- Email: [email protected]
Flammability of gases in focus of European and US standards
Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
Contents lists available at ScienceDirect
Journal of Loss Prevention in the Process Industries journal homepage: www.elsevier.com/locate/jlp
Flammability of gases in focus of European and US standards Maria Molnarne*, 1, 2, Volkmar Schroeder Bundesanstalt fuer Materialforschung und -pruefung (BAM), Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 January 2017 Received in revised form 11 May 2017 Accepted 12 May 2017 Available online 13 May 2017
The presentation will discuss the difference between EU and US standards for the determination of explosion (flammability) limits and limiting oxygen concentration. Small differences observed in measured values can be traced back to the different test apparatuses and criteria. The discrepancies can be much greater in the case of limiting oxygen concentration because of the high amount of inert gases and the corresponding low laminar burning velocities. The paper describes some examples and the influence of the chosen criteria on the results. The European and US standards use the criteria of flame propagation in open test vessels and of pressure rise in closed ones. The examples discussed show that flame propagation is still possible at very small pressure rise values, as observed much below the pressure rise criterion of usual standards. However, flame propagation in a process plant can cause an accident or explosion and must be avoided. Therefore, the flame propagation criterion is recommended to be used in chemical safety engineering. The European safety database CHEMSAFE contains expertevaluated safety data for cases where the determination method and criteria are known. Flammability characteristics based on the pressure rise criterion may suffice in certain cases, e.g. for explosion protection in closed vessels without any connecting pipes. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Flammability Limiting oxygen concentration Test methods Standards
1. Introduction
2. Terminology: “flammability limits” vs. “explosion limits”
Flammability of gases is a fundamental material property that entails legal classifications in various instances and requires appropriate labelling. Explosion protection measures are necessary only when flammable gases are present, e.g. in processes, storage and use. In European technical legislation, a gas or a gas mixture is regarded as “flammable” if under atmospheric conditions an explosion range (explosion limits) in a mixture with air exists. Therefore determination methods for flammability and explosion limits are standardized in many countries. Nevertheless, explosion limits are not the type of independent physicochemical material characteristics such as boiling temperature or density of a substance. As most other safety characteristics they are influenced by the test apparatus and the determination procedure applied. The evaluation and standardization of determination methods for flammability limits and limiting oxygen concentrations are particularly important for chemical safety engineering.
The expression “explosion limit” often is called “flammability limit” in the literature. But “flammable” is an ambiguous term: it means combustible in air on the one hand and explosively reacting without any further addition of air or another oxidizer on the other. Typical examples are flammable gases, e.g. pure hydrogen and flammability (explosion) ranges that characterize hydrogen-air mixtures. In the European ATEX directives and standards of explosion prevention the term “explosion” is used for explosively reacting mixtures to avoid misunderstanding. The actual situation in standardization and scientific literature is characterized by the fact that most of the US authors use the term “flammability limit”, while European standardization calls the same material safety characteristic “explosion limit”. Nowadays there are discussions in international standardization committees about a different meaning in the sense of more violent reactions if the “explosion limit” is reached, compared to slower reactions at the “flammability limit”. An informal IEC paper (IEC, 2003) used the direction of flame propagation as a criterion. Bureau of Mines measurements on hydrogen-air mixtures are often cited, where an upward flame propagation was observed at 4 mol% hydrogen but downward propagation does not begin before 9 mol% hydrogen (Coward and Jones, 1952). Therefore, 4 mol% is proposed as the
* Corresponding author. E-mail address: [email protected] (M. Molnarne). 1 Formerly at BAM, currently: Research Associate at MKOPSC, Texas A&M, USA. 2 Home Address: Brettnacher Str. 17b, 14167 Berlin, Germany. http://dx.doi.org/10.1016/j.jlp.2017.05.012 0950-4230/© 2017 Elsevier Ltd. All rights reserved.
298
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
lower flammability limit (LFL) and 9 mol% as the lower explosion limit (LEL). Nevertheless, such definition is not helpful in the field of explosion protection. There is actually no international standard available for the criterion “downward propagation” on the one hand and it can never be known where a potential ignition source is located in a process plant on the other. Therefore, only the more conservative upward propagation can be the relevant criterion for explosion protection. To avoid misunderstanding it must be pointed out that “flammability limit” in the US and “explosion limit” in Europe represent the same material property, determined by similar standard test procedures using the same criterion “upward propagation” of flames. 3. Comparison of European and US standards for determination of explosion limits With the creation of European directives in the field of explosion prevention, which apply uniformly to all Member States, it became necessary to develop new European standards for the determination of explosion characteristics. The European Technical Committee 305 “Potentially explosive atmospheres e Explosion prevention and protection” (CEN/TC305) is heading this project. Working Group 1 is responsible for the development of new standards for the determination of safety characteristics of gases, vapors and dusts. In this field a new European draft prEN 1839:2015 “Determination of explosion limits and of the limiting oxygen concentration of gases and vapors” is under development. It is a merged revised version of EN 1839:2012 (EN, 1839; 2012) (explosion limits) and EN 14756:2006 (EN 14756, 2006) (limiting oxygen concentration). Parallel to the European standards, corresponding US standards are available (ASTM E681 (ASTM E681-09, 2015), ASTM E2079 (ASTM E2079, 2013)). BAM has performed tests (Schroeder and Daubitz, 2004) to compare US and European standards for a number of years with 5 test substances, 4 gases and a vapor, ethanol, chosen for the tests. The gases, hydrogen, ethylene, methane and ammonia, had different laminar burning velocities. It was believed that differences in the results due to test apparatuses and test procedures could be checked by measurements on the same test substances. 40 mol% of nitrogen was added to the gas-air mixtures to increase sensitivity. Altogether 4 test methods were used (for parameters see Table 1). The parameters of apparatuses and test methods are the same as in the latest versions of the standards (e.g. prEN,
1839:2015 and ASTM E681-09), with the only exception of ammonia. Nowadays in European and US standards, special procedures are available for so-called “difficult-to-ignite” substances such as ammonia and halogenated refrigerants. Tests to compare these test methods should be carried out in the future. Three of the four test methods (DIN 51649 (DIN 51649, 1986), EN 1839-T (tube method), ASTM E681) are so-called “open vessel methods” that apply upward flame propagation as the criterion for a reaction. Another method was the “closed vessel method” EN 1839-B (bomb method) that used the criterion of pressure rise. The following conclusions were drawn from the experiments (see Table 1 and Figs. 1e3): The procedures according to DIN 51649 and EN 1839-T provide identical results in almost all cases. The reason lies in the very similar test apparatuses of the methods. The ASTM method yields similar explosion ranges in many cases. This can be explained by the use of a 5-dm3 flask in connection with the sensitive visual criterion. However, a major disadvantage of this method is the undefined step size in connection with the definition of the explosion limit (mean value between ignition and non-ignition point). This can easily result in unsafe explosion data if the concentration steps are not specified clearly. The closed vessel method EN 1839-B shows the strongest deviations. LELs are higher compared to the other methods and a clearly lower ethylene UEL was observed. The reason might be the pressure threshold criterion used that is obviously less sensitive than the optical criterion. Ammonia was an exception with a significantly higher UEL. The large quenching distance of ammonia may play a role, thus ammonia reactions in the 14dm3 sphere are preferred opposed to smaller volumes. In general, the deviations of difficult-to-ignite gases, e.g. ammonia and nitrogen mixtures, are the greatest. Such reactions are strongly affected by apparatus parameters. 4. Comparison of the flame propagation criterion and pressure rise criterion Determination methods for explosion limits at elevated pressures need another criterion for a reaction. Usually a measured pressure rise (pex/p0) in a closed ignition vessel is used. ASTM E91883 (2011) uses a pressure rise pex/p0 > 7%, EN 1839-B a pressure rise pex/p0 > 5% as the criterion. De Smedt & Berghmans (De Smedt
Table 1 Key characteristics of the most commonly used standard test methods for explosion limit determination. DIN 51649-1 withdrawn in 2004
EN 1839-B:2004
ASTM E681-01
vertical glass cylinder, open ∅ ¼ 60 mm, H ¼ 300 mm Ignition source induction sparks typical 0.5 s, min. 0.2 s Criterion for visual, ignition flame detachment
closed sphere or cylinder H/D ¼ 1 to 1.5, V > 5 dm3 induction sparks or exploding wire, E ¼ 10 Je20 J pressure increase 5% þ ignition pressure in air
flask, spherical V ¼ 5 dm3 (V ¼ 12 dm3) induction sparks or exploding wire
Increment
freely selectable, to be specified in report if x > 10% 10% (rel.) to 0.2 mol% (abs.), depending on the (rel.) of the test substance fraction test substance fraction 4 1
Explosion vessel
EN 1839-T:2004
vertical glass cylinder, open ∅ ¼ 80 mm, H ¼ 300 mm induction sparks typical 0.2 s, max. 0.5 s visual, flame detachment >100 mm or aureole H > 240 mm 0.1e0.2 mol%, depending on the 10% (rel.) to test substance fraction 0.2 mol% (abs.), depending on the test substance fraction 5 4
Number of repetition tests Explosion limit last non-ignition point
last non-ignition point
last non-ignition point
visual, flame propagation up to 13 mm to the wall (horizontal or vertical)
mean value from non-ignition and ignition point
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
299
Explosion range Explosion range at 40 mol% N2 DIN 51649
ASTM E681
EN 1839 T
EN 1839 B
0
10
20
30
40
50
60
70
80
Amount of hydrogen in mol% Fig. 1. Explosion ranges of hydrogen-air and hydrogen-nitrogen-air mixtures (high burning velocity).
