A comparative reliability study of corroded pipelines based on Monte Carlo Simulation and Latin Hypercube Sampling methods. (March 2020)

• Record Type: Journal Article
• Authors: Abyani, Mohsen; Bahaari, Mohammad Reza
• Is Part Of: International journal of pressure vessels and piping. Volume 181(2020)
• Publication Date: 2020-03
• Subjects: Reliability -- Pipeline -- Corrosion -- Monte Carlo -- Latin Hypercube Sampling -- Positive definite; Pressure vessels -- Periodicals; Pipe -- Periodicals; Récipients sous pression -- Périodiques; Tuyaux -- Périodiques; Pipe; Pressure vessels; Periodicals; 681.76041
• DOI: 10.1016/j.ijpvp.2020.104079
• Languages: English
• ISSNs: 0308-0161

Abstract

Abstract: Internal corrosion is categorized as one of the most destructive phenomena for pipeline services. This study compares the reliability analysis of pipelines against internal corrosion based on Monte Carlo Simulation (MCS) and Latin Hypercube Sampling (LHS) methods. Three different failure mechanisms (pitting perforation, local burst and rupture) have been considered for single corrosion defect and only local burst has been investigated for multiple corrosion defects. Furthermore, the effects of correlation between the defect depth and defect length for single defects as well as the correlation between the neighboring segments for multiple defects on the failure probability of the pipelines, have been comprehensively studied. The results show that LHS method could lead to quite acceptable accuracy with considerably less computational effort. Regarding multiple corrosion defects, the correlation between the defect depth or defect length of the neighboring segments can change the system failure probability without any monotonic trend. Highlights: LHS provides acceptable results in reliability analyses of corroded pipelines. Results of LHS are acceptable if the number of simulations is greater than 1000. ISM is use to modify non-positive definite target correlation matrices. The correlation between the defect depth and length alters the system reliability.