Corrosion Prediction Model Of Corroded Pipeline Using Gumbel

Corrosion is an important degradation mechanism that can affect the reliability and integrity of
the pipeline. Offshore pipelines are usually inspected using MFL Intelligent Pigging method; this
is how internal pipeline corrosion can be definitively measured. However, a huge amount of
thickness profile data was not used optimally to predict the corrosion rate. A reliable corrosion
rate model is paramount to determine the re-inspection time interval and corrosion mitigation for
pipelines. The objective of this final year project is to predict and analyze the internal pipeline
corrosion for the chosen case study and develop a corrosion model. The methodology used in
this project includes data gathering, data review, classification into defect type, data analysis,
corrosion modelling, validation and discussion. The IP data was modelled with Gumbel
distribution and result show that the data fits the curve and predicted the time to failure was for
another 60 years. The result from Gumbel was compared to the deterministic approach of
average time to failure of 149 years. The percent error was 40%. The project met the objective
and can be further developed.