Flow assurance is an important part of any development. However, for long distance subsea tiebacks, such as West Delta Deep Marine (WDDM), it is a core discipline which has the highest importance.

The development of the WDDM field has been a significant technical challenge to BG in that the gas was located in deepwater with no existing infrastructure in the area. Another challenge was that the development was to be phased. The production rate of the evacuation system was to start at a relatively low rate, but ultimately had to be capable of transporting large volumes of gas. The WDDM concession lies approximately 62.5 miles (100 km) offshore Alexandria in the Mediterranean Sea. Since the concession was awarded in 1995, BG has drilled 16 successful exploration and appraisal wells in WDDM - a 100% success rate. This exploration program has resulted in the discovery of eight gas fields: Scarab, Saffron, Simian, Sienna, Sapphire, Serpent, Saurus and Sequoia.

The initial two phases of the development Scarab/Saffron (on stream 2003) and Simian/Sienna and Sapphire (on stream 2005/6) are located in water depths from 1,640 ft to 3,280 ft (500 m to 1,000 m). The farthest of the fields, Sienna, is located 84.4 miles (135 km) from the existing terminal, which is a greater distance than any existing subsea tieback.

Pipeline sizing

The original concept for the development of the first phase of the WDDM development was to have a dedicated export pipeline for each phase. However, during the detailed design phase significant additional volumes of gas were discovered which required the system to be re-engineered to transport much larger volumes. The combination of several fields into one pipeline system resulted in significant overall cost saving, but gave the development teams a number of issues that had to be solved.
The system had to have sufficient capacity for domestic gas sales contract plus the supply of two trains of liquefied natural gas (LNG). This equated to a design capacity of 1,750 MMcf/d which was later increased to 2100 MMcf/d. The system also had to be able to meet the Scarab/Saffron turndown production requirement of 200 MMcf/d for an extended period of time; it had to be capable of stable production at all gas production rates between these two rates; and it has to have sufficient flexibility to accommodate fluids with different condensate/gas and water/gas ratios.

Initial work on pipeline options was carried out using a steady state model, since this type of model is relatively easy to set up and the results could be obtained quickly. It was recognized from previous projects that the steady state model was good at comparing the relative volumes of holdup and flowing wellhead pressure required, but poor at estimating the actual volume of liquid holdup in the pipeline. This was better achieved using a dynamic simulator, which takes significantly longer to develop and run.
A three-phase transient model was then generated to examine the most attractive identified options.
In all cases it was the pre-compression (onshore and offshore) case that was studied, since this was the worst case for liquid holdup. Once onshore compression is installed, the system will operate at a lower pressure and therefore the gas velocities will increase. This, in turn, will result in the liquids, water and condensate being swept from the system.

Figure 1 shows typical liquid holdup curves obtained from the steady state model for the export system with different export pipeline diameters. The graph shows the clear relationship between liquid holdup and pipeline diameter. It also shows that liquid holdup in the system dramatically increases at the lower flowrates.

Figure 2 shows a comparison of the steady state and transient simulator results for a 36-in. pipeline. As expected, the transient model predicted more liquid - particularly at lower flowrates. Once the basis system design was frozen, only the dynamic simulator was used due to the inaccuracy of the steady state model at lower flow rates.

Design of the evacuation system was a matter of balancing two opposing requirements. First, to minimize the pressure drop between the wells and the terminal, which required the use of larger diameter lines. Second, to have a system that could flow safety at low production rates, which dictated small diameter lines to minimize liquid holdup.

Meeting these two requirements had to be a compromise. The chosen design was to have a system consisting of unequal diameter pipelines. The smaller 24-in. diameter pipeline, would be used initially and at periods of low demand. The larger 36-in. pipeline would be used once the second phase of WDDM development was complete. The two pipelines would then be used together when the highest flow rates are required.

The two pipelines terminate offshore at a pipeline end manifold (PLEM) in a water depth of 295 ft (90 m). It is feasible to install the larger diameter pipelines into deeper water. However, studies show that extending the 36-in. line into deeper water would compromise the flexibility of the system due to the increase in liquid holdup in the system. Pipelines from the PLEM connect to each of the fields, 26-in. to Simian/Sienna and Sapphire and a pair of 20-in. pipelines to Scarab/Saffron. A schematic of the pipeline network can be seen in Figure 3.

