Common ATEX Misunderstandings
- Carsten Stegelmann | Principal Consultant
- 1 day ago
- 5 min read
In this article:
ATEX is applied for preventing ignition of flammable gases, vapors and dust clouds and thereby avoiding fire and explosions. But often there are misconceptions on the effectiveness of ATEX.
In connection with HAZID workshops and similar it is commonly encountered that people at operators or system suppliers believe the following:
It is not possible to have flammable gas in unclassified areas;
If proper hazardous area classification (HAC) is performed and the Ex equipment is installed and inspected and maintained to codes then the ignition risk is eliminated.
These belief are very wrong and illustrates a general gap in the understanding of what HAC and ATEX actually implies. In this insight the above misconceptions will be investigated with QRA techniques to illustrate how problematic they are.
Hazardous Area Classification
Area classification is a method of analysing and classifying the environment where explosive gas atmospheres may occur, so as to facilitate the proper selection, installation and operation of equipment to be used safely in that environment. The classification also takes into account the ignition characteristics of the gas or vapour such as ignition energy and ignition temperature. Area classification has two main objectives, the determination of the type of any hazardous zone (ATEX zone), and the extent of the zone (ATEX zone).
The aim of area classification is to avoid ignition of those releases that may occur from time to time in the normal operation of facilities handling flammable fluids. It is NOT the aim of HAC to prevent the ignition of major accidental releases of flammable materials that could extend over large distances from the release source.
Hence, HAC addresses expected releases of flammables, but not catastrophic failures or accidents that are not expected to occur. However, the risk of catastrophic failure can normally not be eliminated.
In practice this means that ATEX zones, when following HAC standards, are often assessed for accidental leak sizes smaller than 5 mm (which are expected to occur from time to time even though they shall be avoided). In comparison QRA considers leak sizes all the way up to full rupture of process equipment. Needless to say the reach of flammable gas for QRA scenarios are not covered by the HAC zones. Hence flammable gas can occur outside the ATEX zones!
Ignition probabilities
As the purpose of ATEX is to avoid ignition it is relevant to look into expected ignition probabilities. In Energy Institutes guideline for HAC /1/ the below ignition probabilities are reported for different sources of ignition:
Source of ignition | Description | Ign. Prob. [-] |
Controlled | Where control of sources of ignition extends beyond Zone 2 (e.g. on offshore facilities where ignition sources are linked to fire and gas detection systems). | 0.003 |
Weak | Typical sources of ignition within a Zone 2 area. | 0.01 |
Medium | Ignition due to road traffic, substations, buildings, unclassified electrical equipment, engines, hot surfaces etc. | 0.1 |
Strong | Continuous strong sources of ignitions such as fired heaters, flares etc. | 1 |
Normally ignition source control is designed much more stringent on offshore oil & gas installations than onshore plants processing flammable materials. This is because ignition offshore can have more catastrophic consequences than onshore as personnel cannot just leave the installation the same way as for an onshore plant. This means that offshore it is normal to apply EX-rated equipment even outside the extend of the ATEX zones or alternatively non-EX equipment is isolated in case of gas detection. But even with this design offshore, the ignition probability does not become zero.
For onshore plants within Zone 2 area an ignition probability in the order of 1% is expected. If the flammable gas extends out of Zone 2 and reaches roads, unclassified electrical equipment etc. the ignition probability becomes in the order of 10% and increases to 100% if strong ignition sources are present. This is despite that the plant is designed according to ATEX directive. Hence ATEX will not eliminate the ignition risk.
QRA modelling
For QRA risk modelling correlations have been developed predicting ignition probabilities for flammable gas and liquid releases in different environments. These correlations clearly shows that the ignition probability increases with the size (mass flow) of the release as would be intuitively expected. The correlations also clearly shows that for a similar release the expected ignition probability of an onshore plant is significantly larger than for an offshore plant which can be explained by the differences in ignition source control.
In QRA context it also recommended practice in industry guidance to assume 100% ignition probability for a release of flammable gas extending outside an onshore plant site boundary. In reality, the ignition probability will not be 100% but it reflects that the plant is not in control of ignition sources outside plant site boundary and that these ignition sources also may change with time. But it basically mean that ATEX plays a minor role in the QRA risk modelling for scenarios that can effect 3rd parties outside the site boundary and is more for protection of the workforce.
Considering that ATEX and HAC primarily addresses smaller more normal leak scenarios rather than catastrophic releases it is interesting to look into how risk to personnel is typically distributed on leak sizes. To investigate this some typical offshore QRA results will be presented. An offshore QRA has been selected over onshore QRAs as offshore QRAs normally has a much higher risk resolution than onshore QRA. In offshore QRAs you typically consider a complete range of leak intervals whereas in onshore QRAs you often only consider the leak scenarios that are able to impact 3rd party.
The below pie diagram shows the distribution of leakage sizes (small (1-5 mm), (medium 5-30mm), Large (30-60mm) and Rupture):
As expected a small leakage occurs much more often than a large leakage. And ATEX focuses on the small leakages.
The pie chart below shows how the risk to personnel (measured as potential loss of lives (PLL)) is distributed on leak sizes for the same offshore installation:
From the above pie charts it is the larger leak scenarios, not typically addressed by ATEX, that dominates the risk to personnel. This should not lead to the conclusion that ATEX is not important as one of the reasons that the small leaks does not dominate despite there are many more small releases than large releases is due to a low ignition probability ensured by the ATEX design. What it shows is that ATEX does not significantly reduce the risk from larger releases and this is the dominating residual risk of typical plants.
Conclusions
The below conclusions can be made:
ATEX primarily mitigate the risk of small and expected leakage scenarios but are not effective against large and catastrophic leak scenarios;
Flammable gas will not be limited to be within ATEX zones;
ATEX does not eliminate ignition risk even for small leaks;
It is the large releases that dominate risk to 3rd parties in onshore QRAs and this risk will not benefit significantly from ATEX;
Extending ATEX requirements as done offshore has a positive impact on ignition risk;
ATEX cannot stand alone in risk management.
Hence care should be exercised when crediting ATEX in various risk assessments (e.g. LOPA) to make sure that ATEX is only credited for the scenarios (small ones) where it offers real risk mitigation.
Furthermore, fire and explosion risk management should not be limited to establishing ATEX zones and applying EX rated equipment. But Fire and Explosion Strategies (FES) for managing fire and explosion risk of major equipment failures need to be addressed as well. The latter is also an requirement in the ATEX directive but may not always receive the attention it deserves.
Ref. /1/: Model code of safe practice – Part 15 – Area classification for installations handling flammable fluids, 4th eds)
