SYMPOSIUM SERIES NO.

158 Hazards XXIII # 2012 Crown Copyright

NEW METHODS FOR HAZARDOUS AREA CLASSIFICATION FOR EXPLOSIVE GAS

ATMOSPHERES†

Roger Santon, Mat Ivings, David Webber, Adrian Kelsey

Health and Safety Laboratory, Harpur Hill, Buxton, Derbyshire, SK19 9JN, UK

Hazardous Area Classification (HAC) for explosive gas atmospheres is well established, with
guidance published in various standards and industry codes of practice. One of these documents,
BS EN 60079-10-1:2009 makes use of the concept of a nominal flammable gas cloud volume
Vz to determine the level of ventilation and hence the zone. However, the critical formulae
given in the standard to estimate Vz have no scientific justification. Previous work reported at
Hazards XIX has demonstrated that these formulae over-estimate Vz by several orders of
magnitude.
The present paper summarises an alternative methodology for calculating the Vz in both indoor
enclosures and outdoor situations. The methodology was developed as part of a Joint Industry
Project previously reported at Hazards XXI and as part of more recent work in support of
changes to area classification standards. It is based on an integral free-jet model which has been
validated by comparing predictions to Computational Fluid Dynamics (CFD) simulations, which
were themselves validated against experimental data for gas releases in enclosures. The model for-
mulae are simple to use and easy to apply. For gas leaks within buildings, the ventilation rate is
required as an input to the model. Whilst for forced ventilation this is a known parameter, for
natural ventilation it is necessary to estimate it. A simple and practical methodology to estimate
the natural ventilation rate due to buoyancy and/or the wind is described.
The new methodology is demonstrated using a number of worked examples and compared to the
results from the existing BS EN 60079-10-1:2009 standard.

INTRODUCTION Ang, 1990 and it is now generally accepted by all the

The control of sources of ignition by the use of specially relevant standards and codes that for releases indoors venti-
protected equipment in areas where flammable gases or lation is a critical factor.
vapours may arise has been a fundamental safety measure The current relevant standard, BS EN 60079-10-
for many years. Thus hazardous areas are classified into 1:2009 defines the degree of ventilation as high, medium or
zones based on the frequency of the occurrence and the dur- low on the basis of a calculated hypothetical parameter Vz.
ation of an explosive gas atmosphere. The zones for gases, The volume, Vz, is defined as the volume within which the
vapours or mists are currently defined in the relevant stan- mean concentration of flammable gas arising from a
dard BS EN 60079-10-1:2009 as: release will be 50% (for secondary releases) or 25% (for
primary releases) of the Lower Explosive Limit (LEL). In
Zone 0 – a place in which an explosive gas atmosphere is the standard this definition of the ‘degree of ventilation’ is
present continuously or for long periods or fre- applied to releases outdoors as well as indoors. Given that
quently. the concept of ventilation is meaningless for outdoor
Zone 1 – a place in which an explosive atmosphere is likely releases, the gas cloud volume Vz should be seen as defining
to occur in normal operation occasionally. the degree of ‘dilution’ rather than ventilation.
The standard contains a numerical method for esti-
Zone 2 – a place in which an explosive atmosphere is not mating Vz for indoor and outdoor situations. It also allows
likely to occur in normal operation but, if it the use of alternative methods such as computational fluid
does occur, will persist for a short time only. dynamics (CFD).
There are three defined grades of release; continuous, The value of Vz can then be used to indicate where
primary and secondary. These generally give rise to areas zoning is not required through the concept of negligible
classified as Zones 0, 1, and 2 respectively. extent (NE). If Vz is less than 0.1 m3 thus defining the ven-
tilation (dilution) as ‘high’, it is suggested that, if ignited,
the cloud would produce such small overpressure and
THE CURRENT STANDARD thermal effects that it may be regarded as insignificant.
The application of ventilation to area classification was first Therefore protected equipment and controls over sources
discussed in detail in the seminal publication of Cox, Lees & of ignition are not required under such circumstances.
†
# Crown Copyright 2012. This article is published with the permission of the Controller of HMSO and the Queen’s Printer for Scotland.

