ARTICLE IN PRESS

Journal of Loss Prevention in the Process Industries 20 (2007) 303–309

www.elsevier.com/locate/jlp

Thermo-kinetic modelling of dust explosions

A. Di Benedettoa,, P. Russob
a
Istituto Ricerche Combustione, IRC-CNR, via Diocleziano, 328, 80125 Napoli, Italy
b
Dipartimento di Ingegneria Chimica e Alimentare, Università di Salerno, via Ponte don Melillo, 84084 Fisciano (SA), Italy
Received 29 November 2006; received in revised form 4 April 2007; accepted 4 April 2007

Abstract

The guidelines for protection and mitigation against hazard coming from dust explosion require the knowledge and then the
evaluation either experimentally or theoretically of the thermo-kinetic parameters (i.e. KSt, Pmax). We developed a numerical tool for the
evaluation of the thermo-kinetic parameters of dust explosion. This model is based on the simulations of the combustion reaction by
means of a detailed reaction mechanism assuming that the pyrolysis/devolatilization step is very fast and then gas combustion is
controlling dust explosion. The model allows then the determination of the most conservative values of KSt, Sl, Pmax. In the present paper
we calculated the deflagration index and the laminar burning velocity for dusts utilized in various process industries (i.e. cornstarch,
polyethylene, cellulose) as function of dust concentration. The obtained data were successfully compared with the available experimental
results.
r 2007 Elsevier Ltd. All rights reserved.

Keywords: Dust explosion; Modeling; KSt; Laminar burning velocity

1. Introduction gen, Lemkowitz and Scarlett (1996), Nagy and Verakis

(1983), require the knowledge of the laminar burning
The severity of a dust explosion is a function of a wide velocity (Sl).
variety of parameters. Some of these parameters represent The experimental evaluation of these parameters is
properties that are specific to the chemical properties strongly affected by dust dispersion degree, initial turbu-
of the exploding dust air mixture, while others reflect the lence level, particle size distribution (Cashdollar, 1996)
sensitivity of the explosion to the flow properties and to the which have to be carefully controlled in order to avoid
heat transfer of the dust cloud. It is common practise to scatter of the data. To overcome these difficulties
measure the explosion severity, by means of parameters numerical simulations can be an alternative.
such as the deflagration index (KSt) and the maximum Previous models (Continillo, 1989; Eckhoff, 2003)
pressure (Pmax), in laboratory test vessels (1 m3 vessel, 20 l simulate a single-particle explosion as occurring through
sphere) and to predict what would happen if the same different steps: heating, devolatilization and homogeneous
mixture explodes in an industrial plant unit through the combustion; all of them strongly influenced by dust
cubic relationship. This methodology assumes that labora- properties (i.e. particle mean diameter, size distribution).
tory test data can be considered to be applicable to Some of these models also simulated the occurrence of the
accidental explosions in plant units and to represent a heterogeneous combustion which may also occurs in
conservative case even when the actual industrial circum- parallel to the homogeneous combustion. More sophisti-
stances are not reproduced in the laboratory experiments. cated models are based on CFD simulations (van
Furthermore, many models of industrial dust explosion, as Wingerden, Arntzen & Kosinski, 2001). These models are
those of Bradley and Mitcheson (1976), Dahoe, Zevenber- not well established as CFD for gas explosion and rely on
the use of many sub-models and then are still empirical.
Corresponding author. Tel.: +390817682947; fax: +390817622915. It has been showed that each dust has a critical diameter
E-mail address: [email protected] (A. Di Benedetto). under which thermo-kinetic parameters reach a plateau

