FE analysis controls subsea pipeline buckling

Current subsea hardware has technical limits of around 15,000 psi and 250° F. Pressure and temperature envelopes are further constrained by system sizing or bore, fluid properties, and water depth. As system diameter or bore increases, the qualified pressure rating for components such as subsea valves tends to reduce. The record for a 10-in. subsea gate valve is believed to be 13,050 psi. Design, procurement welding, and installation difficulties for pipelines and risers generally increase with increasing diameter, and therefore, system diameters are also constrained by practical limits.

The industry is continually working to stretch HPHT qualification further. BP's 20KTM project, for example, includes development and qualification of subsea production systems, including HIPPS, to 20 ksi and 350° F. The 20KTM project will enable major deepwater developments including Kaskida and Tiber in the Gulf of Mexico's Paleogene play (Fig. 1), as well as developments in the Mediterranean and Caspian seas.

Designing pipeline systems for deepwater HPHT conditions requires addressing flow assurance and operability (given the difficult nature of HPHT reservoir fluids), control of corrosion and material behavior under extreme conditions, management of pressure containment, mitigation of extreme thermal loads, and installation of heavy systems in deepwater, among other issues. Selecting the optimum system requires effective option screening of all possible field layouts and numerous pipeline configurations.

View Image Gallery>>

Flow assurance, operability

Intervention to remove a wax or hydrate blockage is extremely difficult and costly depending on water depth, requiring the designing and building in of high levels of operability and availability. HPHT reservoir fluids generally need to be kept hot, dictating use of high levels of insulation to prevent wax deposition and avoid hydrate formation during shutdown and restarts. Advanced wet thermal insulation systems have been developed and are continuing to be developed for service conditions of high temperature, external pressure, and long design life.

Pipe-in-pipe systems can provide improved thermal performance when needed as they allow the use of materials such as aerogel, which have extremely low heat-transfer properties but which may have little or no structural strength and need to be maintained dry at atmospheric (or in some cases partial vacuum) conditions. Pipe-in-pipe systems can also affect structural response under axial and bending loads, potentially reducing loading in the hydrocarbon containing inner pipe. Both the selection of a pipe-in-pipe system and the sleeve-pipe dimensions affect the overall pipeline system, as well as the loading in the sleeve and inner pipes.

A number of proprietary and generic pipe-in-pipe systems have been developed and deployed, each with its own parameters of thermal performance, required sleeve-pipe diameter, connectivity of inner to sleeve pipe, structural response, welding requirements, installation, cost, and track record. In addition to these factors, pipe-in-pipe systems can be more difficult to inspect and repair when in operation than single-pipe systems.

Active heating systems that use either electrical power or hot liquid circulation to maintain fluid temperatures during shut-downs provide another option and increasingly are deployed for both wet insulated and pipe-in-pipe systems.

Dual pipelines are commonly used to enable round trip pigging and hot oiling. Dual pipeline systems also enable greater operational flexibility including lower turndown rates and reduced slugging tendency. For deepwater HPHT developments, system sizing is a problem with bores limited by subsea hardware qualification and pipeline and riser practicalities, leading to dual or multiple pipeline systems of smaller diameter as a potential project solution.

System design should consider options of dual or multiple pipeline systems of smaller diameter and single systems of larger diameter, as well as the various options for passive and active thermal management.

View Image Gallery>>

Material selection

Yield and tensile strengths derate at elevated temperatures, and because the material yield point is commonly exceeded under HPHT loading, it is important to understand and model post yield-strain hardening behavior.

Carbon steel will generally require a corrosion allowance, though it may be possible to pressure derate the system so that end-of-life corrosion need not be considered in conjunction with start-of-life pressure. X65 or X70-grade pipes are the highest strength carbon steels used for subsea pipelines due to concerns regarding yield-to-tensile strength ratio, weldability (including overmatching), and fracture toughness (especially in sour service).

