RAMS and Railways: Preventing Human Error in Onboard Systems

April 3, 2019

RAMS analysis is a key part of the CENELEC standards like EN5026 and EN5029 governing rail system safety in Europe. Learn how human error's impact on system failure is not suitably defined within these standards, and Critical Software's suggestions can help mitigate this issue.

train cockpit with driver

Railway accidents are often the result of a combination of factors, though they frequently involve human error in some shape or form. The CENELEC European standards, which define RAMS analysis (particularly standards EN50126/EN50129), clearly specify that human factors must be considered throughout the system life-cycle. For instance, EN50126 identifies some human factors that should be addressed because they can influence the system development process.

However, both EN50126 and EN50129 do not clearly define some concepts related to the way human factors can influence different system life-cycle stages. Here are two examples:

  • There is clear evidence, from studies performed in other domains, that the most significant factor in reducing hazardous failure rates is having better domain knowledge. It is important to realise that human behaviour can vary, despite being presented with similar situations, and that human understanding and interpretation about a concept that is not well specified will also vary substantially. So, the standards must be more effective in defining the degree of human competence required for each stage in the system life-cycle and competence requirements for each role, whilst specifying the need for domain knowledge.
  • There is also a gap in connecting human performance and human culture in evaluating how each individual will realise that they are facing a hazardous situation, and how they will react to that situation. Two different people will invariably behave differently. This leads to the need for better human behaviour modelling.

Although several techniques have been employed to quantify human behaviour influence in situations that lead to accidents (both quantitative and qualitative approaches), none of them have proven to be effective in complex human behaviour modelling. As the systems are becoming more and more complex, the need for considering the connection between human performance, human culture and external factors - which have impact on human behaviour - increases significantly.

These interactions may generate accidents that are not caused by any component failure, and therefore are hard to model with traditional RAMS methodologies.

It is fair to conclude that new strategies should be adopted in order to manage human errors and guarantee that human factors are properly assessed by ensuring that hazardous situations are detected long before they can transform into accidents.

In this free white paper, we present suggestions on how rail standards need to evolve in order to properly assess complex socio-technical systems, focusing on identifying human errors. We also touch on creating appropriate safety barriers in the system design that reduce the impact of human factors in the final evaluation of risk.