Introduction
Accident models are tools to evaluate the sequential events leading to an accident, provide impetus to prevent future mishaps and make assessment on the systems’ risks. With the advent of technology, the etiology of the accidents also changed from simple mishaps to catastrophic disasters. As a result the models also progressively advanced, from viewing an accident from a sequence of isolated event(s), to making an assessment of interactive physical, human and social components, which were based on socio-technological evaluation of constrains rather than the events per se. The objectives for modeling also changed, from simple assessment as to whether to blame a person whose failure has resulted in a mishap in that event, for e.g., a driver overlooks a signal and dashes with a tree, to making an holistic understanding of the constrain(s) that led to a system failure, so that this can be averted in future.
Main Text
According to Helander (2006), in the Energy Exchange Model, the unexpected physical or chemical damages were correlated with energy transfer. This is believed to be based on Newton’s laws of motion, in which a sudden halt of a vehicle will give enough momentum to the driver that he can break the windscreen and injure himself. This model does not justifies the cause of an accident, rather points out towards the preventive measures to cope with the energy transfer, say by way of distributing it to a larger area to reduce the impact. In the Chain of Events model, multiple physical (equipment), environmental and task factors were attributed to cause distraction to the operator, who did not respond in time and met with an accident. Human errors like inability to visually discriminate, memory failures and communication breakdown etc. were attributed to vehicular and aviation related accidents. Reason’s model of human errors (as cited in Helander, 2006), is based on knowledge- and rule-based mistakes. Inadequate knowledge prevents the operator to interpret the information and he makes mistake of an existing rule, and confidently cause an error. Lapses, slips and mode errors (mostly memory failures) are the other causes. According to Ramsey (as cited in Helander, 2006), a crucial component was the human (victim’s) response to the hazardous situations, such as, assessment of gravity of the problem, avoidance and correction measures. Human behavioral factors like reflexes, alertness, prior memory of another event, experience, risk-taking tendencies, swift and accurate organ motion etc. would determine whether accident is to happen or not. Although human error is a major cause of most vehicular/aviation accidents but the person need not be an operator, but may be those running the system, like signaling, or road safety, vehicle design etc. For e.g. if an airliner crash lands in a bad weather just because the Air traffic controller cleared landing, then why pilot is to be blamed? Organizations designing equipment or manage operations also need to be blamed for a failure. Manufacturers violate safety standards, do not train workers to handle equipment, or do not prevent risks are equally liable.
Further advancement in the models, especially for industrial sectors, was described by Leveson (2004). The Interactive Model of Accidents relies on interactivity of several factors, and none is to be blamed in isolation. In such models, where social and developmental (including systems and software) are included, besides the prevailing event trees (forward sequence of events), fault trees (backward sequence of feedback) are used. In Event models there was no defined “starting” cause of the chain. Many factors could be the starting cause, for e.g. in Bhopal disaster of Union Carbide, management issues like cost cutting to refrigeration, reducing the workforce, poor equipment maintenance and sociopolitical reasons, overpopulated slums around the factory were the interactive causes resulting in the worst industrial accident (Borgard, 1989). A slip disc was mistakenly left in washing pipe in the fateful night which triggered the event. This could have happened in any other time, or by any other operator. Operator, managers, engineers, regulating and governing agencies automatically fall in a hierarchical fashion in an improvised model, socio-technical model of Rasmussen and Svedung (as cited in Leveson, 2004). Accordingly, systematic migration of organizational failures results into accidents, and chain of events is just one of the many causes to trigger that event. In this model, fault trees and event trees were combined and there was a feedback loop at every stage following an accident event or upon forecasting constrains. In the most recent Systems-Theoretic Accident model system, development and system operations are two streams under socio-technical hierarchies. The final operating process interacts and gives feedback to both the control structures. Though the hierarchical control systems work independently, they also interact at different channels. Legislatures, law makers, insurance agencies, enforcing agencies, company’s operations and design management, and operating process running the process, are regulated in a systematic hierarchical manner. The human error is looked from a point of view of human to system or software interaction. In Bhopal disaster, such models would have been helpful. The gas leakage was periodically experience by the workers but despite complains, never seriously taken up by the agencies sitting in upper hierarchies. As no feedback chain was established, the process design and maintenance went faulty and only one mistake led to this havoc. With automation, new kinds of operational problems are expected and accordingly the accident models should be upgraded.
References
Borgard, W.(1989). The Bhopal Tragedy. Boulder, Colo: Westview Press.
Helander, M.(2006). A Guide to Human Factors and Ergonomics (2nd Ed.). Boca Raton, London, New York: Taylor & Francis Group.
Leveson, N.(2004). A new accident model for engineering safer systems. Safety Science, 42(4), 237-270.