Fighting Corrosion on an Ageing Pipeline

Proceedings Publication Date:

01 Sep 2016
Markus Seitz
Markus Seitz
Part of the proceedings of
Case study on a DN350 onshore natural gas pipeline, constructed in Western Australia in 1971. The purpose of the integrity management program is to maintain containment and maintain supply. Integrity Management on this pipeline consists of a continuous cycle of surveying, verifying, repairing, forecasting, reviewing and adjusting. Data is gathered by in-line inspection (ILI) and direct assessment. The forecast is a combination of the current known state of the pipeline and a projection into the medium term (up to 5 years). The forecast requires the calculation of a growth rate that can be applied to known corrosion features. The corrosion dataset numbers in excess of 100,000 features. Analysis of the data therefore requires a statistical approach that recognises differences in severity and growth rate between pipeline segments and within pipeline segments. Corrosion Growth rates are determined by comparing in-line inspection data. When ILI datasets with relatively broad error distributions are compared, a Corrosion Growth Rate distribution is produced for each pipeline segment. Summary statistics can be extracted from this distribution and used in forecasts. The wide uncertainties in the available data present a challenge to the integrity engineer: The use of an average rate is clearly not conservative. Use of statistics on the extreme of the distribution can mislead the operator in his perception of risk and lead to unsound decisions by the business. It is therefore important to review the forecast regularly and adjust it as appropriate. When pressure limits are calculated, safety factors can be adjusted to compensate for under or over-conservatism in the forecast. The forecast can then be used to plan repairs, adjust pressure limits or optimise the date for the next ILI.

To view the video or download the paper please register here for free

You already have access? Sign in now.