Analysis of ignition risk on mechanical equipment in

ATEX
Claire Petitfrere, Christophe Proust

To cite this version:

Claire Petitfrere, Christophe Proust. Analysis of ignition risk on mechanical equipment in ATEX. 4.
Petroleum and chemical industry Conference Europe ”Electrical and instrumentation applications”,
Jun 2007, Paris, France. pp.NC. �ineris-00973260�

HAL Id: ineris-00973260

https://hal-ineris.archives-ouvertes.fr/ineris-00973260
Submitted on 4 Apr 2014

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 ANALYSIS OF IGNITION RISK ON MECHANICAL EQUIPMENT IN ATEX

Claire Petitfrère Christophe Proust

INERIS INERIS
Parc Alata BP 2 Parc Alata BP 2
F-60550 Verneuil en Halatte F-60550 Verneuil en Halatte
France France

Abstract - Until the implementation of the ATEX directive In the first part of this paper, the actual requirements for
94/9/CE, the certification of equipment intended to explosive mechanical equipment are given. The main results concerning
atmospheres was only dedicated to electrical equipment. the ignition by friction and impact are presented in the second
Since the 1st July 2003, the non-electrical sources of part and some practical implications in the last section.
inflammation are also to be looked at before putting an ATEX
equipment on the market (pumps, couplers, reducing gears, II. REQUIREMENTS AND HAZARD ASSESSMENT
…).
Among these sources, mechanical friction and impacts are a A. Requirements of Directives and Standards
main cause of ignition of explosive atmosphere. The risk
analysis of the equipment consists in considering the failures, Non electrical equipment are dealt with in directive 94/9/CE
which can lead to ignition. If the failure involves a friction when placed on the market for use in hazardous areas
between two parts or an impact, we shall estimate if this (concerns firstly the supplier) and by directive 1999/92/CE
friction or impact dissipates a sufficient amount of energy to [3] when intended to be installed in hazardous area
ignite the surrounding explosive atmosphere. (concerns firstly the user).
As part of the European program MECHEX, we have The essential health and safety requirements of the directive
studied the process of degradation of the mechanical energy 94/9/CE demand among others that the manufacturer take
into heat during friction and impacts and we have examined measures to prevent the ignition of explosive atmospheres,
the mechanisms of ignition at the contact zone. taking into account the nature of every electrical and non-
An extensive experimental program is presented and some electrical sources of ignition. Some supplementary
“simple” modelling is proposed on purpose of practical requirements are specific of the category of equipment :
applications. For frictional situations, a critical rubbing power is • For category 1 (equipment intended to zone 0) :
calculated without any limitations as for a potential lower Equipment must be so designed and constructed that
boundary concerning the rubbing velocity. For “impacts”, the sources of ignition do not so become active, even in the
relevant parameter for ignition is not the kinetic energy of the event of rare malfunctions.
projectile but its velocity and the nature of the materials. • For category 2 (equipment intended to zone 1) :
Equipment must be so designed and constructed as to
Index Terms — ATEX, Mechanical ignition source, Non- prevent ignition sources arising, even in the event of
electrical Equipment frequently occurring disturbances or equipment
operating faults, which normally have to be taken into
I. INTRODUCTION account.
• For category 3 (equipment intended to zone 2) :
Since the implementation of the ATEX directive 94/9/CE [1] Equipment must be so designed and constructed as to
in July 2003, the non-electrical sources of inflammation shall prevent foreseeable ignition sources which can occur
also be examined before putting an ATEX equipment on the during normal operation.
market (pumps, couplers, reducing gears, …).
European harmonized standards have been developed by
CEN/TC 305 WG2 to give requirements for non-electrical The directive 1999/92/CE demands that :
equipment. The basic standard for non-electrical equipment is • The employer shall ensure that a document, called
the EN 13463-1 [2], which requires an ignition hazard ‘explosion protection document’, is drawn up and kept
assessment of the equipment. up to date. This document shall contain the
Concerning mechanical equipment, this assessment often demonstration that work equipment is designed,
underlines the possibility of mechanical friction or impacts as operated and maintained with due regard for safety.
an ignition source during foreseeable or rare malfunction. • Work equipment for use in hazardous area which was
When such a source appears, the manufacturer shall estimate made available in the undertaking or establishment for
if the friction or impact dissipates a sufficient amount of energy the first time after 30 June 2003 shall comply with the
to ignite the surrounding explosive atmosphere. directive 94/9/CE (with the category adapted to the
The European program MECHEX was dedicated to the zone).
study of the process of ignition during friction and impact. • Work equipment for use in hazardous area which was
already in use or was made available in the undertaking