339
SYMPOSIUM SERIES NO. 158 Hazards XXIII # 2012 Crown Copyright

The numerical method in the standard has no valid as an alternative to CFD to allow a classification of Zone
scientific basis. It is actually based on the assumption that 2 NE for a wide range of natural gas installations operating
the ratios between (a) the actual ventilation rate and the ven- at pressures below 10 barg. The implications of this work
tilation rate required to dilute the release down to a specified were that many installations will be able to use a lower
level, and (b) the enclosure volume and Vz are equal. classification of zone whilst identifying those that require
additional measures, thereby concentrating resources on
genuine hazards. The work also confirmed the validity of
NATURAL GAS the 0.1 m3 limit for Vz as the criterion for Zone 2 NE
Area classification studies carried out at a large site in classification through practical experiments. This work
2005 indicated that the numerical methods in BS EN was subsequently incorporated into a revision of IGEM/
60079-10-1 provided grossly conservative values of Vz. SR/25 (2010), the gas industry’s standard for the area
Subsequent work presented to Hazards XIX (Gant et al., classification of natural gas installations. The concept of
2006) showed that the values of Vz calculated from the Zone 2 NE was introduced into this document for the
standard were typically between 100 and 3000 times first time.
greater than values obtained by using CFD. Table 1
shows the results for methane from a range of hole sizes
and pressures. Significantly, all the results from the CFD OTHER GASES
model for Vz are below the criterion of 0.1m3, whilst all Having developed a scientifically-based solution for natural
the results calculated from the methodology in the standard gas, it became clear that a generic alternative to the numeri-
are above 0.1m3. cal methods in BS EN 60079-10-1:2009 was required for the
Following this work, and in view of the requirement area classification of other gases including releases both out-
for valid viable area classification methods for the natural doors and indoors.
gas industry for the purposes of compliance with the Integral models of gas jets have been used routinely
ATEX Directives, a major Joint Industry Project (JIP) was for the assessment of major hazards for many years. In
set up, jointly funded by HSE and representatives from their simplest form they comprise a few equations describ-
the European gas industry and regulatory bodies. The JIP ing the axial and radial profiles of concentration and vel-
was carried out between 2006 and 2007 and is reported by ocity. More sophisticated versions take into account
Ivings et al., 2008. The results were presented to Hazards complex source terms, dense and buoyant clouds, rainout
XXI (Santon and Ivings, 2009). etc. Despite their essential simplicity, their scientific credi-
The work showed, for natural gas releases below 10 bility is well established and it is therefore appropriate to
barg, that if the average gas concentration at the ventilation apply them to the hazardous area classification of pres-
outlets is kept below 10% LEL then the gas cloud volume surised flammable gases.
Vz would typically be less than 0.1 m3. Calculation of the Here we describe a simple, scientifically-based, inte-
average gas concentration at the outlet is straightforward, gral model for the release of a flammable gas in a ventilated
based on the volumetric release rate divided by the venti- enclosure. The model, called QUADVENT, is described in
lation rate, and therefore provides a simple means of detail by Webber et al. (2011) including the derivation of the
demonstrating compliance with the Zone 2 NE criterion. approach. That paper includes solutions for plume releases
This simple approach to area classification can be used as well as jet releases.
Firstly it is necessary to distinguish between
sonic (choked) and subsonic releases. Unchoked flow will
result if:
Table 1. Comparison of values of Vz for methane from BS EN
60079-10-1:2009 and CFD (from Gant et al., 2006) P
 1.9 (1)
3 Pa
Case Leak Conditions Gas Cloud Volume, Vz (m )
where P is the gas storage pressure and Pa is atmospheric
Pressure Area BS EN 60079-
pressure.
(barg) (mm2) CFD 10-1:2009
It is also necessary to define a pseudo source hole
1 5.0 5.0 0.0936 12.3 – 61.7 radius when the flow is sonic.
2 5.0 2.5 0.0326 6.17 – 30.8 
3 5.0 0.25 0.0012 0.62 – 3.08  
P
4 2.5 5.0 0.0433 7.22 – 36.1 rs = r0 1 + 0.5 − 1.9 (2)
Pa
5 2.5 2.5 0.0148 3.60 – 18.0
6 2.5 0.25 0.0005 0.36 – 1.80
7 0.5 5.0 0.0147 2.98 – 14.9 where rs is the pseudo source radius and ro is the hole radius,
8 0.5 2.5 0.0054 1.50 – 7.48 in metres.
9 0.5 0.25 0.0002 0.15 – 0.75 The derivation of the following equations is given in
Webber et al. (2011).

