Statistical analysis methods for corrosion mapping inspection data

Abstract

Corrosion mapping, in which wall thicknesses over large areas are measured by 0 degree compression probe ultrasonics at closely-spaced points, is being used increasingly in oil and gas applications. This paper describes a new approach to statistical analysis based on wall thickness distributions derived from corrosion mapping data. The analysis methods developed are described and cumulative thickness distributions obtained in a wide range of field applications are presented. A range of situations in which corrosion of carbon steel is active are covered and the results include analysis of data from pressure equipment with CO 2 corrosion, O 2 corrosion, under deposit corrosion, naphthenic acid corrosion and corrosion under insulation. The results show that there are many situations in which the wall thickness distributions display strongly ordered behaviour. In many cases it is observed that the wall loss can be represented by an exponential distribution. Examples of wall loss distributions other than exponential are also provided. It is shown that the distributions established can be a useful basis for estimates for the uninspected areas when less than 100% coverage has been achieved. A summary covering applications of such analyses to integrity management practice is provided. This highlights the benefits of the use of underlying thickness distributions.