Journal of Loss Prevention in the Process Industries 26 (2013) 1106e1112

Contents lists available at SciVerse ScienceDirect

Journal of Loss Prevention in the Process Industries

journal homepage: www.elsevier.com/locate/jlp

Limiting oxygen concentration for coal dusts for explosion hazard

analysis and safety
Manju Mittal
Fire Research Laboratory, CSIR-Central Building Research Institute, Roorkee 247667, Uttarakhand, India

a r t i c l e i n f o a b s t r a c t

Article history: Investigation of explosion characteristics of coal dust was undertaken as a part of regular research
Received 7 February 2013 program at CSIR-CBRI, Roorkee, India, for designing explosion safety measures for coal dust handling
Received in revised form installations. This paper presents results of detailed experimental work on determination of Limiting
22 April 2013
Oxygen Concentration (LOC) and influence of reduced oxygen levels on explosion severity data for two
Accepted 25 April 2013
types of coals with varying volatile matter as 27.18% (coal A) and 19.69% (coal B) from Jharia coalfield of
India determined at ambient conditions with 20-L Spherical Vessel established at CSIR-CBRI. The effects
Keywords:
of coal particle size and moisture content were evaluated. Data presented will be used for hazard
Dust
Explosion
analysis, designing explosion preventive measures, and explosion severity reduction by involving the use
Coal of inert gases for installations handling pulverized coal with similar nature. The importance of ignition
source energy in determining LOC data is highlighted. The data collected lead to an extension of the
current data for coal dusts as found in the literature. Limiting oxygen concentrations were found as 7% for
coal A and 8% for coal B for the size representative to that used in pulverized coal boilers and moisture
content w4%.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction of the dispersed dust cloud which will just not allow an explosion of
a dust/air/inert gas mixture.
Handling of pulverized coal suspension in power plants, coal The possibility of dust explosions in process equipment can in
mines, cement manufacturing plants and other related activities principle be effectively eliminated by substituting the air by a gas
is associated with the risk of coal-dust explosions. The current which makes flame propagation in the dust cloud impossible. This
environmental regulations have introduced confined or partially technique is commonly employed in industry to prevent dust cloud
confined systems for transportation and storage of pulverized coal ignition. Since the use of large quantities of inert gas in a plant may
increasing the probability of coal dust explosions. Such installations be expensive, it is important to limit the inert gas consumption to
should conform to relevant regulations like ATEX 95 related to the extent possible. For most dusts it is not necessary to substitute
explosive atmosphere (ATEX 95, EU Directive 94/9/EC, 2009), the entire atmosphere in the actual area by inert gas e.g. nitrogen
which in turns needs the evaluation of explosion parameters for and it is essential to know the critical gas composition for inerting
safety, hazard analysis and for designing appropriate explosion the dust. In some cases it may even be of interest to use smaller
safety measures for different circumstances under which explo- fractions of inert gas than required for complete inerting, because
sions are likely to occur depending on the particular technological this will reduce both the ignition sensitivity of the dust cloud, and
application. The data needed for explosion hazard analysis and the maximum explosion pressure and rate of pressure rise.
safety include the basic explosibility parameters e maximum The purpose of determining LOC is to estimate the highest
pressure and rate of pressure rise, as well as explosibility limit permissible oxygen concentration below which dispersed dust
parameters such as dust concentration, oxidant concentration, ignition is not possible at ambient temperature and pressure. This
ignition energy and minimum ignition temperature. The dust parameter is measured in a 20-L Sphere (BS EN 14034-4, 2005;
concentration limit, often referred to as Minimum Explosible Siwek, 1985) which is internationally accepted for use in deter-
Concentration (MEC) is the lowest concentration of dust dispersed mining explosion severity data of dusts and is well suited to
in air that can propagate explosion. Another limit measurement of determine LOC of dusts. As a part of research program on deter-
use in hazard analysis and safety is the Limiting Oxygen Concen- mination of explosion characteristics of various dusts an exhaustive
tration (LOC) e the experimentally determined oxygen concen- investigation was undertaken on coal dust explosibility data using
tration at the boundary between propagation and non-propagation CSIR-CBRI 20-L Spherical Vessel (similar to Siwek 20-L Sphere)