147
 or establishment for the first time before 30 June 2003 known that the amount of heat evolved is in proportion of the
shall comply from that date with the minimum applied force and sliding speed [6].
requirements laid down in Annex II, Part A. So all
necessary measures must be taken to ensure that the Further, if the friction is sufficiently intense, some material
work equipment have been designed, constructed, may be torn away resulting in fragments being expelled. This
assembled and installed, and are maintained and event is sometimes called “friction sparks” or “spark shower”.
operated, in such a way as to minimize the risks of an
explosion. Experimental equipment

For a given heat power produced in the rubbing zone [7],

To comply with the requirements of the Directive 94/9/CE we traditional thermodynamical laws tell that the maximum
may generally use the harmonized standard EN13463-1. We temperature may depend on some geometrical sizes (area of
can also use the frame of this standard to carry the hazard the contact zone, diameter of the machine,…) of the rubbing
assessment of the equipment that is already installed and to equipment. Mainly for this reason, two experimental set-ups
include this assessment to the explosion protection document have been used (figure 1). The configuration of a fixed slider
demanded by the directive 1999/92/CE. rubbing onto a rotating wheel has been selected because a
The standard EN 13463-1 is based on the assessment of rather precise control of the rubbing parameter is amenable.
the ignition hazards of equipment depending on the category The larger device [8] is a 30 cm diameter wheel, 22 mm
of this equipment. It can be used as a stand-alone standard thickness rubbing against a 40 mm high, 25 x 25 mm square
but for mechanical equipment the standard EN 13463-5 [4] is cross section slider for most tests. The sliding velocity can be
generally used in addition. This standard defines the protection adjusted between 0.2 and 20 m/s and the normal load
concept “constructional safety” with marking “c”. The principle between 0 and 5000 N.
of this type of protection is to apply constructional measures so The smaller device is a 10 cm wheel, 8 mm thick rubbing
as to protect against the possibility of ignition from hot against a 15 mm high, 7 x 7 mm square cross section slider.
surfaces, sparks and adiabatic compression generated by The sliding velocity can be adjusted between 0.05 and
moving parts. The requirements are based on good 15.7 m/s and the normal load between 0.2 and 1000 N.
engineering principles, so that risk of mechanical failures likely Temperatures have been measured by optical methods and
to create incendive temperatures or sparks, are reduced to a thermocouples. During the tests, the normal load, friction
very low level. coefficient and speed were controlled [5, 8, 10].

B. Ignition Hazard Assessment

The ignition hazard assessment consists in reviewing all

potential ignition sources of the equipment and in
distinguishing normal operation, expected malfunction and rare
malfunction. Then will be considered the measures applied to
prevent each source becoming effective.
Concerning mechanical equipment with moving parts the
hazard assessment will often underline the risk of friction or Figure 1 : rubbing experimental set-ups (Large scale 30 cm
impact as an expected or rare malfunction. That could be the dia. wheel- left and small scale 10 cm dia. wheel –right)
case for instance for pumps when there is no more fluid and
where parts could enter in contact in rare malfunction. We can
The evolution of the rubbing temperature as function of the
also think to the bearings or to the friction of a roller on a rail.
nominal power expressed in terms of normal load (N) times the
When the malfunction introducing a risk or friction or impact
sliding velocity (V) is shown on figure 2 for the various sliders
is identified, the problem is then to determine if this friction or
used and the two machines. There is a clear trend of
impact could become an effective ignition source or not. The increasing the rubbing temperature as function of the nominal
type of source could then be a hot surface or sparks but it is power. There is also a clear incidence of the nature and of the
difficult to know if the phenomena are sufficient to ignite the width of the slider and globally of the size of the rubbing
explosive atmosphere. machine.