0950-4230/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jlp.2007.04.001
 ARTICLE IN PRESS
304 A. Di Benedetto, P. Russo / Journal of Loss Prevention in the Process Industries 20 (2007) 303–309

corresponding to the highest value (Eckhoff, 2003; Peukert, plateau value which is characteristic of each dust and that
1996). When the plateau is reached the dust explosion can be reached when the dust with a diameter lower than
does not depend anymore on the dust diameter. On the the critical value explodes.
other side, at very low values of the diameter the dust We performed simulations of the homogeneous reactions
explosion is mainly controlled by homogeneous combus- of the dust volatiles by using a detailed reaction kinetic
tion rather than by heating, heterogeneous combustion and mechanism. The composition of the volatiles gas produced
devolatilization. from pyrolysis/devolatilization of dust was obtained by
Cashdollar, Hertzberg, and Zlochower (1988) assessed experimental values present in the literature.
that in the case of coal and organic materials pyrolysis/
devolatilization always precedes combustion, which pri- 2.1. Pyrolysis/devolatilization data
marily occurs in the homogeneous gas phase. Accordingly,
Yoshizawa and Kubota (1982) related the explosivity of The pyrolysis/devolatilization data used in our simula-
coal to the occurrence of gas-phase reaction and then to its tions are those available in the literature, obtained in
volatiles content. In the present work we developed a different experimental apparatus and conditions. In the
model of dust explosion based on a detailed reaction following the literature experimental procedures used for
mechanism assuming that the pyrolysis/devolatilization determining volatiles gas composition of cornstarch,
step is very fast and gas combustion is controlling dust polyethylene and cellulose dusts are given.
explosion which is true at diameter lower than the critical
value (Eckhoff, 2003). Therefore, the calculated values of
2.1.1. Cornstarch
KSt, Sl, Pmax are the most conservative and may be assumed
Encinar, Beltran, Gonzalez, and Moreno (1997) studied
as reference values. We run simulations for cornstarch,
the pyrolysis of cornstarch in a fixed bed reactor in
cellulose and polyethylene and we successfully compared
nitrogen flow, particle size between 1 and 1.6 mm diameter,
the computed values of KSt and Sl with the experimental
reaction time of 30 min, in the temperature range
values.
400–700 1C. All experiments were carried out at isothermal
conditions possibly in a fast pyrolysis regime. The authors
2. Description of the mathematical model observed that gas distribution varies strongly with tem-
perature: increasing temperature the formation of H2, CO,
Dust explosion occurs via two main paths: the hetero- CH4 increases, while the production of CO2 reaches a
geneous and the homogeneous combustion path (Fig. 1). maximum at an intermediate temperature which suggests
As deeply discussed by Eckhoff (2003) many models the consumption of CO with carbon.
have been proposed in the literature which simulate all or The heating rate and temperature values at which
part of these steps. All these models are based on one-step explosion occurs, are very high and then, for the explosion
reaction rate to describe both the heterogeneous and the simulations we used only the pyrolysis data obtained at the
homogeneous reactions. highest temperatures (600 and 700 1C).
Heterogeneous combustion, heating and devolatilization Mazurkiewicz, Jarosinski, and Wolanski (1993) reported
strongly increase when the ratio between the dust powder results of tests carried out by heating to different
external surface to volume ratio (1/d) increases. As a temperatures (300, 450, 550 1C) cornstarch dust, at
result, on decreasing the diameter, they may become stechiometric concentration (233 g/m3), in a cylindrical
extremely fast with respect to the homogeneous combus- steel container located in an oven. The composition of the
tion. Accordingly, Peukert (1996) reports that the max- product gases was analysed by means of a gas chromato-
imum explosion pressure exhibits a plateau at about 50 mm graph after thermal decomposition of the dust. The
particle size for flour and about 40 mm for methylcellulose cornstarch used consists of particles nearly spherical in
and polyethylene. shape and fairly uniform in size with a mass mean diameter
The model here developed is based on the assumption of 15 mm. The results of measurement of the gas composi-
that the homogeneous reaction kinetic (path 4) is the tion showed that, at a temperature of 300 1C the reactions
controlling step of dust explosion while the other steps are of decomposition of dust produces chiefly CO2 and a small
assumed extremely fast. Hence, the value of parameters quantity of CO. At higher temperatures the relation
obtained from the model correspond to the (maximum) between CO and CO2 becomes inverse, some methane

Heat 3 4
Dust Devolatilization Gas products Explosion
2
1 5
Heat

Heterogeneous 1Combustion

Fig. 1. Schematic representation of the paths occurring during dust explosion.

