Singh, Simarpreet2020-05-142020-05-142020-05-142020-04-21http://hdl.handle.net/10012/15857The objective of this this thesis is to develop a probabilistic model for assessing the life cycle performance of safety valves used in the nuclear piping system. The life cycle performance is quantified in terms of reliability and life cycle cost in a given operating interval of the plant. A key input to the probabilistic lifecycle analysis is the lifetime distribution of the component in question. The second important element is the estimation of costs of inspection and in-service testing of components as well as costs of repairs and replacement of failed components. Based on this information, the life cycle analysis aims to predict the reliability and expected cost of operating a component in a future time interval. This study illustrates how to develop methods and algorithms for probabilistic assessment of the life cycle of safety valves used in the nuclear piping system. For statistically estimated parameters of the probability distributions of lifetime and various costs, historical operating data are required. This study uses about 20 years of historical data obtained from a group of temperature control valves used in the moderator system of a reactor. A maximum likelihood-based method is developed to estimate parameters of the lifetime distribution of a valve. The lifetime is defined as the time of first leakage in the valve since the time of installation. The ML method is based to consider complete and censored lifetime information. The distribution of repairs cost is also estimated by the ML method. The proposed method is applied to predict reliability and life cycle cost for various operating interval. This model can also be used to optimize the overhaul interval of the valve.endata analysisnhpppower lawrepairable systemvalvemleweibullprobabilitystatisticslife cycle costpreventive maintenanceoverhaulMaster Thesis