CNS 시뮬레이터를 이용한 운전원과 학생 간의 인적수행도 비교

Abstract

Nuclear power plants (NPPs) are operated or managed by operators capable of taking appropriate measures in the event of an accident. However, any human action carries the potential for error, which is why human reliability analysis (HRA) researchers systematically analyze and predict the potential risks due to human factors in order to eliminate or reduce such errors. In this area, much research is still needed to improve the safety of NPPs. Typically, human error probabilities (HEPs) are estimated by collecting HRA data inferred from other industries, expert judgment, or training simulator research. However, the lack of adequate data has been highlighted as a major challenge in the field of HRA. Currently, to accumulate HEP data, the U.S. Nuclear Regulatory Commission and the Korea Atomic Energy Research Institute focus on collecting data via full-scope simulators. These studies are conducted using the Scenario Authoring, Characterization, and Debriefing Application (SACADA) database and the Human Reliability Data Extraction (HuREX) framework, respectively, to collect data from full-scope digital main control room (MCR) simulations of actual NPP environments. HRA data collection using a full-scope simulator is beneficial in terms of more accurate results, but there are problems with using the data collected from such simulators to support the data required for the HRA quantification process. Full-scope simulator studies conducted to collect HRA data are expensive, due to the need for a full-scope simulator facility and numerous operators, experts, and researchers. For this reason, researchers at Idaho National Laboratory (INL) developed the Simplified Human Error Experimental Program (SHEEP) as a method of overcoming the drawbacks of full-scope research. The purpose of the SHEEP study is to infer the HEPs from full-scope simulator data by comparing student and operator performance using a simplified simulator. Similar to a previous study conducted in the SHEEP project, the current study aims to perform experiments conducted on operators and students to identify differences in human performance data using the Compact Nuclear Simulator (CNS). To achieve this goal, the study analyzes correlations between the measured human performance of students and operators, via a randomized factorial experiment design. Investigate five human performances: 1) workload, 2) situation awareness, 3) error, 4) time and 5) eye movement measurements. In addition, several scenarios and related procedures have been developed for CNS experiments. The data collected in the experiment were compared with human performance using several statistical analysis techniques such as the ANOVA test and Pearson correlation analysis.