An alternative to carbon steel is to employ solid corrosion-resistant-alloy (CRA) pipe or carbon steel pipe clad with a 3-mm CRA layer. Selection of the most suitable CRA depends on corrosiveness of reservoir fluids, as well as other issues such as procurement, welding, installation, and costs, and project-specific testing may be required. Solid CRA pipes tend to have both yield strengths and increased strain hardening higher than carbon steel but may suffer from greater levels of strength derating at elevated temperatures. These material characteristics coupled with potentially reduced WT can lead to differences in structural response compared with carbon steel pipelines.

View Image Gallery>>

Pressure containment

Subsea production pipelines conventionally are designed to the maximum wellhead shut-in pressure (SIP), with the hoop stress limited to a percentage of specified minimum yield strength (SMYS) and the pipeline sized so that ultimate burst pressure exceeds design pressure with a sufficient level of safety, as defined by design codes such as DNV OS-F1012 or API 1111.3 At increasing well shut-in pressures, required WT increases and becomes increasingly difficult to design to code. Practical limitations arise including manufacturing, offshore welding and inspections, installation, and in-situ response.

Reduced bores also lead to increased pressure drops and potentially lower production rates. Diameters may need to be increased as an alternative. More stringent codes further increase riser WT, placing additional limits on available options.

High-integrity pressure protection systems (HIPPS) can help address these situations. HIPPS are safety instrumented systems (SIS) that monitor pressure and isolate downstream equipment from overpressure conditions that might develop upstream. Overpressure can result from events either upstream (loss of well control) or downstream (process trip, inadvertent valves closure, hydrate or other blockage). HIPPS are autonomous and have a low probability of failure on demand (PFD). In a subsea system, HIPPS are as close as possible to the pressure source (well) and allow use of lower-rated downstream equipment including pipelines and risers.

Downstream of a subsea HIPPS, pipeline design can cover a range of approaches. These include, in order of reducing WT and increasing cost savings:

• Pipeline is designed not to yield at the SIP should the HIPPS fail. Hence the pipeline should not be damaged by the HIPPS failure. This approach is sometimes referred to as "no yield."

• Pipeline is designed to yield but to have a low probability of burst at SIP in the event of a HIPPS failure. This may cause permanent deformation of the pipeline requiring extensive inspections, fitness-for-purpose assessments, and possible repairs before being returned to service. This approach is sometimes referred to as "no burst." The pipeline must also be designed to code at a reduced design pressure, but this is not generally the limiting case.

• Pipeline design pressure is defined as marginally more than HIPPS trip pressure and the hoop-stress design limit is to code at the reduced design pressure. Should a HIPPS fail, the pipeline will yield and is likely to burst. This approach is sometimes referred to as "burst critical."

Downstream of a HIPPS, the pressure-containment philosophy for the on-bottom pipeline section is generally no burst or burst critical. Although risers are not the subject of this article, the pressure-containment design of the pipelines and risers must include them. Riser design must ensure that the probability of failure close to manned sites is extremely low and therefore should be fortified relative to the on-bottom pipelines. Using a lower design factor on hoop stress ensures this for a fully rated pipeline and riser system. Systems incorporating HIPPS should use a riser as strong as practical and ideally fully rated. Riser thickness, however, may be excessive, and riser feasibility may be the primary driver in decided whether to use HIPPS.

For a no burst or burst critical pipeline, the riser will either need to be fully rated, no yield, or no burst with a lower probability of failure than the pipeline. A more fortified riser will always be preferable, if practical. Fully rated and burst-critical WTs are calculated by standard deterministic methods in accordance with design codes. No burst WT needs to be determined with probabilistic or structural reliability methods.

The required probability of failure-on-demand of a no burst pipeline is an iterative process including safety integrity level (SIL) and pipeline assessments. A pipeline failure probability of around 10-2 is reasonable for the pipeline. At lower failure probabilities the design would tend towards a fully rated system and much of the benefit of employing a HIPPS system would be eroded; any higher and the system tends towards burst critical. A conditional failure probability of 10-2 was also proposed by a DNV led JIP on subsea HIPPS.4

For a no burst riser, a lower probability of failure is required, especially if the pipeline is no burst. The HIPPS JIP proposed a failure probability of 10-3.