Onshore the two pipelines are connected to a pair of 5,000 bbl slugcatchers whose main function is to act as a liquid storage device during ramp up. It is not expected that the onshore facilities will see slugs on liquid during normal operation although there will be liquid slugging in the larger diameter (26-in.) deepwater pipelines.

Turndown and ramp-up

Early in field life, while pipeline pressures are high and gas velocities are relatively low, liquid holdup in the flowlines and pipelines will increase markedly at low pipeline flow rates. Ramp-ups from turndown conditions need to be carefully controlled to keep liquid delivery rates within the liquid handling capacity of the receiving and processing equipment.

The dynamic multi-phase simulation studies were used to demonstrate that the system had a high degree of flexibility to turndown and ramp-up from circa 50% of peak production. After a 25% turndown or below, ramp-up will have to be extensive, especially if the turndown period has been prolonged to enable a substantial increase in liquid hold-up.

The evaluation of ramp-up scenarios, such as the restart of a LNG train, requires a model not only of the onshore facilities but also of the onshore liquid handling facilities. Under most scenarios, the limiting factor is liquid storage capacity (slugcatcher volume) or liquid processing capacity (condensate stabilization capacity). This modification of the dynamic model had the advantage that not only could the onshore equipment be sized correctly, but also the model now predicted the flow of gas, water and condensate and not gas/liquid.

Hydrate prevention

One of the key aspects of flow assurance on a gas development is the avoidance of hydrates forming in the flowlines, manifolds or pipelines. The primary strategy for hydrate control is the continuous injection of inhibitor at rates consistent with maximum potential gas flow rates, a certain level of water production, and the lowest expected temperature in the subsea system and the maximum potential flowline shut-in pressure.

For long distance subsea tiebacks, the most practical solution to this problem is to use either methanol or glycol (MEG) to avoid the formation of hydrates. Hydrate formation is prevented by continuously injecting the chemical into the gas stream at each tree whenever the well is producing. MEG was chosen for the WDDM since the losses in the gas phase were considerably less than for methanol.

For WDDM, the hydrate inhibitor system comprises a number of small diameter pipelines from onshore plant to each of the fields. Within the field the chemical is distributed inside the infield umbilical to each of the injection units located on each of the subsea trees. Corrosion inhibitor is injected with the MEG to avoid the need to supply and distribute this chemical using a separate system.

For the initial phase of WDDM the injection rate was set using remotely operated vehicle-operated valves but, on the most recent phase, the injection rate can be altered from the control room. This allows the operators to react much more quickly to unexpected increases in water production, thereby reducing the risk of hydrate blockage. An important feature that had to be designed into the system was subsea filters, since, although the MEG is filtered before being pumped from onshore after travelling + 62.5 miles (+100 km) in a carbon steel line, its cleanliness is reduced.

Methanol is used to inhibit against hydrates during start up until flow has warmed to help remediate hydrate plugs and to purge vent lines after use. Methanol is transported to the fields and distributed to the wells and manifolds via tubes in the main and infield umbilicals.

Malfunction of the hydrate inhibitor injection system could result in hydrate formation somewhere in the production system. Flowing conditions are monitored for indication of the onset of hydrate formation in order to avoid the downtime that depressurization to melt a hydrate plug would require. Practical indications of hydrate formation include an increase in the wellhead pressure drop, large pressure fluctuations and/or a decrease in the flow of gas. In the WDDM system, there are two downhole pressure measurements, two upstream and two downstream (of choke) pressure measurements, one downhole venturi meter and the inference meter system. All are monitored by the control system, which will indicate abnormal pressure and flow conditions arising from the initial hydrate formation.

The first phase of WDDM, Scarab/Saffron, has now been in production for over 12 months. To date, the system has reacted as per the transient model which allows the Simian/Sienna and Sapphire development team to be confident in using the results from the enhance model.

Operational experience

To assist the operators in managing the offshore system, a dynamic production system simulator (PSS) has been provided to monitor flowing conditions and to predict future response to system control actions. The system is able to predict the volume and timing of liquid surge resulting from a change in production rate and alert the operator to wellhead pressure increases and/or production rate decreases due to hydrate accumulation. This system will be upgraded each time a well is added and will also be benchmarked against actual measured response of the system.

Conclusion

The WDDM development scheme has generated new challenges being one of the first of the new generation of high gas volume long distance tiebacks.
Significant effort has been applied in examining the flow assurance aspects. Although this has been time consuming, it has given BG confidence in the systems performance and sufficient assurance to support the investment in further phases of development.