About the Accident Analysis Benchmarking Exercise (AABE) Project

MAIN OBJECTIVE

EXPECTED OUTCOME

OPERATING PROCEDURES

ANALYTICAL METHODS TO BE APPLIED

CASE STUDIES

PLANNED DURATION

PARTICIPATING ORGANISATIONS

REFERENCES (PROJECT TEMPLATES, CASE STUDIES, METHODOLOGIES)

WORKSHOP 2015

WORKSHOP 2018

PROJECT RESULTS

A number of inspectors and industry experts are charged directly with the responsibility for investigation. Their challenge is generally more in terms of following a structured approach, understanding the strengths and weaknesses of different methodologies, and what approach to use depending on the objectives of the investigation, the nature of the accident, and resource limitations. This community has many tools (methodologies) but often inadequate resources to test and explore the advantages and disadvantages of each tool.

On the other hand, a broad challenge to competent authorities and industry, in particular smaller sites and companies, is identifying and applying accident analysis and investigation that are suitable to their resources, competences and objectives. Under the Seveso Directive, for example, companies are obliged to submit major accident reports to the competent authorities and the competent authorities must report these major accidents to the EU’s eMARS accident database. The purpose of this requirement is to facilitate learning and disseminating lessons from these cases. Yet it is increasingly clear that in many cases industry and competent authority personnel lack adequate experience and/or training in accident analysis, because of limited resources (prioritized to ensure competence in more routine safety requirements) and limited to exposure to accident events. This community often is often not well-supported by existing tools.  Moreover, the two communities often have linked roles (one produces the report and the other analyses the report) and they can have an influence on the others’ work, particularly in improving the result.

Against this background, the Benchmarking Project was launched at a workshop hosted by the Major Accident Hazards Bureau of the European Commission's Research Centre on 5-6 November 2015 in Ispra, Italy.  The JRC invited a cross-section of competent authorities, researchers, institutes, OECD representatives and third parties participate in the workshop with the view to taking part in the exercise.  The exercise was projected to take place over the course of 18 months following this inception workshop.   The purpose of the exercise is to apply selected analytical methods to chemical accident case studies and evaluate them in terms of the types of information they help to generate, user-friendliness, and other relevant strengths and weaknesses associated with their application.  The results of the benchmarking will be used to produce a handbook for the chemical process safety community in reviewing, analysing and communicating the results of accident investigations.

MAIN OBJECTIVE

The main objective of the exercise is to compare the results produced by application of different accident analysis methodologies (selected by participants) to analyse a past accident (or possibly, more than one accident).

EXPECTED OUTCOME

• A final workshop hosted by the JRC to present results of each team's results in applying different methodologies.
• A handbook on accident analysis for safety experts who require guidance in performing accident analyses, generally, or in terms of methods that can be applied to address specific types of cases and problems.  It is hoped that this handbook will be broadly useful to government and industry stakeholders.  The EU will in particular use this handbook to assist users in EU and OECD countries who report to the eMARS database, but it will also be freely available in the public domain.

OPERATING PROCEDURES

Expert volunteers will analyse case studies in teams of two or more members applying a number of methodologies as appropriate to the interests and objectives of the team to a case or case studies selected by the team.  (A number of case studies were proposed at the November 2015 workshop.)

ANALYTICAL METHODS TO BE APPLIED

The exercise divides accident analysis into three phases: 

• Phase 1:  Chronology (Examples:  Step/ECFA)
• Phase 2:  Causal (Examples:  Bow Tie, Change Analysis)
• Phase 3:  Root causes (Examples:  Accimap, MTO)

CASE STUDIES PROPOSED AT THE 2015 WORKSHOP

The workshop participants agreed that the benchmarking could include analysis of data-rich and data-poor cases.  It was considered that the advantage of data-rich cases could be that they show the full versatility and limitations of the different methodologies.  However, in reality, the majority of accident investigations are conducted with limited resources and limited objectives resulting in incomplete investigation reports.  Therefore, it was determined that both types of cases should be analysed to address a broad spectrum of issues that may arise in analysing a particular accident, and to ensure that an evaluation of the weaknesses and strengths of each methodology takes into accoun limitations imposed when the accident information is far from complete.

• Fire at an oil storage depot - Buncefield, United Kingdom (10 December 2005)
• 15 deaths from vapour cloud explosion at a petroleum refinery - "BP Texas City" - Texas City (TX), USA (23 March 2005)
• Two fatalities in a toxic cloud - Belgium
• Two injuries from chemical plant explosion - "Shell Moerdijk" - Moerdijk, The Netherlands  (3 June 2014)
• Accident at a chemical packaging and distribution site - "Chemie Pack" - Moerdijk, The Netherlands (5 January 2011)

Teams are welcome also to propose additional cases for analysis.

DURATION

The project was launched at the meeting that took place in Ispra, ITALY on 5-6 November 2015. In 2018 a workshop was held in December in which the results of team exercises were presented and ideas for implementation of project findings were discussed.

The results of the team exercises are presented here.

A report from the project is pending.

PARTICIPATING ORGANISATIONS