340
SYMPOSIUM SERIES NO. 158 Hazards XXIII # 2012 Crown Copyright

JET RELEASE IN A VENTILATED ENCLOSURE VALIDATION

The gas cloud volume Vz (m3) can be calculated as Equation (3) above has been validated (Webber et al., 2011)
  against detailed CFD simulations which themselves have
    been validated against experimental data (Ivings et al.,
9p rs3 rb 3/2 1−xb 3
VZ = min , V0 ; xb , xcrit 2008). The validation data includes simulations of a range
16a rs xcrit −xb
of flammable gas release rates in enclosures of various
= V0 ; xb ≥ xcrit sizes at a range of different ventilation rates. All of the simu-
lations are for unobstructed releases of methane in a venti-
(3) lation controlled chamber. The agreement between the
QUADVENT model and the CFD simulations is surpris-
where rb (kg/m3) is the density of the background (which ingly good considering how simple the QUADVENT calcu-
normally approximates to that of air), rs (kg/m3) is the lation is. Figure 1 below shows an example of the model
density of the source gas, a is the entrainment coefficient validation data, which in this case was for releases in a
(recommended value 0.05), xb (v/v) is the background con- 44.7 m3 enclosure. The qualitative behaviour of gas cloud
centration, xcrit (v/v) is the concentation of interest (50% build up in a ventilated enclosure is very clearly captured
LEL for secondary releases) and Vo (m3) is the net enclosure by the QUADVENT model. Where the average gas concen-
volume. The background concentration of flammable gas in tration at the outlet is low, for example for a small release in
the enclosure is a large enclosure, the exact value of the ventilation rate has
qs little effect on the gas cloud volume. In such cases the critical
xb = (4) factors are the source of the release (hole size and pressure)
1q1
and the flow rate of fresh air to the leak location, i.e. the local
ventilation effectiveness. As the average concentration at the
where q1 (m3/s) is the ventilation rate, qs (m3/s) is the outlet tends towards the concentration defining the gas cloud
source gas volume flowrate and 1 is the efficiency of back- volume, xcrit in this case, the ventilation rate becomes critical
ground mixing (see below). The leak rate qs can be derived to the gas cloud build up and the gas cloud volume rapidly
from standard methods for the estimation of leak flowrates. becomes equal to the enclosure volume.
Appropriate methods are included in BS EN 60079-10- The cases considered in the validation study are all for
1:2009, Annex A. unobstructed releases, as this is what QUADVENT is essen-
tially modelling. The CFD simulations presented by Ivings
et al. (2008) considered a wide range of cases to assess
JET RELEASES OUTDOORS
the effects of obstructions near to the leak source on the
In this case there is zero background concentration, xb ¼ 0,
gas cloud build up. Some of the cases were artificially con-
and the background density is that of pure air in which case
trived to try and find the largest reasonable increase in gas
equation (3) reduces to
cloud volume that could be obtained for a given release
    rate and ventilation rate.
9p rs3 ra 3/2 1 3 To apply QUADVENT to an obstructed release the
VZ = (5)
16a rs xcrit only free parameter is the efficiency of mixing 1. Using a
value 1 = 13 is sufficient to account for the effects of the
obstructions on the gas cloud build up for all of the con-
ZONE EXTENT FROM A JET RELEASE sidered releases in the 44.7 m3 enclosure. However, a
The axial distance z to a concentration xzone may be derived value of 1 = 10 1
is required for releases in the 400 m3 enclo-
as an approximation to the zone extent. In this case, an sure. This simply demonstrates that the degree to which the
appropriate value of xzone should be chosen. BS EN ventilation is well-distributed throughout the enclosure, rep-
60079-10-1 uses 100% LEL for example. resented by 1, is just one factor that affects the gas cloud
build up. The other key factor, which is particularly impor-
rs (1 − xzone ) tant for releases in large enclosures is the effectiveness of
z= (6)
m (xzone − xb ) the local ventilation. An assessment of the local ventilation
effectiveness will be needed to confirm that fresh ventilation
where air is effectively reaching the leak source through visual
inspection, smoke tests or other effective means.