Explosion range Explosion range at 40 mol% N2
DIN 51649
ASTM E681
EN 1839 T
EN 1839 B
3
5
7
9
11
13
15
17
Amount of methane in mol% Fig. 2. Explosion ranges of methane-air and methane-nitrogen-air mixtures (normal burning velocity).
et al., 1999) proposed 2% as the best pressure rise criterion. Until now only ASTM E918 provides a standard test method for evaluated pressures. In 2015 the European standardization committee TC 305 also proposed to develop a standard for explosion limits at elevated pressures and temperatures to improve process plant safety. Therefore BAM and PTB in Germany started a research project in €der et al., 2013). The focal point was the evaluation this field (Schro of apparatus parameters and criteria for higher temperatures and pressures. Therefore, a special test autoclave (volume ¼ 11 dm3) was built, which was equipped with two pressure resistant windows. This autoclave enables the observation of flame propagation and pressure rise for the same explosion up to a pressure of 100 bar (see Fig. 4). It could clearly be observed that e particularly at higher initial pressures e the flame can propagate at a pressure rise below 5%. Table 2 shows that flame propagation exceeding 100 mm until to
the top was observed at pressure rises of 3.6% in a period of time of 0.1e1s. This is in agreement with De Smedt & Berghmans’ criterion of 2%. Summing up, most of the international standards use the flame propagation criterion for atmospheric conditions. The aim of explosion protection is to avoid the spreading of flames through gas or vapor atmospheres. Therefore, a pressure rise criterion that corresponds to flame propagation should also be applied to nonatmospheric conditions in process plants. 5. Explosion ranges and limiting oxygen concentration Explosion limits are concentration limits of flammable gases or vapors in mixtures with air or air and an inert gas at which flame propagation just failed. The limiting oxygen concentration is defined as the maximal permissible oxygen content in a flammable
300
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
Explosion range Explosion range at 20 mol% N2 DIN 51649
ASTM E681
EN 1839 T
EN 1839 B
13
17
21
25
29
33
37
Amount of ammonia in mol% Fig. 3. Explosion ranges of ammonia-air and ammonia-nitrogen-air mixtures (low burning velocity).
Fig. 4. (a) Experimental device with windowed autoclave, camcorder and (b) frontal view into the autoclave with igniter and a propagating flame.
Table 2 €der et al., 2013). Most evident flame propagation, depending on the methane amount, UEL methane-air mixture, p0 ¼ 5 bar (Schro Flame Propagation
CH4 Time after Ignition pex/p0
19.0 mol% 0.24 s 1.2%
18.2 mol% 0.96 s 17.4%
21.0 mol% 0.24 s 0.8%
18.3 mol% 0.40 s 5.9%
23.0 mol% 0.24 s 0.8%
gas-inert gas-air mixture therefore a so-called explosion range must be specified for a 3-component system (flammable gaseinert gaseoxidizer). In the narrower sense (two-component systems) the
18.4 mol% 0.40 s 3.8%
25.0 mol% 0.24 s 0.7%
18.6 mol% 0.40 s 3.6%
27.0 mol% 0.12 s 0.2%
explosion range is the range between lower and upper explosion limit of a flammable gas in a mixture with air. Flammable gas-air/ oxidizer mixtures within this concentration range are explosive.
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
In a broader sense (3-component systems) the explosion range is the range inside the border line of lower and upper explosion limits. All mixtures at concentrations inside this range are explosive (see Fig. 5). The points on the left axis (Fig. 5) are the upper and the lower explosion limit (UEL, LEL) of methane in air. SCO is the line of stoichiometric concentration of the combustion reaction. LAC is the limiting air concentration. Such triangular diagrams enable the determination of the socalled limiting air concentration (LAC). LAC (red line) passes through the apex (tip) of the explosion range running parallel to the inert axis. The limiting oxygen concentration (LOC) can be calculated from the limiting air concentration according to Eq. (1) considering the oxygen fraction in air:
100
301
0 10
90
UEL
20
80 30
70
40 x
60
N2
50
x H2
50 60
40
70
30
LOC ¼ 0:209 LAC
(1)
The European standard EN 14756 uses for LOC determination the EN 1839 “T” or “B” apparatus. The authors recommend using the “T” method if possible. As shown in Chapters 3 and 4, flame propagation is often more sensitive than the 5% pressure rise criterion. Flame velocities at the apex of an explosion range are very low and the criterion and apparatus parameters can have a significant influence on the measured LAC. EN 14756 specifies two methods for the determination of LOC. The so-called short method is based on the typical shape of explosion ranges as shown in Fig. 5: most hydrocarbons have a sharp tip at the maximum inert gas concentration. The UEL line is not parallel to the inert axis in such cases. Therefore, it is sufficient to measure LEL, UEL and the apex of the explosion range. The apex can be determined by measuring the explosion limits belonging to the stoichiometric line SCO or, in the case of nitrogen, an inert gas, belonging to the line x ¼ 1.2 * LEL. In the case of broader explosion ranges, e.g. hydrogen-nitrogen-air mixtures, this is not sufficient. Here the so-called extended procedure must be used. Since it is not known where the LAC runs, the whole diagram must be measured (see Fig. 6). The criterion of EN 14756 telling when the extended procedure has to be applied, is given by Eq. (2):
UEL > 0:8 100 xair; L
(2)
x air, L is the air concentration at the explosion limit measured at
100 90
LEL
LAC
100
0
100
80
90
90
20 30
UEL 20
0
100
70
30
70
60
30 50 40 Air in mol%
10
N2 60 in mol% 70
Explosion range
80 90
LEL 100
20
50
10
90 SCO
80
40
20
80
LAC 90
(Limiting concentration for stability)
ETO 50 in mol% 40
60
10
'UEL'
60
in mol%
50
LEL
20
70
Inert gas
Explosion range
0
10
30
30
10
20
0
80
50
30
40
the stoichiometric line. In contrary to the use of different test methods (bomb method or tube method, see Fig. 8), the choice of the determination method of LOC (short method or extended method) does not lead to different results. However, as described above, the short method is not applicable to all test substances. In the case of chemically unstable substances, e.g. acetylene, ethylene oxide (ETO), tetrafluoroethylene etc., LOC cannot be determined. These gases can react without any amount of air and so the LOC would be zero. In the case of gas mixtures that contain unstable components one should be careful and take the extended procedure. The same is recommended for substances that have double bonds or other high-energy bonds in the molecule. Fig. 7 shows the ternary diagram of ethylene oxide as an example. At lower amounts of ETO, a combustion reaction takes place
100
60
50
xair
Fig. 6. Explosion range of hydrogen-nitrogen-air mixture at 20 C and 101 kPa. The points on the left axis are the upper and lower explosion limits (UEL, LEL) of hydrogen in air. LAC is the limiting air concentration.
