Lower explosion limit of hybrid mixtures Emmanuel Kwasi Addai, Dieter Gabel, Ulrich Krause E-mail: [email protected]Otto-von-Guericke-University, Magdeburg, Germany Abstract Hybrid mixtures - mixtures of burnable dusts and burnable gases - pose a special problem to industries, as their combined Lower Explosion Limit (LEL) can lie below the LEL of the sin- gle substances. Various mathematical relations to predict the Lower Explosion Limit of hybrid mixtures (LEL hybrid ) by the LEL of the components have been suggested in literature. The aim of this work is to prove the validity or limitations of these formulas for various combinations of dusts and gases. The experiments were executed in a standard 20 liter vessel apparatus used for dust explosion testing. As ignition source, a permanent spark with an ignition energy of 10 Joule was used. The results obtained so far show, that there are combinations of dust and gas where the proposed mathematical formulas to predict the lower explosible limiti of hybrid mixtures are not safe enough. Keywords: explosion, LEL, MEC, dust, gas, hybrid mixture 1. Introduction Safety related design and a safe operation of equipment requires knowledge about the potential of the ignition and explosion behavior of substances. They are based on the safety-related pa- rameters, including in particular explosion limits. These Values describe the ignition behavior and the possible consequences of explosions. Usually these measures refer to a defined material, generally a pure substance or a well charac- terized dust. However, in industrial production mixtures of material, either of the same state of matter or different occur, often also in different configurations during a production process. For these mixtures, the corresponding parameters are not available and their behavior can only be predicted inadequately. From the perspective of practicality it is not accomplishable to measure all possible mixing ratios. To obtain at least a partial prediction of the behavior of mixture is the aim of this research efforts. In particular, the behavior of combustible dusts and flammable gases near the lower explosion limit (LEL) is examined. As it is well known, a mixture of dust and gas can be ignitable at concentrations that lie below the LEL of the individual substance alone. For this reason, a lower Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial Explosions Bergen, Norway, 10-14 June 2014 995
Hybrid mixtures - mixtures of burnable dusts and burnable gases - pose a special problem toindustries, as their combined Lower Explosion Limit (LEL) can lie below the LEL of the sin-gle substances. Various mathematical relations to predict the Lower Explosion Limit of hybridmixtures (LELhybrid) by the LEL of the components have been suggested in literature. The aimof this work is to prove the validity or limitations of these formulas for various combinations ofdusts and gases. The experiments were executed in a standard 20 liter vessel apparatus used fordust explosion testing. As ignition source, a permanent spark with an ignition energy of 10 Joulewas used. The results obtained so far show, that there are combinations of dust and gas wherethe proposed mathematical formulas to predict the lower explosible limiti of hybrid mixtures arenot safe enough.
Safety related design and a safe operation of equipment requires knowledge about the potentialof the ignition and explosion behavior of substances. They are based on the safety-related pa-rameters, including in particular explosion limits. These Values describe the ignition behaviorand the possible consequences of explosions.
Usually these measures refer to a defined material, generally a pure substance or a well charac-terized dust. However, in industrial production mixtures of material, either of the same state ofmatter or different occur, often also in different configurations during a production process. Forthese mixtures, the corresponding parameters are not available and their behavior can only bepredicted inadequately.
From the perspective of practicality it is not accomplishable to measure all possible mixing ratios.To obtain at least a partial prediction of the behavior of mixture is the aim of this research efforts.In particular, the behavior of combustible dusts and flammable gases near the lower explosionlimit (LEL) is examined. As it is well known, a mixture of dust and gas can be ignitable atconcentrations that lie below the LEL of the individual substance alone. For this reason, a lower
Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial ExplosionsBergen, Norway, 10-14 June 2014
995
Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial ExplosionsBergen, Norway, 10-14 June 2014
explosion limit for mixtures is to be derived, as a function of the gas (XGas) and the dust (Cdust)concentration, as generally defined in equation 1:
LELhybrid = f(Cdust, XGas) (1)
Research on this problem has been published for example by Bartknecht (1981) and Glassmannand Yetter (2008) and led to different mathematical relationships for the LELhybrid further de-scribed in chapter 2. Based on the investigations on different material combinations, the validityof these formulas is to be questioned.
The starting point for this investigation was the thesis of Zarour (2013) with a first series of mea-surements of mixtures of brown coal and natural gas. Within that, the basic suitability of the exist-ing apparatus for the investigation of the explosion characteristics could be proofed.
The 20 liter sphere in the version of ”Kuhner Safety” which is equipped with the necessarysupply connections for gases and with a permanent spark as the ignition source was used to carryout out the experiments.
