This Master Thesis is framed within a collaboration agreement between the Nuclear Engineering Research Group (NERG) of Universitat Politècnica de Catalunya (UPC) and a Spanish Pressurized Water Reactor (PWR) Nuclear Power Plant (NPP). The main objective of the collaboration is to study and apply the Probabilistic Safety Assessment (PSA) methodology to risk-informed decision making.
This study is part of the NPP Independent Spent Fuel Storage Installation (ISFSI) PSA requested by the Spanish NPP. It is the continuation of the final degree Project “Estudio piloto para el análisis del riesgo asociado a un Almacén Temporal Individualizado. Aplicación de la metodología APS” [1], which developed a pilot ISFSI’s PSA model without Human Reliability Analysis (HRA). The objective of this thesis is to apply a HRA to the Spanish NPP Independent Spent Fuel Storage Installation and then implement the HRA results into the ISFSI’s PSA model in order to evaluate the impact of human performance on the ISFSI’s Risk. In consequence, the project is divided in two different parts, the HRA development and the HRA implementation into the PSA model. The first part is more research related. On the other hand, the second part is more engineering related.
The ISFSI’s HRA application is based on the methodology described in the regulatory guide NUREG-1880, “ATHEANA User’s Guide Final Report”, as recommended by several Nuclear Regulatory Commission (NRC) publications [2][3]. This methodology requires the participation of ISFSI’s Subject-Matter Experts (SMEs). SMEs are not available since the Spanish NPP has little experience in ISFSI operations. Therefore, it has been decided to develop an HRA methodology which does not need SMEs to be carried out. To do so, the contribution of the SME’s has been replaced with the use of other HRA methodologies, namely THERP and SPAR-H. In consequence, an experimental hybrid ATHEANA-based HRA methodology has been used to perform the analysis. The HRA results should be considered illustrative rather than definitive since several assumptions have been taken to apply the methodology and describe human actions. Furthermore, the HRA results cannot be compared with Nuclear Industry data since no ATHEANA-based ISFSI HRA has been published yet.
The most important operations from a Risk point of view [1] are the human actions performed inside the Spent Fuel Storage Building (SFSB). Therefore, the HRA has been limited to the analysis of these human actions. Some examples are loading the canister with wrong Spent Fuel Elements (SFEs), also known as Misload, and different canister drop scenarios.
The introduction of the HRA results as Human Error Probabilities into the PSA model implies the modification of the Initiating Events (IEs) treatment. Fault Tree (FT) models have been
Pág. 2 Report
developed to include both Human Failure Events (HFEs) and Crane components failure into the assessment of IEs frequency. These models have replaced all the Initiating Events previously included in the PSA model.
The comparison of the PSA model results, with and without HRA, in terms of Risk yields that human performance has a substantial impact on the ISFSI’s Risk. Namely, the implementation of the HFEs increases the overall ISFSI’s Risk roughly 2 orders of magnitude. However, the results obtained in this thesis are illustrative since there are many uncertainties and assumptions in the HRA analysis. It cannot be directly concluded that human performance will actually have an important impact in the Risk. However, it is recommended to take into consideration that human performance could be of importance and that further studies should be carried out. |