Comparative Numerical Evaluation of Angra I Auxiliary Feedwater System Reliability by the Method of Suplementary Variables

This paper proposes a methodology for predicting failure probabilities using the method of supplementary variables for nuclear safety systems subject to aging. Particularly, the failure probability of the auxiliary feedwater (AFSW) system of a typical PWR plant was calculated. In this context, it has been necessary to study systems of partial and ordinary first order differential equations in order to represent a system subject to aging. Numerical solutions were obtained by the method of finite differences using Euler’s implicit to solve the equations related to the concepts “as good as new” and using numerical integration for “as bad as old” concept. These numerical solutions were used to model extension of useful life in a nuclear power plant. For testing the proposed model constant failure rate was used to evaluate the failure probability of the auxiliary feedwater (AFSW). This paper proposes a new method which named is Θ-method the results obtained with “as good as new” and “as bad as old” models for 240, 400, 720, 1000, 2000, 3000 and 4000 hours of operations, with rate constant failure. With these data, the failure probability for 5000 hours by was estimated the Θ-method. This numerical solution is very helpful for considering the life extension of nuclear plants, an issue currently under discussion.