Development of an EEG-based human operator’s mental workload evaluation method in NPPs

Abstract

The environment of the main control room has been changed from the conventional analog type to a digital paradigm. With a digitalized advanced main control room, human op-erators conduct highly cognitive work rather than physical work compared to the case of the original control rooms in nuclear power plants, and various digital operating support systems (OSSs) have been developed to reduce the operators’ mental workload and stress. The operat-ing support systems should, however, be evaluated to determine their effects on human per-formance measures such as task accuracy and mental workload. In general, questionnaire sur-veys such as NASA Task Load Index (NASA-TLX) are used as measures of human operators’ mental workload. However, there are some limitations including the possibility of skewed re-sults due to self-assessment of the mental workload and the impossibility of continuous measuring mental workload. In order to solve these problems, physiological techniques have been used to evaluate cognitive workload. However, there are only a few tools for a relatively simple comparison of changes of bio-signals. Accurate measures that can indicate human operators’ mental work-load therefore should be developed. In this study, electroencephalogram (EEG) was used to measure operators’ mental workload because it has been proven to be sensitive to variations of mental workload in other studies, and it allows various types of analysis. Based on various research reviews, the EEG-based Workload Index (EWI) was suggested and the suggested in-dex was validated through a comparison with NASA-TLX and the number of human errors as human performance. As a result, EWI was valid for measuring operators’ mental workload. The effects of OSSs on human performance were evaluated by using EWI. Experi-ments for investigating the effects of OSSs related to both execution and diagnosis on human operators were conducted. Before conducting the experiments, the test subjects were divided into two groups in terms of competence by using a cluster analysis and a t-test (non-skilled group, skilled group). From the results of the two kinds of experiments, there is a large improvement of hu-man performance in the non-skilled group compared to the skilled group after the application of both OSSs related to execution & control and OSSs related to monitoring & diagnosis. The non-skilled group has a higher mental workload for monitoring and diagnosis tasks than for execution and control tasks. Human performance including the mental workload of the non-skilled operators with OSSs was nearly the same as that of the skilled operators without OSSs. In the future, it is expected that the EWI values of the non-skilled group and the skilled group can be used for training operators and designing OSSs.