III. MECHANICAL FRICTION AND IMPACT

A. IGNITION RISK BY FRICTION

Friction between two solid bodies is a process through which

mechanical energy is transformed into heat. Experience
reveals that heat is produced in the rubbing zone where the
material is severely stretched and diffuses outwards. It is well

148
 1000 local temperature is large enough according to the
900 small scale machine mechanism of the hot plate [9]. The experimental
investigation and some theoretical evidence shows that
Rubbing temperature -°C

800
the critical parameter is the temperature of the hot zone,
700
of the order of hundreds of °C;
600
• The power dissipated in friction may easily be of the order
500 hundreds of Watts. If the friction is occurring in a
400 sufficiently confined area, the mean temperature of the
300 mild steel
ATEX may rise and reach the autoignition point. This
200
quartz situation should well be represented by the standard
pure aluminium
ignition temperatures ;
100
0,00 500,00 1000,00 1500,00 2000,00 2500,00 3000,00 3500,00 4000,00 • The sparks may be a cause of ignition. However, since
N . V (W) the initial temperature at the beginning of the flight should
1300
Hard steel - hard steel : slider 25 mm, wheel 22 mm not be different from that of the rubbing zone, the sparks
Hard steel - hard steel : slider 25 mm, wheel 10 to 16 mm
Hard steel - hard steel : slider 10 mm, wheel 22 mm
are likely to constitute a preferential cause of ignition only
1100 Hard steel - hard
Logarithmique steel rubbing simplified
(Série4) if, later in their flight, their temperature increases well
above that of the rubbing zone. This may occur if the
Rubbing temperature -°C-

900
flying particles are able to burn. In this case, the ignition
Large scale machine
Al - bronze
process may be linked to the spark ignition mechanism
700
referring to minimum ignition energies.
500 Mild steel

Ignition around the heated surface

300 aluminium

This ignition mechanism is believed to be the main cause for

100
100 1000 10000 100000 mechanical ignition [8, 12] often invoked but, to our knowledge,
N . V (W) never really proven.
1300
For friction, experiments performed with the large scale
rubbing machine used in explosive atmospheres and filmed
1100
with high speed video [8] demonstrate clearly that ignition
small scale machine proceeds at the rubbing zone. It occurs as soon as some
Rubbing temperature -°C-

900 critical temperature threshold has been reached at the rubbing

zone. These temperatures have been estimated for a number
700
Large scale machine
of atmospheres (hydrogen-air, methane-air, propane-air,
ethylene-air, diethyl-ether-air, acetone-air,…) and compared to
500 critical temperatures determined previously [9] for hot surface
ignition in explosive atmospheres (Tpcrit).It has been realized
300
that both sets of values correlate extremely well demonstrating
that “frictional” ignition is a “hot surface” ignition mechanism. It
100
0 1000 2000 3000 4000 5000
has been however noted that the critical hot surface
N . V (W) temperature does not seem to correlate so well with the
standard autoignition temperature nor better with the minimum
ignition energy [5], indicating that this might be a new type of
Figure 2 : Rubbing temperature as function of N.V for
ignition criterion.
various sliders and the two machines (on the top, only with the
small scale equipment, in the middle, only with the large scale
1000
equipment and on the bottom graph a comparison of the
results with the two machines with a hard steel – mild steel 950

rubbing configuration) 900

850
In some tests, flying fragments (sparks material) have been
Tcrit friction (°C)

800
collected in order to evaluate the size of them and the wear 750
rate. For steel rubbing against steel at a few m/s, the typical
700
particle size seems of the order of a few hundred of mm and
the wear rate, extremely dependent on the experimental 650

conditions, is better scaled in mm3/s [5]. 600

550
Mechanisms of ignition 500
500 550 600 650 700 750 800 850 900 950 1000
Tcrit hot spot (°C)

The analysis reveals that ignition during friction and impacts Figure 3 : tabulated critical hot surface ignition temperature
may result from three different processes [7] : and observed contact temperature upon ignition during friction
• In direct contact with the hot rubbing zone whenever the