2. Safety related parameters of hybrid mixtures
The same parameters used to characterized the safety relevant potential of gases and dusts arealso applied for hybrid mixtures.These parameters are:
• maximum explosion pressure
• KSt-value
• lower explosions limit
• minimum ignition energy
• upper explosion limit
• minimum ignition temperature
• limiting oxygen concentration
The first three parameters can be derived from the experiments in the 20 liter sphere. For hy-brid mixtures the others are seldom measured (e.g. in Khalili et al. (2012)) and no standard-ized procedures are available; thus they are not treated in this work. The main parameter ofinterest in this work is the lower explosions limit, as it is well known that the maximum explo-sion pressure and the KSt-value are underestimatet when the electrical ignition is used. Thesevalues, in the way they are represented in this text, are only comparable with each other andnot with values measured with standard 10 kJ chemical igniters used for dust explosion test-ing.
To estimate the LELhybrid of gas dust mixtures two main suggestions can be found in literature.The empirical formula below was derived from measurements done by Bartknecht (1981) thatwhere used by the author himself to propose a formular for the LELhybrid (equation 2), theso-called Bartknecht curve:
Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial ExplosionsBergen, Norway, 10-14 June 2014
LELhybrid = MECdust
(Xgas
LELgas
− 1
)2
(2)
where
MECdust = minimum explosion concentration of the dust in air (g/qm)Xgas = gas concentration in the hybrid mixture (vol. %)LELgas = lowest explosion level of the gas in air (vol. %)
This relation was derived for the hybrid mixture of methane and PVC only.
A more general approached was chosen by Glassmann and Yetter (2008), taking Le Chatelier’slaw as origin. Where the latter originally describes homogeneous mixtures by considering aconstant flame temperature, it was adopted by the authors for hybrid mixtures as presented inequation 3:
LELhybrid =100
Xgas
LELgas+ Cdust
MECdust
(3)
where
Xgas = gas concentration in the hybrid mixture (vol. %)LELgas = lowest explosion level of the gas in air (vol. %)Cdust = dust concentration in the hybrid mixture (g/qm)MECdust = minimum explosion concentration of the dust in air (g/qm)
Both relations can be displayed in the diagrams used to represent the measurement results. Thefirst forms a curve and the second a straight line to separate the region of explosion from the non-explosive area. This way of representing the result together with the KSt-value is used in chapter4. and was adopted from Agreda (2009) and Sanchirico et al. (2011).
3. Experiments
3.1 Experimental setup and procedure
The experimental setup consist of a 20 liter explosion sphere, that is usually used for dust explo-sion testing. Additionally a gas inlet section is mounted at the vacuum pump connection as showin Figure 1.
To provide the correct gas mixtures for the experiments, the partial pressure method was used.Therefore, an additional pressure transducer that precisely measures the internal pressure wasinstalled, replacing the usually used analogous pressure gauge.
At the beginning of each experiment, the sphere is evacuated to 0.3 bars and then the neededamount of burnable gas is added. Afterwards, the internal pressure is set to 0.4 bars by lettingenvironmental air stream in through the gas panel, also ensuring that the gas is completely flushedthrough the line into the sphere.
997
Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial ExplosionsBergen, Norway, 10-14 June 2014
KSEP 310
KSEP 332KSEP 320
001
PI
002
PI
003
004
PI
PI
100
ZC
Tracker 200
200
UI
H2O,in
H2O,out
002
PI
200
UI
Fuel gas Backup N2
Air AnalysisPressure gauge
Flue gas cleaning system
Compressed airVacuumLOC/Mix
N2/ Air N2
Vacuum,inputAir
Vacuum,ouput
Dustcontainer
Injection valve
Vacuum Air or N2/ Air
Air
Figure 1: Technical drawing of the 20 liter sphere with gas supply panel, measurement and con-trol devices.
The second derivation from the standard dust test is the ignition source. Unlike the chemicaligniters which are usually used, in this work a permanent electric spark triggers the ignition,providing an energy of approximately 10 J. Even if it is well known, that the results of Pmax andKSt for pure dust can highly depend on the strength and type of the ignition source, the permanentspark was chosen for hybrid mixture experiments. As the focus only lies on the explosion limits,no constraint is expected by this decision.