0950-4230/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jlp.2013.04.012
 M. Mittal / Journal of Loss Prevention in the Process Industries 26 (2013) 1106e1112 1107

which covered a comprehensive review and analysis of existing

data on explosion parameters of coal dust and creation of new data
for specific coal dust samples for which such data do not exist. The
objective of present work was to determine the required LOC for
hazard analysis, designing explosion preventive measures, and
explosion severity reduction by involving the use of inert gases.
There is a wide range of explosibility data of coal dusts in
technical literature on various aspects including limiting oxygen
concentration (Cashdollar, 1996, 2000; Continillo, Crescitelli, Fumo,
Napolitan & Russo 1991a, 1991b; Eckhoff, 1991; Going, Chatrathi &
Cashdollar, 2000; Wilen et al., 1999, 117 pp. þ app.8 pp.;
Woskoboenko, 1988). Data reported on limiting oxygen concen-
tration (LOC) of coals using 20-L Sphere differ in particle size and
composition of coals. It may be desirable to determine the explo-
sibility characteristics of an industrial dust by conducting test as the
published data are for a particular size distribution and composi-
tion of dust, usually different from the dust in question. The study
included various parameters like minimum explosible dust con- Fig. 1. Schematic diagram-20-L Spherical Vessel for determining explosion charac-
centration; minimum ignition energy; minimum ignition temper- teristics of industrial dusts.

ature; maximum explosion pressures and maximum rate of

pressure rise; limiting oxygen concentration; and influence of placed in centre of 20-L Sphere. The upper half of the vessel is
reduced oxygen concentration on explosion severity data. How- bolted on. The vessel is partially evacuated to 0.4 bar prior to
ever, this paper covers results of LOC measurement experiments dispersing the dust. This evacuation of the 20-L Vessel by 0.4 bar
and effect of reduced oxygen levels on explosion severity data for together with air or air/nitrogen mixture in storage chamber
two bituminous coals with varying volatile matter from Jharia (þ20 bar; 0.6 l), results in the desired starting initial pressure
coalfield of India with CSIR-CBRI 20-L Sphere at normal initial (1 bar) for the experiment. The ignition source is pyrotechnic ig-
temperature and pressure using air/nitrogen as the suspension nitors (40% zirconium, 30% barium nitrate and 30% barium
media. Effects of ignition source energy, coal particle size, and peroxide). The degree of turbulence in the explosion chamber is
moisture content on LOC data were also studied. mainly a function of the ignition delay time e time between the
The coals selected for this study are classified as coal A (vola- onset of dust dispersion and the activation of the ignition source e
tiles-27.19%) and coal B (volatiles-19.69%). The sizes used are: size standardized to 60  5 ms to maintain a moderate level of turbu-
representative of those used in pulverized coal boilers (90% passing lence at the time of ignition of dust cloud.
through 200 BS Mesh and nothing remaining on 52 BS Mesh), For experiments at varying oxygen levels, 20-L Sphere is filled
74 mm and 38 mm. The moisture contents of all the three particle with air/nitrogen gas mixture of desired oxygen concentration
sizes for both the coals were varied in the range w4e15%. To the using partial pressure method. The oxygen concentration is verified
author’s knowledge, no coal dust data obtained with the 20-L using an oxygen analyzer designed for accurate and reliable mea-