149
 Volumetric ignition • V : rubbing velocity
• q : heat dissipated in friction
Experiments on the small scale rubbing machine with
steel_steel configurationhave shown only a moderate increase For the practical use of this expression within the frame of
of the average temperature throughout a volume in closing the explosion safety, the following aspects need to be considered:
wheel (40°C) and inside the spark shower as compared to the • a proper choice of the critical ignition
temperature reached at the rubbing zone (700°C). So the temperature must be made. For explosive
ignition should occur most probably at at the hot contact zone atmospheres, the relevant parameter would be
before any possibility of ignition throughout volume could arise Tpcrit which is above but not directly related to the
[5]. standard ignition temperature ;
• the area of the contact zone depends on the
Sparks geometry of the device and may be found with a
detailed analysis of the system. It should not be
There was no evidence of any ignition to be undoubtedly smaller than the ratio of the normal load to the
attributable to a flying spark in the present experiments [5]. Brinell Hardness which gives a lower boundary
There was even no evidence of any fragment to burn in the for R ;
surrounding atmosphere even in the extreme situation of • the critical power dissipated in rubbing (q=f.N.V)
aluminum rubbing intensively against hard steel. Is it to be said is the part lost in the friction which, in normal
that only the situation of a light alloy rubbing against some operation, is usually a very low fraction of the
rusty material and leading to “thermite reactions” [12] is to be nominal power, typically a few percent for friction
considered for “spark” ignition during friction ? coefficients of the order of 0.001 to 0.01. Only in
case of severe malfunction (a shaft broken, a
To help and clarify this important point, explosive fused bearing…), the friction power may reach
atmospheres were submitted to a small chemically inert and the nominal power with a fraction coefficient
artificially heated object, simulating a fragment [5]. Globally, between 0.1 and 1..
the required temperature to produce ignition is well above the
critical “hot surface” temperatures apart from the specific case Some comparison between this simplified method and
of sulphur, which need to be further analysed. Thus, it is experimental data is shown on figure 4. Here, the rubbing
confirmed that the “spark” temperature need to be well above situation typically corresponds to a severe malfunction as
those of the rubbing zone to generate an increased risk of defined above with a friction coefficient equated to unity. The
ignition suggesting that the fragments have to burn somehow. model gives a lower estimate of the critical power but is
reasonable in trends and order of magnitude.

10000
Modelling experiments
Simplified equation f=1
These findings have been analyzed and collapsed into a
reasonable model [4]. It is postulated that ignition would occur 1000
Puissance critique (W)

at the hottest point, in the immediate vicinity of the rubbing

zone (we thus leave apart the special case of « thermite » 100
reaction). Following, the ignition parameter is the critical hot
surface temperature defined before, depending on the nature
of the atmosphere. We then have to derive the contact 10

temperature from the frictional process and compare it to the

former. The standard Coulomb-Joule relationship has been 1
30 cm 30 cm 30 cm 30 cm 30 cm 30 cm 30 cm steel 30 cm steel 30 cm steel 30 cm steel 30 cm steel
used to calculate how much heat is produced in the friction,. sandstone
wheel - 2
sandstone sandstone sandstone stainless
wheel - 2 wheel - 2
stainless wheel on
wheel - 2 steel wheel steel wheel 625 mm2
wheel on
625 mm2
wheel on
625 mm2
wheel on
625 mm2
wheel on
625 mm2

This amount of heat is driven out from the contact zone by cm

limestone
cm cm cm - 2,5 cm - 2,5 cm steel slider
limestone limestone limestone hard steel hard steel at 1 m/s in
steel slider
at 0.7 m/s
steel slider
at 0.7 m/s
steel slider
at 5 m/s in
steel slider
at 5 m/s in
slider - slider - slider - slider - slider - slider - propane in ethylene in hydrogen acetone diethylether
thermal conduction and is modeled as a local heat source onto V=1,6 m/s V=5,3 m/s V=10,6 m/s V=16 m/s V=4,6 m/s V=9,2 m/s
From literature Present work
an semi-infinite solid [10]. The final equation reads:

