DSEAR Dust Explosions
DSEAR Dust Explosions
DSEAR Dust Explosions
for DSEAR
& ATEX
Compliance
Dust Testing
for DSEAR
& ATEX
Compliance
Explosion Testing
for industry
Page 1 of 16
September 2010
Foreword
This document provides an introduction to dust explosion testing and its importance in making process
plant safe from fire and explosion. It describes the standard laboratory tests and the parameters they
measure, enabling plant engineers and system designers to make informed judgement on safety matters
such as: what is safety critical information and what is simply interesting to know; whether explosion and
ignition information can be reliably taken from published data; is data required for every dust and powder
being handled, or can worst case materials be identified?
Dust explosions are fortunately rare occurrences, but they have the potential to kill and legislation under
the Health & Safety at Work Act 1974 and the Dangerous Substances and Explosion Atmospheres
Regulations 2002 is in place to ensure safety in the workplace.
Figure 2 shows a breakdown by industry of the explosion incidents. The most common ignition sources
identified were mechanical sparking, smouldering dust deposits and frictional heating.
What is an explosion?
An explosion is a freely travelling combustion front and can be thought of as a fast fire. When confined,
the combustion front moves at a relatively slow speed (typically 10 metres per second, but strongly
material dependent), yet generates a rise in pressure that will travel at the speed of sound (330 metres
per second).
Explosions have a characteristic pressure/time relationship with a substantial portion being uniform i.e. a
constant rate of pressure rise. When the combustion front reached the vessel walls, it has consumed all
the available fuel or oxygen and self-extinguishes, resulting in a maximum pressure attained. Laboratory
experiments demonstrate that this pressure, known as the maximum explosion pressure or Pmax, is
independent of the vessel volume.
Page 2 of 16
September 2010
Rates of pressure rise are dependant on volume; the larger the volume, the slower the pressure rise. The
cube law accurately describes the relations between volume and rate of pressure rise.
Terminology
Dust explosions may be classified as being either primary or secondary in nature. Primary dust
explosions occur inside process plant or similar enclosures and are generally controlled by pressure relief
through purpose-built ducting to atmosphere. Secondary dust explosions are the result of dust
accumulation inside the factory being disturbed and ignited by the primary explosion, resulting in a much
more dangerous uncontrolled explosion inside the workplace.
Historically, fatalities from dust explosions have largely been the result of secondary dust explosions.
The two photographs below show the aftermath of a large secondary dust explosion that occurred at
th
Imperial Sugar, Georgia on February 8 , 2008, killing 14 and injuring 36.
The official investigation into the incident by the US Chemical Safety Board can be downloaded from the
CSB website www.csb.gov.
Secondary explosions may be avoided by good housekeeping and appropriate process control, stopping
dust from leaking, implementing good cleaning regimes and keeping the factory clear of dust.
There is debate about how thick a layer of dust is needed to support a secondary explosion, but a simple
test is to walk through it leaving footprints is an indication of secondary explosion hazard.
Accumulations of dust on ledges, tops of bins etc. should not be tolerated.
Page 3 of 16
September 2010
Laboratory Tests
A number of laboratory tests have been developed and refined over many years to characterize the
explosion and ignition potential of dusts. The tests described here are designed to simulate plant
operation under worst case conditions that will provide a single value for the test material that can be
applied across a wide range of plant processes and equipment. For example, the 20 litre sphere test
apparatus induces a dust turbulence that is similar to that found inside a hammermill. Conditions inside
(say) a hopper will be considerably less onerous, but the lower explosibility is simply absorbed as
additional safety margin to simplify the testing procedure. Occasionally, plant conditions may exceed that
induced in the standard test and a more strenuous test may be required, as in the case of a jet mill
micronizer or a contained explosion forcing jet flame ignition into an adjoining vessel. Under these
circumstances, seeking specialist advice is recommended.
Test standards are continually under development and it is recommended that testing is undertaken to
the latest applicable EN standard wherever possible.
Page 4 of 16
September 2010
Page 5 of 16
September 2010
Page 6 of 16
September 2010
The ambient temperature is also measured locally to the apparatus and must be within the range 15 C to
o
35 C.
Test criterion
Tests are repeated with fresh layers of dust until a minimum ignition temperature has been found. This is
o
the lowest temperature, rounded down to the nearest integral multiple of 10 C, at which ignition occurs in
a layer of given thickness. The highest value at which ignition does not occur is also recorded. The latter
o
temperature must be no more than 10 C lower than the minimum ignition temperature recorded and is
confirmed by at least three tests.
Ignition is considered to have occurred if:
1. visible glowing or flaming is observed; or
o
2. a temperature of 450 C is measured; or
3. a temperature rise of 250 K above the temperature of the heated plate, is measured.
Any thickness of dust layer may be investigated, but 5 mm is generally regarded as representing the
thickest layer of dust likely in a factory having fully enclosed plant and process operations allied to a good
standard of housekeeping. Where there is the possibility that dust layers in excess of 5 mm may be
formed, the maximum permissible surface temperature must be reduced.
By limiting the temperature of any surface to below the layer ignition temperature (LIT) - usually with a 75
o
C safety margin - fires caused by hot surfaces may be avoided.
