Mechanics of Offshore Pipelines, Volume 2: Buckle Propagation and Arrest

patience

Preface

Stelios Kyriakides, Austin, TX, United States

Buckle propagation is a problem unique to offshore pipelines and other tubulars in which, driven by the ambient external pressure, a locally weakened section of the pipe can initiate a collapse that propagates at high speed, quickly flattening large lengths of the line. The term propagating buckle was coined by researchers in the offshore industry in the early 1970s who first encountered the problem and identified its potential catastrophic effects. This book, Vol. 2, deals with the various aspects of buckle propagation, including the initiation of a propagating buckle, the lowest pressure at which an initiated buckle will propagate, the dynamics of buckle propagation, and the practical problem of arrest of an incoming buckle. Included also are the effects of tension on buckle propagation; buckle propagation in pipe-in-pipe systems; and confined buckle propagation, which can affect tubes that are surrounded by a stiff medium such as grouted casing.

This volume is a sequel to Volume 1, Mechanics of Offshore Pipelines: Buckling and Collapse, Elsevier, 2007, an expanded second edition of which is presently under preparation. Although the technical subjects covered in each volume stand alone, the two books are complementary and should be treated as companion texts. Thus for a general description of offshore oil and gas production and the role of pipelines and other tubular structures, the reader must depend on Vol. 1. The same applies to pipe manufacture and pipeline installation methods. To avoid repetition the reader of Vol. 2 is often referred to Vol. 1 for some technical details.

The two books are aimed at the practicing professional, but can also serve as a graduate-level text on the mechanical behavior of offshore pipelines. Entry-level graduate school background in nonlinear structural and solid mechanics, as well as working knowledge of issues in structural stability, numerical methods in general, and nonlinear finite elements in particular should make going through the analytical developments easier. A background in plasticity theory would enhance the brief chapter on metal plasticity at the end of Vol. 1.

Each problem considered is investigated through a combination of experiments and analysis. Careful experiments are used to elucidate and uncover the underlying physics of each aspect of buckle propagation and arrest. The understanding gained is used to analyze the problem, first with simplified models, and followed by large-scale numerical modeling. The experiments and numerical solutions provide sufficient details to enable reproduction by the interested user. Most chapters finish with design recommendations.

My involvement with buckle propagation and arrest started in 1976 through my work with Chuck Babcock at Caltech. The book cites our first fundamental contributions to the subject made at that time. However, the great majority of the results and developments in the subject were made working with several outstanding graduate students whose contributions are reflected in the references after each chapter. I would like to mention in particular Meng-Kao Yeh, Jyan-Ywan Dyau, Yu-Chung Chang, Tae-Dong Park, Tracy Vogler, and Edmundo Corona who is co-author of Vol. 1. All have gone on to become reputable researchers in their own right. Volume 2 is co-authored by Liang-Hai Lee, with whom I have cooperated for more than 20 years. He was instrumental in the study of slip-on buckle arrestors, participated in the execution of many of the experiments reported in several chapters, and has conducted afresh many of the reported numerical simulations. I am indebted to Liang-Hai for his contributions to the book, his long-term interactions with me and several of my graduate students, and for his friendship. Chapters 6 and 8 are co-authored with Theodoro Netto in recognition for his contributions to the study of dynamics of buckle propagation and the investigation of the dynamic performance of buckle arrestors. I thank him for our three-decade technical interactions, his sunny outlook on life, and his close friendship.

I am grateful to the offshore structures industry for their support through Joint Industry Projects as well as through individual contracts. The support by the U.S. Office of Naval Research of part of the work is also acknowledged with thanks. The writing of this book has also received financial support from the John Webb Chair in Engineering at the University of Texas at Austin. Thanks are also due to Julio Roman for his help with some of the more challenging figures in the book, and to Kevin Johnson for proofreading most of the text. Despite the best efforts of the authors, typos and some, hopefully minor, errors surely remain. The authors would be grateful if readers let us know of any typos and errors they come across. The writing of Vol. 2 was announced at the publication of Vol. 1 in 2007. I am solely responsible for the delay in its publication and apologize to the faithful readers and users of Vol. 1 who have been regularly asking about our progress, and thank them for their patience.