4 A
q = f ⋅ N ⋅V = ⋅ (λ1 + λ2 ) ⋅ ⋅ (T f − Tamb ) Figure 4 : performances of the simplified model for frictional
π R ignition against existing data [7] and results issued from the
present program
where
• Tf : temperature of the rubbing zone This physical analysis implies that the maximum
• Tamb: ambient temperature (boundary temperature in the rubbing zone depends only on the friction
temperature of the bodies) power without any limitations about any lower boundary about
the rubbing velocity. For instance, it has been possible to ignite
• R : radius of contact area
propane-air, ethylene-air and hydrogen-air mixtures at rubbing
• A : contact area
velocities equal to or even smaller than 1 m/s, which has been
• λi : heat conductivity of body i in the rubbing zone sometimes given a “safe” limit [10].
• f : friction coefficient
• N : normal load

150
B. IGNITION RISK BY IMPACT
1000

1000
Experimental equipment 900

800

The experimental facility consists of a free fall vertical

Temperature in °C
700

device. The target is a strong mild steel plate (5 mmx70

Température (°C)
600

cmx45 cm) anchored on a rigid thick wood plate (11cm) 500

impeding flexion during impact. The plate is inclined to favor a 400

glancing blow. In most experiments the angle of 60° has been 300
chosen. The projectile (diameter 18 mm of steel, aluminium,
200
copper,…) is propelled by an air cannon on a precise point of
the target at variable velocity between 5 to 50 m/s. 100

0
0,01 0,011 0,012 0,013 0,014 0,015 0,016 0,017 0,018 0,019 0,02

0,01 time in sec.

temps (s)
0,02

reservoir
Figure 6: typical temperature reading and excepts from the
high speed film (18 mm steel rod of 20 cm long, impacting at
10 m/s)
compression chamber
Some data are shown on figure 7 for various experimental
conditions. The impact temperature depends primarily upon
cannon the impact velocity and nature of the material and only very
weakly on the size of the projectile. A further analysis of the
same data shows that the impact temperature is not correlated
target to the incident kinetic energy [5].

1600 mild steel rod - V = 7 +/- 0.5 m/s

mild steel rod - V = 38 +/- 10 m/s
1400
aluminium rod - V = 4.5 +/- 0.5 m/s

1200 copper rod - V = 5.5 +/- 0.5 m/s

Impact temperature (°C)

mild steel rod - V = 20 m/s

1000

800

600

400

200

0
Figure 5 : Experimental impact test equipment 0 0,05 0,1 0,15
Length of the rod (m)
0,2 0,25

The impact was filmed with a high speed camera in order to

estimate the impact speed and to detect the fragments. Figure 7 : Maximum impact temperature of the hot spot for
Simultaneously, the temperature in the contact zone was various conditions
measured with a high speed monochromatic pyrometer [5, 7].
The physical analysis of the results [5,7] suggests that the
A sample of those measurements is shown on figure 6. The efforts, deformations, fragment production… should be the
impact results in a sharp temperature rise (up to hundreds of consequence of the strain waves produced in the very first
°C in a few tens of microseconds) followed from a rapid but moments of the impact and propagated at the speed of sound
more progressive decrease (in milliseconds). Usually a chip is in the materials. These waves are responsible for the rebound.
produced which constitutes a unique fragment of mm size. The experimental evidence suggests also that heat is
produced due to some intense but very short rubbing on the
target during the time of the rebound. while the projectile slides
against the target with the tangential component of the impact
velocity. The amount of heat produced is then driven away by
transient thermal conduction A model has been proposed
elsewhere [5] and the final equation reads:

151
 (Timp − Tamb ) = f ⋅ K m ⋅V 2 ⋅ L
than a ratio of 10 whereas Tpcrit varies only in a ratio of 2 . It is
then tempting to use this formula by setting Km to 20, a rather
large value, and Timp to 550°C, one of the smallest practical
where L stands for the length of the projectile and V the ignition temperature, L being the remaining unique variable
impact velocity. Timp is the peak temperature reached during equal to the largest dimension of the projectile. It is finally very
the impact. Km is a composite parameter dependant only on simple to find an order of magnitude of a critical impact
the intrinsic properties of the bodies like sound speeds, velocity, capable of igniting an explosive atmosphere. A
specific masses and heat capacities, thermal conductivity Km is comparison with data from the literature [7] is presented in
typically of the order of 10 (20 for steel) in SI units. figure 8.