NOTE: Some materials exhibit more than one stage of self-heating, and it may sometimes be necessary
to prolong the test in order to fully explore this possibility. With organic dusts, combustion will usually take
the form of charring followed by the appearance of smouldering with glowing which will progress through
the layer and leave a residue of ash. With dust layers composed of certain divided metals, ignition may be
Page 7 of 16
September 2010
400
Temperature (deg C)
350
300
250
200
150
100
50
0
0
10
15
20
25
Time (hours)
Page 8 of 16
September 2010
Dust sample
LEL (gm )
Grain dust
125
Coal dust
30
Wheat flour
60
Sugar
125
Aluminium dust
100
Sewage sludge
250
Page 9 of 16
September 2010
Test method
The MIKE3 apparatus consists of a borosilicate tube placed over a
dispersion cup and fitted with two electrodes. The electrodes are
connected to a circuit that produces an electrical spark of known
energy.
Sample preparation
Where the particle sizes of the material in the plant are not known, tests are carried out on dust samples
of particle sizes e.g. less than 63 m. Where the moisture content in
the plant is not known, the sample is carefully dried e.g. at 50 C
under vacuum, or at 75 C and atmospheric pressure until the
sample weight has reached a constant value.
Page 10 of 16
September 2010
Page 11 of 16
September 2010
The Kst value is calculated as the equivalent pressure in a 1 m sphere from the cube law (Kst = cube root
of volume x explosion pressure rise), as previously described.
ST class
The ST class is based on the Kst value as follows:
ST class
ST 0
ST 1
ST 2
ST 3
Pmax & Kst data is used to design explosion protection systems, typically explosion venting & explosion
suppression.
Example test results
Sample
Grain dust
Coal dust
Flour
Sugar
Wood dust
Aluminium dust
Sewage sludge
Page 12 of 16
Kst
(bar m/sec)
89
85
63
138
224
515
102
Pmax
(bar)
9.3
6.4
9.7
8.5
10.3
11.2
8.1
ST class
1
1
1
1
2
3
1
September 2010
Self-heating Materials
Many materials exhibit self-heating where biological and/or chemical reactions (usually oxidation) take
place at an imperceptibly slow rate at room temperature. However, the rate of reaction increases rapidly
with temperature and given the right storage conditions may lead to spontaneous combustion. The
prediction of situations where spontaneous fires may occur and the identification of simple precautions to
prevent such fires are essential to manage the risk effectively.
A simple screening test is used to identify materials with self-heating or exothermic decomposition
properties.
Test method
The test is carried out in a mechanically ventilated oven with an inner chamber of sufficient size to
accommodate the test piece and allow free air circulation around it. A cubical wire mesh basket
measuring 100 mm on each edge is filled with the material under test and a 0.5 mm stainless steel
sheathed chromel/alumel thermocouple is inserted into the centre of the sample to monitor the central
temperature. The filled cube is then suspended in the oven together with a second thermocouple to
register oven temperature and both are connected to a data logger so that any tendency for self-heating
can be observed and recorded.
Ramped 100mm basket test
600
500
Temperature (C)
400
Oven temperature
300
Sample temperature
200
100
0
0
10
15
20
25
Time (hours)
Test criterion
If the central temperature of the sample tracks the oven temperature for the whole of the test duration (24
hours), the material may be said to exhibit no propensity for self-heating over the time and temperature
ranges investigated.
Page 13 of 16
September 2010
Sugar
Sugar
Sugar
Sugar
Pmax (bar)
ST class
11
59
138
205
3.9
8.2
8.5
9.1
ST 1
ST 1
ST 1
ST 2
With some materials, it is possible to carry out an assessment of explosion risk through the oxygen
balance thus:
CxHyOz + (x + y/4 z/2)O2 xCO2 + (y/2)H2O
where x,y,z are the number of moles of carbon, hydrogen and oxygen respectively
The oxygen balance, which is a measure of the lack of oxygen in the molecule necessary for
stoichiometric combustion, is derived from the following formula:
-16 x (2x +y/2-z) x 100/(molecular weight)
This approach is useful for chemical compounds, but is of little value for foodstuffs, for example. When
dealing with a multitude of different materials, a combination of form, quantity and chemical group can be
appropriate guidelines. Taking food ingredients as an example, a number of product categories could be
considered; flours, sugars, starches, cocoas, milk powders (including whey), egg powders, glutens,
enzymes, emulsifiers, powdered fats, gelling agents and miscellaneous. Examination of existing
explosibility and ignition data then allow us to make a number of reasonable assumptions:
1.
2.
3.
4.
Flours will tend to have high MIE, whereas sugars will generally be much lower.
MIT of flour is likely to be lower than sugar.
High fat flours, such as soya, may have lower LIT temperatures.
Dusty materials, such as cocoa powder, are likely to be drawn into dust extraction filters. If these
are protected by explosion venting, knowledge of Kst & Pmax is required to validate the protection
design.
In this way, a structured examination of the plant and materials handled can be most useful in shaping the
test requirements for a project.
Page 14 of 16
September 2010
Self-heating
Screening test
Action
Specify equipment surface temperature
limits, including safety margin
Influences the choice of plant materials
(conductors/non-conductors), earth
bonding and personnel antistatic
precautions
Identify materials with the potential for
self-heating. Isothermal basket tests will
be required to specify safe storage
conditions
Page 15 of 16
Action
Verify vent design is adequate for the
materials being handled
Verify suppression design is adequate
for the materials being handled
Verify explosion containment and
explosion isolation design is adequate
for the materials being handled
September 2010
Explosion Hazard Testing Ltd, 8 Dalton Court, Astmoor Industrial Estate, Runcorn, Cheshire, WA7 1PU.
Tel: 01753 855782
www.explosiontesting.co.uk
Page 16 of 16
September 2010