Risk Analysis of Submarine Pipeline Subjected to Impact Loads by Anchor

Jiang, Fengyuan (Ocean University of China) | Zhao, Yuliang (Ocean University of China) | Dong, Sheng (Ocean University of China) | Zhang, Ri (Ocean University of China)

OnePetro 

ABSTRACT

With the increase of human activities at sea, it is inevitable that anchors drop into the water due to operating errors, which may lead to failure of pipelines and cause economic damage and environmental pollution. Previous methods of related analysis are mostly based on the DNV-RP-F107 recommended method (hereinafter referred as DNV method). DNV method hardly considers the variation of anchor's size and weight. And it is insensitive to the pipeline geometry and material properties. Based on reliability theory, DNV method is improved to calculate failure probability under the consideration of the above relevant factors. The efficiency of the proposed method is verified by a practical case. Besides, analysis of the influence of various factors on pipeline failure probability is completed in this paper, including anchor weight, size, pipeline geometry and material properties, the distance from the anchor drop point. Meanwhile, considering the variability, the sensitivities of variables to the failure probability are discussed. Study results indicate that the failure probability calculated by DNV method is underestimated in some situations, which can probably cause a loss for pipeline projects. Whereas the proposed method is able to consider much more influences and leads to reasonable results consistent with the actual situation.

INTRODUCTION

Submarine pipeline is seen as the ‘lifeline’ for offshore oil and gas industry. Pipeline safety is one of the most important problems for engineering practice. Recently, anchors dropping into the sea becomes more frequent due to the increasing human activities at sea. The dropped anchors are likely to impact on pipelines and lead to pipeline failures, which can cause economic damage and environmental pollution. In order to reduce the risk and provide safe design, considerable research efforts have been devoted to risk assessment and reliability analysis of pipelines. In general, methods of the relevant research mainly consist of two categories: one is qualitative analysis, which can study the main influence factors on pipeline failures. Among them, fault tree analysis (FTA) is the most popular methodology and has been extensively applied to pipeline failure analysis. (Wang et al., 2007; Dong et al., 2005; Lavasani et al., 2011). The other one is quantitative analysis, which can determine pipeline failure probability and provide reliable reference for safe design. Katteland et al. (1995) developed a model for risk calculation, and applied it to evaluate the risk of all the installations in the North Sea. Det Norske Veritas (2010) proposed a ubiquitously used method for pipeline risk assessment and failure probability calculation (DNV method). Based on statistics of crane accidents, Det Norske Veritas (2013) also gave the falling probability for typical loads and various objects, which provided abundant references for pipeline risk assessment. On the basic of the above research, Liu et al. (2005) proposed a model to calculate the probability of pipeline being impacted under various anchorage conditions. Ding et al. (2010) modified DNV method and made a risk assessment of pipelines due to third-party activities. Yan et al. (2014) proposed a procedure to estimate the pipeline failure probability caused by anchoring activities. Up to now, to the best of the author's knowledge, quantitative analysis methods are mainly based on DNV method. In some situations, this method is hardly to consider the effect of anchor size and weight on pipeline failure probability. What's more, it is insensitive to the effect of pipeline geometry and material properties, which is not consistent with practice and may cause errors. In order to give an insight into those effects, a method based on reliability theory to calculate pipeline failure probability is proposed.