 ARTICLE IN PRESS
A. Di Benedetto, P. Russo / Journal of Loss Prevention in the Process Industries 20 (2007) 303–309 305

and hydrogen occurring in addition. However, the authors time, which also influences the product yields. Therefore,
highlight that, being the products of thermal decomposi- data at reaction times of 2.2.–3.0 ms and different
tion dependent on the rate of temperature rise, the results temperatures of 1300, 1500 and 1700 1C were used for
of the measurements reported, which were obtained under simulations.
conditions of very slow heating of a sample, cannot be Using the University of Western Ontario’s isothermal
conclusive for processes going on under real conditions, in ultra pyrolysis reactor system, Graham, Bergougnou, and
an atmosphere of chemically active flame. In our calcula- Freel (1994) have obtained total gaseous fast pyrolysis
tions, we tested all the pyrolysis composition obtained by yields and individual gaseous yields for cellulose (100 mm
Encinar et al. (1997) at 600 and 700 1C and those of average particle size) as a function of residence time
Mazurkiewicz et al. (1993) at 550 1C. Further, we run (50–500 ms) and temperature (650–900 1C). The data of
simulations with gas composition evaluated from elemen- highest total gas production were obtained at temperature
tary composition of cornstarch (Proust, 1993). of 900 1C and residence times higher than 300 ms, where
the main product was CO (yield ¼ 70 wt% of cellulose)
2.1.2. Polyethylene followed by CO2, H2, CH4, C2H4, C2H2. These values were
Font, Marcilla, Garcia, Caballero, and Conesa (1995) used for simulations.
studied the decomposition of polyethylene in a fluidized
bed reactor and analysed the obtained gas-by-gas chroma-
tographic analysis. The sample (HDPE, molecular weight 2.2. Model
22,000, particle size 0.5 mm) was discharged in to the hot
bed (800 1C) made of sand and using helium as inert gas The model consists of three modules. The first module is
operating at atmospheric pressure. In this conditions the the model of the flame propagation to evaluate the laminar
authors observed a 95% polyethylene total conversion, burning velocity, Sl. The second allows the evaluation of
mainly to ethylene and methane. the adiabatic pressure (Pmax), the third consists of the
Mastral, Esperanza, Garcia, and Juste (2002) carried out evaluation of the maximum pressure rise (dP/dtmax) and
experiments with polyethylene (mean particle size then the deflagration index (KSt).
0.225 mm) using a stainless steel fluidized bed reactor. The homogeneous combustion of the gas volatiles was
Measurements were made at different temperatures ran- simulated by using a detailed reaction mechanism, the
ging from 650 to 850 1C and at residence time ranging GRI-Mech3.4. This mechanism was implemented in the
between 0.81 and 1.45 s. The gas yields of pyrolysis CHEMKIN module to calculate the laminar burning
products were measured by gas chromatography. velocity as a function of the dust concentration. The
The authors observed that temperature mainly influences deflagration index (KSt) was calculated by using the
the yield in gas respect to oil and wax: when the pyrolysis formula of Lewis von Elbe here reported
temperature increases above 730 1C the gas production  
dP
increases while oil and wax production decreases. At K St ¼ V 1=3 , (1)
780 1C the gas production stabilises. The temperature also dt max
affects the composition of the gas: an increase of H2, CH4, where the maximum pressure rise (dP/dtmax) is calculated
C2H4 and C6H6 yields is observed when the temperature according to the formula recently proposed by Dahoe and
increases. de Goey (2003)
For our simulations, we used the data obtained at the
highest temperature (850 1C). Moreover we also performed   "  1=g #2=3
dP 3ðPmax  P0 Þ P0 Pmax  P
simulations by assuming that all polyethylene gives rise to ¼ 1
ethylene. dt max Rvessel P Pmax  P0
 1=g
P
2.1.3. Cellulose  Sl . ð2Þ
P0
The thermal decomposition of cellulose was studied by
Öztürk and Merklin (1995) in pure argon under conditions Pmax is the maximum pressure reached in a closed vessel
of high temperature (700–2200 1C), very short reaction time which has been calculated by using the equilibrium module
(0.3–3.0 ms), high heating rates and rapid quenching in a of the CHEMKIN code; P0 is the initial pressure (assumed
single-pulse shock tube. Commercial Sigma Cell-Type 20 equal to 1 bar); Rvessel is the radius of the reference
microcrystalline cellulose with an average particle size of spherical vessel (1 m3) and Sl is the laminar burning
20 mm was used. Gas samples from the experimental velocity which is calculated by means of the CHEMKIN
section of the shock tube were collected and analysed on module which simulates the laminar flame propagation.
a gas chromatograph. The results showed that biomass can In this paper, preliminary results of the model, by
be completely converted to gaseous products under simulating the deflagration of pyrolysis gas products
conditions of high temperature, high heating rate, short in a closed vessel without initial turbulence, are reported.
reaction time and rapid cooling. In particular, the onset of The laminar flame is assumed to be the basic element
evolution of pyrolysis products is affected by the reaction of flame propagation and, hence, the laminar burning
 ARTICLE IN PRESS
306 A. Di Benedetto, P. Russo / Journal of Loss Prevention in the Process Industries 20 (2007) 303–309