Pressure-containment options are considered for a range of shut-in pressures, water depths, and typical HPTH parameters (Table 1). Two diameters are considered to illustrate the selection of multiple smaller diameter pipelines vs. a reduced number of larger diameter pipelines.

The 7,500-psi design pressure downstream of the HIPPS is based on an assumption of a typical operating pressure of around 2,500-3,000 psia at the wellhead. HIPPS trip pressure must be sufficiently higher to avoid spurious trips and the pipeline design pressure higher again.

The analysis of the no burst pipeline is based on the Stewart and Klever burst model enacted in a Monte-Carlo assessment,5 which takes account of the length of the pipeline and predicted corrosion morphology and depth.6 Fig. 2 shows the no burst results for the 10.75-in. OD pipeline with 6-mm corrosion allowance and SIP of 17,500 psi.

A typical no burst failure probability of 10-2 requires 40-mm WT for a pipe-in-pipe system and 35-mm for a single pipe in 2,000-m water depth; reducing to 32 mm and increasing to 37 mm in 3,000 m and 1,000 m water depths, respectively. A similar plot to Fig. 2 can be made for the riser system.

Fig. 3 shows required WT for the three pressure-containment philosophies, and for pipe-in-pipe and single pipe, for pressures from 15,000 to 20,000 psi. Fully rated 10.75-in. OD wet insulated pipelines are close to or exceed practical limits over the range of pressures. Practical WTs are typically 38-42 mm. While no burst design does reduce WT requirements, WTs are still relatively large, and approach practical limitations at higher pressures. The burst critical approach, by contrast, gives conventional and practical WTs.

The use of pipe-in-pipe adds to the WT requirement; e.g., the external pressure at 2,000-m water depth reduces the pressure differential across the wall of a single pipe by about 3,000 psi, which equates to about 4 and 5 mm of WT for 8.625-in. and 10.75-in. OD pipelines, respectively, for the fully rated and no burst cases, and greater still for burst critical cases. This requirement pushes the fully-rated 10.75-in. PIP WT beyond present practical limits for the range of pressures and makes no burst WT requirements similarly problematic over the majority of the pressure range. The differential between WT requirements for single and PIP systems further increases in deeper water; 3,000-m depths would require PIP WT 6-8 mm greater than for a single-pipe system.

All of these examples use a corrosion allowance of 6 mm. For fully rated and no burst designs it may be feasible to derate pipeline design pressure over life so that the pipeline need not be designed for start-of-life pressure in combination with end-of-life corrosion. This approach can lead to WT savings, particularly for the fully rated case, and has been applied on several projects.

Fig. 4 shows WT requirements for 10.75-in. and 8,625-in. OD risers. For a riser of constant nominal WT the greatest through-wall pressure differential, and therefore greatest hoop loading during an overpressure event, is at the base of the riser for a PIP system and at the top of the riser for a wet-insulated system.

In comparison with the pipeline, fully rated riser WTs are greater again and are at or exceed the practical limits of 38-42 mm. No burst WTs are generally within practical limits except for the highest pressure 10.75-in. OD cases.

Table 2 shows no burst riser WTs at a shut-in pressure of 17,500 psi for a 10.75-in. OD inner pipe within a PIP system for riser target failure probabilities of 10-3-10-5, which represent increasing levels of fortification compared with the no burst pipeline (target failure probability = 10-2). The pipeline and riser WTs are rounded up to the nearest millimeter.

Table 2 gives the true probability of failure for a given WT. For the pipeline this is the probability of failure given an overpressure event. But for the riser this is the probability of riser failure given that the pipeline survives the overpressure event and is roughly equal to the probability of riser failure in isolation, as there is only a 1% chance of pipeline failure (or less if the pipeline WT is rounded up to the nearest millimeter). The table also presents the probability of where a system failure would occur. All probabilities of failure given in Table 2 are in the corroded condition (end of life) and are conditional; i.e., an overpressure event has occurred and HIPPS has failed to close on demand.