Sphere for the coal samples selected have been published. surement of oxygen in air, nitrogen, etc. and calibrated using
standard calibration samples. A 3.5 digit LCD display provides
2. Experimental set-up and procedure direct percent oxygen level (by volume) in the range 00.0e25.0%
with least count 0.1% with accuracy of 2%. To maintain the same
The 20-L chamber used for determination of limiting oxygen actual oxygen concentration in initial stages of dust dispersion and
concentration of dusts is a hollow sphere made of steel. The ignition in 20-L Sphere, the gas storage chamber for dust dispersion
permissible working pressure for the sphere is 30 bar. The dust to is pressurized to desired pressure with premixed air/inert gas
be tested is dispersed from a pressurized storage chamber (0.6 l) by mixture having the same oxygen concentration as in 20-L Sphere.
means of outlet valve and a perforated annular nozzle using com- At least three dust dispersion tests are conducted without igniting
pressed air or airenitrogen mixture supply at 20 bar. Schematic the dust cloud to check the oxygen concentration in the explosion
diagram for instrumentation of 20-L Spherical Vessel is shown in chamber at the time of ignition for each oxygen level.
Fig. 1. Chamber top is hinged and opens across the full chamber Typical pressureetime curve recorded during the experiment is
diameter allowing easy access to the interior for positioning in- shown in Fig. 2. The pressure trace starts at the partially evacuated
struments and for cleaning. During routine operation at 2e4 tests value of 0.4 bar(a). The blast of air that disperses the dust starts at
per hour, the thick walls of the chamber provide sufficient heat sink 40 ms and ends at 90 ms on the pressureetime trace. The ignitor is
for the post explosion gases and particles to cool down to room activated at 100 ms at a chamber pressure of 1.0 bar(a). Pex is the
temperature without need for water cooling of the chamber. There maximum explosion pressure (above the pressure in the vessel at
are several ports in the chamber for vacuum creation, dust intro- the time of ignition). The value of Pex, for a test at a given con-
duction, installing various instruments for measuring pressures, centration, is the highest deflagration pressure (absolute) minus
etc. The dynamic pressure during explosion is measured using the pressure at ignition (normally 1 bar). (dP/dt)ex is the maximum
piezoelectric and strain gauge type pressure transducers provided rate of pressure rise reached during the course of a single explosion
at two ports. The strain gauge pressure transducers measure the experiment. Pmax is the maximum pressure (above pressure in the
explosion pressure and can also be used during partial evacuation vessel at the time of ignition) and (dP/dt)max is the maximum value
of the chamber prior to dispersion of dust and for adding gases to for the rate of pressure increase per unit time reached during the
the chamber by partial pressure since it is an absolute pressure course of explosion for the optimum concentration of the dust
gauge. The pressureetime curve is recorded by storage oscilloscope tested and equals maximum slope of a tangent through the point of
or high speed chart recorder. For experiments to measure explosion inflexion in the rising portion of the pressure vs. time curve. The
data, first the known quantity of dust sample is placed in dust values for Pmax and (dP/dt)max are the averages of the highest values
holder at the bottom of the dispersion nozzle and ignitors are (over the range of dust concentrations). The explosibility dust
1108 M. Mittal / Journal of Loss Prevention in the Process Industries 26 (2013) 1106e1112