Chapter 1: Introduction

Contents

1.1Buckle Propagation in Offshore Pipelines

1.2Examples of Other Propagating Instabilities

1.3Buckle Initiation and Propagation Scenarios From Offshore Pipeline Operations

1.4Dynamic Buckle Propagation

1.5Influence of the Propagation Pressure on Pipeline Design

1.6Buckle Arrestors

1.6.1Arrestor performance

1.7Review of Buckling and Local Collapse of Pipelines

1.7.1Buckling and collapse under external pressure

1.7.2Buckling and collapse under combined tension and external pressure

1.7.3Buckling and collapse under combined bending and external pressure

1.7.4Collapse of dented pipes under external pressure

1.7.5Collapse of partially worn of corroded pipe under external pressure

1.8Brief Review of the Contents of the Book

References

Pipelines constitute a major means of transporting commodities like oil, natural gas, chemicals, various slurries, water, etc., which are very important to modern economic activity and trade. They have been in use since the beginning of the 20th century but have undergone significant increase in use from 1950 onwards so that thousands of miles of land pipelines exist today with many more being planned. A major driver behind this ever-increasing use of pipelines is that they constitute a safer, more economical, and energy efficient means of transportation than alternative means. Furthermore, although pipelines require a steep initial investment, they can have a lifespan of up to 40 years during which they can operate with relatively modest additional and maintenance costs. Oil and gas pipelines are usually the longest and most contentious, often encountering local resistance due to environmental and safety concerns. Accidents that result in spilling of oil, although rather rare considering the size of the industry, have prompted the imposition of more regulations by government agencies on the construction and operation of pipelines. With thousands of miles of pipelines in operation worldwide, onshore pipeline engineering is reasonably well established. Design can, however, be more demanding for pipelines in earthquake-prone areas, the artic, or environmentally sensitive areas.

In the early 1970s, hydrocarbons were discovered in relatively shallow waters on the continental shelf and in the seas beyond. This spawned new technologies and supporting industries that address underwater exploration, surveying, drilling, various underwater and floating production facilities, remotely controlled vehicles, underwater welding techniques and facilities, and many others. Offshore production introduced new tubular structures like conductors, well casing, flowlines, risers, pipelines, etc., which now operate under the more hostile environment of the sea that includes external pressure loads, high temperature, wave and current loads, unstable seabeds, corrosion issues, etc.

This book series is concerned with the mechanical behavior and limit states of the family of tubular structures used in offshore applications. Volume 1, an expanded 2nd Edition of which is currently under preparation (Kyriakides & Corona, 2007), deals with some of the main limit states encountered during installation and operation of the lines that can result in buckling and collapse of the pipe. The book also includes a review of offshore facilities and pipeline installation methods, which connect the limit states studied to field applications. A chapter on steel and linepipe manufacture helps the reader appreciate differences in the mechanical performance and geometric tolerances between seamless pipes and different types of seam-welded pipes rolled from plate. Because most of the limit states of interest are influenced by the inelastic behavior of the material, elements of incremental plasticity and major elastic-plastic constitutive models are reviewed in a dedicated chapter. Mechanical testing required for the design of pipelines and for the analysis of the limit states of interest is addressed in several appendices.

Volume 2 is dedicated to the problem of buckle propagation and its arrest. This is a problem unique to offshore pipelines and other tubulars where, driven by the ambient external pressure, a locally weakened section of the pipe can initiate a collapse that propagates at high speed, quickly flattening large lengths of the line. The extent of the catastrophic damage suffered is usually limited by the introduction of buckle arrestors to the pipeline. These are devices that locally strengthen the pipe circumferentially so that they halt the propagation of an incoming buckle, reducing the damage to the length between two arrestors. The book addresses several aspects of buckle propagation including its initiation, establishment of the lowest pressure at which an initiated buckle will propagate, known as the propagation pressure, the dynamics of buckle propagation, and various methods of arresting propagating buckles. Buckle propagation in pipe-in-pipe pipeline systems is examined in a dedicated chapter. Grouted casing or liner shells supported circumferentially by a stiff medium exposed to external pressure can collapse within the confinement in a characteristic U-shaped mode. The conditions under which such a collapse can evolve into a confined propagating buckle are examined and a confined propagation pressure is quantified. Although the technical subjects covered are independent of limit states discussed in Vol. 1, the two volumes present exposés of different aspects of mechanical behavior of pipelines. They are complementary and should be treated as companion texts. The following are areas of direct intersection:

(a)Offshore production consists of several wells connected to a central production platform with flowlines and risers. After processing, gas and oil are transported to shore via pipelines. Several types of platforms and satellite facilities have been outlined in Vol. 1 through which the related nomenclature used in industry has been introduced. Facilities for installing the required tubulars to the sea floor, and the associated installation and operation loads, are also introduced. Consequently, these subjects will only be discussed as the need arises in Vol. 2.

(b)A propagating buckle is usually initiated from a section of the pipeline whose resistance to external pressure has been reduced by local collapse caused by one of the limit states discussed in Vol. 1. Therefore buckling and collapse under external pressure, bending, and tension and combinations of these will be very lightly touched upon here, and the reader will be referred to the relevant chapters of Vol. 1 for details.

(c)The material properties of pipeline steels are produced to industry specifications. They are influenced by the manufacturing process and by any subsequent process that results in thermomechanical loads. The geometric characteristics of pipes are similarly influenced by their manufacture and are again held to specified standards. These issues, as well as material and geometric measurements routinely performed on pipelines, have been discussed in Vol. 1. Accordingly, the present text assumes that the reader has a working knowledge of these subjects.

1.1: Buckle Propagation in Offshore Pipelines

The first concerns about the potential catastrophic effects of a propagating buckle to an offshore pipeline came in the early 1970s from a study conducted at Battelle Laboratories in Columbus, Ohio, dealing with a variety of mechanical problems affecting offshore pipelines. At the time, the continental shelf and the seas beyond were becoming a new frontier for hydrocarbon exploration, and offshore structural engineering was an emerging technology. The study was cosponsored by a group of companies with offshore interests and was conducted at the Battelle Laboratories in collaboration with technical personnel from the sponsors. The first public airing of the name propagating buckle came in a brief description of the problem given by Mesloh, Sorenson, & Atterbury in July 1973. Parallel to the work at Battelle, various aspects of the problem were being investigated by researchers employed by the sponsors. For example, in April 1972, Lochridge & Gibson (Brown & Root, Inc., 1973) and Broussard, Ayers, & Walker (Shell Oil Company, 1973) applied for patents for a number of buckle arrestor designs. In their applications, they demonstrated significant insight into the problem (the patents were awarded in 1973). During approximately the same period, British Petroleum (1972) was conducting full-scale tests on 32 in (0.81 m) pipes in the sea to determine the efficiency of buckle arrestors (Palmer, 1980). Although reports from the time are rather sketchy, propagating buckles seem to have been first encountered during experiments to establish the collapse capacity of tubes under combined bending and pressure. It was observed that, below a certain pressure, buckling tended to remain local, whereas above it, it tended to spread and collapse the whole length of the test specimen. Buckle propagation was officially introduced to the scientific community by Palmer & Martin (1975) in their groundbreaking brief publication in Nature, which also included the first estimate of the propagation pressure. The principal author came to the problem in 1976 and published some of the early technical results on buckle propagation and arrest in Kyriakides & Babcock (1979). Many developments have followed since; these contributions will be cited in the relevant sections of the book.

