David Kaye, James Ingram
Andrew Palmer & Associates Ltd.David N. Galbraith, Russell Davies
Mobil North Sea ltd.
Aberdeen
A new procedure for assessing and rectifying subsea pipeline freespans was successfully used in 1992 in Mobil North Sea Ltd.'s Beryl field and the Scottish Area Gas Evacuation (SAGE) pipeline.
The span assessment method is in two parts, each part having two stages, and consists of preliminary stress and vibration frequency checks followed by detailed strain and fatigue life checks where appropriate.
Comprehensive software, automatically linked to an inspection data base, has been written to allow efficient use of the methodology.
Results from the freespan assessment indicate that the assessment procedure, and in particular the strain based and fatigue analyses, gave significant savings in terms of reduced number of freespans for rectification.
Critical freespans were stabilized by grout bags positioned by an ROV. The ROV based system enabled both risks and costs to be reduced in a normally hazardous and costly environment, and utilized technology already on board a pipeline inspection vessel.
The overall freespan assessment and rectification program represents a significant step forward for Mobil North Sea in terms of reducing costs, while simultaneously improving the speed and simplicity of freespan assessment.
The system provides the possibility that future freespan rectification works may be performed in a single offshore program, which includes pipeline inspection, survey, assessment, engineering, and repair of all freespans from a single survey vessel.
SAGE AREA
The Beryl A platform was installed in Block 9/13 in 1973. Subsequent developments in the block include the Beryl B platform; the Ness, Bwiss, Linnhe, and NESS II multiwell subsea developments; and four single subsea wells.
A total of 25 pipelines are laid in the block. Most are 6-in. OD flow lines from the subsea wells to the platforms but include one 16 in. and one 20 in. hydrocarbontransfer line between the production platforms and two short 36 in., oil export lines to loading buoys.
All the facilities noted are operated by Mobil North Sea Ltd. on behalf of the Block 9/13 co venturers Amerada Hess Ltd., Enterprise Oil plc, BG North Sea Holdings Ltd., and OMV (UK) Ltd.
Additionally, the recently installed SAGE (OGJ, Mar. 8, 1993, p. 37) gas export line runs 323 km (200 miles) from Beryl A to St. Fergus and is operated by Mobil North Sea on behalf of the Beryl and Brae groups.
The fines are inspected annually in compliance with regulatory and Mobil North Sea requirements and for continued fitness for purpose. As with most pipelines, the annual surveys identify numerous freespans which must be assessed and, if necessary, rectified.
Mobil North Sea has now implemented several improvements to its methods for collecting and recording pipeline inspection data and for automatically assessing the significance of freespans in relation to the service, pressure, temperature, and orientation of pipelines.
The methods of data storage and freespan assessment are based on PC programs for use onshore and at sea.
The basis for recording and displaying the inspection results is a data base system (Coabis). Survey data can be entered into the system in real time on the inspection vessel as the inspection is performed.
Further checks can be undertaken on the data in the office and include graphical comparison of all features from one year to another.
The benefits include the ability to correct mismatches iii datum locations between surveys by reference to fixed features (for example, flanges, anodes, and major debris) and a visual representation of how freespans, buried lengths, and any deterioration vary from year to Near.
The freespan assessment procedures are incorporated into a PC based suite of freespan assessment programs. The data base and freespan-assessment software are linked so that there is no manual transcribing of data following their entry into the data base on the vessel.
This extent of automation eliminates the previous errors made during manual transcribing of data and also reduces the time required for freespan assessment.
The freespan assessment procedure consists of two separate stages. The first stage is designed to be undertaken by Mobil North Sea's inspection representative on the vessel and can be used to identify the more critical freespan@ where additional measurements (natural frequency, for example) may be beneficial.
This enables more data to be collected on these freespans, optimizes utilization of the inspection vessel, and may also reduce the extent of freespan rectification that is subsequently recommended, using the techniques described presently.