explosions no longer occur at any dust concentration. Step

changes in the oxygen concentration are made in multiples of 1%
V/V at reduced oxygen levels. The highest oxygen concentration
where no ignition occurs in three consecutive tests is reported as
the LOC. The difference between the highest permissible oxygen
concentration, at which dust explosion no longer occurs for any
dust concentration and the lowest O2 concentration that gives
explosion should not exceed 1% V/V.
One of the on-going concerns in experimental determination of
LOC is that of overdriving the system by a large ignition source in
20-L Sphere. When the ignition source is too weak, LOC measured
in 20-L Sphere will be higher than the true value i.e. the LOC
measured with 1 m3 Vessel using 10 kJ ignitor (Going et al., 2000;
ISO, 1985). The system is underdriven and the results are based
more on ignitability than flammability. As the ignition energy is
increased, at some point, the energy level is excessive for the size of
the vessel and the system is overdriven. In this situation, the energy
contributed by the ignitor is sufficient to combust enough dust so
that the result appears to be an explosion although there is no real
propagation beyond the ignitor flame. An overly strong ignitor
can markedly change the initial test conditions by raising the
overall temperature of the system, which in turn would lower the
Fig. 2. Typical pressure-verses-time trace during a dust explosion experiment in 20-L apparent limits and a non-explosible system would appear to be
Spherical Vessel.
explosible. The LOC will be lower than the true value in overdriven
system. Effect of ignition source on LOC was studied in present
constant, KSt, characterizes the explosibility of the material. KSt is investigation.
maximum (dP/dt) normalized to a 1.0 m3 volume measured at the To prepare suitable dust samples of coal A and coal B for LOC
optimum dust concentration and defined in accordance with the tests it was decided to mill and sieve the coal to a particle size
following relationship (Eckhoff, 1991): representative of those used in pulverized coal boilers (90% passing
through 200 BS Mesh and nothing remaining on 52 BS Mesh) and
  dry them to a moisture content of less than 5%. The particle size
dP distribution was determined by normal sieving procedure. The
KSt ¼ $V 1=3 (1)
dt max method of mechanical sieving was evaluated by sieving a number
of coals for 10, 15 and 20 min. The results suggest a minimum
where, sieving time of 15 min. The particle size distributions of the pul-
verized coals used for the main part of study are shown in Fig. 3.
P e Pressure, bar However, some coal samples with particle sizes 74 and 38 mm were
t e Time, s prepared for studying effect of particle size on LOC. The moisture
V e Vessel volume, m3 content for each particle size was varied between w4 and 15% for
KSt e Explosibility dust constant, bar m/s observing the effect of moisture. The proximate analysis data and
calorific values of coals determined as per standard procedures (IS,
KSt values rounded to the nearest integer are used. 1970, 1984) are given in Table 1.
The LOC is determined by experiments conducted at various Experimental results for the tests carried out for coal A and coal
oxygen levels over a wide range of dust concentration as per B as per procedure described above were collected as experimental
standard procedure (BS EN 14034-4, 2005). The experiments are
started with an oxygen concentration of 21% to determine
the most severe explosion and for measurement of explosion data
at ambient conditions. Further tests are carried out at reduced
oxygen concentrations. The oxygen concentration is lowered in
steps of 3% initially and 1% when it came closer to LOC. At
the commencement of dust dispersion the pressure in 20-L
chamber is atmospheric. The pressureetime data for each test is
recorded. An ignition/explosion of dust is considered to have
taken place, when the measured overpressure (influence of
chemical igniters included) relative to the initial pressure Pi is
0.5 bar (Pex  (Pi þ 0.5 bar)). From pressureetime curve the
explosion pressure, Pex, and rate of explosion pressure rise, (dP/
dt)ex, can be determined for each dust concentration as explained
above. Optimum dust concentration, Cex, at which the highest
explosion pressure occurs at any given oxygen concentration, is
determined. Maximum explosion pressures, Pmax, and maximum
rate of pressure rise, (dP/dt)max are determined by varying the
dust concentrations at various oxygen levels. By increasing step by
step the ratio of inert gas to air and varying the dust concentra-
tion, the oxygen concentration is reduced to a level at which Fig. 3. Particle size distribution of pulverized coals.
 M. Mittal / Journal of Loss Prevention in the Process Industries 26 (2013) 1106e1112 1109

Table 1
Proximate analysis and calorific values of coals.

Property Coal A Coal B

Moisture (%) 3.92 3.80

Volatility (%) 27.18 19.69
Fixed carbon (%) 50.70 40.72
Ash (%) 18.20 35.79
Calorific value, cal/g 6011.76 3684.91

curves shown in Fig. 2. These were analysed and predicted data are
presented and discussed in next section.