To ensure an evenly distribution of the dust in the sphere and a constant level of turbulencebefore the ignition usually an ignition delay time of 60 ms is used with the chemical ignitersand pure dusts. For the application of the permanent spark, this value is set to 0 ms by themanufacturer of the apparatus. As the spark is active for about one second, this ensures anactive ignition source to be present when the dust enters the chamber. In some cases, the gas airmixture present in the sphere might be ignited at a low pressure before the dust is completelydriven into the explosion chamber. To get a better comparison to the pure dust procedure theignition delay time should be adopted to be 60 ms. As this was not possible at the beginningof the investigation, the result presented up to now refer to the immediately acitvated ignition.Meanwhile all measurements arel repeated to reproduce the results with an ignition delay timeof 60 ms.
998
Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial ExplosionsBergen, Norway, 10-14 June 2014
3.2 Tested materials
To be able to give a more profound answer to the question whether the suggested relations topredict the LELhybrid hold a variety of dust and gas combination was tested. Table 1 lists thetested substances together with the used concentration for already tested materials. The lowerpart lists future substances of interest.
Table 1: Tested gases and dusts with the applied concentrations.
gas concentration dust concentration[vol. %] [g/qm]
Following, the results of our experiment of the 8 combinations of gases and dusts listed in Table1 are represented. The diagrams have the relative concentration of gas and dust on either axis.The x-axis is the gas concentration used in the experiment divided by the LELgas and the y-axis is the dust concentration used in the experiment divided by the LELdust. For the LELs thevalues measured for the pure substances in our experimental setup was used, not a literaturevalue.
A non-ignition is symbolized by a small green dot, whereas a red circle indicates an ignition.Their area is proportional to the KSt-value (whereas Agreda (2009) used the diameter for herdiagrams). Hybrid mixtures are colored red, pure substances brown. The curves of the relationsrepresented in chapter 2. are drawn as well, where the solid curve represents equation 2 and thedashed line represents equation 3.
Even if the KSt-value opainted with the electrical ignition is not directly comparable to the oneof the same experiment with chemical igniters, a general trend in the dependency on the concen-tration can be shown.
4.1 Results of hybrid mixtures with hydrogen
The results for toner and hydrogen in Figure 2 show no ignitions below the LEL of the puresubstances and no specific dependency on the concentrations. This might be due to the veryviolent reactions of the toner at low concentrations that lowers the effect of the added hydro-gen.
Also, the combination of lycopodium and hydrogen produced no explosions below the LEL ofthe pure substances as shown in Figure 3. At least a rise of the KSt-value along the axis of thegas concentration can clearly be seen.
999
Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial ExplosionsBergen, Norway, 10-14 June 2014
-0,1
0,1
0,3
0,5
0,7
0,9
1,1
-0,1 0,1 0,3 0,5 0,7 0,9 1,1 1,3 1,5
CD
ust/M
EC
Dus
t
XGas/LELGas
TonerHydrogen
no ignition
Hydrogen
Toner
Figure 2: Diagram of hybrid mixtures of toner and hydrogen.
-0,1
0,1
0,3
0,5
0,7
0,9
1,1
-0,1 0,1 0,3 0,5 0,7 0,9 1,1 1,3 1,5
CD
ust/M
EC
Dus
t
XGas/LELGas
LycopodiumHydrogen
no ignition
Hydrogen
Lycopodium
Figure 3: Diagram of hybrid mixtures of lycopodium and hydrogen.
1000
Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial ExplosionsBergen, Norway, 10-14 June 2014
Polyethylene added to a hydrogen atmosphere increases the KSt-value but produces no ignitionbelow the single LELs, as represented in Figure 4.
-0,1
0,1
0,3
0,5
0,7
0,9
1,1
-0,1 0,1 0,3 0,5 0,7 0,9 1,1 1,3 1,5
CD
ust/M
EC
Dus
t
XGas/LELGas
PE-HdHydrogen
no ignition
Hydrogen
PE-Hd
Figure 4: Diagram of hybrid mixtures of polyethylene and hydrogen.
So far, no ”specific” hybrid explosion at concentrations below the ignition limit of the respectivepure substances was recognized. The diagrams show a gap between the vertical values 1 and0.5 where the concentration of the dust is divided by 2 due to the given values of the computerprogram that controls the experiment. Here investigations for dust concentrations at 0.75 of theMEC will follow.
Finally, the mixture of starch and hydrogen lead to a hybrid explosion below the LEL of the puresubstances, as shown in Figure 5.
-0,1
0,1
0,3
0,5
0,7
0,9
1,1
-0,1 0,1 0,3 0,5 0,7 0,9 1,1 1,3 1,5
CD
ust/M
EC
Dus
t
XGas/LELGas
StarchHydrogen
no ignition
Hydrogen
Starch
Figure 5: Diagram of hybrid mixtures of starch and hydrogen.
