Flow accelerated corrosion (FAC) is a major degradation form of carbon steel and low alloy pipes in the secondary circuit of pressurized water reactor (PWR) plants, which has great impact on plant safety and reliability. For the purpose of effectively monitoring FAC in nuclear power plants, a statistical model for accessing FAC wall thinning rate using plant inspection data is proposed in this paper. The presented model is developed based on Gaussian stochastic process models. Wall thinning rate is considered as a function of key factors that have important influence on the FAC process (i.e., temperature, pH, mass transfer coefficient, etc.). The Kriging method, which has been widely applied in the domain of spatial analysis, is used to model the relationship between wall thinning rate and its impact factors. Model parameters are determined through maximum likelihood estimation using the inspection data. Since the likelihood function of the Kriging model is usually complicated in form, the genetic algorithm is employed to find parameter values that maximize this function. From the presented model, residual lifetime distributions of pipes affected by FAC can be derived, and conditions that may lead to high FAC rate can be found, which provides decision-making support for maintenance strategies optimization in life cycle management of the feed water system. Wall thinning data simulated from a physical-chemical mechanism model presented in literature are used to verify the presented model. Results of validation show that reasonable wall thinning rates and lifetime distributions can be obtained using this model.
C206 Prediction trend of existing evaluation method of wall thinning rate caused by flow accelerated corrosion in two phase flow
