This thesis addresses safety in the railway industry with a focus on safety culture, defined by the United Kingdom’s Health and Safety Executive as ‘the behavioural aspects (i.e. what people do) and the situational aspects of the company (i.e. what the organisation has)’. Current safety management systems do not appropriately incorporate safety culture. This has the potential to cause serious harm to human life.
As the definition implies, safety culture is not easily measured or quantified. It involves factors that influence human behavior in safety critical and technology dense environments such as the railway environment. Furthermore, as railways become more advanced in their operational capabilities and integrated across European countries, safety culture will become increasingly important.
Therefore, safety culture should be a key component of an organisation’s safety management system. However, research to date has shown its integration to be piecemeal. To address this problem, this thesis specifies an enhanced safety case that uses safety culture as an integral part of the process. This provides an improved approach towards safety management.
The key findings from this research show that railways are inherently safe. This is primarily due to the regulations across technical and operational disciplines. Regulations and procedures typically relate to the three possible operational states that can occur: normal, degraded and emergency. An example of a degraded operational state includes a signal failure where a train driver may be given the permission to proceed at caution. The variability between the states can affect a humans understanding of the various technical interfaces and their emergent properties. This in turn can affect the type of behaviour exhibited by a driver, signaller, controller or maintainer. System architecture is therefore an essential tool to identify functional and physical relationships and can be used as a training tool. Training was found to be an effective measure to practically test and evaluate safety culture behaviours. Specifically, the use of a simulated environment has shown to be efficient for learning and training exercises and can be used to improve an organisations safety management system.
The safety case derived in this thesis is therefore, driven by the safety management system and is optimised by an understanding of the particular environment and the user interfaces. The process of integrating safety culture is shown through the improved and derived safety assessment process developed in the thesis.