Evaluating the Characteristics of Marine Pipelines Inspection Data Using Probabilistic Approach

This paper presents a probabilistic approach on analyzing internal corrosion defects formed in a Malaysian gas pipeline using an intelligent pigging (IP) device. The significant parameters of IP data and their characteristics in the pipeline are analysed through correlation and multivariate regression methods. Some insensitive parameters like the defect circumferential width which is normally ignored in the current design codes is given high priority in this analysis. These governing parameters are later evaluated to produce a new dimensionless limit state function that can best describe the characteristics of the de-fects. Results from the probabilistic analysis are compared to the results from the existing pipeline failure assessment codes.