EQUIPMENT/ENGINEERING: Computer simulation helps increase flow rate in three-phase separator

Sept. 1, 2001
Computer simulation has contributed to increased production in a North Sea oil field by helping to conceive and validate design changes that increased the output of the three-phase separator that was previously a limiting element.

Computer simulation has contributed to increased production in a North Sea oil field by helping to conceive and validate design changes that increased the output of the three-phase separator that was previously a limiting element. This was the first study of fluid flow and separation behavior in a real separator.

The separator on the Gullfaks A platform prevented production increases because of an excessive water cut in oil coming out of the separator above a certain production volume. Engineers suspected that the design of the internals of the separator was to blame, but were limited in evaluating design modifications by high cost and lead time of physical testing as well as scalability concerns.

Instead, Sintef researchers used computational fluid dynamics (CFD) to analyze velocity, pressure, and concentration of multiphase fluid flow within the separator. The insight gained by this analysis made it possible to specify design changes that allowed the production volume to be substantially increased while actually reducing the water cut.

Three-phase separators are huge pressurized vessels that play a critical role in offshore oil production. The mixture of oil, gas, and water emerging from the ground enters the separator as a high momentum jet and proceeds through a cyclone designed to separate solids from the fluid. The liquid splashes into a pool in the lower part of the vessel and the gas, together with liquid drops, flows to the upper part. A liquid pool forms near the inlet where the multiphase fluid flows as a dispersion with low horizontal velocity.

The released gas rises to the outlet at the top of the vessel while oil and water flow to separate outlets on the bottom of the vessel. In the upper part of the vessel, oil and water particles together with condensed gas flow down to the liquid interface as the multiphase fluid flows to the outlet.

The Gullfaks field was licensed in Norway's fourth offshore licensing round in 1979. The Norwegian firm Statoil is the operator and majority owner while Norsk Hydro and Saga have smaller shares. The Gullfaks field features three giant, concrete-base production platforms and associated subsea wells and pipelines.

The Gullfaks A platform, an integrated drilling, production, and accommodation facility, came online in 1986. After a number of years of operation, several problems were experienced including emulsion problems inside the separator, sand accumulation, water-level control failure, and most importantly, a rising water cut (water content) in oil exiting the separator. These problems stood in the way of a planned production increase for the platform.

Free surface analysis

Sintef researchers performed a number of computer simulations on the existing separator in order to determine the reason for the problems and to experiment with possible solutions. CFD involves the solution of the governing equations for fluid flow at thousands of discrete points on a computational grid in the flow domain. Researchers used Flow-3D (Flow Science - Los Alamos, New Mexico) because of its capabilities in this area. Flow-3D uses the volume of fluid (VOF) method to predict free-surface fluid motions, surface tension, and other flow complexities.

In particular, this package provides algorithms that track sharp liquid interfaces through arbitrary deformations and applies the correct normal and tangential stress boundary conditions, an accuracy feature that distinguishes it from other CFD programs. Sintef researchers require these capabilities for modeling sloshing that occurs in separators placed on top of floating facilities for offshore production of oil and gas.

When properly validated, a CFD analysis allows engineers to obtain flow velocities and pressures at any location in the problem domain. This is far more information than can be obtained from physical experiments and brings far more useful information to bear on the design process.

The geometry of the vessel internals, depth of the oil-water mixture, size of the disturbances, or any other condition can be changed quickly on the computer and re-analyzed to determine the effect of the change. In addition, scaling is not an issue with CFD because the simulation can easily be performed at actual size. Of course, simulation is not a substitute for testing, but rather a useful tool that, once validated, can reproduce conditions that would be impossible or impractical to duplicate in physical testing.

Separator modeling

The inlet section of the separator in which all three phases are present was modeled as a two phase gas/liquid flow system. A gas/liquid jet was modeled to flow against the cup-shaped momentum breaker. A splash plate below the inlet and jet inside the liquid pool was also included in the model. The calculation of the two-phase flow behavior in this lower area provides information on the distributed velocity field in the liquid pool.

The velocities inside the bulk liquid zone are rather low so a homogeneous two-phase flow field was assumed in this area. No information about the liquid/liquid dispersion characteristics were available so the simulations were performed with the assumption of homogeneous fluid properties with no slip between the phases inside the bulk liquid zone.

The liquid velocity distribution, calculated in the inlet zone, was modeled through the separator fenced by the internals. Sintef researchers began modeling complex dispersions or emulsions recently by calculating the slip velocity between the water drops and the oil, and the drop growth in emulsions.

Simulation results

The results of the simulation showed flow patterns and velocities in the inlet zone of the separator for the existing production rate. The velocity vectors indicated that the flow turns after hitting the momentum breaker in the upper part of the separator and the flow in the liquid pool is turned by the splash plate to form a jet flow further downstream.

The flow velocity over the splash plate is uneven and back-flow occurs along the separator walls. The fluid volume below the splash plate is a dead zone with little or no flow. The next step was to re-run the analysis with the production rate that Statoil management was hoping to achieve in the coming year. It showed that velocity distribution was more homogeneous over the splash plate, eliminating backflow, an encouraging result.

The distributed velocity field in the bulk liquid zone shows that the flow field is highly influenced by the internals. In the original design, the internals (mostly baffles) extended to the bottom of the tank and caused rotational flow structures that produced backflows along the separator wall. In the simulation results, the internals could clearly be seen to be acting like mixers to reduce separation efficiency.

When the flow rate through the separator was increased in the simulation, the overall low field behavior was similar but even worse. The rotational flow structures were even more intense and the magnitude of the velocities was higher, resulting in poorer separation efficiency and increased watercut.

Improving existing design

Sintef engineers modified the simulation model to investigate the effect of several design modifications in the separator. The following changes were introduced: The internals, particularly the flow streamers and drop removal system, were reduced in length so that they did not extend down into the liquid pool. In addition, the sand removal system was also redesigned. The measurement devices for the liquid level control were rebuilt to prevent sand from blowing into the tubes. When the new design was simulated, the results showed that it would provide excellent operating results, even when the production rate was increased.

Based on the simulation results, Statoil modified the separator in accordance with Sintef's recommendations. Field test results indicated that the new design was far superior to the old one. In particular, both water level control and watercut were at acceptable levels even with the higher production volume. These results illustrate the benefits that CFD analysis can provide by optimizing the design of new separators and solving problems with existing designs. Computer simulation clearly has the potential to substantially improve separator performance.

The latest trend in offshore production is that oil and gas production are moving to greater water depths and floating production facilities are taking over developing and processing functions from fixed platform systems. The advantages of floating facilities are lower cost and shorter construction lead time.

The disadvantage is unstable production conditions due to wave and drift motion at sea. Sintef is developing new methods for modeling the oil/water interface to determine the effect of sloshing in both horizontal and vertical gravity separators. Sintef has also increased the accuracy of separator analysis by extending FLOW-3D's drift-flux model, which computes the balance between buoyancy and viscous stresses to determine the oil and water separation rate.