Field Test of ATHEANA (A Technique for Human Event Analysis) in Pressurised Thermal Shock Probabilistic Risk Assessments

The U.S. Nuclear Regulatory Commission (USNRC) is currently developing the technical basis for potentially modifying the Pressurised Thermal Shock (PTS) screening criteria specified in Title 10 Part 50.61 of the U.S. Code of Federal Regulations. As part of this effort, PTS probabilistic risk assessments (PRAs) are being performed by the USNRC and its contractors, for two U.S. nuclear power plants, with assistance by the respective utilities. Two other plant analyses will also be performed as part of the PTS work, but are not the subject of this paper. These latter two analyses are being performed by industry and will be reviewed by the USNRC for inclusion into the entire effort. The ATHEANA methodology is being applied to address the human reliability portion of the two USNRC analyses in a way that is also significantly interactive with the accident sequence and system modelling portions of the work. This paper discusses the results to date of the application of ATHEANA in these PTS analyses, including consideration of possible relevant errors of commission. The field test of ATHEANA on this project has served two purposes. One is to use it in developing technical analyses to address and solve a particular issue; i.e., the risk of PTS. The results of these analyses will be used by the USNRC to decide whether or not the PTS screening criteria should be modified. The second purpose is to test the use of ATHEANA in a real application to determine strengths and weaknesses of the method, and how it might be improved. With regard to the first purpose, the PTS project is one where a current screening criteria may be relaxed. Before that can be done, the NRC must be sure that it is safe to do so, and to the extent the specific risk values affect that decision, those values must be able to withstand public scrutiny. Since the ATHEANA process requires the analysts to identify potential vulnerabilities and challenging scenario deviations associated with a human action, and to uncover any particularly error-forcing or challenging contexts under which a human action may need to be performed, the ability to understand and defend the human action probabilities is enhanced. In this work, ATHEANA has addressed many actions, whether acts of commission or omission, potentially detrimental from the PTS point of view and has provided careful descriptions of the context under which the events might occur. Because of the emphasis on understanding context as an integral part of the ATHEANA process, interfaces with the PRA modelling and thermal-hydraulic aspects of the overall analyses have been stronger than has been typically performed in other PRAs. This benefits the PTS study since these aspects of the analyses (i.e., HRA, PRA, thermal-hydraulic) are forced to be more integrated and cohesive. With regard to the second purpose, as a test of ATHEANA, the following observations are made: - We have determined that the search process and the resulting quantification of human actions can be made more efficient by being performed iteratively and in a more integrated manner. - To provide a formal structure for the quantification process, we are working on a framework to accomplish this goal. In the improved process, the specific elements of the context most affecting the probabilities will be made even more explicit and guidance will be provided for how to include these specific elements in determining the probability estimates. - The 'heart' of ATHEANA involves understanding the plant conditions and human performance factors that seemingly affect a specific human action. Therefore, application of ATHEANA may be more resource-intensive than first generation HRA methods. Use of the full power of the method is especially important for those problems where many factors or their complex integration are likely to be important; particularly when these factors include the plant physical response (such as thermodynamic behaviour) as well as the more typical HRA issues such as operator training, the time available, and human-system interface issues. However, application of ATHEANA for the PTS work has been less demanding than the analysts initially expected. Furthermore, as with any newly developed method (e.g., compare the effort to do the first PRA versus what it takes to do one today), it is believed that repeated applications, such as those performed for the PTS analyses, will allow identification of additional ways to make it work more efficiently. (authors)