velocity is used in the evaluation of the maximum pressure than for an adiabatic flame. Besides, the flame will not be
rise (Eq. (2)). uniform over its cross section, and burning velocity and
flame thickness are not always easy to define.
3. Results Moreover, Dahoe, Hanjalic, and Scarlett (2002) found
that the laminar burning velocity varied with flame
Simulations were performed on cornstarch, polyethylene shape, and this was accounted for by introducing the
and cellulose which are classified as St2 classes (GESTIS- ‘‘Markstein length’’ of a dust–air flame. This parameter
DUST-EX; NFPA 68, 2002). In the following the is specific for any given dust cloud. It has a magnitude
parameters KSt, and Sl evaluated by the model are of the order of the laminar flame thickness of that
compared with the experimental values reported in the specific dust cloud and serves as a measure of the sensitivity
NFPA 68 guidelines (2002), in the database GESTIS- of the laminar burning velocity to changes in the flame
DUST-EX and in literature related papers (Bartknecht, shape.
1981; Cashdollar, 1996; Eckhoff, 2003; Krause & Kasch, As a result the experimental values of Sl are significantly
2000; Mazurkiewicz et al., 1993; Nagy & Verakis, 1983; lower than those obtained in an un-stretched flame with
Proust & Veyssiere, 1988; Proust, 1993; van Wingerden & complete combustion.
Stavseng, 1995). Our model is able to predict the laminar burning velocity
in a dust laminar flame. In our model, the flame is un-
3.1. Cornstarch results stretched and complete combustion occurs, hence, the
higher Sl values. Therefore, in Fig. 2, only the simulations
In Fig. 2, calculated values of the laminar burning run with gas composition after thermal decomposition of
velocity as function of the dust concentration are compared starch dust consistent with Sl measurements (Mazurkiewicz
with the experimental data. A significant discrepancy et al., 1993 (T ¼ 550 1C); Encinar et al., 1997 (T ¼ 600 1C))
between model and experimental values appears at high are in agreement.
values of the dust concentration. KSt values both from simulations and literature for
Proust (2006) and Mazurkiewicz et al. (1993) observed cornstarch are shown as function of the dust concentration
that in their experiments incomplete mixing with air and, in Fig. 3. The symbols represent the experimental values of
hence, incomplete combustion of starch dust at flame front KSt obtained by different authors, at different dust
occurs. Indeed, they measured a significant difference concentration and at different diameter values.
(about 400 1C) between the theoretical (adiabatic) and the In the literature it is reported that the plateau KSt
measured values of the maximum flame temperatures. value of the cornstarch dust is KSt ¼ 200 bar m/s, at
Moreover, they found a great sensitivity of the flame to diameter lower than the critical diameter (10 mm),
stretch. where the devolatilization process is not controlling the
As previously highlighted by Lee (1987), a laminar dust explosion behaviour and the homogeneous reactions
flame is difficult to stabilize without causing significant prevail (Eckhoff, 2003). Accordingly, in Fig. 3 the values
cooling of the flame. Therefore, such stabilized flame is of KSt relevant to cornstarch particles with a diameter
non-adiabatic, and average burning velocity will be lower higher than the critical diameter (Eckhoff, 2003; Nagy &