Before implementation of the system described here, selection of freespans for rectification occurred by reference to a "limiting value" for each diameter of pipeline. Calculation of the limiting value was based on the worst possible combination of conditions for any pipeline of that size.
The change in assessment techniques was prompted by, a belief within Mobil North Sea's engineering department that the technique was overly conservative with excessive numbers of freespans being identified for correction. Indeed, for a number of Nears Mobil North Sea applied "engineering judgment" to the requirements for rectification.
FREESPAN ASSESSMENT
Pipeline freespans are at risk from damage by one of two distinct mechanisms: by excessive bending from externally applied hydrodynamic or self weight loading, or by long term fatigue damage, from flow induced vibrations.
The traditional means of ensuring that excessive bending deformation cannot occur is to define a maximum allowable freespan length such that the maximum equivalent stress in the freespan is less than an acceptable fraction of the pipeline steel's yield stress.
This approach has been used extensively for pipelines in the North Sea and elsewhere. The approach is conservative, however, in that it fails to recognize the post yield strength of a pipeline freespan and does not recognize that a pipeline can remain perfectly serviceable and fit for purpose even though the part of the pipeline steel may have exceeded the yield limit.
The traditional means of ensuring that flow induced vibrations cannot occur is to define a maximum allowable freespan length such that the natural frequency of the freespan is too high to allow any flow induced vibrations to develop.
Again, this approach has been used extensively in the North Sea and elsewhere but is also conservative. The approach prevents any form of freespan vibration and does not recognize that modest freespan vibrations may not cause fatigue failure.
More sophisticated and less conservative freespan-assessment methods have been introduced in recent years. Modern pipeline design codes, such as the recently, issued BS 8010 Part 3,1 explicitly allow the more sophisticated approach.
In the freespan analysis program described here, the simple stress and vibration criteria described previously are used for a first pass analysis to identify the more critical freespans which are then examined in more detail.
These freespans are re analyzed with more sophisticated strain based criteria to assess bending deformation and fatigue criteria to assess the extent of fatigue damage induced by freespan vibration.
The overall flow chart for the freespan assessment is described in Fig. 1. The preliminary and detailed freespan assessment methods are described in more detail presently.
The preliminary freespan assessment consists of checks for overstress in the freespan and the possibility that vortex induced vibrations may occur. These are also described presently.
STRESS ASSESSMENT
The preliminary freespan analysis performed in this study is based on a freespan model described by Palmer and Kay.2 The model considers the effect of the axial force in the freespan, tension induced by the sag of the freespan, and partial foundation support at the freespan ends.
Most seabed soils in the North Sea provide only partial foundation support at the ends of a freespan. In fact, the pipeline embeds into the soil at each end of the freespan.
The point at which the freespan is subject to complete end fixity is usually some distance along the seabed beyond the end of the freespan.
This end fixity controls the rotational and vertical stiffness of the freespan supports and affects the bending moments within the freespan itself.
For bending moment calculations in this study, the end supports were assumed to be completely fixed, that is, "clamped crimped" end conditions.
This assumption provides a conservative estimate of end moments over a wide range of typical soil conditions in the North Sea.
The maximum bending moment in the freespan, as a result of both the pipeline's self weight and external hydrodynamic loading, is calculated with beam column theory, taking into account the effective axial force in the freespan. This force includes both the force in the wall of the pipeline and the force resulting from the pressure of the pipeline's contents.3
Additionally, the effect of tension generated by the sag of the freespan is also taken into consideration. This effect can be significant for long freespans and has an important effect on both the axial stresses in the freespan and the freespan's natural vibration frequency.
Hydrodynamic forces are calculated with Morison's equation with appropriate force coefficients. The maximum bending, axial, and hoop stresses in the wall of the pipeline are then calculated, and the greatest von Mises equivalent stress in the freespan is evaluated.
This stress is compared against the yield stress of the pipeline steel:
- If the ratio of the maximum equivalent stress to the specified minimum yield stress (SMYS) is less than 0.96, the freespan is concluded to be safe.