r
m ; 2a s (7)
rb
ENCLOSURE AND BUILDING VENTILATION
The ventilation rate of an enclosure is a key input to an area
This calculation only makes sense if the background classification assessment. Forced ventilation rates can be
concentration is less than the zone extent concentration, established from design or equipment specifications, but in
i.e. xb , xzone, otherwise the zone would extend throughout many cases the enclosure will be naturally ventilated and
the enclosure. hence the ventilation rate will not be specified but must be

341
SYMPOSIUM SERIES NO. 158 Hazards XXIII # 2012 Crown Copyright

Figure 1. Predictions from CFD (filled symbols) and the integral model QUADVENT (hollow symbols) for Vz plotted against the
concentration at the ventilation outlet. Points corresponding to different jet release rates and ventilation rates are included in the figure.

measured or calculated. The natural ventilation rate will opening, driven by the pressure difference, Dpi (Pa). The
vary through time as it is strongly influenced by the sign of the pressure difference across an opening determines
weather conditions. Simple approaches for the estimation the direction of the flow through it. The coefficient of dis-
of ventilation rates, suitable for use as part of HAC method- charge, Cd, is the loss coefficient for an opening, accounting
ologies, are therefore required. for the difference between the ideal and actual flow rates.
Natural ventilation consists of two main elements: For a planned ventilation opening a typical value for the
wind driven and buoyancy driven ventilation. When esti- coefficient of discharge is 0.6.
mating the ventilation rate as part of a HAC assessment The pressure difference at each opening i will be due to
only planned ventilation openings are included. The effect a combination of wind and buoyancy forces. The pressure
on ventilation rate of adventitious openings, due to cracks differences can be added to give the pressure difference
and gaps in a structure, are not considered. across an opening. The pressure difference Dpi at each
The ventilation in an enclosure can be simply opening i will be due to a combination of wind and buoyancy
described from the flow rates, qi (m3 s21), through each of forces:
the i openings. For an enclosure where the interior can be
Dpi = Dp0 − Dr0 gzi + 0.5C p,i rout U 2 (10)
considered to form a single volume, mass continuity tells
us that
where Dp0 (Pa) is a reference pressure difference. The
second term on the right hand side represents pressure differ-
ri qi = 0 (8) ence due to buoyancy, where Dr0 (kg m23) is the difference
between external and internal densities at the reference
where ri (kg m23) is the density of the flow through the height, g (m s22) is the acceleration due to gravity and zi
opening. (m) is the distance of the mid-height of opening i above
There can be multiple planned ventilation openings the reference height. The third term on the right hand side
on any surface of the enclosure. The flow rate through represents the pressure difference due to wind effects for
each opening, i, can be written as small openings, where Cp,i is the coefficient of pressure for
opening i, U (m s21) is the velocity of undisturbed flow at

the reference height used in the measurement of Cp,i, (typi-
2|Dpi |
qi = sgn(Dpi )Cd Ai (9) cally the roof height of a structure is used as the reference
ri height) and rout (kg m23) is the density of the external air.

 Both the coefficient of pressure, determined by a
where Ai (m2) is the area of opening i and 2|Dpi |/ri structure and its surroundings, and the wind speed will
(m s21) is the magnitude of the velocity through the affect the estimate of the ventilation rate.