1.5
van Wingerden & Stavseng (1995)_Experimental
Proust & Veyssiere (1988)_Experimental
Mazurkiewicz et al. (1993)_Experimental
Krause & Krasch (2000)_Experimental
Nagy & Verakis (1983)_Experimental
Dahoe etal. (2002)_Experimental
Encinar et al. (1997) @700°C_Model
1.0 Encinar et al. (1997) @600°C_Model
Mazurkiewicz et al. (1993) @550°C_Model
Sl ( m/s)

Elemental composition_Model

0.5

0.0
0 200 400 600 800 1000
C (g/m3)

Fig. 2. Experimental and model values of the laminar burning velocity as function of the dust concentration—cornstarch.
 ARTICLE IN PRESS
A. Di Benedetto, P. Russo / Journal of Loss Prevention in the Process Industries 20 (2007) 303–309 307

240
Experimental

NFPA (2002)_d=7 m
Encinar et al. (1997) @700°C_Model
200 database_d<10 m Encinar et al. (1997) @600°C_Model
Mazurkiewicz et al. (1993) @550°C_Model
160 Elemental composition_Model
KSt (bar m/s)

120

80 Nagy&Verakis (1983)
Proust (1993)_d=28 m
Eckhoff (2003)_d=37 m

40 Eckhoff (2003)_d=37 m
Eckhoff (2003)_d=37 m

0
100 600 1100 1600 2100
C (g/m3)

Fig. 3. Experimental and model values of the deflagration index as function of the dust concentration—cornstarch.

250
NFPA (2002) & database_Experimental
Bartknecht (1981)_Experimental
200 Mastral et al. (2002)_Model
Font et al.(1995)_Model
All ethylene_Model
KSt (bar m/s)

150

100

50

0
10 100 1000
C (g/m3)

Fig. 4. Experimental and model values of the deflagration index as function of the dust concentration—polyethylene.

Verakis, 1983; Proust, 1993) are lower than the plateau Mazurkiewicz et al. (1993) data are obtained at slow
value reported in the NFPA 68 (2002) and in the database heating rate compared to those achieved during dust
GESTIS-DUST-EX (diameter lower than 10 mm). explosion.
It appears that the simulation results obtained with the It is worth noting that model is able to give a good
product distribution proposed by Encinar et al. (1997) at prediction of experimental deflagration index also when
the highest temperature (T ¼ 700 1C) is able to capture the gas products composition is determined from elementary
maximum value of KSt reported in the database (GESTIS- composition of cornstarch (Proust, 1993).
DUST-EX), in the NFPA 68 guidelines (2002) and by
Roser, Vogl, Radandt, Malalasekera, and Parkin (1999). 3.2. Polyethylene results
Worst results are obtained using the gas products of
Mazurkiewicz et al. (1993) (T ¼ 550 1C) while those In Fig. 4, the computed values of KSt as function of the
relevant to Encinar et al. (1997) at the lower temperature polyethylene concentration obtained with different pyroly-
(T ¼ 600 1C) approaches the experimental KSt relevant to sis gas compositions are compared with the experimental
particle diameter higher than the critical diameter. This values (Bartknecht, 1981; NFPA 68, 2002). The experi-
behaviour can be explained considering that in both cases mental results of Bartknecht (1981) of the database
the experimental temperature is inadequate to complete (GESTIS-DUST-EX) and of NFPA 68 (2002) are the
cornstarch devolatilization; moreover, in the case of maximum values obtained as function of the concentration.
 ARTICLE IN PRESS
308 A. Di Benedetto, P. Russo / Journal of Loss Prevention in the Process Industries 20 (2007) 303–309

Öztürk & Merklin (1995) @1300°C_Model

280
Öztürk & Merklin (1995) @1500°C_Model
Öztürk & Merklin (1995) @1700°C_Model
240
Graham et al.(1994)_Model

KSt (bar m/s) 200

160

120

80

40

0
100 200 300 400 500 600
C (g/m3)

Fig. 5. Experimental and model values of the deflagration index as function of the dust concentration—cellulose.