- If the ratio is greater than 0.96, then the freespan is unacceptable on the preliminary stress criteria and requires further analysis.
REDUCED VELOCITY ASSESSMENT
A pipeline freespan can be subject to flow induced vibration from vortex shedding from the freespan. Vortex induced vibration is predominantly controlled by a dimensionless parameter called the reduced velocity written as shown in Equation 1 in the accompanying equations box.
Generally, relatively small amplitude vibrations of the freespan in line with the flow direction occur at values Of VR greater than 1 and peak at VR between 2 and 3, while larger amplitude vibrations across the flow direction begin at VR around 3 and peak at VR of about 5.
The traditional design approach is to adopt a critical VR which must not be exceeded in design storm conditions. The selection of the critical reduced velocity and the appropriate storm velocity is open to engineering judgment, however, and the application of this approach is often inconsistent.
In the analysis described here, the critical reduced velocity has been taken as equal to 3.5, in which the incident velocity (V) is the sum of the maximum wave-induced velocity and the maximum current velocity.
In order to assess whether vortex induced oscillations will occur, it is necessary to calculate the freespan's natural frequency (fN).
In this analysis, the natural frequency, is calculated by modeling the freespan as a clamped clamped beam under an axial force which yields Equation 2.
In order to model the finite foundation stiffness at each end of the freespan, LE in Equation 2 is taken as 1.1 multiplied by the observed length of the freespan.
The freespan frequency can be a difficult quantity t@ predict accurately. To investigate this further, the natural frequency of longer freespans was measured during the 1992 survey program with an ROV mounted accelerometer package.
Fig. 2 shows the calculated natural frequency for various freespans in (he 30 in. SAGE gas export pipeline plotted against the measured frequency,. The figure shows the calculated frequencies for isolated freespans only.
The general agreement is clearly good, despite some inconsistencies in the data. The agreement for multiple freespans is noticeably worse, however.
Although not shown in Fig. 2, the comparison for multiple freespans suggests that predicted frequencies for multiple freespans strongly depend on the extent of support provided by intermediate touchdown points.
The detailed freespan assessment consists of a check for excessive strain in the freespan and the calculation of the fatigue life of the freespan,
STRAIN ASSESSMENT
The serviceability of a pipeline which fails the overstress check previously described will be unaffected provided that the post yield deflections of the freespan are not excessive.
The BS 8010 design code,1 for example, allows yield due to large bending stresses provided that certain requirements on the plastic strains, weld ductility, and diameter wall ratio are satisfied.
These conditions can often be met by subsea pipeline freespans where the seabed provides a boundary which prevents excessive deformation of the freespan. Once the freespan touches down, it becomes essentially two freespans separated by a single support.
The central touchdown limits the bending deformation of the original freespan and provides a support which reduces the stresses in the subsequent double freespan.
The greatest possible bending strain which could occur can be calculated from the gap below the freespan and the length of the freespan. This maximum possible strain is then compared against two limits, based on the following:
- The maximum allowable strain governed by weld ductility and steel properties
- The buckling strain on the compressive side of the pipeline, which is a function of the pipe diameter/wall thickness ratio.
If either of these limits is exceeded, there is a risk of excessive pipeline deformation and the freespan must be rectified.
FATIGUE ASSESSMENT
A detailed and accurate assessment of the consequences of flow induced vibrations of the freespan can be performed by predicting the amplitude of freespan vibrations over the freespan life and evaluating the subsequent fatigue damage.
The fatigue calculation is performed by first generating a probability distribution of the incident wave and current velocities over the design life of the freespan.
For waves, this is usually provided by a scatter diagram. For currents, the data are often much more difficult to obtain but can often be estimated from the statistical distributions of the tidal and storm current components.
For a given freespan, the incident reduced velocity can then be calculated for each wave and current combination in the environmental data set. The amplitude of freespan vibration can be predicted by interpolation of experimental measurements of freespan vibrations.4
The prediction of the expected amplitude is complicated by the presence of the oscillatory wave component superimposed on a steady current. This can be resolved by reference to more extensive data or (as followed in this study) more detailed modeling of freespan vibrations with steady and oscillatory flow.
