Dr. Shane Mueller of Applied Research Associates conducted a pilot project to determine if operator mental models could be made objective. Two incidents were evaluated relative to operator troubleshooting: (1) indicated loss of steam flow to reboiler and (2) high tower bottoms pressure alarm. Causal chains were developed for each and annotated as to how they are used in the troubleshooting process. The results showed that operators look for or anticipate impact on system, then try to identify ways to verify that impact. This is not just learning associations between pairs of inputs and outputs, as causal chain reasoning is several layers deep, with experts better able to utilize deep layers.
Determine if the mental models of operators can be objectified for use in training and display development.
Is it possible to objectify mental models of operators for comparisons with each other and ideal?
It is not uncommon to describe operators in terms of whether they have a good or bad mental model of the process. However, what exactly that model is and how it differs from the standard we used to judge goodness is rarely defined. If such mental models could be objectified, then gaps in the model could be addressed via training and/or the models of experts transferred to the novices. The Department of Defense has been attempting to quantify mental models to help predict the actions/response of target groups.
Objective mental model of process operator along with limitations and opportunities.