CUBE—A New Technology for the Containment of Subsea Blowouts

After the Macondo incident, the oil industry started a significant review of its capabilities to respond to a subsea well blowout and launched some specific development projects.

jpt-2013-01-cubefig1hero.jpg
Fig. 1: Overall configuration of the Separation Dome: liquid-discharge pipe/riser, LP; gas-discharge pipe/riser, GP; gas-demisting section, GD; perforated head, PH; short pipe, SP; liquid-collection section, LC; and inlet chamber, IC.

After the Macondo incident, the oil industry started a significant review of its capabilities to respond to a subsea well blowout and launched some specific development projects. Eni decided to participate in the principal joint industry projects launched on this topic, and also started internal activities aimed at the development of new containment methods.

Two years after Macondo, it is clear that the major containment method of a subsea blowout relies on the use of a capping stack. This can be deployed in water depths as deep as 3000 m and equipped with multiple outlets that allow liquid and gaseous hydrocarbons to be routed to surface.

However, in some cases well capping cannot be applied when a) well-­integrity issues arise, this may happen when the capping stack causes a substantial pressure increase at the wellhead, with a potential rupture of the casing and the lateral dispersion of well fluids; b) the blowout preventer (BOP) stack is not accessible because of mechanical problems (status of wellhead/BOP or irremovable debris); c) the blowout does not come directly from the well (i.e., from nearby sea lines or mud line). On the basis of these and similar considerations, it is quite clear that the containment of a blowout event cannot only rely upon the use of a capping stack and other methods must be available.

Eni has been working in the last 2 years on a new concept with the support of TEASistemi, a research company of the University of Pisa, developing an alternative to the capping technology. This system is based on the use of an open-sea containment system, which is able to perform at least partial gas/liquid separation. After separation, most of the liquid hydrocarbons plus a minor amount of seawater are pumped to surface. The gas phase can be sent to surface through a separate riser or can be discharged at sea. In order to prevent hydrate formation within the separation system, there must be a low separator-residence time. The system conceived by the authors of this paper has been patented in June 2010 (De Ghetto and Andreussi, 2010) and the first tests were performed in September 2010 at the TEASistemi Laboratory in Pisa.

Of course the concept of subsea phase separation is not new for the oil industry, but up to now it has not been adopted for the containment of subsea blowouts. The reason probably being that up to the Macondo incident, it was believed that subsea collection and transmission to surface of blowout fluids could be easily accomplished by some type of containment dome. Unfortunately, the Macondo incident showed that this is not always the case and that the use of a containment dome presents a number of potential problems which need to be faced, such as the formation of hydrates, the poor control of the system, and finally, issues associated with large fluxes of seawater into the dome.

The system described in this article tries to solve the main problems listed above with the objective of developing a possible alternative to the capping technology. In this respect, the separation dome proposed by Eni presents potential advantages in terms of ease of installation and effectiveness and can be an excellent option for the containment of a subsea blowout; in particular, when a riser from a drill rig can be used and oil arriving at surface can be properly collected and transported to shore.

Separation Dome

The system presented in this article has been named CUBE, which stands for Containment of Undersea Blowout Events. Its shape has evolved into a truncated pyramid. In the two cases, the main separation section is similar.

The basic elements of CUBE are shown in Fig. 1 and can be summarized as follows:

  • The jet-inlet section, made of a short pipe, SP, with a perforated head, PH
  • The gas-demisting section, GD
  • The gas-discharge pipe/riser, GP
  • The liquid-collection section, LC
  • The liquid-discharge pipe/riser, LP

The blowout jet enters into the containment system through the short pipe (SP) which connects the open sea to the separation section. The SP is terminated with a perforated head, PH, which ensures a uniform jet distribution to the separation section and reduces the momentum of the incoming jet. In the truncated pyramid configuration, an inlet chamber, IC, is added to the containment system.
In the upper part of the separation section, most of the entrained liquid is separated from the gas due to gravity. The gas then flows to the upper outlet pipe, GP, from where it can be routed to a gas riser or discharged into the sea by means of one or more control valves. The liquid phase is collected in the lower part of CUBE from where it is extracted by means of one or more variable speed pumps. The internal arrangement of CUBE can vary from the simple scheme shown in Fig. 1 to a more complex scheme which adopts different internals to perform at least partial water separation.

jpt-2013-01-cubefig2.jpg
Fig. 2: Separation Dome under operation in the laboratory.

The overall configuration of the Separation Dome shown in Fig. 2 is quite different from a conventional gas-liquid separator, mainly because there is direct fluid communication between the open sea and the separation section. Hence, it is not possible to design this system with the same methods adopted for a conventional gas-liquid separator. Nonetheless, the system has been shown to be capable of providing simple and stable operations. This is because the pressure within the separator is relatively constant and is equal to the external pressure at sea bed, with only minor variations due to the jet momentum and buoyancy. If the outlet gas flow becomes blocked, the gas accumulates in the separator, pushes the liquid out of the separator and eventually arrives at the liquid pumps. This requires the outlet gas flow to be facilitated, as in this design, by the presence of a gas riser or by means of a valve, which controls gas discharge into the sea.

The separator level is maintained by direct communication between the open sea and the separator: if the liquid flow to the pumps is increased, more sea­water will enter the separator and the level will undergo a minor variation. Or, if the pump speed is reduced, less liquid will enter the separator and eventually some liquid hydrocarbons will be released to sea. Significant level variations are not expected in this case.

Potential problems are foreseen for CUBE during startup and with the risk of hydrate formation. Initial testing, however, showed that the liquid pump was particularly useful during startup, as the pump can be switched on before the containment system is placed above the blowout jet. This allows the liquid and gaseous hydrocarbons to be immediately removed as soon as they enter the CUBE. If a gas riser is used, at startup the riser should be empty. If the gas is discharged into the sea through a control valve, no startup problems are expected.