About ignition, both the fragment, if any, and the hot spot
may contribute to ignition equally because of their similar size.
It has been found also [5], that the ignition process is closer to
a hot spot mechanism as described before than to the
standard spark ignition mechanism. Because of the short
duration of the hot spot during an impact, the critical ignition
temperature appears larger, by a factor between 1 to 2
depending on the atmosphere.

For the practical use of the preceding expression, a friction

coefficient of 1 may be selected. √L may easily vary in more

Figure 8 : performances of the simplified model for impact ignition against existing data
experiments nor in the theoretical approach as Rasuo [11]
The comparison is less favorable than for friction but impact noticed earlier The velocity of the impact appears as a more
ignition tests are difficult to perform because of a rather poor convenient parameter.
reproducibility. Nevertheless, this very simplified approach
seems conservative apart from the specific situations of
“thermite” reactions or large objects, mostly because, we
believe, the exact geometry of the projectile is not precisely IV. PRACTICAL IMPLICATIONS
known.
As a final remark, the kinetic energy of the projectile does The foregoing simplified models are issued from a more
not appear, at any moment, neither in the data reduction of the sophisticated tool in which the various equations are solved

152
numerically. The latter gives less conservative results than the A gasoline pump supplier is concerned by the explosion
hazard raised by a new type of connector to feed LPG in cars.
former with exactly the same trend [5]. Both can be used for The possibilities of an impact of the connector on some solid
practical applications. surface and of a friction due to the rubbing along a moving
The basic parameters to feed into the models are the nature vehicle were identified. Impact and rubbing experiments have
of involved material, the size of the equipment, the explosive been performed. Both the connector and the rubbing/impact
properties of the atmosphere, the velocity (sliding velocity for surfaces are mainly made of steel. The typical diameter of the
friction and incident velocity for impact), the normal load and connector is 5 cm. Given the geometry of the connector, a
frictional area for friction. The temperature in the contact zone typical rubbing size of 10 mm was chosen.
is calculated and compared to the critical hot spot temperature
for the atmosphere. For LPG, the critical ignition hot spot temperature is about
Some examples of applications are presented below. 650°C. At this level of temperature, the steel-steel friction
coefficient is about 0.3.
A. Lifting screw
The use of the foregoing simplified expressions reveals that
Within the scope of the CE certification, the case of a lifting the critical impact velocity would be of the order of 23 m/s and
screw devoted to the transfer of combustible powders was the critical rubbing power of about 800 W (N.V =
studied. The screw is a long helical steel wire turning inside a 800/0.3=2500 W). Experiments have shown that the peak
steel sleeve (inner diameter of the sleeve is about 150 mm, 10 impact temperature is about 630°C at 28 m/s and the rubbing
mm wire diameter). The device is driven by a 5500 W electrical temperature is close to 300°C with a rubbing power of 400 W
engine at 1500 rpm. Hazard assessment reveals a risk of (N.V = 1200 W). Recalling that the rubbing temperature has
friction of the wire against the sleeve. The explosive dusty been observed to be approximately proportional to N.V, the