Buckle propagation in offshore pipelines is an instability driven by external pressure. It is best introduced through a brief discussion of buckling and collapse of a long tube or pipe under external pressure. A pipe with radius R and wall thickness t under external pressure P develops a circumferential membrane stress of PR/t. The stress causes circumferential contraction resisted by a stiffness that is proportional to Et/R², where E is the elastic modulus of the material. Buckling refers to the pressure at which an alternate, oval-shaped mode of deformation becomes energetically preferable to uniform contraction. For pipes with higher diameter-to-thickness (D/t) ratios, the buckling pressure is proportional to 2E(t/D)³. For lower D/t pipes, buckling is influenced by the inelastic properties of the material so E is replaced (approximately) by the instantaneous tangent modulus (Et). In both cases, buckling is accompanied by a drastic drop in stiffness, making the buckling pressure a critical design variable for offshore pipelines. The oval-shaped buckling mode involves bending in addition to the membrane stress, and their combined effect yields the material soon after the onset of buckling. The additional loss of stiffness causes a downturn in the pressure that can be sustained and the buckled structure now has the tendency to collapse. The maximum pressure in such a response is the collapse pressure of the pipe (PCO). As in many other structures, the collapse pressure of pipes can be reduced by small initial geometric imperfections among other factors. It is noted, however, that collapse under external pressure and the factors affecting it are quite well established, so collapse can be avoided by judicious design.

Establishment of the collapse load is usually sufficient for the design of most structures. In the case of pipelines under external pressure, PCO is only one of the required limit states. The propagation pressure is a second limit state that must be considered. To help understand its importance, the full postbuckling behavior of the pipe is examined. For simplicity, consider the uniform collapse of a representative slice of the pipe loaded by external pressure. Figure 1.1 shows schematically the expected pressure-change in volume response (P − δυ). It exhibits the initial stiff prebuckling branch OL mentioned earlier, which terminates into a pressure maximum that represents the collapse pressure. This is followed by a descending branch, LM, during which the section becomes progressively more oval. The deformation gradually morphs into a peanut shape while the pressure drops down to a level that is only a small fraction of PCO. At some point, the two most deformed sectors of the collapsing cross section come into contact, forming a figure 8 configuration. Contact stiffens the cross section, arrests further collapse, and further deformation requires an increase in pressure. In other words, the structure develops a second stable branch, MN, during which contact between the walls of the collapsed section tends to spread outward as shown in the configuration in the ascending branch.

Fig. 1.1 Schematic representation of the underlying pressure-change in volume response of a long tube collapsing under external pressure.

In summary, this up-down-up pressure-change in volume response consists of an initial stiff, stable, and structurally sound branch; a second stable branch that develops once the walls of the collapsed section come into contact; and an unstable branch that connects the two. It will be referred to as the underlying response of a collapsing pipe. As in other structures with such a snap-through response, if the equilibrium of a state along the stable branch OL is disturbed sufficiently, the structure will snap to a collapsed configuration on branch MN. In a long pipeline under constant pressure, the disturbance can be applied locally, leading to local collapse while the rest of the line remains in the relatively undeformed configuration corresponding to this pressure along OL. In other words, the collapsed and uncollapsed configurations can coexist. The coexistence of two deformation regimes at the same load is a characteristic of structures and solids that exhibit a propagating instability (Kyriakides, 1993, 2001).

The snapping from equilibrium A to C at pressure P consumes strain energy corresponding to the integral under the P − δυ response from A to C. Concurrently, the pressure does work due to the reduction in the volume δυ in going from the intact to the collapsed section. The lowest pressure at which the snap-through can take place is at the level that makes the two energies equal. In Chapter 2 it will be demonstrated more rigorously that this is the lowest pressure at which the local collapse will propagate and is known as the propagation pressure, PP. At any pressure higher than PP, Pδυ is larger than the energy required to deform the section, and the difference is converted into kinetic energy. Thus in a constant pressure environment at P > PP, the collapse will propagate along A′B′C′ dynamically.