Fortification associated with single order-of-magnitude increases in failure probability result in relatively small increases in riser WT for the example project (Table 2). A riser target failure of 10-3 would result in a 9% probability of a failure occurring within the fortified riser should a system failure occur (1.1% probability). Because the WTs have been rounded up, there is about a 16% probability of any failure being in the fortified riser should a system failure occur (0.32% probability), the level of WT rounding up being marginally greater in the pipeline, 0.8 mm, than in the riser, 0.3 mm. A single order-of-magnitude difference in failure probability between the pipeline and riser, therefore, does not appear to provide a particularly high level of fortification.

The true probability of failure in the riser and the probability of a system failure occuring within the riser reduce in-line with riser target failure probability. Analysis of the start-of-life condition (not reported in Table 2) shows the probability of a system failure in the riser tends to be higher, as the effect of the unused corrosion allowance is more significant for the pipeline than the thicker riser. The true probabilities of failure in both pipeline and riser, however, are lower than at the end of life.

This study assumes that key distributions of WT and UTS for both riser and pipeline are towards the lower ends of the specified ranges to give a conservative estimate of the probability of burst. If the as-manufactured distributions for the pipeline are relatively stronger than those for the riser (due to different suppliers, manufacturing processes, materials, etc.), the level of fortification for a single magnitude of target failure probability could be greatly, even completely, eroded.

The target probability of failure and fortification of the riser require careful consideration. It is important to consider the possible upper bound burst strength of the pipeline; otherwise there may be little or no riser fortification in the as-manufactured condition.

View Image Gallery>>

External pressure

In increasing water depths, the sleeve pipe of a PIP system will become increasingly thick walled to resist hydrostatic collapse and local buckling under combined bending and external overpressure. The sleeve pipe WT should resist bending during installation and under lateral buckling, so that the sleeve pipe does not become excessively limiting. Bending imposed by pipelay and occurring in operation due to lateral buckling may not be known in early engineering, requiring that sensitivity of WT to likely bending loads be considered.

Sleeve pipe diameter varies between different proprietary systems and with thermal insulation requirements. For the 10.75-in. OD pipeline considered here, a typical sleeve pipe diameter would be around 16 in. but could be reduced to 14-15 in., depending on the system.

Fig. 5 plots required sleeve-pipe WT in accordance with DNV OS-F101 as a function of applied (or allowable) strain. The required WT for hydrostatic collapse is essentially the value at zero bending strain. Given the required WTs across the range of diameters and bending strains, seamless pipe is assumed and the fabrication factor is taken as 1.0.

The 16-in. OD sleeve pipe requires WT around 20 mm for hydrostatic collapse only and increases to 26.5 mm when allowing for 0.5% bending strain. Reducing sleeve pipe diameter to 14 in. saves 2.5-3.5 mm of WT depending on the allowable bending strain sought.

The selection of sleeve pipe diameter and WT can greatly affect fabrication, welding, and installation. Large WT may also reduce annular space available for insulation and fabrication. Sleeve-pipe diameter and WT selection therefore must carefully balance cost against structural response, robustness, and thermal performance.

View Image Gallery>>

Thermal mitigation

Elevated operating pressures and temperatures will tend to induce Euler buckling of deepwater pipelines, which usually occurs in the horizontal plane (lateral buckling). Lateral buckling can induce strains in excess of yield at the crown of the buckle and high cyclic stresses on repeated operational shutdowns and restarts, with resulting limit states including local buckling, fatigue, and weld fracture.

Successful lateral buckling design requires the intentional promotion of regularly spaced buckles to share the loads between the buckles. The design should ensure that the formation of the buckles is as reliable as possible so that all planned buckles form and the probability of unplanned buckles is minimized. Unplanned buckles can become more severely loaded than planned buckles and become the limiting factor in design. Design must understand, quantify and, where possible, reduce inherent uncertainties in lateral buckling and in parameters such as complex pipe-soil interactions.