We run the simulations by using the pyrolysis gas turbulence could be relevant and has to be included in the
distribution obtained by Font et al. (1995) and by Mastral model (Eckhoff, 2003).
et al. (2002). The obtained results (Fig. 4) are quite The difficulty in the experimental evaluation of the
satisfactory. thermo-kinetic parameters useful for sizing protection and
Interestingly, if we assume that all the ethylene contained mitigation systems against hazard coming from dust
in the polymer burns as it is, we obtain the best agreement. explosion can be overcome if numerical simulations are
parallely used.
3.3. Cellulose results The here developed model allows the calculation of the
deflagration index (KSt) and the laminar burning velocity
In Fig. 5, the model values of the deflagration index are (Sl) of a dust, once the composition of pyrolysis gas
plotted vs. the cellulose concentration. The available products is known.
experimental values given in the NFPA 68 (2002) and in The pyrolysis composition to be used in the model has to
the database (GESTIS-DUST-EX) are KSt ¼ 229 and be obtained at the highest temperature and heating rate.
KSt ¼ 126 bar m/s, respectively. Future work will be devoted to include the effect of the
In our simulations the maximum values of KSt are initial turbulence necessary for dust dispersion in the model
reached at concentration values ranging from 300 to to study the sensitivity of the deflagration index to the
500 g/m3 and they assume values from 200 to 245 bar m/s. turbulence intensity.
The comparison with the available experimental results is
then quite good. References

4. Discussion and conclusions Bartknecht, W. (1981). Explosions. Berlin, Heidelberg, New York:
Springer.
From comparison of simulation results and experimental Bradley, D., & Mitcheson, A. (1976). Mathematical solutions for
explosions in spherical vessels. Combustion and Flame, 26, 201–217.
values it turns out that the model is able to predict the
Cashdollar, K. L. (1996). Coal dust explosibility. Journal of Loss
experimental values of KSt. Prevention in the Process Industries, 9(1), 65–76.
However, for cornstarch dust at concentration higher Cashdollar, K. L., Hertzberg, M., & Zlochower, I. A. (1988). Effect of
than that stoichiometric, the experimental values of KSt volatility on dust flammability limits for coals, gilsonite and
result higher than those predicted by the model (Fig. 3). polyethylene. In Proceeding of the 22nd symposium (international) on
This behaviour may be ascribed to the difficulties combustion (pp. 1757–1765). Pittsburgh: The Combustion Institute.
Continillo, G. (1989). A two-zone model and a distributed-parameters model
encountered in the experiments to ensure the same of dust explosions in closed vessels. Archivum Combustionis, 9, 79–94.
turbulence level at all the concentration values. The desired Dahoe, A. E., & de Goey, L. P. H. (2003). On the determination of the
dust concentration is obtained experimentally by adjusting laminar burning velocity from closed vessel gas explosions. Journal of
the pre-injection pressure in the dust storage canister and Loss Prevention in the Process Industries, 16(6), 457–478.
the explosion chamber. Therefore, the initial level of Dahoe, A. E., Hanjalic, K., & Scarlett, B. (2002). Determination of the
laminar burning velocity and the markstein length of powder air
turbulence may be higher for rich than lean mixtures flames. Powder Technology, 122, 222–238.
determining an increased burning velocity and, hence, an Dahoe, A. E., Zevenbergen, J. F., Lemkowitz, S. M., & Scarlett, B. (1996).
increased deflagration index. In this case the effect of Dust explosions in spherical vessels: The role of flame thickness in the
 ARTICLE IN PRESS
A. Di Benedetto, P. Russo / Journal of Loss Prevention in the Process Industries 20 (2007) 303–309 309