In practice, the disturbance of the stable equilibrium can be caused by a third agency that locally weakens the pipe leading to local collapse. This can be a local denting caused by impact of a foreign object, a local buckle caused by excessive bending, or damage such as corrosion or wear (see Section 1.7). If the external pressure is higher than PP, the collapse will propagate. An example of a buckle that partially propagated at the propagation pressure in a steel tube with D/t ≈ 37 is shown in Fig. 1.2a. It consists of a uniformly collapsed length with a figure 8 cross section, and a length with an intact circular cross section connected with a smooth transition zone a few tube diameters long. The image demonstrates the coexistence of the two deformation regimes at the same pressure alluded to earlier. Figure 1.2b shows a sequence of cross sections taken through such a transition zone that illustrate the geometric changes imposed on the pipe by the propagating collapse. It is worth pointing out that the crown points of the collapsed sections in the two images were in contact under pressure; the observed opening develops on unloading due to springback. Furthermore, since a propagating buckle can be initiated at any pressure PP ≤ P ≤ PCO, the severity of flattening suffered by the collapsed pipe increases with pressure (e.g., see Figs. 1.8b and 7.13b).

Fig. 1.2 (a) The profile of a buckle that propagated quasi-statically connecting the collapsed and intact cross sections. (b) Sections through such a profile.

1.2: Examples of Other Propagating Instabilities

The study of buckle propagation in offshore pipelines has revealed that the underlying mechanism has fundamental implications in that it governs the behavior of several other large structures but also several material systems. Collectively, this class of problems is called propagating instabilities. What they have in common is an underlying up-down-up local-deformation response like the one in Fig. 1.1. If the first stable branch is sufficiently disturbed to result in localized deformation, the instability has the potential of spreading over the whole domain. A line of standing dominoes is perhaps the most recognizable problem in this class, where the toppling of one of the dominoes leads to the progressive toppling of all in the line (Bert, 1986; Stronge & Shu, 1988). Figure 1.3 shows images from four problems that exhibit a propagating instability. The first is the propagation of collapse by external pressure in a liner shell inside a stiff contacting cavity that is analyzed in Chapter 4. Figure 1.3a shows the front separating the U-shaped collapsed and intact sections of the liner shell (Kyriakides, 1993, 2002; Kyriakides & Lee, 2005). The second is the propagation of a bulge in an elastic tube under internal pressure. Figure 1.3b shows the front separating the relatively undeformed and highly deformed sections of a latex rubber tube (Kyriakides & Chang, 1991). The third involves the propagation of collapse in a long roof or panel with an arch cross section by external pressure. Figure 1.3c shows the propagation of collapse in a mylar panel under pressure (Power & Kyriakides, 1994). Lüders banding is a material instability most prominently exhibited by low-carbon steels. It leads to localized plastic deformation of 1%–4%, which tends to spread to the rest of the domain. Figure 1.3d shows an inclined band that separates Lüders deformed and relatively undeformed sections during a tensile test on a strip. Under displacement controlled loading, the front propagates at a nearly constant load until the whole specimen is deformed to the Lüders strain (Hallai & Kyriakides, 2013; Kyriakides & Miller, 2000). In the first three problems, the underlying up-down-up behavior can be established by evaluating the response of a characteristic section of the structure. For Lüders strain, the underlying behavior must come from dislocation considerations. However, the macroscopic behavior can be reproduced by introducing a negative slope in the stress-strain response over the extent of the Lüders strain (Ericksen, 1975). A recent review of the localization and propagation of a bulge in a rubber tube, along with movies that demonstrate the events, appears in Appendix A.

Fig. 1.3 Examples of systems exhibiting a propagating instability. (a) Buckle that propagated under external pressure in a tube confined in a circular cavity. (b) A bulge that propagates elastomeric tube under internal pressure. (c) A buckle that propagates in an arch-shaped panel under external pressure. (d) The propagation of Lüders deformation in a steel strip under tension.

1.3: Buckle Initiation and Propagation Scenarios From Offshore Pipeline Operations

