Abstract

The purpose of this paper is to develop, test, and illustrate a simple spreadsheet-based probabilistic procedure that can be used by practicing engineers to determine the remaining useful life (RUL) of pipelines. The suggested reliability assessment together with a suitable linear, or nonlinear, time dependent wall thinning relationship can be used to determine the failure probability as a function of years in service. DNV F101, FITNET FFS and ASME B31G, are used to demonstrate the method, which also permit to compare and contrast results. Accuracy of the proposed approach is proved by replicating results from published literature.

Introduction

Uncertainties related to corroded pipelines are related to imperfect measurement of metal loss, randomness of the pipeline data, and variation of operational data. DNV RP-F101 has incorporated safety factors to account for these uncertainties. However, the way this code is applied is as a deterministic approach similar to other failure pressure equations such as ASME B31G and the recent addition i.e. FITNESS FFS. The reliability of a pipeline is also dependent on the inspection tool accuracy, which is manifested by the dispersion of the corrosion growth rate geometry of the damaged areas. Equations in these codes are based on large databases and is generally on the safe side by an unknown margin. In order to predict the remaining useful life (RUL), the level of uncertainties, which could escalate with time, must be accounted for (Alamilla and Sosa, 2008; Caleyo et al., 2009).

The objective of this paper is to develop a spreadsheet based first order reliability method; test it against published results and demonstrate its application. We use the published result of Qian et al. (2011) and Caleyo et al (2002) as base lines for the validation of the proposed approach. Between them they cover the more popular failure prediction equations. Codified deterministic methods use unfavourable data; e.g. maximum corrosion depth, maximum corrosion rate, maximum design pressure and minimum wall thickness without allowing for uncertainties (Amirat et al, 2006). Thus, such results can be somewhat conservative in terms of probability of failure for pipelines containing extensive corrosion defects. For example, a maximum corrosion growth rate is assumed for the entire pipeline, ignoring the possibility that not all defects will grow equally. The maximum probable rate is used for the sake of simplicity, as well as due to lack of data. An assessment method is required to determine the severity of such defects when they are detected on pipelines.

This content is only available via PDF.
You can access this article if you purchase or spend a download.