PIPELINE TECHNOLOGY Rocky road to Mongstad for Troll pipeline engineers

Project team overcomes seabed topography
and strong currents as laying nears conclusion

J.Breivik
Statoil
R.Bruschi, G.Leopardi
Snamprogetti

Troll Oljeror is one of the most demanding pipeline developments undertaken on the Norwegian shelf. Water depth (up to 540 metres), seabed conditions (rocky outcrops and deep depressions for more than 5km), and considerable on-bottom current velocities (up to a metre per second) at the entrance of the Fensfjorden are the issues that make this such a challenging crossing.

Troll is located in the Norwegian trench at water depths of 310-345 metres, approximately 70 km north west of Bergen. Development is divided into phases, the second of which covers the oil reserves in the Troll West Oil Province.

The concept for Phase 2 is a single floating production unit (FPU) serving as platform and field centre connected to associated subsea systems, with Troll Oljeror carrying the stabilised crude from the FPU to Mongstad. Statoil is the operator of this line both in the development and operating phases. Installation is currently in progress with a planned start-up in November.

Starting point for the 86 km long pipeline will be the tie-in point to the flexible riser near the FPU. It will be routed in a direct course to Mongstad, landing at the toe of a steep slope where it enters a 30-in. borehole to climb up onto the refinery fence. It will thus be the first pipeline to be installed in a deep Norwegian fjord, Fensfjorden.

At Mongstad, there will be a pig-trap, associated valving, and a process flow meter for the oil before it is transferred to an existing dedicated storage cavern.

Design capacity is 25,000 cu m/d which can be expanded to 30,000 cu m/d through use of flow improver. Design pressure is 167 barg for an inlet temperature of 60C while the operating pressure/inlet temperature is currently set at 155 barg/60C (Base Case).

Route stages

The pipeline route can be divided into six sections with different characteristics: Norwegian Trench, Western Slope of the Sill, Top of the Sill, Eastern Slope of the Sill including the Deep Depression where the maximum water depth of 540 metres is reached, Inner Fensjorden and Landfall. Water depth in the Norwegian trench is mainly 320-335 metres with a maximum of 450 metres. At the entrance to the Fensfjorden, the water depth reaches the maximum of 540 metres, declining slightly to 530 metres maximum through the fjord.

Seabed topography of the Sill area and its eastern slopes including the deep depression is very irregular, which necessitated a very detailed, precise survey to identify the optimum pipeline route. The engineering team had to take into account strong bottom currents of up to 1 metre/s, caused by the combination of the density gradient between the Atlantic water and the Norwegian coastal current and the flux balance between the Norwegian coastal current and the inner fjord waters.

The most advanced technology available was used to document the seabed condition along the route, including geotechnical information. A section of the pipeline route (Kp 60 - 65) is on a very uneven seabottom characterised by outcropping bedrock, steep lateral and longitudinal slopes, and seabed varying from soft clay to boulder areas.

A long baseline underwater positioning system (LBL) of more than 30 subsea stations has been established for the most difficult section of the route (Kp 61 - 64.5) and at landfall, where extensive pre-lay intervention works are to be performed prior to pipeline installation.

This LBL array was considered the best possible common positioning reference for documentation of such unusually uneven seabed topography. The precise description of the seabed morphology is the basis for construction of the gravel supports to pipeline installation.

Thorough soil investigations comprising more than 150 theoretical piezocone penetration tests (PCPTs) and 200 soil samples have been carried out along the pipeline route. Additional investigations were performed in November 1993 and June 1994 along the optimised route in the Fensfjorden: the survey covered critical gravel sleeper locations and possible unstable sediment pockets at the steep sides of the Fensfjorden.

Soil uniformity

Based on field and laboratory tests, the main areas have been divided into sections where the soil condition is fairly uniform: the offshore, coastal and fjord areas.

The offshore part of the route including the Norwegian Trench and the Western Slope of the Sill show fairly similar conditions with soft to very soft homogeneous clay. At most, two layers are revealed with depth. At some locations a very thin layer of loose sand is encountered at the seabed.