80
Flammable gas in mol% 40
60
70
10
40
90
10
0
70
80
Explosion range
20
0
Fig. 5. Explosion range of the methane-nitrogen-air mixture at 20 C and 101 kPa.
0
100 100
90
80
70
60
50 40 Air in mol%
30
20
10
0
Fig. 7. Explosion range of ethylene oxide (ETO)-nitrogen-air mixture at 20 C and 101 kPa. The limiting concentration for stability is about 20 mol% nitrogen without any amount of air (CHEMSAFE, 2013e2015).
302
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
100 90
0 10 20
80 30 70
Nitrogen
40
60
50
in mol%
50
Methane
60
40
in mol%
70
30
10 0
90
80
90
LAC 1
Explosion ranges
100
80
LAC 2
20
100 70
60
30 50 40 Air in mol%
20
10
0
Fig. 8. Explosion ranges of methane-nitrogen-air mixtures at 20 C and 101 kPa measured acc. to EN1839 tube method (red, dotted) and bomb method (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
predominantly. As the proportion of ETO increases, a decomposition reaction occurs. At the upper explosion limit the flame propagates without any availability of air. Unlike the European standard EN 14756 (or the new draft prEN, 1839:2015) the ASTM E2079 describes a separate test apparatus and does not use the glass flask of ASTM E681 with open vessel and visual criterion. ASTM E2079 uses a bomb method and a pressure rise of 7% as a criterion, not flame propagation. It can be expected that LOC values are higher than those measured according to EN 14756 using the flame propagation criterion and the EN 1839-T glass tube. One example for the difference of LOC using tube method vs. bomb method of EN 1839 is given in Fig. 8. The LAC values resulted from the explosion range curves differentiate each other significantly. So a LOC 1 of 9.9 mol% correspond to LAC 1 (visual criterion) and a LOC 2 of 11.0 mol% to LAC 2 (pressure rise criterion of 5%). This results correspond to the NFPA 69 value of methane, LOC ¼ 11.2 mol% (Brandes and Ural, 2008), which was measured of according to ASTM E2079 (bomb method). 6. Safety characteristics derived from explosion ranges of CHEMSAFE The CHEMSAFE Database (CHEMSAFE, 2013e2015) contains rated safety characteristics data for fire and explosion protection for approximately 3000 different flammable gases, liquids and dusts and mixtures of gases, vapors, and dusts. The data include explosion limits, flash points, ignition temperatures, limiting oxygen concentrations, minimum ignition energy, maximum experimental safe gap, etc. This type of data is the basis for the classification of flammable substances with regard to testing, choice and use of explosion protected electrical equipment, for the transport of dangerous goods, the Ordinance on Hazardous Substances (classification, labelling and packaging, directive EG 1272/2008) and for many European and ISO standards. CHEMSAFE is a joint project of three partners from Germany: Physikalisch-Technische Bundesanstalt (PTB), Bundesanstalt für Materialforschung und -prüfung
(BAM), and the Gesellschaft für Chemische Technik und Biotechnologie e.V. (DECHEMA). PTB is responsible for the data of flammable liquids and BAM for flammable gases and dusts. The recommendations in the database are given by the experts of these two German federal institutes: BAM and PTB. CHEMSAFE is updated yearly and it has been accessible online since 1989. The database is available as a chargeable version all over the world on the internet (http://www.dechema.de) and as an in-house version in German and English, developed by DECHEMA, Frankfurt/Main. In the future CHEMSAFE will only be available as an online version instead the current different versions of different functionality. This will be implemented by PTB taking over the hosting of the database. The “[email protected]” newsletter will announce future upgrades of the Open Access version to registered customers. This English/German version will be available from the beginning of 2017. In addition to safety characteristic data, the database also delivers substance description (CAS registry number, names) and classification according to German and international regulations, physicochemical properties such as density, boiling point, etc., and bibliographic information about measured data. The database forms the basis for the Data Books “Safety Characteristic Data” €ller, 2008) and 2 (Molna rne et al., 2008) Volumes 1 (Brandes and Mo of PTB and BAM. To avoid an explosive atmosphere, the knowledge of explosion limits of flammable gases and vapors in air are essential. If an explosive atmosphere cannot be prevented, inert gases such as diluents can be applied to reduce or avoid the explosive mixtures, e.g. in the case of inerting/purging. The limiting oxygen concentration (LOC) and the limiting values of flammability (Tci values) are very important safety characteristic values which are derived from explosion diagrams of mixtures containing flammable gas, inert gas and air. These data and the corresponding explosion diagrams are available in CHEMSAFE, in the “Explosion Regions of Gas Mixtures” rne et al., 2008) Data Book, and the limiting values from (Molna these diagrams are also listed in ISO 10156 standard (ISO 10156, 2010). This standard specifies methods to determine whether or not a gas or gas mixture is flammable in air and is more or less oxidizing than air under atmospheric conditions. It is also used for classification of gases and gas mixtures including the selection of gas cylinder valve outlets and for labelling according to international transport regulations and dangerous substances regulations such as in the Globally Harmonized System (GHS). The in-house version and the future Open Access online CHEMSAFE Database will provide a graphical representation of these data for the explosion range of these mixtures in Cartesian or triangular diagrams. A general view of these explosion diagrams using a triangle plot can be seen in Fig. 9. The limiting oxygen concentration (LOC) is the oxygen concentration in a flammable gas-air-inert gas mixture below which an explosion cannot occur under specified conditions or, more generally, the “maximum oxidizing gas concentration” (MOC) is the maximum permissible oxidizing gas content in a 3-component mixture e which can be used in the case of oxidizing gases other than oxygen e and it is valid not only under atmospheric conditions. The “maximum permissible amount of flammable gas concentration” (MXC) generally can be used in an inert gas/flammable gas mixture in mol % or in the special case, if the inert gas is nitrogen, it is called “Threshold for flammability of the flammable component i in the mixture with nitrogen (Tci)” which corresponds to the MXC value with air as oxidizer and nitrogen as inert gas at room temperature and atmospheric pressure. The “minimum required inert gas concentration” (MAI) in mol% in an inert gas-oxidizer mixture is the smallest fraction of the inert gas for which this gas mixture cannot become explosive irrespective of the amount of flammable gas
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
100
of old apparatuses and the different criteria for flammability.
0 10
90
Interpolated curve Experiment
20
80 30 70
Flammable gas 50 in mol%
40
IAR line
60
50 MOC line
40
Inert gas 60 in mol% 70
30
80
20 10 0
C Explosion range
MAI
90
ICR line
90
80
70
50 60 40 Oxidizer in mol%
MXC
100
MOC
100
30
20
10
0
Fig. 9. Safety characteristics derived from the explosion diagram taken from CHEMSAFE Database.
being added. IAR (minimal inert/air ratio) and ICR (minimal inert/ combustible ratio) lines are straight lines drawn to the border of the explosion range and they represent the limiting ratios of inert gas to oxidizing gas and inert gas to flammable gas, respectively. The LOC values can be calculated by Eq. (1) using the MOC values in the case of air (they are equal to the LAC values). Table 3 shows some recommended limiting oxygen concentration data taken from CHEMSAFE. The LOC values stem from the following ternary mixtures: flammable gas-air-carbon dioxide and flammable gasair-nitrogen. The measuring method is the DIN 51649 or EN 1839-T method when not otherwise commented. A comparison of these values with the US Bureau of Mines and the NFPA 69 (2008) edition values from Table 1 in Brandes and Ural (Brandes and Ural, 2008) shows substantially higher values for almost all substances. They explain these differences with the use
Table 3 Limiting oxygen concentration (LOC) of ternary mixtures. Flammable Gas
Ammonia Butane 1-Butene Carbon monoxide (pure and dry) Carbon monoxide 1-Chloro-1,1-difluoroethane 1,1-Difluoroethane Ethane Ethene (Ethylene) Ethyne (Acetylene) Ethylene oxide Hydrogen Hydrogen sulfide Methane 2-Methylpropane 2-Methyl-1-propene Propane Propene Propylene oxide a b
303
Inert CO2
Inert N2
LOC mol%
LOC mol%
13.2a
5.1b
11.8 10.4 6.3 0.0 5.2 10.5a 13.6 12.8 13.2 12.6 12.5 10.3
Test method EN 1839-B. With traces of moisture, hydrogen or hydrocarbons.