The effect of wave induced oscillations is also included in the analysis. In this case, the wave induced displacement was estimated from the oscillatory hydrodynamic force on the freespan with an allowance for any resonance of the freespan.
For each wave current combination, the oscillatory stresses induced by the applied vortex displacement or wave loading can be predicted from beam column theory. The corresponding fatigue damage at that stress range is calculated from S N curve fatigue data.
If the number of cycles to failure for a particular wave-current combination (i) is given by Ni, then the total fatigue damage for that combination is shown in Equation 3.
The total fatigue damage for all wave current combinations can be calculated with Miner's rule as the sum of all particular combinations (Equation 4).
If the total fatigue damage is greater than 1, then the freespan is deemed to be unacceptable and rectification should be considered. If total fatigue damage is less than 1, the freespan does not require correction, but it would be prudent to monitor the freespan during subsequent surveys to determine if it will grow with time.
It should be noted that spans can reduce with time and the rate of fatigue damage may decline.
ASSESSMENT RESULTS
Figs, 3 5 show the results of the inspection and analysis program for the SAGE gas export pipeline and the Beryl B to A 16 in. gas and 20 in. oil transfer lines.
Each figure shows the total number of freespans observed during the survey, the number of freespans that failed the preliminary analysis, and the number of freespans that failed the detailed analysis.
The freespans have been grouped into 10 m lengths to help identify which freespans have failed. Results are summarized in Table 1.
The portion of freespans within each length group which failed the preliminary and detailed analyses increases with increasing freespan length. Several short freespans were observed during the survey, but of these freespans none failed either the preliminary or detailed analysis.
As the freespan's length increases, the total number of freespans observed decreases; of the observed freespans, an increasing proportion failed the preliminary analysis and an increasing proportion of these freespans failed the detailed analysis.
None of the freespans in the 6 in. flow lines failed the detailed analysis or required rectification.
The cost saving that results from using the revised freespan analysis is demonstrated by Figs. 3 5 and by Table 1. For example, Fig. 3 shows the results for the 30-in. SAGE pipeline (Beryl gas export to St. Fergus).
If the freespan failure criteria were based on the worst case limiting freespan length from the preliminary analysis, which is typical of the previous Mobil North Sea approach, then all freespans greater than 30 m would require rectification.
This gives a total of 166 freespans. Clearly not all these freespans would be stabilized, but the selection of appropriate freespans for stabilization would then be based largely on intuitive engineering judgment rather than engineering calculation.
If the freespan failure criteria were based on the preliminary analysis but applied to each individual freespan, the total number of freespans requiring rectification would be reduced significantly to 20. Of these freespans, only 11 failed the detailed analysis described previously.
The total number of freespans requiring some form of stabilization is therefore reduced from 166 to only 11. Many of these freespans failed assessment by fatigue only. Where the remaining life extends over several years, rectification could be delayed and advantage taken of any natural backfilling of the line.
Similar conclusions can be made for the Beryl B to A oil and gas transfer pipelines in Figs. 4 and 5. The trends are rather less clear because of the lower total number of freespans in these two pipelines, but the revised assessment procedure again provides a better and less conservative basis for the selection of freespans requiring some form of rectification.
STABILIZATION TECHNIQUES
The normal techniques for stabilizing spanning pipelines are rock dumping, grouted support, or gravelcement bag supports. These techniques are normally installed and deployed by either a custom built vessel, such as rock dump vessels, or diving support vessels.
Both types are extremely costly with associated high mobilization costs.
The decision was made to consider not only these types of vessels but also new techniques such as ROV systems. The cost of running an ROV based vessel capable of remote maintenance operation is approximately 25 33% that for a diving vessel.
In addition, it has been a Mobil North Sea initiative over the past few years to remove the diver from as many subsea applications as possible to reduce hazards and costs.
