Evaluation of human error probabilities for post-initiating events

Abstract

The United States Nuclear Regulatory Commission is responsible for the safe operation of the United States nuclear power plant fleet, and human reliability analysis forms an important portion of the probabilistic risk assessment that demonstrates the safety of sites. Treatment of post-initiating event human error probabilities by three human reliability analysis methods are compared to determine the strengths and weaknesses of the methodologies and to identify how they may be best used. A Technique for Human Event Analysis (ATHEANA) has a unique approach because it searches and screens for deviation scenarios in addition to the nominal failure cases that most methodologies concentrate on. The quantification method of ATHEANA also differs from most methods because the quantification is dependent on expert elicitation to produce data instead of relying on a database or set of nominal values. The Standardized Plant Analysis Risk Human Reliability Analysis (SPAR-H) method uses eight performance shaping factors to modify nominal values in order to represent the quantification of the specifics of a situation. The Electric Power Research Institute Human Reliability Analysis Calculator is a software package that uses a combination of five methods to calculate human error probabilities. Each model is explained before comparing aspects such as the scope, treatment of time available, performance shaping factors, recovery and documentation. Recommendations for future work include creating a database of values based on the nuclear data and emphasizing the documentation of human reliability analysis methods in the future to improve traceability of the process.