The remaining Sill and fjord area, which differ greatly in water depth and topography, show more variation in the soil condition: mostly soft to very soft clay, but with thin sand/silt layers at some locations.

From the deep depression and westward, very small sediment thickness is encountered mainly less than a metre. Rock outcrops at the seabed along parts of the pipeline route in this area.

Several studies have been performed to identify and quantify potential geotechnical hazards in the Fensfjorden. These relate primarily to potential instability of subsea soil deposits along the pipeline route and in the adjacent fjord sides where released slide masses may affect the pipeline integrity.

To evaluate the submarine slide hazard, the following factors were considered: seabed topography in the fjord; sedimentation during deglaciation period; present situation with respect to sedimentation; development of highly sensitive clays (quick-clays) due to leaching of salt; presence of loosely-deposited silt or sand layers; level of future potential triggering factors; and earlier slide activities.

Earthquakes are considered the most likely potential triggering factor for a submaring slide in the Fensfjorden. From the analyses, for the 10(-2)/yr earthquake event the margin of safety for instability-induced settlements is in general high.

For the 10(-4)/yr earthquake event, permanent displacement and relative slippage of the soil deposit may occur, probably of the order of 5-20 cm for the slopes in the shore approach area and 10-60 cm for the outer part. The slopes will remain stable after the 100/yr earthquake event and the NPD requirements concerning overall stability are satisfied.

Results suggest that major submarine slides are unlikely to affect the pipeline during an earthquake. Some areas will, however, be excluded from anchoring to avoid any potential triggering event for slides. The risk of the pipeline being hit by a sliding/rolling boulder is negligible except for one section where the pipeline will be protected by gravel berms.

Current monitoring

Two current meter moorings were installed for two months in spring 1992 to monitor the current regime in the Fensfjorden. Conditions in the Norwegian Trench were fairly well known from other current measurement programmes.

Measurements at Mongstad showed very low speeds and for long periods no current at all, but in outer Fensfjorden just inside the sill where the water depth is 540 metres, bursts of water inflow were experienced. These inflows had a duration of two to three days and peaked over 0.5m/s. The flow direction was East during the inflows.

Many questions were raised after these first measurements. Was it a seasonal phenomena? Was the measured speed a maximum, or could one expect higher speeds? Was the direction absolutely topographically steered? What would the 100-year design current be? To provide answers to these questions a comprehensive programme of further current measurements and model studies was established.

Altogether 16 rigs with 28 current meters have been operating from one to over 18 months. The main concentration of instruments was in the Sill area, where the highest current speeds were expected and where the pipeline is designed to have several free spans. Maximum measured sea bed current was 0.8m/s in an easterly direction and a m/s in a westerly direction over the Sill top.

The time series of the measurements were analysed with different methods to find the best fit for calculating the design current normal to the pipeline sections. Since the current was mainly topographically steered and the pipeline follows the topography, the design current normal to the pipe was much lower than along the pipeline.

The complex current regime in the entrance to the Fensfjorden has been modelled both in the laboratory and by numerical techniques to determine the maximum possible/probable bottom current velocity along the pipeline route.

Both sets of simulations are based on a hydraulic model concept with three layers of density as observed in the site. The model is forced by specifying the in- and outflow in the various layers. Actual topography, and real time forcing, are included in both models.

The strong current in-flow events which have been observed through the current measurement programme have been successfully reproduced by the models. The source for the heavy water that flows over the sill is the Atlantic water in the bottom of the Norwegian Trench. This water is lifted dynamically when whirls in the surface layer pass from south to north in the Norwegian coastal current.

Spanning boundaries

Due to the very rough rocky seabed conditions at the entrance to Fensfjorden for a distance of 5km, the expected pipeline profile is characterised by a series of free spans with intermediate supports, or `multi-spanning'. In this situation, boundary conditions appropriate to each individual span vary considerably and there is no simple relationship between span length and natural frequency as is generally the case for isolated free spans.

The uncertainty in the estimated natural frequency is mainly due to uncertainties in the bending stiffness, mass, span length, effective axial force and in the heights, angles and support lengths at span shoulders. Each of these parameters is modelled statistically (in terms of mean, standard deviation and statistical distribution) and model uncertainties are introduced. The relative importance of each of these parameters is assessed through the initial deflection - amplitude frequency relationship.

