Abstract:
Hydrocarbon release due to Vibration Induced Fatigue (VIF) costs millions of dollars each year in inspection, maintenance and replacement activities, as well as lost prod...Show MoreMetadata
Abstract:
Hydrocarbon release due to Vibration Induced Fatigue (VIF) costs millions of dollars each year in inspection, maintenance and replacement activities, as well as lost production for oil and gas operators. Therefore, it is vital to assess the reliability of topside piping against VIF. In this context, the authors have utilized the fundamentals of probability, reliability and statistical methods, using Dynamic Bayesian Networks (DBNs) to estimate the reliability of process piping. Firstly, different sources of uncertainty, such as physical variability, statistical uncertainty, etc., in the crack growth process (Paris law is used to model the crack growth) are identified and quantified, with suitable distributions and parameters obtained from literature. Thereafter, a DBN is developed to obtain the distribution of the Remaining Fatigue Life (RFL). The results (in terms of crack size) are validated against experimental data. Thereafter, statistical methods are used to obtain the reliability/PoF curve from the RFL distribution derived previously, which can be used to set up an inspection schedule, as illustrated in the case study. The advantage of using DBNs for reliability analyses lies in the ease of updating the prior information to obtain the posterior distributions.
Published in: 2019 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM)
Date of Conference: 15-18 December 2019
Date Added to IEEE Xplore: 03 February 2020
ISBN Information: