Probabilistic modelling of free spanning pipelines considering multiple failure modes

Abstract

Free Spanning occurs due to unevenness of seabed, artificial supports and etc. Based on DNV-OS-F101, free spanning may cause Vortex-Induced Vibration (VIV), local buckling and pipeline bursting. The purpose of this study is to present a methodology for assessing free spanning pipelines in terms of Probability of Failure (POF). In order to assess VIV fatigue of subsea pipelines, a Finite-Element Reliability based method with aim of Artificial Neural Network (ANN) is proposed. For training ANN, 560 different ABAQUS FE scenarios with different pipeline configurations are defined, and their results are transferred to MATLAB. According to the recommended criteria, ANN is optimized. For determining reliability level of VIV fatigue, Monte-Carlo Sampling (MCS) is used to optimize ANN, and POF is determined consequently. In addition, using recommended method by DNV-RP-F105, Fatigue Life Capacity (FLC) of pipeline is determined for different conditions. For assessment of pipeline bursting and local buckling, recommended models by DNV-OS-F101 is adopted to determine their relevant POF using First-Order Reliability Method (FORM) and MCS. Furthermore, the effect of different operating pressure on reliability level of pipeline is investigated. Finally, in order to determine the effect of each parameter on POF of pipeline bursting and local buckling sensitivity analysis is carried out, using alpha index.

Publication
Engineering failure analysis