Another major challenge for the project team has been to design intervention works either to correct the seabed before the pipelaying or to freeze the as-laid configuration after laying.

Seabed preparation works before pipelaying have been identified to achieve an as-laid configuration of the air-filled pipeline such that static stress levels are within allowable values, and no-cross flow oscillations occur during the air-filled condition of the pipeline. Seabed preparation before pipeline installation has also been necessary to avoid overstresses in operating condition due to an excessive curvature of the as-laid pipeline.

Post-laying intervention works consist of supporting pipe sections along the as-laid (air-filled) pipeline to avoid unacceptable dynamic response of the residual suspended sections when exposed to current.

Static and dynamic analysis have been performed to identify the location and extension of the free span correction work. About 30 gravel sleepers need to be installed before laying the pipeline. Maximum heights are 3.5 metres and the required total volume of gravel is around 20,000 cu m.

The pipeline route within Fensfjorden includes five horizontal curves with a radius of less than 1,200m and with a minimum radius of 650m. Installation analysis has concluded that counterblocks should be installed along these curved sections of the route to prevent possible horizontal sliding of the pipeline during installation.

A low on-bottom residual tension in the pipeline is necessary to reduce free spanning lengths and seabed clearance. For this purpose, the stinger of the laybarge Castoro Sei was extended 13m to achieve low residual laypull.

Extensive computerised on-bottom analyses, static and dynamic, were performed on the performance of the pipeline installation on gravel supports, pre-built to remove the longest free spans after laying envisaged at the design stage.

Of paramount importance were studies to optimise the pipeline configuration in the deep depression located at the eastern toe of the Sill. These studies allow a special laying procedure to be specified for a pipeline configuration consisting of fine tuned overweight pipeline sections.

A thorough assessment was made of the effect of variation in lay tension to define a lay equipment setting which allows excessive movements of the pipe to be avoided at the touchdown point. It was concluded that any intereference could be avoided, limiting the lower value of the lay pull dead band and installing pre-lay gravel berms over which the pipeline will be laterally supported to change laying direction.

A guided laying procedure is required to install the pipeline on the centreline of the theoretical route with very high accuracy. Normal monitoring of the on-bottom position during laying was specified as achievable by a Super Short Base Line positioning system, except where a Long Base Line system is ROV-assisted for continuous monitoring of the touchdown point.

Mongstad approaches

The topography of Fensfjorden at Mongstad is quite uneven in the deepest central part with soft sediment extending up to the feet of the steep side slopes. The slope to Mongstad features a very steep route cliff interrupted at a few locations by steep canyons heading laterally with regard to the fjord direction and towards shallower waters. The bed of the canyons is generally bedrock sparsely covered with sediments.

The recommended solution was directional drilling from land to route the last section of the pipeline directly to Mongstad. Advantages included reduced environmental impact; improved pipeline on-bottom configuration in the borehole approach area; reduction of the overall route length; no requirement for diverless tie-ins.

The borehole is being drilled through hard bedrock, and runs from the subsea exit at 314 meters water depth to an onshore point of two meters elevation at the terminal site. Installation of the pipeline is due to take place with a conventional pull-in operational with the laybarge positioned in front of the borehole exit and laying the pipeline which, in turn, is pulled to Mongstad through the borehole by means of an onshore winch.

A gravel embankment is envisaged in front of the borehole exit to achieve an artificial pull-in ramp aligned with the borehole axis. The embankment will ease the entrance of the pull-in head and the pipeline and will avoid high bending of the pipeline free spanning outside the borehole.

Due to the steepness of the seabed slope and weak nature of the surficial sediments, the stability of the embankment is a major concern and the complete removal of the soft clayey layer of about 2m thicknesses down to the bedrock before gravel dumping is required to meet the stringent criteria for such a load-bearing structure.

This is an edited version of a paper delivered at OMC Ravenna in March. Early in June, a section of the 16-in. line was accidentally dropped, an incident which is likely to delay completion of the pipeline.