342
SYMPOSIUM SERIES NO. 158 Hazards XXIII # 2012 Crown Copyright

Tables of coefficients of pressure for structures are CFD simulations, may be necessary to determine the venti-
available in BS5925:1991 and Liddament (1996). For lation rate, but many factors, including the weather con-
more complicated structures and openings it may be necess- ditions, must be taken into account. Clearly the amount of
ary to make experimental measurements or perform CFD effort that is put in to determining the ventilation rate
simulations. should be in proportion to the degree of accuracy required
A realistic worse case condition for the wind speed is for the ventilation rate estimate.
the speed exceeded for 80% of the time. For sites in the UK
BS5925:1991 provides information on wind speed that can
be used to calculate this value. A map of the UK provides
LIMITATIONS AND RECOMMENDATIONS
the wind speed, measured at a height of 10 m, exceeded
There are certain limitations that should be noted when
for 50% of the time. Information is given to transform this
applying this work.
value to the reference wind speed exceeded for 80% of
the time. Finally, the wind speed exceeded for 80% of the . The biggest uncertainty lies in estimating the size of the
time at 10 m can be transformed to an effective undisturbed hole. The value of Vz depends on the cube of the radius
wind speed at the building height, that can be used with the and is therefore sensitive to it. The size of expected hole
pressure coefficients. The available correlations describing must not, therefore, be underestimated. A minimum of
the variation with height account for the effect of terrain, 0.25mm2 is recommended unless sensitive leak testing
e.g. rural, open or city, and height. procedures have been applied to test the security of
The system of equations, (8) –(10), can be solved by the system at construction and regularly thereafter in
using an iterative approach to find the value of Dp0 that which case a minimum value of 0.025mm2 may be
enforces mass continuity for the enclosure. A solution can used for pressures up to 100 mbarg. Further guidance
be found with spreadsheets using solvers to find the value on hole sizes is given in Cox, Lees, and Ang, 1990.
obeying the continuity constraint. An example of a spread- . The build up of flammable gas following a release from
sheet containing this simple model of wind and buoyancy a pressurised system can be strongly affected by local
driven ventilation, developed for the purpose of ventilation obstructions to the resulting jet and to the ventilation
design, is available from the Chartered Institution of Build- flow. Both have a very significant effect on the dis-
ing Services Engineers: www.cibse.org/venttools. This persion of the flammable gas and the resulting size of
does not use an iterative solution but demonstrates how gas cloud. It is therefore very important that the effect
the model could be described in a spreadsheet. of this congestion/confinement is accounted for in an
The calculation of natural ventilation described above area classification methodology. Further guidance is
is a general solution for a structure that can be represented given in IGEM/SR/25, 2010.
as a single enclosure. The contribution to the ventilation . A preliminary assessment of the effects of confinement/
from multiple openings, on any surface of the enclosure, congestion on the gas cloud build up can be carried out
due to the influences of wind and buoyancy can be con- by assigning an appropriate value to the efficiency of
sidered. Both simpler and more complex cases can also be mixing, 1. A value of 1 ¼ 1 represents an unobstructed
considered. release, 1 = 12 represents a moderate degree of obstruc-
In BS5925:1991 simpler examples of natural venti- tion and 1 = 13 represents a significant obstruction to
lation, where openings are only on the upwind and downwind the ventilation flow. These values are taken from an
surfaces of a single enclosure, are presented. For these analysis of the data produced by the JIP (Ivings et al,
examples analytical solutions can be found for either wind 2008). For larger enclosures, say over 100 m3, this
driven, or buoyancy driven, ventilation, but not their com- approach is not sufficient on its own and an assessment
bined effects. The approach suggested in BS5925:1991 is of local ventilation effectiveness will be needed to
to calculate the flow rates due to buoyancy and wind separ- confirm that fresh ventilation air is effectively reaching
ately, then use whichever of these values is larger as the the leak source through visual inspection, smoke tests or
flow rate. The approach suggested here is to calculate the flow other effective means.
due to both buoyancy and wind and treat the larger flow as . Uncertainty in the model accuracy needs to be
representing their combined effects. The areas of openings accounted for in applying this area classification meth-
can be combined in the calculations, allowing some flexi- odology. This can be done by applying a safety factor
bility in representation, but not the flexibility available by of two to the estimated ventilation rate or by ensuring
using an iterative solution. that the hole size is over-estimated.
More complex enclosures, for example, where treat- . The indoor release result assumes that the effects of ven-
ing a structure as containing a single volume is inappropri- tilation permeate the whole room. If the release is likely
ate, due to internal partitioning, can also be examined. In to be within a sub-chamber which is largely isolated
these cases multizone models, such as COMIS (Haas from the effects of ventilation in the main room, then
et al., 2002) or CONTAM (Walton and Dols, 2010), can Vz should be computed for the volume of the sub-
be used to calculate ventilation flows. chamber with a lower ventilation rate, or simply set to
For more complex enclosures, where the above be the volume of the sub-chamber, which is likely to
approach is unsuitable, experimental measurements, or be greater than 0.1 m3.