12.2 9.6 9.7 6.1 4.7b 14.1 9.8 8.7 7.6 6.2 0.0 4.3 8.9a 9.9 10.3 10.6 9.3 9.4 7.7
7. Discussion and conclusion The explosion limits and limiting oxygen concentration are affected by the chosen test apparatus, e.g. the type and size of the ignition vessel (closed, open, dimensions, material), ignition source (type, power, duration and location) and the test criterion (flame propagation or pressure rise) and accuracy. In the international standardization there are test methods for the experimental determination of explosion limits using open and closed ignition vessels. The flame propagation criterion is usually used for open ignition vessels and the pressure rise criterion for closed ignition vessels. The European standard EN 1839 describes both methods, i.e. the open ignition vessel (tube method, “T”) and the closed ignition vessel (bomb method, “B”). A flame spread of 10 cm and a pressure rise >5% of the initial pressure have been selected as the criterion for atmospheric conditions. Investigations in an autoclave having viewing windows performed at BAM showed that flame propagation still can be observed even if pressure rise less than 5%. In conformity with De Smedt & Berghmans’ suggestions it is proposed to reduce the pressure threshold criterion to 2%. The examples discussed show that flame propagation is still possible at very small pressure rise values, as observed much below the pressure rise criterion of most standards. However, flame propagation in a process plant can cause an accident or explosion and must be avoided. Therefore, the flame propagation criterion is recommended to be used in chemical safety engineering. Experimental studies of BAM showed that the two determination methods of EN 1839 provide somewhat different results. The explosion areas that were determined in open vessels are usually a little larger than those measured in closed vessels. This is especially significant when explosion limits of mixtures with a high proportion of inert gas are determined. Therefore, it is expected that especially the LOC values, which are determined using the “tip” of the explosion range can be strongly influenced by the selected measuring method. It is not surprising that LOC values which were determined according to EN 14756 using the tube method, are usually smaller than those measured according to ASTM E2079 (bomb method). CHEMSAFE includes triangle diagrams of many explosion ranges of flammable gas-inert gas-air mixtures. The data were evaluated by BAM and PTB experts. In addition to the safety characteristics and specifications, the database also describes the measuring methods. The explosion areas enable the determination of important safety characteristics such as LEL, UEL, LOC, Tci, etc. The database is currently still fee-based, but it will be available for free of charge on the Internet from 2017. It is generally recommended to select data of explosion limits and LOC of flammable gas/vapor, air, and inert gas mixtures determined by the same experimental methods used for safety concept, risk analysis and estimation methods. The international standardization is useful in harmonization of different national safety concepts and in understanding the different safety cultures of individual countries. Many efforts have been taken in the past 20 years to analyze the differences of the published explosion/flammability limits and limiting oxygen concentration data of both European and US researchers. The Safekinex (Safekinex, 2009) €der and Molnarne (Schro €der research project, publications of Schro and Molnarne, 2005), Brandes and Ural and Zlochower (Zlochower and Green, 2009) have to be mentioned here. Britton (Britton et al., 2016) has recently published a paper about LOC which also discusses the differences between the US and European experimental data.
304
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
As it was mentioned earlier the European standards define LEL, UEL and LOC as the concentration points at which flame propagation has just not occurred. In the US standards the border concentration points between flammable and not-flammable concentrations are used in the case of visual flame propagation criterion (see Table 1). When comparing the data, it is essential to know the chosen concentration step size in the test procedure. Therefore, it is useful to follow the discussion between the US and European standardization authorities aimed at establishing a global standard for the determination of these important safety characteristics. Acknowledgements The authors would like to thank BAM and PTB for the presentation materials they provided. References ASTM E2079-07, 2013. Standard Test Methods for Limiting Oxygen (Oxidant) Concentration in Gases and Vapors. ASTM International, American Society for Testing and Materials. ASTM E681-09, 2015. Standard Test Method for Concentration Limits of Flammability of Chemicals (Vapors and Gases). ASTM International American Society for Testing and Materials. ASTM E918-83, 2011. Standard Practice for Determining Limits of Flammability of Chemicals at Elevated Temperature and Pressure. ASTM International American Society for Testing and Materials. €ller, W., 2008. Safety characteristic data. In: Flammable Liquids and Brandes, E., Mo Gases, vol. 1. NW-Verlag (Publishing Co.), Bremerhaven, Germany. Brandes, E., Ural, E.A., 2008. Towards a global standard for flammability determination. In: Proceedings of the 42nd Annual Loss Prevention Symposiumeglobal Safety Congress, Paper 2E, April 6, 2008. Britton, L.G., Clouthier, M.P., Harrison, B.K., Rodgers, S.A., 2016. Limiting oxygen
concentrations of gases. Process Saf. Prog. 35, 107e114. CHEMSAFE d Database for Recommended Safety Characteristics, 2013-2015. BAM, PTB, and DECHEMA, In-house Version (INTERNET: dechema.de/en/ chemsafe.html). Coward, H.F., Jones, G.W., 1952. Limits of flammability of gases and vapors, US Bureau of Mines. Bull 503. De Smedt, G., de Corte, F., Notele, R., Berghmans, J., 1999. J. Hazard. Mater. A70, 105e113. DIN 51649 Teil 1, 1986. Bestimmung der Explosionsgrenzen von Gasen und Gasgemischen in Luft. Deutsches Institut für Normung, Berlin withdrawn in 2004. EN 14756, 2006. Determination of the Limiting Oxygen Concentration (LOC) for Flammable Gases and Vapours. CEN European Committee for Standardization, Brussels. EN 1839, 2012. Determination of Explosion Limits of Gases and Vapours. Sections 4.1, Method T (EN 1839-T) and 4.2, Method B (EN 1839-B). CEN European Committee for Standardization, Brussels. IEC Inf Paper 31/460/INF, 2003-03-07. “Explosive” Limits and “Flammable” Limits. ISO 10156:2010, 2010. Gases and Gas Mixtures d Determination of Fire Potential and Oxidizing Ability for the Selection of Cylinder Valve Outlets. rne , M., Schendler, Th, Schro €der, V., 2008. Safety characteristic data. In: Molna Explosion Regions of Gas Mixtures, vol. 2. NW-Verlag (Publishing Co.), Bremerhaven, Germany. Safekinex, 2009. SAFe and Efficient hydrocarbon oxidation processes by KINetics and Explosion eXpertise. In: Energy, Environment and Sustainable Development. Programme (Contract Number EVG1-CT-2001e00098), EU Project. €der, V., Molnarne, M., 2005. Flammability of gas mixtures. Part I: fire potential. Schro J. Hazard. Mater. 121, 37e44. €der, V., Tschirschwitz, R., Brandes, E., 2013. BAM Research Report No. 2-3408/ Schro € ßen 2013 „Sicherheitstechnische Kenngro bei nicht-atmosph€ arischen Bedingungen“. Schroeder, V., Daubitz, R., 2004. Evaluation of standard test methods for the determination of explosion limits of gases and vapours. In: Proceedings 11th Symposium on Loss Prevention 2004, pp. 2109e2117, 31 May-3 June, Praha, Czech Republic, Them. Sect. B. Zlochower, I.A., Green, G.M., 2009. The limiting oxygen concentration and flammability limits of gases and gas mixtures. J. Loss Prev. Process Ind. 22, 499e505.