atmosphere has a standard ignition temperature of 240°C experimental observations seem coherent with the predictions.
suggesting a critical hot spot ignition temperature of about
700°C [9]. On the practical side, those values seem high with regard to
the normal way of operating the device and the considered risk
With a thermal conductivity of steel of the order of 15 W.m/K might only appear in case of rare events.
and a rubbing size of the order of 10 mm (diameter of the
wire), the critical temperature could be reached for a frictional
power as low as 400 W according to the simplified friction
model presented above. This finding may be compared with
the experimental data of figure 2 if we consider that this lifting V. CONCLUSION
screw is somewhat comparable to the experimental rubbing
machines in size. At 700 °C, the friction coefficient is about 0.3 Since the implementation of the ATEX directive 94/9/CE in
so that N.V for the lifting screw should be about 1500 W, in line July 2003, the non-electrical sources of inflammation shall also
with the data obtained with the small scale machine. be examined before putting an ATEX equipment on the market
(pumps, couplers, reducing gears, …).
Although rather indicative, this result suggests that the power European harmonized standards have been developed by
of the machine is larger enough to generate immediately a CEN/TC 305 WG2 to give requirements for non-electrical
dangerous situation in case of steel-steel friction. The equipment. The basic standard for non-electrical equipment is
simultaneous presence of a dust atmosphere inside the sleeve the EN 13463-1, which requires an ignition hazard assessment
should then be avoided by ensuring that the device is always of the equipment with the list of all potential ignition sources.
full with bulk dust at full power for instance. Depending on the category of the equipment (1, 2 or 3),
inducing the area where it can be installed (zone 0, 1, 3 or 20,
B. Turbine 21, 22), the assessment has to be carried considering only
normal use or also foreseeable or rare malfunction. The same
We have examined a turbine with a mobile part in aluminium work can be done to ensure that the equipment already in use
and an envelope in alloy of steel-aluminium. fulfills the requirements of the directive 1999/92/CE and can be
The diameter of the wheel was 340 mm. The nominal speed described in the explosion protection document.
-1
was of 3500 tr.min and the peripherical speed was of 62 m/s. Concerning mechanical equipment, this assessment often
We have considered the case of a piece of alloy steel- underline the possibility of mechanical friction or impacts as an
aluminium blocked between the rotor and the casing of the ignition source during foreseeable or rare malfunction. When
turbine generating a strength of 3,6 N. such a source appears, the manufacturer shall estimate if the
The use of the foregoing simplified expressions reveals that friction or impact emits a sufficient energy to ignite the
the rubbing temperature at the contact point will be about surrounding explosive atmosphere.
416°C, which can be compared to the critical ignition hot spot The European program MECHEX has permitted to study the
temperature of an explosive atmosphere. In the practical process of ignition during friction and impact. The process of
studied case it has been demonstrated that there was non degradation of the mechanical energy into heat during friction
ignition risk. or impacts and the mechanisms of ignition at the contact zone
have been studied.
C-Gasoline feeder An extensive experimental program is presented in this paper
but many points have only been surveyed. It was found that,

