Constructing Neural Network Model to Evaluate and Predict Human Error Probability in Nuclear Power Plants Based on Eye Response, Workload Rating, and Situation Awareness

Abstract

The accurate assessment of human error probability (HEP) has an important impact on the safety of nuclear power plants. Therefore, it is necessary to develop a HEP model. This study analyzes the validity, sensitivity, and relationship between HEP and the indices of eye response and the subjective rating method. The analysis result showed that there is a correlation between HEP and the indices of eye response, subjective workload, and situation awareness level. Therefore, a back propagation neural network model was developed based on these indices. The correlation coefficient is more than 0.95 between the predicted data of the developed model and the target data. Also, the root mean square error was 0.0073, 0.0083, and 0.0077, and the determination coefficient was 0.965, 0.933, and 0.931 for the training, validation, and testing data sets, respectively. Therefore, the developed back propagation neural network model has reliable prediction accuracy for HEP.

Keywords:

• Human error probability
• nuclear power plant
• eye response
• subjective rating method
• neural network model

Acknowledgments

The authors thank the reviewers for their valuable comments. Also, the authors are grateful to the participants who helped with the research.

Disclosure Statement

No potential conflict of interest was reported by the authors.