3. Results & discussions

Experiments were carried out for pulverized coal A (particle

size distribution as per Fig. 3, moisture-3.92%) at various oxygen
concentrations in the range 5e15% for ignition energies e 1 kJ,
2.5 kJ, 5 kJ and 10 kJ for dust concentrations 60, 80, 100, 120, 150,
200, 250, 300, 350, 400, 450 and 500 g/m3, to examine the effect
of ignition source on LOC data. The pressureetime curves similar
to Fig. 2 were recorded for all the experiments. The experiments,
in which recorded explosion pressure was 0.5 bar, were
considered as indicating ignition/explosion. The results showing
the effect of ignition source energy on LOC data are presented in
Figs. 4 and 5. The ignitions/ explosions are shown as the solid data
symbols and the non-ignitions/non-explosions as the open data
symbols. The data indicate that ignition source has significant
effect on LOC data. As ignition energy increases lower quantities of Fig. 5. Effect of ignition energy on LOC data for pulverized coal A e experimental
the dust and oxygen are required to create explosion. A compar- results.
ison of 12% oxygen and 120 g/m3 coal coordinates on graphs for
different ignition energies indicates that 12% oxygen and 120 g/m3
point is outside the flammability envelope for 1 kJ ignition source, it is just inside the flammable zone for 2.5 kJ and is well within
the flammability envelope for 5 kJ and 10 kJ. Thus, increasing the
ignition energy increases the size of the flammability zone. The
boundary between oxygen concentration that supports combus-
tion and that does not support combustion is the limiting oxygen
concentration, LOC, which decreased from 11% at 1 kJ to 7% at
2.5 kJ, 6% at 5 kJ and 5% at 10 kJ. Some experiments were con-
ducted for measuring LOC of the same pulverized coal A in 1 m3
Chamber using 10 kJ ignitor. The resultant value of LOC was 7%.
The data from this study and the previous work (Going et al.,
2000) demonstrate that overdriving can occur when using
strong chemical ignitors and apparent LOC values are lower than
true value. 5 and 10 kJ energies are too strong and overdrive the
explosion for the 20-L chamber and inappropriate for LOC mea-
surement. 1 kJ ignition energy causes underdriving of the system.
The present results and literature data support that the 2.5 kJ
ignition source may be the most appropriate ignition source for
LOC measurement in the 20-L vessel to match the LOC data from
1 m3 vessel using 10-kJ igniter. This energy was therefore used for
further LOC experiments.
Based on above results, LOC for the pulverized coal A (particle
size distribution as per Fig. 3, moisture-3.92%) was considered 7%
corresponding to 2.5 kJ ignition energy. Further systematic ex-
periments were conducted to determine LOC for pulverized coal B
(particle size distribution as per Fig. 3, moisture-3.80%) and to
study the effect of reduced oxygen on explosion severity data for
these pulverized coal A and coal B at oxygen levels 21, 18, 15, 12, 9,
6, 8 and 7% for dust concentrations e 80, 90, 100, 125, 250, 500,
750, 1000 g/m3 using 2.5 kJ ignition energy. Each experiment was
repeated thrice. The experimental curves were analysed and
values of Pex, and (dP/dt)ex were estimated for each experiment.
Fig. 4. Effect of ignition energy on LOC data for pulverized coal A e experimental Pmax and (dP/dt)max and KSt were estimated at each oxygen
results. concentration.
1110 M. Mittal / Journal of Loss Prevention in the Process Industries 26 (2013) 1106e1112

Fig. 6. LOC data for pulverized coal B e experimental results.

Fig. 8. Effect of oxygen concentration on maximum explosion pressure for pulverized
coal B for various dust concentrations.

