DSTO Publications Online >
DSTO Publications Online Repository >
DSTO Formal Reports >
Please use this identifier to cite or link to this item:
|Title: ||A review of accident modelling approaches for complex critical sociotechnical systems.|
|Report number: ||DSTO-TR-2094|
|AR number: ||AR-014-089|
|Report type: ||Technical Report|
|Authors: ||Qureshi, Z.H.|
|Issue Date: ||2008-01|
|Division: ||Command, Control, Communication and Intelligence Division|
|Release authority: ||Chief, Command, Control, Communication and Intelligence Division|
|Task sponsor: ||DMO|
|Task number: ||C3ID DMO 07-007|
|Pages or format: ||66|
|DSTORL/DEFTEST terms: ||Accidents|
|Other descriptors: ||Safety-critical systems|
|Abstract: ||The increasing complexity in highly technological systems such as aviation, maritime, air traffic control, telecommunications, nuclear power plants, defence and aerospace, chemical and petroleum industry, and healthcare and patient safety is leading to potentially disastrous failure modes and new kinds of safety issues. Traditional accident modelling approaches are not adequate to analyse accidents that occur in modern sociotechnical systems, where accident causation is not the result of an individual component failure or human error. This
report provides a review of key traditional accident modelling approaches and their
limitations, and describes new system-theoretic approaches to the modelling and analysis of accidents in safety-critical systems. It also discusses current research on the application of formal (mathematically-based) methods to accident modelling and organisational theories on safety and accident causation. This report recommends new approaches to the modelling and analysis of complex systems that are based on systems theory and interdisciplinary research, in order to capture the complexity of modern sociotechnical systems from a broad systemic
view for understanding the multi-dimensional aspects of safety and accident causation.|
|Executive summary: ||Highly technological systems such as aviation, maritime, air traffic control,
telecommunications, nuclear power plants, defence and aerospace, chemical and
petroleum industry, and healthcare and patient safety are exceedingly becoming more
complex. Such complex systems can exhibit potentially disastrous failure modes.
Notable disasters and accidents such as the Bhopal toxic gas release disaster
(Srivastava, 1992), the NASA Challenger shuttle explosion (Vaughn, 1996), the US
Black Hawk fratricide incident during the 1994 Gulf War Operation Provide Comfort
(AAIB, 1994), the Royal Australian Air Force F-111 chemical exposure of maintenance
workers (Clarkson et al., 2001), the Esso Longford gas plant accident (Hopkins, 2000),
and a number of critical aviation and train accidents such as the 1993 Warsaw accident
(Höhl & Ladkin, 1997) and the Glenbrook NSW Rail accident (Ladkin, 2005) respectively, are clear examples of system failures in complex systems that led to serious loss of material and human life.
Large complex systems such as the Bhopal chemical plant and the Operation Provide
Comfort Command and Control System are semantically complex (it generally takes a
great deal of time to master the relevant domain knowledge), with tight couplings
between various parts, and where operations are often carried out under time pressure
or other resource constraints (Woods et al., 1994). In such systems, accidents gradually
develop over a period of time through a conjunction of several small failures, both
machine and human (Perrow, 1984; Reason, 1990). This pattern is generally found in
different industrial and aerospace accidents, despite the fact that every sociotechnical system is unique and each accident has many different aspects.
It is important to understand the causes of accidents in complex systems in order to
enhance the safety of such systems, and to develop preventative strategies to mitigate
the occurrence of future similar accidents. Accident models provide a conceptualisation of the characteristics of the accident, which typically show the relation between causes and effects. They explain why accidents occur, and are used as techniques for risk assessment during system development, and for post hoc accident
analysis to study the causes of the occurrence of an accident. Most of the engineering models originated before the introduction of digital technology; these models have been updated but have not kept pace with the fast change in technological revolution.
Modern technology is having a significant impact on the nature of accidents, and this
requires new causal explanatory mechanisms to understand them and in the development of new risk assessment techniques to prevent their occurrence (Leveson, 2003).
Traditionally, accidents have been viewed as resulting from a chain of failure events,
each related to its “causal” event or events. Almost all safety analysis and risk
assessment techniques are based on this linear notion of causality, which have severe
limitations in the modelling and analysis of modern complex systems. As opposed to
conventional engineered systems, modern complex systems constitute different kinds
of elements, intentional and non-intentional: social institutions, human agents and technical artefacts (Kroes et al., 2006). In these systems, referred as sociotechnical systems, humans interact with technology to deliver outcomes that cannot be attained by humans or technology functioning in isolation. In sociotechnical systems human agents and social institutions are integrated, and the attainment of organisational objectives is not met by the optimisation of technical systems alone, but by the joint optimisation of the technical and social aspects (Trist & Bamforth, 1951). Thus, the study of modern complex systems requires an understanding of the interactions and interrelationships between the technical, human, social and organisational aspects of systems. These interactions and interrelationships are complex and non-linear, and traditional modelling approaches cannot fully analyse the behaviours and failure modes of such systems.
In this report, we provide a review of key traditional accident modelling approaches
and their limitations in capturing accident causality and dynamics of modern complex
systems. We discuss new approaches to safety and accident modelling of sociotechnical
systems that are based on systems theory and cognitive systems engineering. Systems
theory includes the principles, models, and laws necessary to understand complex
interrelationships and interdependencies between components (technical, human,
organisational and management) of a complex system. Cognitive systems engineering
(Hollnagel & Woods, 1983) provides a framework to model the behaviour of joint
human-machine systems in the context of the environment in which work takes place.
We also review the current research in formal (mathematically-based) methods for the
modelling of complex system accidents. In addition, organisational sociologists have
made significant contributions to the understanding of accidents in complex
sociotechnical systems. Vaughn (1996) rejects the prevalent explanations (provided by traditional safety engineering techniques) of the cause of the Challenger shuttle accident and presents an alternative sociological explanation that explores much deeper cause of the failure.
The findings of this survey recommend new approaches to the modelling and analysis
of complex systems that are based on systems theory. The sociotechnical system must
be treated as an integrated whole, and the emphasis should be on the simultaneous
consideration of social and technical aspects of systems, including social structures and cultures, social interaction processes, and individual factors such as capability and motivation as well as engineering design and technical aspects of systems.
Interdisciplinary research is needed to capture the complexity of modern sociotechnical systems from a broad systemic view for understanding the multi-dimensional aspects of safety and modelling sociotechnical system accidents.|
|Appears in Collections:||DSTO Formal Reports|
Items in DSTO Publications Online are protected by copyright, with all rights reserved, unless otherwise indicated.