In order to better connect the problem of buckle propagation to offshore pipeline engineering practice, we consider a couple of scenario in which this limit state develops. The first takes place during the installation of a pipeline on the seabed. The most commonly used methods of installing offshore pipelines, namely the S-lay, J-lay, and reeling, are outlined in Chapter 2 of Kyriakides & Corona (2007). On the way down to the sea floor, the line usually acquires either an S or a J shape. The vessel must support the hanging pipeline with enough tension to keep the curvature at allowable levels, while simultaneously following a prescribed route encountering sea currents and waves. In S-lay, shown schematically in Fig. 1.4a, the hanging line acquires a characteristic S-shape. Near the surface, it experiences high tension and bending over the relatively stiff surface of the stinger, which can lead to buckling. As it moves further down, it experiences increasing pressure and decreasing tension. In the sagbend, close to the seabed, it undergoes high bending under the highest pressure. Since the pipeline is usually installed empty in order to take advantage of buoyancy, buckling under combined bending and external pressure is of concern. The bending curvature is kept at a safe level by the tension applied at the vessel. Modern installation vessels are equipped with powerful thrusters and computer-controlled navigation systems. Despite this technology, accidental loss of tension is possible causing an increase in the sagbend curvature and local buckling as shown in Fig. 1.4a. Buckling under combined bending and pressure is initially local, but if the external pressure is high enough a propagating buckle will be initiated as illustrated in Fig. 1.4b. It is worth pointing out that although bending is the cause of local buckling, once the propagating buckle is initiated, the local bending moment is reduced by as much as an order of magnitude, and thus the external pressure becomes the primary loading driving the propagation of collapse.

Fig. 1.4 (a) Initiation of a propagating buckle from a section that buckled under combined bending and pressure during the installation of a pipeline. (b) Photograph of a bending buckle that initiates a propagating buckle.

Propagating buckles can also be initiated during the operation of a pipeline from a locally weakened section of pipe. For example, in many offshore fields, on the sea surface above a pipeline there is traffic from shipping, fishing, and other commercial activities. The dropping of equipment or an anchor on the pipelines, as shown in Fig. 1.5, is not uncommon. Impact can induce a dent, which reduces the local resistance of the pipe to collapse by the ambient pressure. If the line is operating at a high enough internal pressure, catastrophic failure may be averted. However, the pipe can subsequently collapse during a shutdown or during a depressurization for inspection of the line using a PIG. The local collapse can initiate a buckle that propagates dynamically with the aforementioned catastrophic consequences.

Fig. 1.5 Initiation of a propagating buckle in an offshore pipeline damaged by impact by a foreign object.

1.4: Dynamic Buckle Propagation

The scenarios outlined in Section 1.3 describe possible causes of local collapse during the installation and operation of an offshore pipeline. Once the circular geometry of the pipe is thus locally compromised and a propagating buckle is initiated at a pressure P > PP, the buckle will propagate dynamically and has the potential of quickly destroying the whole line. Dynamic buckle propagation in a constant pressure environment is a fundamental problem on its own accord. In addition to inertial effects, it adds to the problem of the interaction of the water with the receding walls of the collapsing pipe, material rate effects, more severe self-contact of the collapsing walls of the pipe, much larger deformations, and potentially rupture.

Since in practice any initiated buckle will propagate dynamically, the problem commands fundamental understanding. The first systematic experimental study of dynamic buckle propagation was reported in Kyriakides & Babcock (1979). The experiments were conducted in a constant pressure environment imitating that of the sea using Al-alloy tubes. In order to illustrate the role of the fluid-structure interaction on the problem, dynamic propagation experiments were conducted using air as well as water as pressurizing media. They reported that once initiated, collapse accelerated to steady state within a few tube diameters and continued at that value until the whole length of the tube was collapsed (typical length ~ 75D). Figure 1.6a plots the measured steady-state velocities vs. pressure ranging from just above PP to just under PCO for tubes with D/t ≈ 35.7. Velocities reaching several hundred meters per second were reported with the induced flattening of the cross section increasing with pressure. For both pressurizing media, the velocities experience an initial sharp increase with pressure and then follow a nearly linear increasing trajectory. The velocity in air is nearly double that in water pointing to the influence of the surrounding fluid. The authors observed a significant shortening of the profile of buckle propagation as the velocity increased. They conducted dynamic propagation experiments on tubes with D/t ranging between 30 and 50 and reported an increase in velocity as the D/t was reduced. This is illustrated in Fig. 1.6b where velocities measured at P/PP = 1.7 are plotted against the tube D/t (from Kyriakides,