder was calibrated by seeding the particles into the air flow for a specified time and measuring the mass collected 25 Kolbe 47 published results from a new steady state dust burner in 2001 The typical experimental time span was approximately 5 6 minutes from which a stable flame could be achieved for up to four minutes The cylindrical steel hopper in which the dust is contained guided the piston which pushed the dust sample upwards The piston speed was controlled by means of an electro mechanical actuator Another cylindrical housing in which air is fed surrounded this contraction As both dust and air traveled separately upward they were mixed when the air is forced into a circumferential cha
193. velocity a ct 17 ct_17 1 end if particleSize 75 amp amp flowRate a 40 dust_conc_10_75_40_dst_sze ct_18 dust_conc a burn_vel_10_75_40_dst_sze ct_18 Burning_velocity a ct 18 ct_18 1 end end if phi a 1 2 if particleSize 106 amp amp flowRate a 10 dust_conc_12_106_10_dst_sze ct_19 dust_conc a burn_vel_12_106_10_dst_sze ct_19 Burning_velocity a ct 19 ct 19 1 end if particleSize 106 amp amp flowRate a 30 dust 12 106 30 dst sze ct 20 dust burn vel 12 106 30 dst sze ct 20 Burning velocity a ct 20 ct 20 1 end if particleSize 106 amp amp flowRate a 35 dust conc 12 106 35 dst sze ct 21 dust conc a burn vel 12 106 35 dst sze ct 21 lt Burning velocity a ct 21 lt ct 21 1 end if particleSize 106 amp amp flowRate a 40 171 dust_conc_12_106_40_dst_sze ct_22 dust_conc a burn_vel_12_106_40_dst_sze ct_22 Burning_velocity a ct_22 ct_22 1 end if particleSize 75 amp amp flowRate a 10 dust_conc_12_75_10_dst_sze ct_23 dust_conc a burn_vel_12_75_10_dst_sze ct_23 Burning_velocity a ct 23 ct_23 1 end if particleSize 75 amp amp flowRate a 30 dust_conc_12_75_30_dst_sze ct_24 dust_conc a burn_vel_12_75_30_dst_sze ct_24 Burning_velocity a ct 24 ct 24 I end if particleSize 75 amp amp flowRate a 35 du
194. ving to look at the flowmeter Between the output of the flowmeter and the burner nozzle there is a section of 1 8 copper tubing as shown below This is used to keep any potential pressure buildup in the system in the sink area rather than at the nozzle to avoid water leaks un the on the apparatus itself Plug in 114 Black chord for battery power strip Black chord for mass flow controllers The HFA has two different power strips to provide power one is located on the 2 shelf the other is located on the lab bench behind the computer Turn on exhaust fan Turn on cooling fan for shadowgraph light point source Turn on Camera power Turn on heat sink if using gas analysis it will require time to reach steady state temperature For ease of use the power for these parts of the experiment are routed through a set of switches as shown below in ovver Hood Turn on computer Plug in timing hub usb port 115 Plug in gas analyzer USB port Turn off driver for the assumed mouse if needed Start hyperterminal Turn on Mass flow controllers allow 15 min to warm up Turn on gas bottle valves valves Air set to 25psi and open shutoff valve all the way Methane 10 psi Oxygen 10 psi 116 Gas analyzer is usually left on 117 118 Calibration Gas Analyzer Nitrogen for zero Specialized tanks for various analyzer Depending on the analyzer the gas sensors need to be calibrated onc
195. where ds is the particle diameter and is the integral length scale The corresponding change is turbulent intensity is shown in Fig 4 8 13 As observed in Fig 4 8 when the ratio of d l is above 0 07 the presence of particles increases the turbulent intensity In the current experiments for the 75 90 um range dy l varies between 0 06 0 08 while for the 106 125 um range 4 varies between 0 07 0 11 Thus it can be concluded that if only fluid dynamics no combustion effects are are considered the particles will tend to increase the turbulent intensity Further Crowe 14 has shown that the increase in turbulent intensity becomes more pronounced as concentration of particles is increased cf Fig 3 in 14 YoAu d Range of current experiments Figure 4 8 Change in turbulent intensity as a function of length scale ratio 13 93 4 2 3 Effect of turbulence on burning velocity of a hybrid dust CHa air flame Figures 4 9 a f show relationships of the experimentally measured turbulent burning velocity for the different equivalence ratio and particle size ranges tested dy and A represent the particle size range and particle concentration respectively The Y axis is nondimensionalized by the laminar burning of the corresponding dust flame Ss while the X axis 15 nondimensionalized using the laminar burning velocity of the gas flame S at corresponding equivalence ratio In the following f
