Work Domain Analysis of a Pasteurization Plant: Using Abstraction Hierarchies to Analyze Sensor Needs

This paper discusses the use of Rasmussen's abstraction hierarchy (AH) in performing an analysis of the work domain of a Pasteurization plant. Our goal is to examine the strengths and weakness of ecological interface design (EID) for systems in which critical variables are unreliable, faulty, or not measurable. In this paper we report our use of AHs to analyze the functioning of a pasteurization plant and the impact of unreliable or faulty sensors on the intelligibility of information available to the human operator. Although there is considerable current interest in EID, building thorough AHs for complex systems is a large task and there are currently very few detailed published examples to help human factors professionals wishing to take this approach. In this paper we present details of how we built the Pasteurization plant AH and show we are using the AH in our research. We argue that because AHs can indicate information that should be displayed for a process to be intelligible, techniques like EID that use AHs are in a position to bring about a profoundly user-centered approach to system design—an approach in which the ultimate information needs of human controllers will drive the engineering agenda of sensor and instrumentation design in a feedforward manner.