Diagnosis is one of the most complex and mental resource-demanding tasks in nuclear power plants (NPPs), especially, to main control room (MCR) operators. Diagnosis is a crucial part of disturbance control in NPPs, since it is a prerequisite task for initiating operating procedures.
In order to design a control room feature for NPPs, three elements need to be considered: 1) the operational tasks that must be performed, 2) a model of human performance for these tasks, and 3) a model of how control room features are intended to support performance. The operational tasks define the classes of performance that must be considered. A model of human performance makes more explicit the requirements for accurate and efficient performance and reveals potential sources of error. Finally, the model of support allows the generation of specific hypotheses about how performance is facilitated in the control room. The model of support needs to be developed based on the human performance model.
This paper proposes three approaches for the system design of operator support systems to aid MCR operators’ diagnosis tasks in NPPs, considering the above three elements. This paper presents 1) a quantitative approach to modeling the information flow of diagnosis tasks, 2) strategy-based evaluation of information aids for diagnosis tasks, and 3) quantitative evaluation of NPP decision support systems.
As an analysis of diagnosis tasks, this paper presents a method to quantify the cognitive information flow of diagnosis tasks, integrating a stage model (a qualitative approach) with information theory (a quantitative approach). The method includes: 1) constructing the information flow model, which consists of four stages based on operating procedures of NPPs; and 2) quantifying the information flow using Conant’s model, a kind of information theory. Then, three experiments were conducted to evaluate the effectiveness of the proposed approach to predicting human performances, especially i...