Reliability and Safety Assessment of a Passive Containment Cooling System in Advanced Heavy Water Reactors

Passive Safety Systems (PSSs), which rely on natural forces and processes, such as natural circulation, gravity, internal stored energy, etc., are increasingly utilized in generation 3+ and generation 4 advanced nuclear power plants to increase inherent safety features of the nuclear reactor design. Although PSSs should considerably increase the safety of nuclear power plants, it is still challenging to systematically assess the reliability of passive systems because of the lack of data and uncertainties associated with phenomenon involving natural forces that underlies their safety functions. In this study, the Fault Tree Analysis (FTA) was used to assess the reliability and safety of the Passive Containment Cooling System (PCCS) in Advanced Heavy Water Reactor (AHWR). The failure probability of PCCS was calculated from the failure probabilities of Basic Events (BEs). Using the data for the failure probabilities of Top Event (TE) and BE from the FTA model, two Artificial Neural Network (ANN) models were proposed for the reliability analysis of PCCS to supplement the FTA model. Rectified Linear Unit (ReLU) and Sigmoid activation functions were utilized to build ANN models, and an Adaptive moment estimation (Adam) optimizer was used to train the ANN models to make these models computationally efficient. The results of the FTA model were compared with the predictions of the ANN models to find out the ANN model performance.

Keywords: Reliability, Passive safety systems, Passive containment cooling system, Fault tree analysis, Artificial neural networks.

Download PDF