153
apart for the very special situation of “thermite” reactions,
which has not been investigated in this program, the “sparks VII. REFERENCES
shower” resulting from the production of fragments in the
contact zone is less dangerous than the high temperature of
the friction or impact area. It has been noted that the critical
parameter is the temperature of the hot surface, a property of [1] Directive 94/9/EC of the European Parliament and the
the mixture, which does not seem to be closely linked to any Council of 23 March 1994 on the approximation of the
known standard ignition parameters. laws of the Member States concerning equipment and
Some “simple” modeling is proposed on purpose of practical protective systems intended for use in potentially
applications. For frictional situations, a critical rubbing power is explosive atmospheres, Official Journal no. L100, 1994-
calculated without any limitations about any lower boundary 04-19, 1-29.
concerning the rubbing velocity. For instance, it has been [2] EN 13463-1 : 2001, Non-electrical equipment for
possible to ignite propane-air, ethylene-air and hydrogen-air potentially explosive atmospheres – Part 1 : Basic
mixtures at rubbing velocities equal to or even smaller than 1 method and requirements.
m/s, which has been sometimes unduly given a “safe” limit. [3] Directive 1999/92/EC of the European Parliament and of
The main variable is the friction coefficient the prediction of the Council of 16 December 1999 on minimum
which is not easy: it varies from 0.001 to 1 depending on the requirements for improving the safety and health
lubrication and materials. protection of workers potentially at risk from explosive
For “impacts”, the relevant parameter for ignition is not the atmospheres, Official Journal no L23/57, 2000-01-28
kinetic energy of the projectile but its velocity and the nature of [4] EN 13463-5 : 2003, Non-electrical equipment for
the materials. The experimental data are not sufficiency to potentially explosive atmospheres – Part 5 : Protection
valid the model and more experiences need to be lead. by constructional safety “c”.
As practical implications, some results of MECHEX have [5] C. PROUST, S. HAWSKWORTH, R. ROGERS, M.
already been taken into account in the redaction of the second BEYER, D. LAKIC, D. RAVEAU, P. HERVE, V. PINA, C.
edition of EN 13463-1. Some limited couples of parameter PETITFRERE, X. LEFEBVRE, “Development of a
speed/torque of rubbing have been introduced instead of the method for predicting the ignition of explosive
only speed limit of 1 m/s which was given in the EN 13463-5. atmospheres by mechanical friction and impacts
We have presented in this paper some practical cases (MECHEX),“ HPMIE, Halifax, 2006.
where the simplified model has been used. The materials in [6] KRAGELSKII I.V. (1965), Friction and Wear,
contact, the considered atmosphere, the speed and torque Butterworths, London
shall be known. The surface of contact is estimated and the [7] PROUST C., RAVEAU D., 2004, « Toward modelling of
model calculates the temperature of rubbing or impact zone frictional ignition Loss prevention and safety promotion in
and compares it to the critical ignition temperature of the the process industries », Proceedings of the 11th
explosive atmosphere. It may help for the hazard assessment. International Symposium, 31 May - 3 June 2004, Praha,
On the other hand a certain time is necessary to reach the Czech Republic
maximum temperature at the rubbing surface and some
equipment may use this parameter for implementing a [8] HAWKSWORTH S., ROGERS R., PROUST Ch.,
detection of the failure that will induce the rubbing. BEYER M., SCHENCK S., GUMMER J., RAVEAU D.
The model should still be tested on other practical case. (2004a), “Mechanical ignition hazards in potentially
Some complementary studies would be necessary to explosive atmospheres – EC project MECHEX”,
determine the time to reach maximal temperature by rubbing. communication to the international ESMG symposium,
It would also be interesting to have a better knowledge of Nürnberg, Germany, 16th-18th of March 2004
the value of the critical ignition temperature as a new data for [9] CARLETON F, BOTHE H., PROUST Ch.,
the different explosive atmospheres as it seems not to be HAWKSWORTH S. (2000), « Prenormative Research on
correlated with the known data of autoignition temperature the Use of Optics in Potentially Explosive Atmospheres »,
(AIT) and minimal ignition energy (MIE). Final Report, European Contract SMT4-CT96-2104
[10] HAWKSWORTH S., ROGERS R., PROUST Ch.,
BEYER M., SCHENCK S., GUMMER J., RAVEAU D.
VI. ACKNOWLEDGEMENTS (2005), “Ignition of explosive atmospheres by mechanical
equipment », I.Chem. Symp. Series 150
[11] RASUO V., ZIVKOVIC R. (1990), « Development of
The authors are grateful to European communities with device and method for testing of mechanical sparks
supported the MECHEX project between 2001 and 2005. They which might ignite explosive mixtures », HSE translation
thank also the others sponsors like HSE, the German n°13471
authorities, the French ministry for research and the French [12] POWELL F. (1969), « Ignition of gases and vapors »,
ministry of industry. A warm congratulation to the colleagues Industrial and Engineering Chemistry, vol. 61, pp. 29-37
throughout Europe who took an active part to the project : S.
Hawksworth (coordinator), M. Beyer, D. Lakic, R. Rogers, D; VIII. VITA
Raveau, X. Lefebvre, V. Pina and P. Hervé
Dipl. Ing. Claire Petitfrere leads a certification team
concerning equipment intended to explosive atmospheres in
INERIS since 2002. She is a member of standardisation group

154
 VIII. VITA

Dipl. Ing. Claire Petitfrere leads a certification team

concerning equipment intended to explosive atmospheres in
INERIS since 2002. She is a member of standardisation group
WG2 from TC305 dealing with standards for non electrical
equipment in explosive atmospheres.
Dr. Christophe PROUST, is a senior scientist specialised in
explosions and has managed the work of INERIS in MECHEX
project.

155