Reliability and safety assessment of submarine pipeline stopper based on Fuzzy Comprehensive Dynamic Bayesian Network

The submarine pipeline stopper is an emergency device that can quickly seal damaged pipelines. Investigating the stopper's dependability and safety is vital to ensure that subsequent maintenance activities go smoothly. However, in production environments, it is challenging to get failure data for complex systems due to high experimental costs. This work proposes a fuzzy comprehensive dynamic Bayesian network (FCDBN) based on fault tree, fuzzy evaluation, and dynamic Bayesian network. Using this method, the failure rate of the stopper can be obtained, thus solving the problem of difficult data acquisition. The time slice-based qualities are taken into account while evaluating the reliability and safety. By controlling variables, each failure rate's effect is quantified. Finally, the process of failure prediction is completed. The sealing device is least reliable and most possible to fail, according to the results. The most significant influence on reliability comes from rubber barrel shoulder upwarping. The failure rate of the stopper is highest if the sealing device fails. Based on the aforementioned findings, appropriate control measures are suggested, which can greatly lower the stopper failure risk.