Reducing Accident Causation in Complex Plants by Identifying Mutual Misconceptions Between Designers and Operators

Abstract

Catastrophic accidents in complex plants arise from an unforeseen combination of a number of factors. Although much effort has been invested in both improving the reliability of components, and the design of user interfaces, human error and complex plant failures still occur. One contributing factor to plant accidents is the nature of understanding that design engineers and operators have of each other, or rather the mutual misconceptions that arise between them. For example, operators may adopt practices that do not reflect the demands and limitations inherent within the design of a plant. Similarly, the design engineer may prescribe practices that cannot be successfully completed due to limitations inherent within operators. This paper describes the development of a database that attempts to capture these mutual misconceptions. The database has been produced from causal analyses of case studies of previous accidents involving complex plants. In addition, the database forms the basis of the development of an agenda-generating mechanism for use by designers and other decision makers. The tool provides cues to key decision points and managerial activities that influence the design and operation of a plant. It lets the decision makers choose the level of abstraction at which they are cued by the agendas of misconception type. For example, in writing shut down instructions, it is important that the authors can see all the main types of operator misconception that are associated with shutdown activity, and that they can navigate to more detailed sub-types, or even specific accident accounts. The tool is designed to help decision makers avoid the types of mutual misconceptions that have been implicated in previous plant accidents, and hopefully increase their understanding of the demands they place upon operators.