Prediction of corrosions in Gas and Oil pipelines based on the theory of records

TL;DR

This paper introduces a statistical approach using the theory of records and non-homogeneous Poisson processes to predict corrosion locations in pipelines, aiding preventive maintenance and performance comparison.

Contribution

It applies the theory of records and Poisson processes to model pipeline corrosion, providing a novel quantitative method for corrosion prediction and performance evaluation.

Findings

Effective prediction of corrosion locations in pipelines.

Demonstrated application on real pipeline data.

Provides a basis for preventive maintenance strategies.

Abstract

Predictions of corrosions in pipelines are valuable. Based on the available data sets, it is critical and useful, for example in preventive maintenance. This paper deals with this problem by two powerful statistical tools, i.e. theory of records and non-homogeneous Poisson process, for modeling location of corrosions. These methods may be used for comparing performances of different pipelines via a quantitative approach. For illustration purposes, we applied the obtained results for a real pipeline in order to prediction the next corrosions based on the available data. Finally, some concluding results and further remarks are given.