343
SYMPOSIUM SERIES NO. 158 Hazards XXIII # 2012 Crown Copyright

. Consideration must also be given to the integrity of the Since Vz . 0.1 m3, Zone 2 NE cannot be adopted,
vessel or pipe-work containing the hazardous gas. If the irrespective of the value of f, the factor (≥1) to allow for
pressure is sufficiently high, any small hole which impeded flow. Vz does not exceed 3400 m3, the value
appears may rapidly enlarge, so that, for example, a pin- assumed by the standard to represent outdoor circumstances,
prick hole at 100 barg would not be considered credible. and therefore in accordance with B.5.3 the ventilation is
For this reason, it is recommended that the classification classed as medium. The availability of outdoor ventilation
of NE should be limited to a gas storage pressure of is classed as good, and therefore according to Table B.1 the
10 barg. This constraint may be raised to 20 barg area classification for secondary sources is Zone 2.
based on risk assessment taking the consequences of
ignition, i.e. the risk of injury into account. (The method-
(B) Applying QUADVENT
ology for the estimation of exposure and ignition prob-
Equations (2) and (5) can be applied using
ability in EI15 (2005) Annex C may be used to
ro ¼ 0.28 mm (for a hole area of 0.25 mm2)
determine the hole size to be used based on release fre-
quency level.)
rs ¼ 2.47 kg/m3
ra ¼ 1.204 kg/m3
. For enclosure volumes of less than 10m3 the criterion of
xcrit ¼ 50% LEL ¼ 0.0075 v/v
0.1 m3 for Vz should be reduced to 1% of the enclosure
Therefore
volume. This constraint is taken from BS EN 60079-10-
1:2009 and is based on Ivings et al. (2008). 
 
