Analysis and Classification of Operators Demands for System improvements

To uncover human-system mismatches in system operation, a methodology is proposed for the analysis and classification of operators opinions about operation problems and their demands for system improvements. Previous methodologies had several limitations for dealing with human-related accident reports, that is, human error data. To be able to process operators opinions, a clustering method, “the modified affinity diagram,” is used to analyze opinions. Clustered opinions are classified into eight categories based on Rasmussens qualitative model of human behavior. The proposed methodology has several advantages because it uses the operators opinions as dataödata that can be collected at any time even in the absence of reported accidents. These data are intended to include information related to human-system mismatch situations. The methodology prevents unconscious or deliberate bias from influencing data collection. An application of the methodology to the analysis of network systems is presented. The proposed methodology is useful to clarify the subjective human-system mismatch areas in terms of human behavior in system operation.