Identifying the Random Field Propertiesof Wall Thinning Due to Flow-accelerated Corrosion

Abstract

Flow accelerated corrosion (FAC) is a degradation mechanism which causes wall thinning of carbon steel piping in power plants. Apart from the uncertainties involved in the factors leading to FAC, the inspection data of wall thinning measurements also involves uncertainties because of the lack of fixed spatial referencing. The concept of a single random variable cannot capture the real nature of this problem since the wall thinning is a spatially distributed mechanism and pointing out the exact location and measure of minimum wall thickness is not possible due to the uncertainties. Hence a random field approach is to be applied to effectively account for the uncertainties involved. This paper depicts the study of correlation pattern of wall thickness loss due to FAC and fits an appropriate autocorrelation function to model FAC wall thinning as a homogeneous random field.