validity of the ‘cube-root law’. Journal of Loss Prevention in the Process NFPA 68. (2002). Guide for venting of deflagration. Quincy: National Fire
Industries, 9(1), 33–44. Protection Association.
Eckhoff, R. K. (2003). Dust explosions in the process industries (3rd ed.). Öztürk, Z., & Merklin, J. F. (1995). Rapid pyrolysis of cellulose with reactive
Gulf Professional Publishing. hydrogen gas in a single-pulse shock tube. Fuel, 70(11), 1658–1663.
Encinar, J. M., Beltran, F. J., Gonzalez, J. F., & Moreno, M. J. (1997). Peukert, W. (1996). Trends in the development of solids process
Pyrolysis of maize, sunflower, grape and tobacco residues. Journal of engineering. Chemie -Ingenieur-Technik, 68(10), 1254–1263.
Chemical Technology & Biotechnology, 70, 400–410. Proust, C. (1993). Experimental determination of the maximum flame
Font, R., Marcilla, A., Garcia, A. N., Caballero, J. A., & Conesa, J. A. temperatures and the laminar burning velocities for some combustible
(1995). Comparison between the pyrolysis products obtained from dust–air. Archivum Combustionis, 13(3-4), 175–188.
different organic wastes at high temperatures. Journal of Analytical and Proust, C. (2006). Flame propagation and combustion in some dust–air
Applied Pyrolysis, 32, 41–49. mixtures. Journal of Loss Prevention in the Process Industries, 19(1),
GESTIS-DUST-EX /http://biaonline.hvbg.de/EXPLOE.htmS. 89–100.
Graham, R. G., Bergougnou, M. A., & Freel, B. A. (1994). The kinetics of Proust, C., & Veyssiere, B. (1988). Fundamental properties of flames
vapor-phase cellulose fast pyrolysis reactions. Biomass and Bioenergy, propagating in starch dust–air mixtures. Combustion Science and
7, 33–47. Technology, 62, 149–172.
Krause, U., & Kasch, T. (2000). The influence of flow and turbulence on Roser, M., Vogl, A., Radandt, S., Malalasekera, W., & Parkin, R. (1999).
flame propagation through dust–air mixtures. Journal of Loss Investigations of flame front propagation between interconnected
Prevention in the Process Industries, 13(3–5), 291–298. process vessels. Development of a new flame front propagation time
Lee, J. H. S. (1987). Dust explosions: an overview. In: Proceedings of the prediction model. Journal of Loss Prevention in the Process Industries,
international symposium on shock tubes and waves (pp. 21–38). Aachen, 12(5), 421–436.
Germany. van Wingerden, K., & Stavseng, P. L. (1995). Measurement of fundamental
Mastral, F. J., Esperanza, F., Garcia, P., & Juste, M. (2002). Pyrolysis of burning velocity of dust–air mixtures in industrial situations. Chr.
high-density polyethylene in a fluidised bed reactor. Influence of the Michelsens institutt (CMI), Report nr. CMR-95-F20018.
temperature and residence time. Journal of Analytical and Applied van Wingerden, K., Arntzen, B. J., & Kosinski, P. (2001). Modelling of
Pyrolysis, 63, 1–15. dust explosions (vol. 1601, pp. 411–421). VDI-Berichte, /www.gexcon.
Mazurkiewicz, J., Jarosinski, J., & Wolanski, P. (1993). Investigations of com/download/DustCFD1.pdfS.
burning properties of cornstarch dust–air flame. Archivum Combus- Yoshizawa, Y., & Kubota, H. (1982). Experimental study on gas-phase
tionis, 13(3–4), 189–201. ignition of cellulose under thermal radiation. In Proceeding of the 19th
Nagy, J., & Verakis, H. C. (1983). Development and control of dust symposium (international) on combustion (pp. 787–795). Pittsburgh:
explosions. New York: Marcel Dekker. The Combustion Institute.