1001
Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial ExplosionsBergen, Norway, 10-14 June 2014
In this case a clear explosion at 0.5 MECstarch and 0.66 LELH2 could be recorded, as well as weakignition at 0.5 MECstarch and 0.33 LELH2. This violates the suggested LEL line by Glassmannand Yetter (2008).
4.2 Results of hybrid mixtures with methane
The same dusts were also tested with methane as flammable gas.
The plastic material polyethylene shows at least one hybrid explosion at a gas concentration nearthe LEL of methane, as represented in Figure 6.
Figure 6: Diagram of hybrid mixtures of polyethylene and methane.
The diagram of the mixture of toner with methane in Figure 7 shows a series of ignitions at 50%of the MEC. The one with the lowest gas concentration actually lies in the ”no explosions area”defined by formula 3.
The same behavior can be shown for lycopodium, as represented in Figure 8. Additionally therewas another ignition close to Glassmann’s line at a lower dust concentration.
1002
Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial ExplosionsBergen, Norway, 10-14 June 2014
Figure 7: Diagram of hybrid mixtures of toner and methane.
-0,1
0,1
0,3
0,5
0,7
0,9
1,1
-0,1 0,1 0,3 0,5 0,7 0,9 1,1
CD
ust/M
EC
Dus
t
XGas/LELGas
LycopodiumMethane
no ignition
Methane
Lycopodium
Figure 8: Diagram of hybrid mixtures of lycopodium and methane.
1003
Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial ExplosionsBergen, Norway, 10-14 June 2014
Finally, the mixtures with starch again show the widest explosion rage. As illustrated in Figure9 an ignition still occurred at 50 % MECstarch and 20% of the LELmethane, thus lying in the ”noexplosions area” defined by Bartknecht’s curve.
Figure 9: Diagram of hybrid mixtures of starch and methane.
While the KSt-value along the y-axis is only influenced by the higher turbulence due to theaddition of dust and else only changes slightly with the concentration, the values clearly show adependency with the gas concentration. Different from the other dust this behavior is obvious atthe MECstarch as well as at 50% of it.
As starch showed the highest potential to create explosions at very low combined concentrations,the research first will be continued with this dust, taking into account the necessary improvementsof the system and procedure, which include the following:
• delayed ignition of 60 ms; comparable to the standard dust procedure
• additional experiments at 75% of the MEC
• more accurate generation of the gas mixture by new pressure sensor
Moreover, the experiments are all to be repeated at least 5 times to ensure reproducibility of theexperimental results. This not only to answer the general question whether an ignition is possibleat given concentrations but also to show the deviation in the maximum explosion pressure and thederived KSt-value for the used ignition source the permanent spark with only 10 Joule ignitionenergy.
1004
Tenth International Symposium on Hazards, Prevention, and Mitigation of Industrial ExplosionsBergen, Norway, 10-14 June 2014
5. Conclusions
Investigations on hybrid mixture were carried out in a standard 20 liter sphere showing the poten-tial of mixtures of dust and gas to be ignited at concentrations below the limit defined for the puresubstance alone. The results so far show, that the suggested relations to predict a common lowerexplosion limit do not seem to be completely reliable. The dependency on the specific materialis obvious and leads to the conclusion that a simplified formula based on the lower explosionlimits of pure substances might never be sufficiently safe. Here more knowledge on the reactionmechanism is necessary to better understand and predict the behavior.
Furthermore the 20 liter sphere used here and by other researchers provides an experimentalprocedure that makes results comparable and allows to create a foundation of data for furtheractivities to decipher the underlying process of hybrid mixture ignitions.
References
Agreda, A. (2009). Study of hybrid mixtures explosions. PhD thesis, Universita degli Studi diNapoli Federico II.
Bartknecht, W. (1981). Explosions: Course, Prevention, Protection. Spinger Verlag.
Glassmann, I. and Yetter, R. (2008). Combustion. Elsevier Inc.
Khalili, I., Dufaud, O., Poupeau, M., Cuervo-Rodriguez, N., and Perrin, L. (2012). Ignitionsensitivity of gasvapor/dust hybrid mixtures. Powder Technology, 217(0):199 – 206.
Sanchirico, R., Benedetto, A. D., Garcia-Agreda, A., and Russo, P. (2011). Study of the sever-ity of hybrid mixture explosions and comparison to pure dustair and vapourair explosions.Journal of Loss Prevention in the Process Industries, 24(5):648 – 655.
Zarour, Y. (2013). Characteristics of explosions of gases and hybrid mixtures. Master’s thesis,Otto-von-Guericke-University.