Fig. 6 summarizes ignition/explosion and non-ignition/non-

explosion data for all the experiments for pulverized coal B. In air,
the dust ignites and burns at all coal concentrations above MEC KSt) with oxygen concentration, for each series of Figs. 7 and 8 for
w90 g/m3. As the oxygen concentration decreases the minimum coal A and coal B, respectively. The maximum explosion pressure
explosible concentration increases and explosions occur down to in Fig. 9 varies almost linearly with oxygen concentration upto
9% O2. At 8% O2 the coal dust could not be ignited. LOC for pulver- certain oxygen level, in agreement with the behaviour of a simple
ized coal B is 8% in comparison to 7% for coal A. This shows that high gas-dynamic model. However, this linear relationship changes
volatile coals have lower LOC. as the LOC is approached. Near the LOC, the pressure would
The experimental results showing the variation of maximum decrease very rapidly with decreasing oxygen concentration,
explosion pressure with dust concentration (Pex) for each oxygen until mixture would no longer be explosible. As expected, ex-
level are presented in Figs. 7 and 8 for the pulverized coal A and coal plosion overpressure decreases with decreasing oxygen concen-
B, respectively. The maximum value of explosion severity data oc- tration. Fig. 10 indicates that the rate of pressure rise varies
curs at a dust concentration 500 g/m3 for both the coals. In 21% almost exponentially with oxygen concentration on this semi-
oxygen explosions could be observed upto a dust concentration logarithmic plot. Explosion severity and ignition sensitivity of
80 g/m3 for coal A and 90 g/m3 for coal B. These dust concentrations dust clouds decrease with decreasing oxygen content of the gas in
are equal to the minimum explosible concentrations (MEC) for which the dust is suspended. It is worth mentioning that a
these coals determined in the same experimental set-up using reduction of the oxygen content from that of air to 12% caused a
same ignition source. Explosible behaviour was observed and flame reduction of the maximum rate of pressure rise by a factor of five,
propagation and minor explosions were recorded for oxygen con- whereas the maximum pressure was reduced by less than a factor
centrations down to 8% for coal A and 9% for coal B. The results also of two. This illustrates the strong influence of the oxygen content
confirm the existence of rich limits, rarely observable at higher dust on the kinetics of the combustion process. The reduction of the
concentrations, and made evident at reduced oxygen concentration maximum pressure is approximately proportional to the reduc-
(8% curves in Fig. 7). Similar observations were found in literature tion of the oxygen content, as would be expected from thermo-
(Continillo et al., 1991a, 1991b) in experimental determination of dynamic considerations, whereas the maximum rate of pressure
LOC of South-African bituminous coal (volatile matter e 23% on dry rise falls much more sharply. There appears to be a correlation
basis) dust with 20-L Sphere where existence of rich limit was between the two variables. This is in agreement with the idea of
observed at 8% oxygen. The maximum explosion pressure varied the oxygen being the limiting reactant at optimum conditions. In
between 0.5 and 6.4 when oxygen concentration was increased fact, measurements of the residual oxygen in the after-gases
from 8 to 21% for pulverized coal A. shows that oxygen is almost totally consumed in the conditions
Figs. 9 and 10 show variation of maximum explosion pressure of maximum reactivity for each series of Figs. 7 and 8, down to
(Pmax) and normalized rate of pressure rise ((dP/dt)max$V1/3 i.e.

Fig. 7. Effect of oxygen concentration on maximum explosion pressure for pulverized Fig. 9. Variation of maximum explosion pressure at optimum dust concentration with
coal A for various dust concentrations. oxygen concentration for pulverized coals.
 M. Mittal / Journal of Loss Prevention in the Process Industries 26 (2013) 1106e1112 1111