. 4.5
The guidance in BS EN 60079-10-1:2009 on the avail- rs = 0.28 1 + 0.5 − 1.9 = 0.42 mm
ability of ventilation should be observed. 1
. Whilst this methodology is valid for all gases, it should    3
be noted that the validation of the criterion of 0.1 m3 for 9p 0.423 1.204 3/2 1
VZ = = 0.0021 m3
the value of Vz leading to an NE classification has only 16 × 0.05 × 109 2.47 0.0075
been carried out for natural gas.
Vz is less than 0.1 m3 and the area classification is
therefore Zone 2 NE providing other constraints such as
EXAMPLES freedom from local confinement and congestion are met.
For nomenclature and clause references in the calculations
below, see IGEM/SR/25 and BS EN 60079-10 where
relevant.
EXAMPLE 2
A plant room contains 3 × 1400 kW boilers fired by natural
EXAMPLE 1 gas, together with gas boosters, metering and pipework.
Outdoor butane gas pipework, secondary releases Natural ventilation is provided by 3 louvered vents which
LEL ¼ 1.5% v/v are all in one wall. The total ventilation area is 5.83 m2.
Molecular weight ¼ 58.123 kg/kmol The room net volume is 2660 m3. There are no primary
Pressure P ¼ 4.5 bara (4.5 × 105 Pa) sources of release. Maximum gas pressure is 76 mbarg.
Hole size ¼ 0.25 mm2 (Typical value for screwed From Santon and Ivings (2009) applying BS
pipework with small valves) 5925:1991 the ventilation rate is 0.243 m3/s, giving an air
change rate of C ¼ 0.243/2660 s21 ¼ 0.0000913 s21. The
(A) Applying BS EN 60079-10-1:2009
ventilation was checked using an anemometer and the esti-
The release rate calculated from clause A.3.2.1 for choked
mate was confirmed as conservative. It was also confirmed
flow is 0.000344 kg/s.
by smoke tests that there were no significant stagnant areas
Calculation of Vz:
or areas of recirculation.
From B.5.2.2 Note 1, LELm ¼ 0.416 × 1023 × M The boosters are treated as giving rise to an adverse
× LELv environment, so that leak sizes of 2.5 mm2 are used.
LELm ¼ 0.416 x 1023 x 58.123 x 1.5 ¼ 0.0363 kg/m3 Using Cd of 0.8 and standard flow calculation equations
  from BS EN 60079-10-1:2009, leak rate at 76 mbarg ¼
dV (dG/dt)max T
From equation B1: dT = × 0.25 g/s.
min k × LEL m 293
Applying IGEM/SR/25 (2010), A7.1.1, the total
0.000344 numbers of gas fittings is approximately 8 regulators,
=
0.5 × 0.0363 24 valves, 88 flanges and 22 screwed joints. The room is
= 0.019 m3 /s normally visited daily, but for the purposes of frequency
of inspection it is assumed to be visited at worst every
  2 weeks. Applying the calculationmethodology in section
dV
f× A7.1.2 of IGEM/SR/25 (2010), Tfin ¼ 0.0068, so that
dt min
From equation B.5:Vz = = 0.019/0.03 the number of simultaneous secondary releases that should
0.03 be taken into account for the purposes of IGEM/SR/25
= 0.63 m3 (taking f = 1) (2010) is two based on Table 7. For comparative zoning