Contents lists available at ScienceDirect
Journal of Loss Prevention in the Process Industries journal homepage: www.elsevier.com/locate/jlp
Flammability of gases in focus of European and US standards Maria Molnarne*, 1, 2, Volkmar Schroeder Bundesanstalt fuer Materialforschung und -pruefung (BAM), Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 January 2017 Received in revised form 11 May 2017 Accepted 12 May 2017 Available online 13 May 2017
The presentation will discuss the difference between EU and US standards for the determination of explosion (flammability) limits and limiting oxygen concentration. Small differences observed in measured values can be traced back to the different test apparatuses and criteria. The discrepancies can be much greater in the case of limiting oxygen concentration because of the high amount of inert gases and the corresponding low laminar burning velocities. The paper describes some examples and the influence of the chosen criteria on the results. The European and US standards use the criteria of flame propagation in open test vessels and of pressure rise in closed ones. The examples discussed show that flame propagation is still possible at very small pressure rise values, as observed much below the pressure rise criterion of usual standards. However, flame propagation in a process plant can cause an accident or explosion and must be avoided. Therefore, the flame propagation criterion is recommended to be used in chemical safety engineering. The European safety database CHEMSAFE contains expertevaluated safety data for cases where the determination method and criteria are known. Flammability characteristics based on the pressure rise criterion may suffice in certain cases, e.g. for explosion protection in closed vessels without any connecting pipes. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Flammability Limiting oxygen concentration Test methods Standards
1. Introduction
2. Terminology: “flammability limits” vs. “explosion limits”
Flammability of gases is a fundamental material property that entails legal classifications in various instances and requires appropriate labelling. Explosion protection measures are necessary only when flammable gases are present, e.g. in processes, storage and use. In European technical legislation, a gas or a gas mixture is regarded as “flammable” if under atmospheric conditions an explosion range (explosion limits) in a mixture with air exists. Therefore determination methods for flammability and explosion limits are standardized in many countries. Nevertheless, explosion limits are not the type of independent physicochemical material characteristics such as boiling temperature or density of a substance. As most other safety characteristics they are influenced by the test apparatus and the determination procedure applied. The evaluation and standardization of determination methods for flammability limits and limiting oxygen concentrations are particularly important for chemical safety engineering.
The expression “explosion limit” often is called “flammability limit” in the literature. But “flammable” is an ambiguous term: it means combustible in air on the one hand and explosively reacting without any further addition of air or another oxidizer on the other. Typical examples are flammable gases, e.g. pure hydrogen and flammability (explosion) ranges that characterize hydrogen-air mixtures. In the European ATEX directives and standards of explosion prevention the term “explosion” is used for explosively reacting mixtures to avoid misunderstanding. The actual situation in standardization and scientific literature is characterized by the fact that most of the US authors use the term “flammability limit”, while European standardization calls the same material safety characteristic “explosion limit”. Nowadays there are discussions in international standardization committees about a different meaning in the sense of more violent reactions if the “explosion limit” is reached, compared to slower reactions at the “flammability limit”. An informal IEC paper (IEC, 2003) used the direction of flame propagation as a criterion. Bureau of Mines measurements on hydrogen-air mixtures are often cited, where an upward flame propagation was observed at 4 mol% hydrogen but downward propagation does not begin before 9 mol% hydrogen (Coward and Jones, 1952). Therefore, 4 mol% is proposed as the
* Corresponding author. E-mail address: [email protected] (M. Molnarne). 1 Formerly at BAM, currently: Research Associate at MKOPSC, Texas A&M, USA. 2 Home Address: Brettnacher Str. 17b, 14167 Berlin, Germany. http://dx.doi.org/10.1016/j.jlp.2017.05.012 0950-4230/© 2017 Elsevier Ltd. All rights reserved.
298
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
lower flammability limit (LFL) and 9 mol% as the lower explosion limit (LEL). Nevertheless, such definition is not helpful in the field of explosion protection. There is actually no international standard available for the criterion “downward propagation” on the one hand and it can never be known where a potential ignition source is located in a process plant on the other. Therefore, only the more conservative upward propagation can be the relevant criterion for explosion protection. To avoid misunderstanding it must be pointed out that “flammability limit” in the US and “explosion limit” in Europe represent the same material property, determined by similar standard test procedures using the same criterion “upward propagation” of flames. 3. Comparison of European and US standards for determination of explosion limits With the creation of European directives in the field of explosion prevention, which apply uniformly to all Member States, it became necessary to develop new European standards for the determination of explosion characteristics. The European Technical Committee 305 “Potentially explosive atmospheres e Explosion prevention and protection” (CEN/TC305) is heading this project. Working Group 1 is responsible for the development of new standards for the determination of safety characteristics of gases, vapors and dusts. In this field a new European draft prEN 1839:2015 “Determination of explosion limits and of the limiting oxygen concentration of gases and vapors” is under development. It is a merged revised version of EN 1839:2012 (EN, 1839; 2012) (explosion limits) and EN 14756:2006 (EN 14756, 2006) (limiting oxygen concentration). Parallel to the European standards, corresponding US standards are available (ASTM E681 (ASTM E681-09, 2015), ASTM E2079 (ASTM E2079, 2013)). BAM has performed tests (Schroeder and Daubitz, 2004) to compare US and European standards for a number of years with 5 test substances, 4 gases and a vapor, ethanol, chosen for the tests. The gases, hydrogen, ethylene, methane and ammonia, had different laminar burning velocities. It was believed that differences in the results due to test apparatuses and test procedures could be checked by measurements on the same test substances. 40 mol% of nitrogen was added to the gas-air mixtures to increase sensitivity. Altogether 4 test methods were used (for parameters see Table 1). The parameters of apparatuses and test methods are the same as in the latest versions of the standards (e.g. prEN,
1839:2015 and ASTM E681-09), with the only exception of ammonia. Nowadays in European and US standards, special procedures are available for so-called “difficult-to-ignite” substances such as ammonia and halogenated refrigerants. Tests to compare these test methods should be carried out in the future. Three of the four test methods (DIN 51649 (DIN 51649, 1986), EN 1839-T (tube method), ASTM E681) are so-called “open vessel methods” that apply upward flame propagation as the criterion for a reaction. Another method was the “closed vessel method” EN 1839-B (bomb method) that used the criterion of pressure rise. The following conclusions were drawn from the experiments (see Table 1 and Figs. 1e3): The procedures according to DIN 51649 and EN 1839-T provide identical results in almost all cases. The reason lies in the very similar test apparatuses of the methods. The ASTM method yields similar explosion ranges in many cases. This can be explained by the use of a 5-dm3 flask in connection with the sensitive visual criterion. However, a major disadvantage of this method is the undefined step size in connection with the definition of the explosion limit (mean value between ignition and non-ignition point). This can easily result in unsafe explosion data if the concentration steps are not specified clearly. The closed vessel method EN 1839-B shows the strongest deviations. LELs are higher compared to the other methods and a clearly lower ethylene UEL was observed. The reason might be the pressure threshold criterion used that is obviously less sensitive than the optical criterion. Ammonia was an exception with a significantly higher UEL. The large quenching distance of ammonia may play a role, thus ammonia reactions in the 14dm3 sphere are preferred opposed to smaller volumes. In general, the deviations of difficult-to-ignite gases, e.g. ammonia and nitrogen mixtures, are the greatest. Such reactions are strongly affected by apparatus parameters. 4. Comparison of the flame propagation criterion and pressure rise criterion Determination methods for explosion limits at elevated pressures need another criterion for a reaction. Usually a measured pressure rise (pex/p0) in a closed ignition vessel is used. ASTM E91883 (2011) uses a pressure rise pex/p0 > 7%, EN 1839-B a pressure rise pex/p0 > 5% as the criterion. De Smedt & Berghmans (De Smedt
Table 1 Key characteristics of the most commonly used standard test methods for explosion limit determination. DIN 51649-1 withdrawn in 2004
EN 1839-B:2004
ASTM E681-01
vertical glass cylinder, open ∅ ¼ 60 mm, H ¼ 300 mm Ignition source induction sparks typical 0.5 s, min. 0.2 s Criterion for visual, ignition flame detachment
closed sphere or cylinder H/D ¼ 1 to 1.5, V > 5 dm3 induction sparks or exploding wire, E ¼ 10 Je20 J pressure increase 5% þ ignition pressure in air
flask, spherical V ¼ 5 dm3 (V ¼ 12 dm3) induction sparks or exploding wire
Increment
freely selectable, to be specified in report if x > 10% 10% (rel.) to 0.2 mol% (abs.), depending on the (rel.) of the test substance fraction test substance fraction 4 1
Explosion vessel
EN 1839-T:2004
vertical glass cylinder, open ∅ ¼ 80 mm, H ¼ 300 mm induction sparks typical 0.2 s, max. 0.5 s visual, flame detachment >100 mm or aureole H > 240 mm 0.1e0.2 mol%, depending on the 10% (rel.) to test substance fraction 0.2 mol% (abs.), depending on the test substance fraction 5 4
Number of repetition tests Explosion limit last non-ignition point
last non-ignition point
last non-ignition point
visual, flame propagation up to 13 mm to the wall (horizontal or vertical)
mean value from non-ignition and ignition point
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
299
Explosion range Explosion range at 40 mol% N2 DIN 51649
ASTM E681
EN 1839 T
EN 1839 B
0
10
20
30
40
50
60
70
80
Amount of hydrogen in mol% Fig. 1. Explosion ranges of hydrogen-air and hydrogen-nitrogen-air mixtures (high burning velocity).