has a volatile matter of 23% on dry basis e a value closer to that

for coals in present study. As reported in Table 2, the LOC for this
coal is <<10% (Continillo et al., 1991a, 1991b). In this work rich
limit of coal was observed at 8% oxygen level, while pressuree
time data at 6% oxygen show Pmax << 0.4 bar which may not be
considered as explosion. The LOC was therefore 6% for this coal.
Comparable results should in principle be obtained by all labo-
ratories when using the standard 20 l procedures. The variation
of the LOC values obtained by the laboratories reflects the diffi-
cult nature of the coal dust samples in respect of uniform test
conditions. A substantial part of variations in the explosion data
must also be ascribed to heterogeneous nature of samples.
Classification of coarse and fine particles during handling and
transportation and minor changes in the moisture content during
storage of the rather voluminous dust samples at different test
Fig. 10. Variation of normalized rate of explosion pressure rise ((dp/dt)max.V1/3 ¼ KSt) facilities could not be avoided. Significant differences in LOC
with oxygen concentration for pulverized coals. values were measured by LOM and TNO in the 20-L Sphere for
the more conventional lignite dusts (Wilen et al., 1999, 117 pp. þ
app.8 pp.).
12%. In the experiments conducted at initial oxygen content The USBM determined LOC for Pittsburgh seam high volatile
lower than 12%, measurements of residual oxygen in the after- bituminous coal (moisture 1%, volatility 37%, 80% minus 200 mesh
gases indicate that complete combustion is never achieved, (<75 mm) and a mass median particle diameter of 48 mm) from the
although considerable overpressures are still measured. It is experiments conducted in their 20-L vessel using 2.5 kJ ignitors as
remarked that these data do not support the idea of the existence 11  0.5% O2 (Cashdollar, 1996). The turbulence level in these tests
of a limit value of oxygen concentration. However, for safe side, is lower than that in Siwek 20-L Sphere as there is a total ignition
the LOC may be taken as 7% for pulverized coal A and 8% for delay of w0.4 s from the start of dispersion until ignition for the
coal B. standard test procedure in this study in comparison to ignition
The experimental data obtained by present study and various delay of w0.06 s and a reservoir pressure of 20 bar in standard
researchers in a 20-litre sphere similar to that of Siwek are procedure for the Siwek 20-L chamber, resulting in higher level of
summarized in Table 2. The data from present study can be turbulence. The LOC data reported by this work were not affected
compared with those for a South-African bituminous coal which by turbulence.
Further experiments were conducted for determining LOC for
other coal samples prepared in present study. The experimental
Table 2 results showing the influence of particle size and moisture
Summary of LOC of coal results with 20-L Sphere e literature data/CSIR-CBRI study. content on LOC of coals are shown in Table 3. It has been
S. No. Coal classification Volatile Moisture Particle LOC observed that dust particle size has a comparatively small in-
matter (%) (%) size, Dm or (%) fluence on the limiting oxygen concentration e.g. LOC was found
D50 (mm) to vary from 7% to 6% when particle size was reduced from
1 Morewell brown coal 49.5 14.1 22b 13 74 mm to 38 mm for coal A with moisture content w4%. Similar
(Woskoboenko, 1988) observation was made in a study for brown coal inerting. At an
2 German lignite-TNO 46.10 9.85 113a 9 initial temperature of 50  C and nitrogen as inert gas, the values
(Wilen et al., 1999,
were 11.8% for a median particle size of 19 mm, and 12.4% for
117 pp. þ app.8 pp.)
3 Spanish lignite-TNO 36.60 16.65 66a 8 52 mm (Eckhoff, 1991). LOC data for two sizes of Pittsburgh coal
(Wilen et al., 1999, are 11% for median particle size of 48 mm and 10.5% for 11 mm
117 pp. þ app.8 pp.) (Cashdollar, 1996). The moisture has a significant influence on
4 German lignite-LOM 46.10 9.85 70a 10 LOC value. The LOC of coal increases from 7 to 11% on increasing
(Wilen et al., 1999,
117 pp. þ app.8 pp.)
the moisture from w4 to 15% for coal A (particle size e 74 mm).
5 Bituminous coal 29 16.4b 14 Similar observations were found for other particle sizes and for
(Eckhoff, 1991) 17 coal B.
6 Brown coal (Eckhoff, 43 17.5b 12
1991) 42,63,66
7 S. Africa coal (Continillo 23.13 0 <<10 Table 3
et al., 1991a, 1991b) Effect of particle size and moisture content on LOC of coal e CSIR-CBRI study.
b
8 Pittsburgh pulverized 37% 1% 48 11
coal (Cashdollar, 1996) S. No. Coal A Coal B
9 Coal A, Jharia coalfield, 27.18 3.92 90% passing 7
Particle Moisture, LOC, Particle Moisture, % LOC, % O2
India e CSIR-CBRI study through 200 BS
size, mm % % O2 size, mm
mesh, nothing
remaining on 1 D 3.92 7 D 3.80 8
52 BS Mesh 2 D 9.3 9 D 9.1 10
10 Coal B, Jharia coalfield, 19.69 3.80 90% passing 8 3 D 15.2 11 D 15.3 11
India, e CSIR-CBRI study through 200 BS 4 74 3.89 7 74 3.9 8
mesh, nothing 5 74 9.1 9 74 9.3 10
remaining on 6 74 15.1 11 74 15.2 11
52 BS Mesh 7 38 4.0 6 38 3.80 7
a 8 38 9.2 8 38 9.0 9
D50 e mean particle size e mesh of the sieve dividing the fuel in two equal parts.
b 9 38 15.34 9 38 15.1 10
Dm e mass median diameter, TNO e TNO Prins Maurits Laboratories (TNO), The
Netherlands, LOM e Laboratorio Oficial J.M. Madariaga (LOM), Spain. D e 90% passing through 200 BS Mesh, nothing remaining on 52 BS Mesh.
1112 M. Mittal / Journal of Loss Prevention in the Process Industries 26 (2013) 1106e1112