Project design requires a buckle mitigation considering a range of variables including pipe dimensions and configuration, pipe material, buckle initiator type and configuration, and initiator-buckle spacing. Buckle mitigation design is further complicated by factors such as pipelay bends, bathymetry, mid-line and end structures and connections, and pipeline anchors.

Engineered buckle initiators improve the reliability of buckle formation by reducing the magnitude and uncertainty of critical buckling force, instead creating controlled imperfections in the pipeline and lessening dependency on pipe-soil interaction. Initiators also reduce both the magnitude and level of uncertainty of operational loading in the buckle, again by lessening dependence on the pipe-soil interaction.

Sleepers, buoyancy, and zero-radius bends are the three most commonly used initiators in deepwater operations (Fig. 6). The deepwater environment reduces such concerns as over spans, instability, and fishing interactions. Sleepers, however, may not be suitable for multiphase lines because the regular passage of slugs can induce fatigue loading. Incorporating a push mechanism in sleeper design further improves the reliability of buckle formation.

Buckle mitigation strategy will include selection of the preferred engineered initiator type and configuration, considering the relative merits of the different systems with regard to buckle formation reliability, loads in buckles, installation requirements, and costs. The relative merits vary with pipe weight, water depth, pipelay method, and vessel and must be assessed on an individual project basis.

The Safebuck JIP developed a structured design methodology incorporating probabilistic assessment of buckle formation, both planned and unplanned, enhancing design confidence.7 Safebuck requires that allowable buckle spacing or virtual anchor spacing, at which all limit states are not exceeded, be less than characteristic buckle spacing or virtual anchor spacing, which has a low probability of exceedance defined by the methodology.

A probabilistic assessment of buckle formation is outside the scope of this article, but the influence of pipeline configuration (pressure containment, single vs. PIP, sleeve WT, etc.) and buckle spacing on the response within a buckle is considered, limiting the examination to conventional sleepers. Given the discontinuous support conditions along the buckle as it crosses the sleeper, the assessment uses finite-element analysis as presented in the Safebuck design guideline.

View Image Gallery>>

Limit states

As all the pipelines have relatively low diameter-WT (D/t) ratios, including the burst critical cases, the local buckling limit state is not critical for either the wet insulated system or PIP inner pipes. Only the more critical limit states of fatigue and local buckling of the sleeve pipe under external overpressure are presented. A knockdown factor of 10 is applied in the assessment of fatigue of the pipeline weld root due to the detrimental effects of sour service, and a knockdown factor of 9 is applied to the assessment of the weld cap of the single wet insulated pipe and sleeve pipe of the PIP due to the accelerated fatigue associated with low frequency loading in a seawater environment. The fatigue assessment considers a total of 120 full shut-down cycles over the life of the development and an allowable damage ratio of 0.18, sufficient to address some fatigue damage during installation.

Fig. 7 shows unity checks against the critical limit states for the fully rated 10.75-in. OD wet insulated pipeline and 10.75 +16 in. PIP. The tolerable buckle spacing is the maximum spacing at which all unity checks are less than 1.0. For buckle initiation employing sleepers, a tolerable buckle spacing of at least 1,000-1,500 m is typically required for a robust and practical design strategy.

For the example project, fatigue at the weld root is the limiting criterion for the wet insulated fully rated pipeline (Fig. 7), but there is little difference between the root and cap due to the thickness correction factor applied to the cap for the thick-walled pipe.

Within the PIP system, the presence of the sleeve pipe reduces the fatigue damage in the inner pipe root and the sleeve weld cap becomes the most critical weld location. Increasing sleeve-pipe WT to 27 mm from 23.8 mm reduces fatigue damage in the sleeve pipe welds, but the level of damage remains unacceptable for all cases except for the larger WT at the shortest buckle spacing. Fatigue performance of the sleeve pipe could be increased if it can be demonstrated that the pipeline coatings protect the weld cap from accelerated fatigue in the seawater environment. If not, the sleeve may require further WT increases to achieve robust thermal mitigation design. Inner pipe root fatigue is not greatly affected by small changes in sleeve WT, and therefore only a single trace is shown. Local buckling of the sleeve pipe under external overpressure is not as critical as fatigue.