Explosion range Explosion range at 40 mol% N2
DIN 51649
ASTM E681
EN 1839 T
EN 1839 B
3
5
7
9
11
13
15
17
Amount of methane in mol% Fig. 2. Explosion ranges of methane-air and methane-nitrogen-air mixtures (normal burning velocity).
et al., 1999) proposed 2% as the best pressure rise criterion. Until now only ASTM E918 provides a standard test method for evaluated pressures. In 2015 the European standardization committee TC 305 also proposed to develop a standard for explosion limits at elevated pressures and temperatures to improve process plant safety. Therefore BAM and PTB in Germany started a research project in €der et al., 2013). The focal point was the evaluation this field (Schro of apparatus parameters and criteria for higher temperatures and pressures. Therefore, a special test autoclave (volume ¼ 11 dm3) was built, which was equipped with two pressure resistant windows. This autoclave enables the observation of flame propagation and pressure rise for the same explosion up to a pressure of 100 bar (see Fig. 4). It could clearly be observed that e particularly at higher initial pressures e the flame can propagate at a pressure rise below 5%. Table 2 shows that flame propagation exceeding 100 mm until to
the top was observed at pressure rises of 3.6% in a period of time of 0.1e1s. This is in agreement with De Smedt & Berghmans’ criterion of 2%. Summing up, most of the international standards use the flame propagation criterion for atmospheric conditions. The aim of explosion protection is to avoid the spreading of flames through gas or vapor atmospheres. Therefore, a pressure rise criterion that corresponds to flame propagation should also be applied to nonatmospheric conditions in process plants. 5. Explosion ranges and limiting oxygen concentration Explosion limits are concentration limits of flammable gases or vapors in mixtures with air or air and an inert gas at which flame propagation just failed. The limiting oxygen concentration is defined as the maximal permissible oxygen content in a flammable
300
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
Explosion range Explosion range at 20 mol% N2 DIN 51649
ASTM E681
EN 1839 T
EN 1839 B
13
17
21
25
29
33
37
Amount of ammonia in mol% Fig. 3. Explosion ranges of ammonia-air and ammonia-nitrogen-air mixtures (low burning velocity).
Fig. 4. (a) Experimental device with windowed autoclave, camcorder and (b) frontal view into the autoclave with igniter and a propagating flame.
Table 2 €der et al., 2013). Most evident flame propagation, depending on the methane amount, UEL methane-air mixture, p0 ¼ 5 bar (Schro Flame Propagation
CH4 Time after Ignition pex/p0
19.0 mol% 0.24 s 1.2%
18.2 mol% 0.96 s 17.4%
21.0 mol% 0.24 s 0.8%
18.3 mol% 0.40 s 5.9%
23.0 mol% 0.24 s 0.8%
gas-inert gas-air mixture therefore a so-called explosion range must be specified for a 3-component system (flammable gaseinert gaseoxidizer). In the narrower sense (two-component systems) the
18.4 mol% 0.40 s 3.8%
25.0 mol% 0.24 s 0.7%
18.6 mol% 0.40 s 3.6%
27.0 mol% 0.12 s 0.2%
explosion range is the range between lower and upper explosion limit of a flammable gas in a mixture with air. Flammable gas-air/ oxidizer mixtures within this concentration range are explosive.
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
In a broader sense (3-component systems) the explosion range is the range inside the border line of lower and upper explosion limits. All mixtures at concentrations inside this range are explosive (see Fig. 5). The points on the left axis (Fig. 5) are the upper and the lower explosion limit (UEL, LEL) of methane in air. SCO is the line of stoichiometric concentration of the combustion reaction. LAC is the limiting air concentration. Such triangular diagrams enable the determination of the socalled limiting air concentration (LAC). LAC (red line) passes through the apex (tip) of the explosion range running parallel to the inert axis. The limiting oxygen concentration (LOC) can be calculated from the limiting air concentration according to Eq. (1) considering the oxygen fraction in air:
100
301
0 10
90
UEL
20
80 30
70
40 x
60
N2
50
x H2
50 60
40
70
30
LOC ¼ 0:209 LAC
(1)
The European standard EN 14756 uses for LOC determination the EN 1839 “T” or “B” apparatus. The authors recommend using the “T” method if possible. As shown in Chapters 3 and 4, flame propagation is often more sensitive than the 5% pressure rise criterion. Flame velocities at the apex of an explosion range are very low and the criterion and apparatus parameters can have a significant influence on the measured LAC. EN 14756 specifies two methods for the determination of LOC. The so-called short method is based on the typical shape of explosion ranges as shown in Fig. 5: most hydrocarbons have a sharp tip at the maximum inert gas concentration. The UEL line is not parallel to the inert axis in such cases. Therefore, it is sufficient to measure LEL, UEL and the apex of the explosion range. The apex can be determined by measuring the explosion limits belonging to the stoichiometric line SCO or, in the case of nitrogen, an inert gas, belonging to the line x ¼ 1.2 * LEL. In the case of broader explosion ranges, e.g. hydrogen-nitrogen-air mixtures, this is not sufficient. Here the so-called extended procedure must be used. Since it is not known where the LAC runs, the whole diagram must be measured (see Fig. 6). The criterion of EN 14756 telling when the extended procedure has to be applied, is given by Eq. (2):
UEL > 0:8 100 xair; L
(2)
x air, L is the air concentration at the explosion limit measured at
100 90
LEL
LAC
100
0
100
80
90
90
20 30
UEL 20
0
100
70
30
70
60
30 50 40 Air in mol%
10
N2 60 in mol% 70
Explosion range
80 90
LEL 100
20
50
10
90 SCO
80
40
20
80
LAC 90
(Limiting concentration for stability)
ETO 50 in mol% 40
60
10
'UEL'
60
in mol%
50
LEL
20
70
Inert gas
Explosion range
0
10
30
30
10
20
0
80
50
30
40
the stoichiometric line. In contrary to the use of different test methods (bomb method or tube method, see Fig. 8), the choice of the determination method of LOC (short method or extended method) does not lead to different results. However, as described above, the short method is not applicable to all test substances. In the case of chemically unstable substances, e.g. acetylene, ethylene oxide (ETO), tetrafluoroethylene etc., LOC cannot be determined. These gases can react without any amount of air and so the LOC would be zero. In the case of gas mixtures that contain unstable components one should be careful and take the extended procedure. The same is recommended for substances that have double bonds or other high-energy bonds in the molecule. Fig. 7 shows the ternary diagram of ethylene oxide as an example. At lower amounts of ETO, a combustion reaction takes place
100
60
50
xair
Fig. 6. Explosion range of hydrogen-nitrogen-air mixture at 20 C and 101 kPa. The points on the left axis are the upper and lower explosion limits (UEL, LEL) of hydrogen in air. LAC is the limiting air concentration.
