File(s) not publicly available
Human factors in rail regulation: Modelling a theory of non-linear rail accident and incident networks using the contributing factors framework
thesisposted on 06.12.2017, 00:00 by Karen KlocknerKaren Klockner
In early 2009, the Contributing Factors Framework (CFF), a manual providing a data set for the coding of systemic factors contributing to rail safety occurrences, was introduced nationally to the rail industry in Australia. It was developed specifically for the rail industry, endorsed by the Rail Safety Regulators’ Panel, and developed by a working group made up of representatives from Australian Rail Safety Regulators, industry and investigation bodies. The CFF aligns with current systemic safety occurrence investigation models (Bird & Germain, 1985; Reason, 1990b) and was developed to capture and code information about all factors that may have contributed to a rail safety occurrence. The three main coding headings used in the CFF to capture this information are: Individual/Team Actions; Technical Failures; and Local Conditions & Organisational Factors. Whilst the CFF captured contributing factors under these three stand-alone headings, this research is interested in modelling the previously unseen non-linear network interactions and relationships of all the contributing factors across these three major headings. It is also interested in looking at the strength of the relationships amongst the contributing factors as a more complex systems oriented way of understanding accident taxonomy. The aim of the research program is to develop a method that identifies the trends, patterns and relationships, through accident modelling, of the unique interaction of Copyright © Karen Klockner 2015 iv factors that contribute the most to different types of rail safety occurrences. To obtain the data necessary for this accident modelling to occur, major rail safety occurrence reports have been analysed for the 5 year period 2006 to 2010 using the CFF tool. The contributing factors for sub-types of major rail safety occurrences were modelled, including: 1) Collisions, 2) Derailments, 3) Safe Working Breaches, and 4) Signals Passed at Danger. This research began by reviewing the contributing factors data from a traditional linear analysis, and then progressed to the use of the newly developed Safety And Failure Event Network (SAFE-Net) methodology to investigate how the contributing factors are interlinked, and to further show the contribution that the factors have on accident phenomenology. The visualisation of the network metrics reveals important properties about the individual contributing factors in the network. This can allow a risk management focus to be directed to mitigating the main contributing factors with a view to controlling and breaking the identified relationships that seem to be present in typical, and all too often reoccurring, railway safety occurrence scenarios. The outcome of this project has been the development of a new methodology for accident modelling called SAFE-Net. The SAFE-Net method has been shown to generate an accident model for each of the various railway safety occurrences under investigation. This is in keeping with the recent call for modern day accident models to be representative of actual accident phenomenology, and to better represent the complex socio-technical systems in which they occur. The SAFE-Net methodology is able to show where maximum benefit can be obtained by directing preventative measures towards those combinations of contributing factors which are having the most influence, in an attempt to prevent the reoccurrence of that type of event. For the first time, contributing factors can be understood as a network of interacting factors and safety improvements can be made by focusing on the critical co-occurrence of various factors discovered across multiple occurrences. The method provides an original means of understanding the complexity of human factors and technical systems that are instrumental in causing safety occurrences.