Step Complexity Measure for Emergency Operating Procedures - Determining Weighting Factors

Abstract In complex systems, such as nuclear power plants (NPPs) or airplane control systems, human error has been regarded as the primary cause of many events. Therefore, to ensure system safety, extensive effort has been made to identify the significant factors that can cause human error. According to related studies, written manuals or operating procedures are revealed as one of the important factors, and the understandability is pointed out as one of the major reasons for procedure-related human errors. Many qualitative checklists have been suggested to evaluate emergency operating procedures (EOPs) of NPPs so as to minimize procedure-related human errors. However, since qualitative evaluations using checklists have some drawbacks, a quantitative measure that can quantify the complexity of EOPs is indispensable. From this necessity, Park et al. suggested the step complexity (SC) measure to quantify the complexity of procedural steps included in EOPs. To verify the appropriateness of the SC measure, averaged step performance time data obtained from emergency training records of the loss-of-coolant accident (LOCA) and the excess steam demand event were compared with estimated SC scores. However, although averaged step performance time data and estimated SC scores show meaningful correlation, some important issues such as determining proper weighting factors have to be clarified to ensure the appropriateness of the SC measure. These were not properly dealt with due to a lack of backup data. In this paper, to resolve one of the important issues, emergency training records are additionally collected and analyzed in order to determine proper weighting factors. The total number of collected records is 66, and the training scenarios cover five emergency conditions including the LOCA, the steam generator tube rupture, the loss of all feedwater, the loss of off-site power, and the station blackout. From these records, average step performance time data are retrieved, and new weighting factors are determined by using a nonlinear regression analysis. The results show that the SC scores quantified by the new weighting factors show statistically meaningful correlation with averaged step performance time data. Thus, it can be concluded that the SC measure can represent the complexity of procedural steps included in EOPs.