80
Flammable gas in mol% 40
60
70
10
40
90
10
0
70
80
Explosion range
20
0
Fig. 5. Explosion range of the methane-nitrogen-air mixture at 20 C and 101 kPa.
0
100 100
90
80
70
60
50 40 Air in mol%
30
20
10
0
Fig. 7. Explosion range of ethylene oxide (ETO)-nitrogen-air mixture at 20 C and 101 kPa. The limiting concentration for stability is about 20 mol% nitrogen without any amount of air (CHEMSAFE, 2013e2015).
302
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
100 90
0 10 20
80 30 70
Nitrogen
40
60
50
in mol%
50
Methane
60
40
in mol%
70
30
10 0
90
80
90
LAC 1
Explosion ranges
100
80
LAC 2
20
100 70
60
30 50 40 Air in mol%
20
10
0
Fig. 8. Explosion ranges of methane-nitrogen-air mixtures at 20 C and 101 kPa measured acc. to EN1839 tube method (red, dotted) and bomb method (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
predominantly. As the proportion of ETO increases, a decomposition reaction occurs. At the upper explosion limit the flame propagates without any availability of air. Unlike the European standard EN 14756 (or the new draft prEN, 1839:2015) the ASTM E2079 describes a separate test apparatus and does not use the glass flask of ASTM E681 with open vessel and visual criterion. ASTM E2079 uses a bomb method and a pressure rise of 7% as a criterion, not flame propagation. It can be expected that LOC values are higher than those measured according to EN 14756 using the flame propagation criterion and the EN 1839-T glass tube. One example for the difference of LOC using tube method vs. bomb method of EN 1839 is given in Fig. 8. The LAC values resulted from the explosion range curves differentiate each other significantly. So a LOC 1 of 9.9 mol% correspond to LAC 1 (visual criterion) and a LOC 2 of 11.0 mol% to LAC 2 (pressure rise criterion of 5%). This results correspond to the NFPA 69 value of methane, LOC ¼ 11.2 mol% (Brandes and Ural, 2008), which was measured of according to ASTM E2079 (bomb method). 6. Safety characteristics derived from explosion ranges of CHEMSAFE The CHEMSAFE Database (CHEMSAFE, 2013e2015) contains rated safety characteristics data for fire and explosion protection for approximately 3000 different flammable gases, liquids and dusts and mixtures of gases, vapors, and dusts. The data include explosion limits, flash points, ignition temperatures, limiting oxygen concentrations, minimum ignition energy, maximum experimental safe gap, etc. This type of data is the basis for the classification of flammable substances with regard to testing, choice and use of explosion protected electrical equipment, for the transport of dangerous goods, the Ordinance on Hazardous Substances (classification, labelling and packaging, directive EG 1272/2008) and for many European and ISO standards. CHEMSAFE is a joint project of three partners from Germany: Physikalisch-Technische Bundesanstalt (PTB), Bundesanstalt für Materialforschung und -prüfung
(BAM), and the Gesellschaft für Chemische Technik und Biotechnologie e.V. (DECHEMA). PTB is responsible for the data of flammable liquids and BAM for flammable gases and dusts. The recommendations in the database are given by the experts of these two German federal institutes: BAM and PTB. CHEMSAFE is updated yearly and it has been accessible online since 1989. The database is available as a chargeable version all over the world on the internet (http://www.dechema.de) and as an in-house version in German and English, developed by DECHEMA, Frankfurt/Main. In the future CHEMSAFE will only be available as an online version instead the current different versions of different functionality. This will be implemented by PTB taking over the hosting of the database. The “[email protected]” newsletter will announce future upgrades of the Open Access version to registered customers. This English/German version will be available from the beginning of 2017. In addition to safety characteristic data, the database also delivers substance description (CAS registry number, names) and classification according to German and international regulations, physicochemical properties such as density, boiling point, etc., and bibliographic information about measured data. The database forms the basis for the Data Books “Safety Characteristic Data” €ller, 2008) and 2 (Molna rne et al., 2008) Volumes 1 (Brandes and Mo of PTB and BAM. To avoid an explosive atmosphere, the knowledge of explosion limits of flammable gases and vapors in air are essential. If an explosive atmosphere cannot be prevented, inert gases such as diluents can be applied to reduce or avoid the explosive mixtures, e.g. in the case of inerting/purging. The limiting oxygen concentration (LOC) and the limiting values of flammability (Tci values) are very important safety characteristic values which are derived from explosion diagrams of mixtures containing flammable gas, inert gas and air. These data and the corresponding explosion diagrams are available in CHEMSAFE, in the “Explosion Regions of Gas Mixtures” rne et al., 2008) Data Book, and the limiting values from (Molna these diagrams are also listed in ISO 10156 standard (ISO 10156, 2010). This standard specifies methods to determine whether or not a gas or gas mixture is flammable in air and is more or less oxidizing than air under atmospheric conditions. It is also used for classification of gases and gas mixtures including the selection of gas cylinder valve outlets and for labelling according to international transport regulations and dangerous substances regulations such as in the Globally Harmonized System (GHS). The in-house version and the future Open Access online CHEMSAFE Database will provide a graphical representation of these data for the explosion range of these mixtures in Cartesian or triangular diagrams. A general view of these explosion diagrams using a triangle plot can be seen in Fig. 9. The limiting oxygen concentration (LOC) is the oxygen concentration in a flammable gas-air-inert gas mixture below which an explosion cannot occur under specified conditions or, more generally, the “maximum oxidizing gas concentration” (MOC) is the maximum permissible oxidizing gas content in a 3-component mixture e which can be used in the case of oxidizing gases other than oxygen e and it is valid not only under atmospheric conditions. The “maximum permissible amount of flammable gas concentration” (MXC) generally can be used in an inert gas/flammable gas mixture in mol % or in the special case, if the inert gas is nitrogen, it is called “Threshold for flammability of the flammable component i in the mixture with nitrogen (Tci)” which corresponds to the MXC value with air as oxidizer and nitrogen as inert gas at room temperature and atmospheric pressure. The “minimum required inert gas concentration” (MAI) in mol% in an inert gas-oxidizer mixture is the smallest fraction of the inert gas for which this gas mixture cannot become explosive irrespective of the amount of flammable gas
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
100
of old apparatuses and the different criteria for flammability.
0 10
90
Interpolated curve Experiment
20
80 30 70
Flammable gas 50 in mol%
40
IAR line
60
50 MOC line
40
Inert gas 60 in mol% 70
30
80
20 10 0
C Explosion range
MAI
90
ICR line
90
80
70
50 60 40 Oxidizer in mol%
MXC
100
MOC
100
30
20
10
0
Fig. 9. Safety characteristics derived from the explosion diagram taken from CHEMSAFE Database.
being added. IAR (minimal inert/air ratio) and ICR (minimal inert/ combustible ratio) lines are straight lines drawn to the border of the explosion range and they represent the limiting ratios of inert gas to oxidizing gas and inert gas to flammable gas, respectively. The LOC values can be calculated by Eq. (1) using the MOC values in the case of air (they are equal to the LAC values). Table 3 shows some recommended limiting oxygen concentration data taken from CHEMSAFE. The LOC values stem from the following ternary mixtures: flammable gas-air-carbon dioxide and flammable gasair-nitrogen. The measuring method is the DIN 51649 or EN 1839-T method when not otherwise commented. A comparison of these values with the US Bureau of Mines and the NFPA 69 (2008) edition values from Table 1 in Brandes and Ural (Brandes and Ural, 2008) shows substantially higher values for almost all substances. They explain these differences with the use
Table 3 Limiting oxygen concentration (LOC) of ternary mixtures. Flammable Gas
Ammonia Butane 1-Butene Carbon monoxide (pure and dry) Carbon monoxide 1-Chloro-1,1-difluoroethane 1,1-Difluoroethane Ethane Ethene (Ethylene) Ethyne (Acetylene) Ethylene oxide Hydrogen Hydrogen sulfide Methane 2-Methylpropane 2-Methyl-1-propene Propane Propene Propylene oxide a b
303
Inert CO2
Inert N2
LOC mol%
LOC mol%
13.2a
5.1b
11.8 10.4 6.3 0.0 5.2 10.5a 13.6 12.8 13.2 12.6 12.5 10.3
Test method EN 1839-B. With traces of moisture, hydrogen or hydrocarbons.
