Probabilistic Modeling of Pitting Corrosion in Steam Generators

Abstract

Pitting corrosion is a serious form of degradation in steam generator tubing of the CANDU reactors. The initiation and propagation of the pitting process is fairly random, and it requires inspection and cleaning to minimize the risk of tube leakage. The paper presents an in-depth statistical analysis of pitting data collected over the years through the inspection of several steam generators. The paper describes the probability distributions of the number and size of pits generated over the service life and correlates them with the history of outages and chemical cleaning and water lancing campaigns. Based on statistical analysis, the paper develops a risk-based model for the life-cycle management of stem generators. This model is applied to evaluate the effectiveness of the cleaning and forecast the risk and cost over the remaining service life of a generator.