The wet-insulated fully rated system appears to offer greater tolerable buckle spacing-roughly 1,700 m vs. PIP's 1,000 m-but this depends on the relative knockdown factors appropriate for the pipeline root and sleeve cap. Both systems, however, use very thick pipe, and it may not be practical to procure fully rated risers or install the PIP system.

Fig. 8 presents corresponding results for burst critical 10.75-in. OD wet-insulated and PIP systems. Fatigue at the weld root is again the limiting criterion for the wet-insulated burst critical pipeline, and sleeve weld cap fatigue the limiting factor for PIP. Tolerable PIP buckle spacings are much better than those of the fully rated PIP systems. The pipe-in-pipe system now appears to perform better than the corresponding wet-insulated system, and any improvement to the sleeve cap weld fatigue performance would further increase tolerable buckle spacing. Similar trends and results occurred for an 8.625-in. OD system.

View Image Gallery>>

Subsea cooling

Engineered subsea cooling spools so far only have been used in shallow water basins, partly due to higher temperatures (Fig. 1). A cooling spool reduces the thermal load on the downstream pipeline, reduces strength derating, and potentially increases downstream material and insulation options. Engineered cooling may not always be suitable due to operability considerations, especially during start-up and turn-down, and also due to bathymetry.

The example project shows pipeline design temperatures of up to 125° C. as achievable without engineered subsea cooling. Development of deepwater reservoirs with extreme wellhead temperatures, however, may require engineered cooling, and the shallow water experience may transfer to the deepwater environment.

View Image Gallery>>

Pipeline walking

Pipelines subject to lateral buckling may also suffer from pipeline walking. If walking is predicted it may be necessary to install expensive mitigation measures such as pipeline anchors. Modifying pipeline configuration can also reduce the severity of walking. The increased weight of a PIP system, for example, will increase resistance to walking without adding to the thermal transient-induced driving force as the steel cross-section of the sleeve pipe remains at near ambient temperature.

View Image Gallery>>

Installation

Thick wall requirements for pressure containment and to resist external overpressure and bending, especially for pipe-in-pipe systems, result in heavy pipelines which in combination with deep and ultra-deep water in turn lead to increasing lay tension requirements. Lay tensions may need to be increased further due to dynamic effects and to allow for accidental flooding. The latest J-lay vessels have tension capacities of 2,000 tonnes and the latest reel-lay vessels 400-800 tonnes, allowing heavy pipe to be installed in deeper water.

Pipeline design, however, must still consider pipeline configuration regarding installation feasibility, vessel availability, and costs. And even at an early design phase, system behaviors and interactions under severe HPHT loading in a deepwater environment must be fully understood and captured.

View Image Gallery>>

References

1. Maldonado, B., Arrazola, A., and Morton, B., "Ultradeep HP/HT Completions: Classification, Design Methodologies, and Technical Challenges," Offshore Technology Conference (OTC), Houston, May 1-4, 2006.

2. "Submarine Pipeline Systems," Offshore Standard, DNV-OS-F101, October 2013.

3. "Design, Construction, Operation and Maintenance of Offshore Hydrocarbon Pipelines (Limit State Design)," API RP 1111, 4th Ed., December 2009 and errata, May 2011.

4. Politis, N., Banon, H., and Curran, C., "HIPPS-Based Design of Flowlines and Risers," OTC, Houston, Apr. 30-May 3, 2012.

5. Stewart, G., Klever, F.J., and Ritchie, D., "An Analytical Model to Predict the Burst Capacity of Pipelines," Offshore Mechanics and Arctic Engineering (OMAE) Conference, Houston, Feb. 27-Mar. 3, 1994.

6. Stewart, G., Roberts, C., Matheson, I., and Carr M., "Reliability Based Design Optimization of a "No Burst" High Pressure Pipeline," OMAE Conference, Oslo, June 23-28, 2002.

7. "Design Guideline," SAFEBUCK JIP, January 2015 (confidential to JIP participants).

The authors