Hydrate formation is of course possible, as it happened in Macondo. According to a kinetic study performed by the Colorado School of Mines (Zach et al. 2012), hydrate formation is not likely to occur within the separation section of CUBE, due to the low residence time of the gas-liquid mixture within the system. However, hydrates may form in either of the risers and this requires special attention by flow assurance experts.

The initial development of CUBE foresees discharging gas into the sea, and to limit water flow to surface by performing at least partial separation of the water phase. The limited presence of water and gas in the liquid riser allows the hydrate formation issue to be dealt with a conventional procedure (riser insulations, inhibitors, etc.).

Experiments

The Containment Dome has been tested in the Multiphase Flow Laboratory of TEASistemi. A water tank, 5 m high, with a square 1 m×1 m cross section, has been set up. The tank is equipped with glass windows and the water level in the tank is kept constant. At the bottom of the tank a gas-liquid jet is generated by injecting oil (diesel) or water and air in a short vertical pipe of varying size. The Containment Dome is placed at different distances above the jet. Different internals and inlet sections have been tested.

The initial Containment Dome was designed for a nominal oil-flow rate of 1,500 B/D and a gas-flow rate of 40,000 scf/D. This resulted in a dome size of 0.5 m×0.5 m×0.8 m with an overall volume 2,000 times less than the one used during the Macondo incident, which according to available information was designed for a 10-times larger volume than the Macondo oil-flow rate. A picture of the system under operation is shown in Fig. 2.

The main parameters measured were the inlet gas- and liquid-flow rates, the liquid entrainment by the gas in the gas riser, the liquid-pump flow rate, the liquid height, and the gas pressure within the dome. All measurements were performed by conventional field instruments. The liquid content of the gas has been measured downstream of a gas-liquid separator installed at gas outlet by determining the time required to fill a known volume of the separator with liquid.

The separated gas was discharged through a 3-in.-diameter riser. A control valve in the gas riser was used to set the pressure within the dome at such a value that the liquid entrainment by the gas was at a minimum value. At this pressure the liquid level in the separator is stable. When the pressure in the separator is increased by closing the valve, the liquid level will decrease and eventually the separator will empty. If the valve is opened, the liquid level will rise and the liquid entrainment increases appreciably. In all the experiments it has been clear that under laboratory conditions, the system is very stable and does not require any type of control. Nonetheless, a pressure-control system based on the value of the liquid height in the separator has been developed and it has been found to be effective.

The main results obtained are shown in Figs. 3 and 4. These figures show the liquid level (Fig. 3) and the pressure differential (Fig. 4) between the gas space in the separator and the hydrostatic pressure outside of the dome (at the top section of the dome) plotted versus the overall pump flow rate (oil and water), for three different gas-flow rates and an oil-flow rate of 1,500 B/D.

Fig. 3: Liquid level in the separator.
jpt-2013-01-cubefig4.jpg
Fig. 4: Pressure differential in the separator.

As can be seen from Fig. 3, the liquid level is relatively constant at increasing pump flow rates, up to the maximum value of the liquid flow applied. Under this condition, an appreciable gas entrainment into the liquid is observed. The pump used in the present experiments did not sustain a gas fraction above 1% in volume and when this happened the pump turned down.

As shown in Fig. 4 also, the gas pressure in the dome is fairly constant at increasing pump flow rates, but, as expected, it is affected by the gas-flow rate (and by the oil-flow rate). It is interesting to note that the pressure inside the dome is slightly above the external pressure. This is due to the contribution of the jet momentum and buoyancy.

Conclusions

A set of experiments for a containment dome scaled down to operate with a-1,500-B/D oil jet with associated gas has been carried out at a water depth of 4 m in a specially designed experimental facility. The size of the dome was 0.5 m×0.5 m×0.8 m. This provides an overall volume that is 2,000 times lesser than the containment dome used in the Macondo incident and designed for a jet-flow rate of 15.000 B/D. The main result of the present experiments has been that, notwithstanding the small size of the separator, stable operation of the containment/separation system has been demonstrated for a wider range of conditions than the design ones.

The successful operation of the small-scale gas-liquid separator developed in the current phase of the project has encouraged Eni to test a different set of internals and to investigate the possibility of separating most of the seawater entrained by the hydrocarbon jet. This may provide a substantial help to the solution of the hydrate problem. Testing the system at higher operating pressures is also being considered, as gas-liquid separation systems are appreciably affected by the gas density and also because the control system, which has been used to date (liquid level in the separator controlled by the gas pressure), needs to be tested at higher operating pressures. For this purpose, a new test loop able to operate with a gas density equivalent to that of methane at a water depth above 1000 m has been set up.

The main conclusion of the present paper is that a major step on the feasibility of the gas-liquid separation concept for the containment of a subsea blowout has been established. This approach, which can be classified as open-sea separation, may represent a solution when the capping-stack technology cannot be used. The main limitation of the proposed technology appears to be hydrate formation, although in many scenarios the problem is overcome due to the extremely low resident time of the fluid in the separation dome. In any case, it is proposed that this issue can be tackled by minimizing seawater flow into the separation dome and the separator residence time.


References

  • De Ghetto, G. and Andreussi, P. 2010. Apparatus for Carriage and Recovery of Hydrocarbons From a Subsea Well in Conditions of Uncontrolled Release. Patent Nos. 17.06.2010 IT MI2010A001101; 10.06.2011 PCT/IB11/001326.
  • Zach, A., Grasso, G., Lafond, P., et al. 2012. CUBE Analysis, Final Report.  Center for Hydrate Research, Colorado School of Mines, Colorado, US, 2012.