344
SYMPOSIUM SERIES NO. 158 Hazards XXIII # 2012 Crown Copyright

calculations the leak rate of 0.5 g/s is therefore used in all    3

9× p × 0.8923 1.204 1.5 1−0.0059
examples below. Vz =
16 × 0.05 × 109 0.698 0.022 − 0.0059
= 0.013 m3
(A) Applying BS EN 60079-10-1:2009
Vz is less than 0.1m3 and the area classification is
Calculation of Vz:
therefore Zone 2 NE.
From B.5.2.2 Note 1, LELm ¼ 0.416 × 1023 × M
× LELv SUMMARY FROM EXAMPLES
M ¼ 17 kg/kmol, LELv ¼ 4.4% v/v In both cases the use of BS EN 60079-10-1:2009 results in a
LELm ¼ 0.416 × 1023 × 17 × 4.4 ¼ 0.0311 kg/m3 classification of Zone 2, whilst QUADVENT results in a
From equation B1: classification of Zone 2 NE. In the second example, the
  use of the industry code IGEM/SR/25 (2010) also results
dV (dG/dt)max T 0.5 in a classification of Zone 2 NE.
= × = 3
dT min k × LELm 293 10 × 0.5 × 0.0311
= 0.032 m3 /s  
dV CONCLUSIONS
f× We have developed QUADVENT for the purposes of
dt
From equation B.4: Vz = min
area classification for gases and as an alternative for estimat-
C ing the parameter Vz as defined in BS EN 60079-10-1:2009.
= 0.032/0.0000913 The methodology in the standard has no basis in science and
gives arbitrary results. QUADVENT is based on simple well
= 350 m3 (taking f = 1) established jet model theory, and has been validated against
Since Vz . 0.1 m3, Zone 2 NE cannot be adopted, experimentally validated CFD. It has been shown to provide
irrespective of the value of f. Vz does not exceed V0, and solutions that are significantly lower than those resulting
therefore in accordance with B4 the ventilation is classed from the use of the methodology in the standard, and
as medium. From table B.2 the area classification for sec- which are compatible with the gas industry standard
ondary sources is therefore Zone 2. IGEM/SR/25 (2010). Future work to extend the principles
to flashing liquids and other scenarios is planned.
This publication and the work it describes were
(B) Applying IGEM/SR/25 (2010)
funded by the Health and Safety Executive (HSE). Its con-
Applying A7.2.1, to assess the adequacy of the ventilation:
tents, including any opinions and/or conclusions expressed,
Calculating the value of the bulk gas concentration
are those of the authors alone and do not necessarily reflect
Cout
HSE policy.
Gas density ¼ 0.666 kg/m3
From Ivings et al. (2008)
Cout ¼ 0.5 × 1023/(0.243 × 0.666) ¼ 2.47 × 1023 REFERENCES
v/v ¼ 0.31% v/v BS 5925:1991, Code of practice for Ventilation principles and
LEL ¼ 4.4 % v/v, so 10 % LEL ¼ 0.44% v/v designing for natural ventilation.
BS EN 60079-10-1:2009, Electrical apparatus for explosive gas
Cout , 10 % LEL criterion is met. The ventilation is atmospheres – Part 10: Classification of hazardous areas.
“more than adequate”. It meets the other criteria in A7.2 so Cox A.W., Lees F. P. and Ang M. L., Classification of Hazar-
that, in accordance with 5.3.1 (a) the zone distances in dous Locations, IChemE, 1990.
Table 1 apply and Zone 2 NE applies. EI15, Model code of safe practice, Part 15, Area classification
code for installations handling flammable fluids, 3rd
(C) Applying QUADVENT edition, Energy Institute, 2005.
Equations (3) and (4) are applicable using Gant S.E., Ivings M.J., Santon R.C., and Jones A., Hazardous
Area Classification of Low Pressure Natural Gas Systems
qs ¼ leak rate ¼ 0.5/698 m3/s ¼ 0.000716 m3/s using CFD Predictions, Hazards XIX, Manchester 2006,
(equivalent to 0.5g/s) IChemE.
1 ¼ efficiency of background mixing ¼ 12 (in this Haas A., Weber A., Dorer V., Keilholz W. and Pelletret
case) R. (2002) Comis 3.1 simulation environment for multizone
q1 ¼ ventilation rate ¼ 0.243 m3/s air flow and pollutant transport modelling Energy and Build-
rs ¼ 0.892 mm (for a hole area of 2.5 mm2) ings 34, 873-882.
rs ¼ 0.698 kg/m3 IGEM/SR/25, Edition 2, Hazardous area classification of
rb ≈ 1.204 kg/m3 (density of air) (a more accurate natural gas installations, Institution of Gas Engineers and
value can be calculated if necessary) Managers, 2010.
xcrit ¼ 50% LEL ; 0.022 v/v Ivings M.J., Clarke S., Gant S.E., Fletcher B., Heather B., Pocock
xb ¼ 0.5 × 2/698 × 0.243 ¼ 0.0059 v/v D.J., Pritchard D.K., Santon R.C., and Saunders C.J., Area

345
SYMPOSIUM SERIES NO. 158 Hazards XXIII # 2012 Crown Copyright

classification for secondary releases from low pressure natural Walton G. N. and Dols W. S. (2010) CONTAM User Guide and
gas systems, HSL Research Report RR630, 2008. Program Documentation NIST, NISTIR 7251
Liddament M. W. (1996) A Guide to Energy Efficient Venti- Webber D.M., Ivings M.J. and Santon R.C., Ventilation theory
lation AIVC and dispersion modelling applied to hazardous area classifi-
Santon R.C. and Ivings M.J., Area Classification of Natural Gas cation, Journal of Loss Prevention in the Process Industries
Installations, Hazards XXI, Manchester 2009, IChemE. Volume 24, Issue 5, September 2011, Pages 612-621.

346