Statistical analysis for reliability assessment of corroded structures: A pipeline case study
Abstract
Corrosion defects are a major concern for pipeline operators, and in-line inspections (ILI) are essential to detect and identify such defects and to optimize maintenance actions to avoid costly losses. However, it is crucial to conduct statistical analysis to estimate and monitor the corrosion evolution over time and enhance the accuracy of pipeline reliability assessments. This paper presents a statistical study of corrosion defect parameters (length and depth) obtained from ILI. The study evaluates statistical moments, considers the Pearson correlation coefficient, and determines the probability of failure using the Monte Carlo method based on the burst criterion. Moreover, we examine the influence of the coefficient of variation of corrosion defects on the probability of bursting by analyzing the sensitivity of key design parameters, such as operating pressure and the depth-to-wall thickness ratio. The paper employs a real case study of a corroded gas pipeline in Algeria to illustrate these findings. The results of this study can provide crucial information to pipeline operators to make informed decisions regarding maintenance activities throughout the pipeline's operational life.