4. Conclusion Pex [bar] Maximum explosion pressure for the dust concentration
tested
The Limiting Oxygen Concentration data and explosion severity Pmax [bar] Maximum explosion pressure
data at reduced oxygen levels for coals determined with CSIR-CBRI t [s] Time
20-L Sphere reported in this paper may be used for hazard analysis V [m3] Vessel volume
and practical design of inerting systems/other explosion safety
measures aimed at preventing dust explosions or suppressing the
disastrous effects of explosion. The data on explosion severity References
together with limiting oxygen concentration, minimum explosible
ATEX 95, EU Directive 94/9/EC. (2009, September). Equipment and protective systems
concentration, minimum ignition temperature, and minimum intended for use in the potentially explosive atmosphere (ATEX). Version 3.
ignition energy, appear to be sufficient to characterize the protec- BS EN 14034-4. (2005). Determination of explosion characteristics of dust clouds e
tive measures. The LOC measurement is not normally used directly Part 4: Determination of the limiting oxygen concentration LOC of dust clouds.
British Standards Institution.
to provide inerting levels as a reasonable safety factor should be Cashdollar, K. L. (1996). Coal dust explosibility. Journal of Loss Prevention in the
applied to account for the sensitivity, accuracy and reliability of the Process Industries, 9, 65e76.
plant monitoring system to establish safe inerting levels in indus- Cashdollar, K. L. (2000). Overview of dust explosibility characteristics. Journal of Loss
Prevention in the Process Industries, 13, 183e199.
trial processes and to set oxygen concentration alarms or interlocks
Continillo, G., Crescitelli, S., Fumo, E., Napolitano, F., & Russo, G. (1991a). Experi-
in internal vessels. NFPA 69 (NFPA, 2008) recommends keeping the mental results of coal dust explosions in a spherical bomb. In Proceedings of the
system oxygen concentration at least 2% lower than measured LOC 4th international colloquium on dust explosions (November 4e9, 1990, Porabaka-
value when protecting equipment. Gases other than nitrogen can Kozubnik, Poland) (pp. 195e207).
Continillo, G., Crescitelli, S., Fumo, E., Napolitano, F., & Russo, G. (1991b). Coal dust
also be used to reduce oxygen concentration. The LOC values re- explosions in a spherical bomb. Journal of Loss Prevention in the Process In-
ported above are only for the specific coal dusts tested and may not dustries, 4(4), 223e229.
be applicable to other particle sizes of the same coals. It is also Eckhoff, R. K. (1991). Dust explosions in the process industries. Oxford: Butterworth
Heinemann.
important to recognize that these LOC data are for nitrogen inerting Going, J. E., Chatrathi, K., & Cashdollar, K. L. (2000). Flammability limit measure-
of air. Inerting with other gases such as carbon dioxide may give ments for dusts in 20-L and 1m3 vessels. Journal of Loss Prevention in the Process
different results as LOC of a given dust cloud is dependent on the Industries, 13(3e5), 209e219.
IS. (1984). Methods of test for coal and coke e Part 1: Proximate analysis, IS 1350,
type of inert gas that is used to replace the oxidant of the atmo- second revision. New Delhi: Bureau of Indian Standards.
sphere as well as some process conditions such as temperature. IS. (1970). Methods of test for coal and coke e Part 2: Determination of calorific value,
Therefore, LOC testing should simulate the process conditions and IS 1350, first revision. New Delhi: Bureau of Indian Standards.
ISO. (1985). Explosion protection systems e Part 1: Determination of explosion indices
be performed by using inert gas that is representative of inert gas of combustible dusts in air, ISO 6184/1. Switzerland: International Organization
used in practice. for Standardization.
NFPA. (2008). Standard on explosion prevention systems, NFPA 69. Quincy, MA: Na-
tional Fire Protection Association.
Symbols used
Siwek, R. (1985). Development of a 20 litre laboratory apparatus and its application for
the investigation of combustible dusts. Basel, Switzerland: Ciba Geigy AG.
Wilen, C., Moilanen, A., Rautalin, A., Torrent, J., Conde, E., Lodel, R., et al. (1999). Safe
Cex [g/m3] Optimum dust concentration handling of renewable fuels and fuel mixtures. Technical Research Centre of
Finland: VTT Publications 394. Espoo 1999.
KSt [bar m/s] Explosibility index Woskoboenko, F. (1988). Explosibility of Victorian Brown coal dust. Fuel, 67(8),
P [bar] Pressure 1062e1068.