A technical and cost comparison was made of the three options and of remote maintenance using ROV systems.
The costs (Table 2) are indicative and based on a Mobil North Sea cost estimate of critical freespans on both the SAGE and Beryl B to-Beryl A pipelines. It was assumed that a total of 12 freespans required support, split equally into two locations approximately 200 km apart.
On the basis of the cost comparison, the ROV based techniques were eventually chosen and used successfully in 1992.
STABILIZATION PROCEDURE
The critical freespans requiring rectification in the 1992 pro,ram varied greatly in length and height off the seabed. In particular, the height of freespans ranged 75 250 mm off the seabed.
The system of deploying and installing the grouted support had to deal with both extremes of freespan height.
The chosen method of rectification was to deploy a grout bag and position it below the pipeline. The bag was then inflated with grout which, when cured, provides a rigid support to the line.
The vessel used for the freespan rectification program was the Kommandor Subsea operated by SubSea Offshore Ltd. This vessel had already been charted for the SAGE pipeline inspection program and was ideally suited for ROV operations.
The only additional equipment required was a grouting spread and launch and-recovery system for the ROV installed grout bags.
A standard grouting spread was employed to mix and pump a grout mixture specified by Mobil North Sea.
The grout design was based on achieving a minimum compressive strength within a 28 day period.
The grout bag support was designed to ensure that it be inflated correctly and that it be easy to use for the ROV and deployment frame system. The bag itself was built from industry standard fabric.
The operation was carried out in five operations:
- Grout support location. The ROV' was deployed to survey the freespan and mark the location on the pipeline where the proposed grouted support should be placed.
This marking was made by use of a white marker chain visible by ROV sonar and cameras.
- Grout-bag deployment. The grout bag was deployed on a circular swivel deployment frame, launched from an A frame over the side of the Kommandor Subsea.
The grout bag deployment frame and rout hose umbilical were then lowered to the seabed surface, landing approximately 5 10 m from the pipeline support location.
- Grout bag installation. The deployment frame was turned on its swivel base by ROV to ensure that the grout bag's leading edge was facing towards the pipeline. The ROV was used to push a needle attached to the grout bag under the freespan and then to pull the bag through on the other side of the pipeline and pull the bag away from its deployment frame.
This operation was carried out slowly to ensure the accurate placement of the grout bag under the pipeline.
- Grout bag inflation. Grout was pumped through a 58 mm bore grout hose which was attached to the grout bag via a quick release connector.
Once the desired support shape was achieved and grout was seen venting from the top the bag, grouting was stopped and the quick-release connection broken by the ROV.
- Deployment frame recovery and post installation survey. The deployment frame was pulled away from the location to allow flushing of the grout hose; the deployment frame was then recovered back to the vessel.
The ROV then made its final post installation survey, of the freespan and the newly, installed support, recording positional and visual data.
The total duration for these five operations, based on a grout bag volume of 1.5 cu m, was approximately 3 hr depending on weather. Approximately four grout bags could be placed in close proximity, within a 12 hr shift.
ACKNOWLEDGMENTS
The authors wish to acknowledge the assistance of Stephen Booth and Andrew, Palmer for advice on the work described in this article.
REFERENCES
- British Standard, Institution, BS 8010 "Code of Practice fir Pipelines, Part 3 Pipelines Subsea: Design Construction and Installation," February 1993.
- Palmer, A,C., and Kaye, D., "Rational Assessment Criteria for Pipeline Spans," European Seminar Offshore Pipeline Technology,, 1991.
- Palmer, A.C., and Baldry, J.A.S., "Lateral Buckling o( Axially Compressed Pipelines," journal of Petroleum Technology, No. 26 (1974), pp. 1283 84.
- "Vibration of Pipeline Spans," Report No. EX1268, December 1984, Hydraulics Research, Wallingford
- "Rules for Submarine Pipeline Systems," Det norske Veritas, DnV 1981.
Copyright 1994 Oil & Gas Journal. All Rights Reserved.
