This website uses non-intrusive cookies to improve your user experience. You can visit our cookie privacy page for more information.

Risk analyses or ‘predictive’ aspects of comah safety reports guidance for explosives sites

The COMAH Safety Report Process for Predictive Assessment of Explosives Sites

Step 4 : Detailed Assessment & Extent and Severity

4.1 Proportionality

The depth of analysis of each major accident scenario needs to be proportionate to the hazards and risks. These analyses can reasonably be banded into three levels of risk assessment:

  1. Qualitative risk assessment
    • is a comprehensive identification and description of hazards from a specified activity, to people on and off­ site or to the environment, and a qualitative assessment of whether the measures in place make the risks ALARP.
    • can represent the range of possible events by a broad classification of the likelihood and consequences for comparison purposes and the identification of priorities.
    • should be informed by a representative selection of specific examples to demonstrate compliance with relevant operational standards and good practice.
    • must quantify the extent and severity of the "worst-case" accident in each consequence category.
    • requires a proportionate consideration of additional risk reduction measures when the risks are not broadly acceptable. Even if the risk is at the broadly acceptable level, there is an obligation to look for further measures and apply if reasonably practicable.
  2. Semi-quantitative risk assessment
    • is the systematic identification and analysis of hazards from a specified activity
    • is represented by means of qualitative and quantitative descriptions of the frequency, and extent and severity of the consequences, to people (on and off-site) or the environment. If the frequency is not generic, then it will be necessary to justify the value used.
    • will involve the use of some mathematical models and levels of harm to people.
    • requires a proportionate consideration (e.g. using simple cost benefit analysis) of the practicability of further risk-reduction measures over and above those representing good practice.
    • judges the significance of the risk estimates by a comparison with specified criteria.
    • must demonstrate gross disproportion for options ruled out.
    • should consider partial implementation of the most effective options when appropriate.
  3. Quantitative risk assessment
    • produces a full numerical representation of the frequency, and extent of a specified level of exposure or harm, to specified people on and off-site, or the environment, from a specified activity.
    • may quantify individual risk for hypothetical people on and off-site.
    • judges the significance of the risk estimates by a comparison with specified relevant criteria.
    • requires a proportionate consideration of all possible further risk reduction measures. The most effective option must be considered first to assess whether it is reasonably practicable.
    • must demonstrate gross disproportion for options ruled out
    • should consider partial implementation of the most effective options when appropriate.

As the proportionality increases from a low level to the highest level, the form of risk assessment is likely to change from qualitative, through semi-quantitative to, in a few cases, full quantitative risk assessment. As described above, a risk matrix can be used as a guide in determining the appropriate depth of assessment required from the distribution of major accidents on the matrix. The risk matrix diagrams (Figures 2 & 3) also indicate the tolerability of risk framework for ALARP decision-making. When risks are in the ‘tolerable’ region, they must also be reduced ALARP. The greater the risks are the greater the degree of proportionality that applies and hence the greater effort or amount of detail required in the risk analysis. Although QRA is not a legal requirement, some operators (and certainly those who choose not to comply with the quantity-distance rules) may find it expedient to use quantified arguments when demonstrating that all measures necessary are in place and further risk reduction is not reasonably practicable.

For explosives sites the proportionality will usually be driven by the on-site risks to workers from accidents with the highest frequencies which can cause the greatest numbers of fatalities. Even with limited worker exposure to certain explosives operations, these typically constitute the highest individual risk events.

For sites where there are significant risks resulting from non-explosives hazards (toxic substances for example), proportionality may be driven by those other substances. Where this is the case then off-site risks must be considered more explicitly.

For those sites with a significant toxic hazard, guidance on the determination of the proportionality is described in the Chlorine SRAG which is available on the HSE web site.

Information must be provided in the safety report, which will allow the CA to assess the basis for the company’s adopted approach to risk assessment of

the site. Further information can be found on establishing the proportionality of risk assessment for the site and the appropriate depth of analysis in HSE’s/HID’s Safety Report Assessment Guide (SRAG), and in HSE’s/HID’s Semi Permanent Circular SPC/Permissioning/12 ‘Guidance on ALARP Decisions in COMAH’ 

4.2 Detailed Analysis of MAS’s in the Representative Set

For each MAS in the representative set, both the consequences and the likelihood of the scenario need to be considered in detail. This includes;

The safety report should describe the above process and identify the key outcomes of the analyses for each element, particularly where they are sensitive to assumptions made.

4.3 Extent and Severity Determinations

Quantification is required of the extent and severity of each MAS in the representative set (i.e. how many people could be:- (i) injured, (ii) hospitalised and (iii) killed). This is established from the detailed analysis by combining the hazard ranges, harm criteria and the specific population in the vicinity.

2010-09-12