12.2 9.6 9.7 6.1 4.7b 14.1 9.8 8.7 7.6 6.2 0.0 4.3 8.9a 9.9 10.3 10.6 9.3 9.4 7.7
7. Discussion and conclusion The explosion limits and limiting oxygen concentration are affected by the chosen test apparatus, e.g. the type and size of the ignition vessel (closed, open, dimensions, material), ignition source (type, power, duration and location) and the test criterion (flame propagation or pressure rise) and accuracy. In the international standardization there are test methods for the experimental determination of explosion limits using open and closed ignition vessels. The flame propagation criterion is usually used for open ignition vessels and the pressure rise criterion for closed ignition vessels. The European standard EN 1839 describes both methods, i.e. the open ignition vessel (tube method, “T”) and the closed ignition vessel (bomb method, “B”). A flame spread of 10 cm and a pressure rise >5% of the initial pressure have been selected as the criterion for atmospheric conditions. Investigations in an autoclave having viewing windows performed at BAM showed that flame propagation still can be observed even if pressure rise less than 5%. In conformity with De Smedt & Berghmans’ suggestions it is proposed to reduce the pressure threshold criterion to 2%. The examples discussed show that flame propagation is still possible at very small pressure rise values, as observed much below the pressure rise criterion of most standards. However, flame propagation in a process plant can cause an accident or explosion and must be avoided. Therefore, the flame propagation criterion is recommended to be used in chemical safety engineering. Experimental studies of BAM showed that the two determination methods of EN 1839 provide somewhat different results. The explosion areas that were determined in open vessels are usually a little larger than those measured in closed vessels. This is especially significant when explosion limits of mixtures with a high proportion of inert gas are determined. Therefore, it is expected that especially the LOC values, which are determined using the “tip” of the explosion range can be strongly influenced by the selected measuring method. It is not surprising that LOC values which were determined according to EN 14756 using the tube method, are usually smaller than those measured according to ASTM E2079 (bomb method). CHEMSAFE includes triangle diagrams of many explosion ranges of flammable gas-inert gas-air mixtures. The data were evaluated by BAM and PTB experts. In addition to the safety characteristics and specifications, the database also describes the measuring methods. The explosion areas enable the determination of important safety characteristics such as LEL, UEL, LOC, Tci, etc. The database is currently still fee-based, but it will be available for free of charge on the Internet from 2017. It is generally recommended to select data of explosion limits and LOC of flammable gas/vapor, air, and inert gas mixtures determined by the same experimental methods used for safety concept, risk analysis and estimation methods. The international standardization is useful in harmonization of different national safety concepts and in understanding the different safety cultures of individual countries. Many efforts have been taken in the past 20 years to analyze the differences of the published explosion/flammability limits and limiting oxygen concentration data of both European and US researchers. The Safekinex (Safekinex, 2009) €der and Molnarne (Schro €der research project, publications of Schro and Molnarne, 2005), Brandes and Ural and Zlochower (Zlochower and Green, 2009) have to be mentioned here. Britton (Britton et al., 2016) has recently published a paper about LOC which also discusses the differences between the US and European experimental data.
304
M. Molnarne, V. Schroeder / Journal of Loss Prevention in the Process Industries 48 (2017) 297e304
As it was mentioned earlier the European standards define LEL, UEL and LOC as the concentration points at which flame propagation has just not occurred. In the US standards the border concentration points between flammable and not-flammable concentrations are used in the case of visual flame propagation criterion (see Table 1). When comparing the data, it is essential to know the chosen concentration step size in the test procedure. Therefore, it is useful to follow the discussion between the US and European standardization authorities aimed at establishing a global standard for the determination of these important safety characteristics. Acknowledgements The authors would like to thank BAM and PTB for the presentation materials they provided. References ASTM E2079-07, 2013. Standard Test Methods for Limiting Oxygen (Oxidant) Concentration in Gases and Vapors. ASTM International, American Society for Testing and Materials. ASTM E681-09, 2015. Standard Test Method for Concentration Limits of Flammability of Chemicals (Vapors and Gases). ASTM International American Society for Testing and Materials. ASTM E918-83, 2011. Standard Practice for Determining Limits of Flammability of Chemicals at Elevated Temperature and Pressure. ASTM International American Society for Testing and Materials. €ller, W., 2008. Safety characteristic data. In: Flammable Liquids and Brandes, E., Mo Gases, vol. 1. NW-Verlag (Publishing Co.), Bremerhaven, Germany. Brandes, E., Ural, E.A., 2008. Towards a global standard for flammability determination. In: Proceedings of the 42nd Annual Loss Prevention Symposiumeglobal Safety Congress, Paper 2E, April 6, 2008. Britton, L.G., Clouthier, M.P., Harrison, B.K., Rodgers, S.A., 2016. Limiting oxygen
concentrations of gases. Process Saf. Prog. 35, 107e114. CHEMSAFE d Database for Recommended Safety Characteristics, 2013-2015. BAM, PTB, and DECHEMA, In-house Version (INTERNET: dechema.de/en/ chemsafe.html). Coward, H.F., Jones, G.W., 1952. Limits of flammability of gases and vapors, US Bureau of Mines. Bull 503. De Smedt, G., de Corte, F., Notele, R., Berghmans, J., 1999. J. Hazard. Mater. A70, 105e113. DIN 51649 Teil 1, 1986. Bestimmung der Explosionsgrenzen von Gasen und Gasgemischen in Luft. Deutsches Institut für Normung, Berlin withdrawn in 2004. EN 14756, 2006. Determination of the Limiting Oxygen Concentration (LOC) for Flammable Gases and Vapours. CEN European Committee for Standardization, Brussels. EN 1839, 2012. Determination of Explosion Limits of Gases and Vapours. Sections 4.1, Method T (EN 1839-T) and 4.2, Method B (EN 1839-B). CEN European Committee for Standardization, Brussels. IEC Inf Paper 31/460/INF, 2003-03-07. “Explosive” Limits and “Flammable” Limits. ISO 10156:2010, 2010. Gases and Gas Mixtures d Determination of Fire Potential and Oxidizing Ability for the Selection of Cylinder Valve Outlets. rne , M., Schendler, Th, Schro €der, V., 2008. Safety characteristic data. In: Molna Explosion Regions of Gas Mixtures, vol. 2. NW-Verlag (Publishing Co.), Bremerhaven, Germany. Safekinex, 2009. SAFe and Efficient hydrocarbon oxidation processes by KINetics and Explosion eXpertise. In: Energy, Environment and Sustainable Development. Programme (Contract Number EVG1-CT-2001e00098), EU Project. €der, V., Molnarne, M., 2005. Flammability of gas mixtures. Part I: fire potential. Schro J. Hazard. Mater. 121, 37e44. €der, V., Tschirschwitz, R., Brandes, E., 2013. BAM Research Report No. 2-3408/ Schro € ßen 2013 „Sicherheitstechnische Kenngro bei nicht-atmosph€ arischen Bedingungen“. Schroeder, V., Daubitz, R., 2004. Evaluation of standard test methods for the determination of explosion limits of gases and vapours. In: Proceedings 11th Symposium on Loss Prevention 2004, pp. 2109e2117, 31 May-3 June, Praha, Czech Republic, Them. Sect. B. Zlochower, I.A., Green, G.M., 2009. The limiting oxygen concentration and flammability limits of gases and gas mixtures. J. Loss Prev. Process Ind. 22, 499e505.