196. with cellulose based dusts like those found the food processing industry While varying the types of dust the effect of inert particles should be examined which would help explain the effects of the coal dust used in this work Suppressants should also be added to the flow stream to see what their effects are on the turbulent hybrid flame The dust concentration should be varied the range can be increased by changing the size of the helix in the dust feeder Different dust sizes should be tested a new helix will be required for the small dusts as they will not feed through the current design The effect of the integral length scale should be examined next To do this the author recommends 106 matching turbulent intensity values with different perforated plates thereby decoupling the effect of the length scale with the turbulent intensity Radiations effects should be examined using this apparatus as well It is known that radiation plays a much larger role in dust flame than gas flames though this was not discussed in this work Information about both the fundamental combustion behavior and the risk of flash fires involving condense phase fuels should be studied This while all of these tasks are important building confidence in the apparatus to show that the measurements are applicable to use in industry and thus a benefit to the fire field is most important Several modifications to the HFA are also proposed The author recommends adding an emerg
197. xygen The presence of 43 these particles is mainly due to their larger sizes and or slow vaporization rate It is important to note that this case results in increase of the temperature in the convection zone Case Cy represents a condition where 2 1 All the condensed phase particles are vaporized in the preheat zone as shown in Fig 2 3 Cp However in this case only part of the gas phase fuel is burned in the flame zone due to the fuel richness of the mixture and there is fuel vapor left over in the convection zone Oxygen is the limiting reactant in this case It should be also noted that the temperature remains constant in the convection zone Case Cr represents the conditions where 21 and the condensed phase fuel is not completely vaporized in the preheat zone as shown in Fig 2 2 Similar to case Cr oxygen is the limiting reactant for this case also However as the condensed fuel continues to vaporize the mass fraction of the fuel vapor increases and the temperature in the convection zone decreases The inset labeled T in Fig 2 2 shows a close up of the convection zone which occurs in case Cy where fuel particles continue to vaporize but do not burn due to oxygen limitation This continued vaporization increases the fuel vapor mass fraction and decreases the convection zone temperature Figures 2 2 and 2 3 summarize the dust problem and shows the significant diversity in situations which can occur in
198. y the flame speed in a dust cloud with a pilot stabilized flame in 2006 Because of the large density of dust and the high particle loading of the cloud they used a vertical downward facing flow Dust stored in a hopper was fed continuously to the burner using a vibrator A stable particle flow was obtained regulating the opening exit in the hopper and the vibration frequency Oxygen and air were fed in the upper part of the burner each flow was measured and adjusted to get the desired concentration The mixture passed through an annular space formed by the burner tube and the ignition gas pilot tube and was discharged downward to the combustion chamber with an Acetylene and air pilot burner in the centre The pilot s function was to initiate the reaction of the dust air mixture and stabilize the flame 1 4 1 2 Non stationary flames Palmer et al 49 published results from a flame propagation apparatus using a long vertical tube in 1971 as described by Eckhoff 21 The dust was introduced at the top of the tube by a screw feeder and dropped into a vibrating 20 cm diameter and 15 cm high dispersing cylinder hanging immediately underneath the screw exit After having passed through the perforated bottom of the cylinder the dispersed dust settled freely under gravity through the entire length of the tube until finally being collected in a bin at the bottom end Dust concentration and flame propagation could not be measured in the same test but

Influence of Coal Dust on Premixed Turbulent Methane

Contents

Download Pdf Manuals

Related Search

Related Contents