Reducing “Separation Anxiety”
with powerful 3D Flow & Thermal Simulation

 

The largest single product of the global oil and gas industry is neither oil nor gas, but water: produced at a rate of approximately 3 barrels to every barrel of oil, in 1999 the oil and gas industry was responsible for extracting 77 billion barrels of water.

Separating reservoir fluids into streams of oil, water and gas is a major concern to the global oil and gas industry, and has been almost since its inception. Historically, the major driver for effective separation was economics – extracting the maximum amount of usable hydrocarbon from the reservoir fluids. However, environmental concerns now mean that oil and gas producers are also increasingly bound by legislation that strictly controls the levels of pollution in discharged produced water – this combination is the growing ‘separation anxiety’.

 

Fig:01 The largest single product of the global oil and gas industry is neither oil nor gas, but water: produced at a rate of approximately 3 barrels to every barrel of oil, in 1999 the oil and gas industry was responsible for extracting 77 billion barrels of water.

 

Designing separators to meet these demands remains a significant engineering challenge. Critically, separators do not come in a ‘onesize- fits-all’ specification. They must be carefully chosen to not only account for the unique composition of fluids produced from a given reservoir, but also for the likely changes in composition that will occur over the lifetime of the well. Separator technology that is effective in early production might become less effective, or even fail, as the well matures or because of some temporary and unexpected change in the reservoir fluids. The increasing cost of platform real estate also means that there is also constant demand either to reduce the size of offshore separators or else to move them off the platform altogether, turning to newly developed subsea separation technologies.

Whatever type of separation technology is employed, or retrofits and adjustments made, the cost of getting it wrong can be immense. The production capacity of any facility depends, to an extent, on the effectiveness of its separation process. Although most facilities employ at least two independent separation trains, diverting production while diagnostic analysis is performed on a poorly performing separator inevitably results in a reduction of throughput. With oil prices topping $60/b, even a 5% drop in production from a 50,000 b/d installation will cost in excess of $150,000/d. Worse, if significant problems occur in the separation process, the cause of which cannot easily be diagnosed, the only alternative is to stop production altogether, or ship the reservoir fluids for processing at another facility.

Separator flow simulation
CFD has been applied at every stage of the oil, gas and petrochemical production process and can provide insight into any problem involving fluid flow (whether liquid or gas or a mixture of both) or structural components that are influenced by flow, and thus is particularly suitable for separator analysis. CFD simulation can help both in the design of new separator technology and in determining the range of operating conditions under which existing technology might be successfully deployed.

Data from CFD calculations can also be used to assist other types of analysis, for example, the forces acting on the separator internals can be calculated, either directly within the CFD code or via an external stress-analysis software package. In extreme cases, where fluid forces cause large deflections of components, the CFD simulation can be coupled directly with the stress simulation tool and both stress and fluid simulations can be performed simultaneously, each simulation feeding new boundary conditions to the other.

Case study 1
Sloshing in a free water knockout drum
The free water knockout (FWKO) drum is perhaps the crudest form of separator. FWKO drums work on a gravitational principle, relying on the fact that oil has a lower specific gravity than water and, if allowed to settle, will float to the top, forming a layer than can easily be skimmed off and extracted. Water is extracted through a valve at the bottom of the tank, while in the example shown in Figure 4, the oil trickles over a weir plate at the left hand side of the drum into the oil stream outflow.

Under normal operating conditions, this system provides a very effective means of preliminary. However, when deployed aboard an FPSO (floating production, storage and offloading vessel), there is a risk of the tank being disturbed by the motion of a passing wave, causing sloshing within the tank and leading to significant amounts of water passing over the weir plate or oil-emulsion contaminating the water outtake and possibly damaging downstream separation equipment.

Figure 5 shows a large sloshing motion that has developed in the vessel due the disturbing motion of a passing wave (as predicted by the CFD calculation). The simulation predicts that, under these conditions, a significant amount of water will slosh over the weir plate into the oil outflow.

In order to prevent sloshing, separator manufacturers typically insert a series of permeable vertical baffles into the tank, which act to damp the motion of the fluid within the vessel, preventing large-scale sloshing motions from developing. CFD simulation allows separator designers to make informed decisions early in the design process, before even the first prototypes are available, allowing them to answer questions such as ‘How many baffles do I need?’, ‘How do the baffles influence separator performance?’, ‘What sort of forces are acting on the baffles and on the vessel walls?, and ‘Under what range of wave conditions can the separator safely operate?’

Case study 2
Redesign of a gas phase separator
The aim of a gas phase separator is to remove small particles of hydrocarbon condensate (and other well-fluids) from a stream of natural gas. To be effective, the separator needs to be able to remove the wide variety of droplet sizes transported in a typical gas stream, from large visible droplets of hydrocarbon, to individual mist particles measuring just a few microns in diameter.

Exactly which fate each particle eventually meets depends largely on its size, but for the separator to work effectively, all but the smallest particles should be caught by one of the first three mechanisms. The vane pack demister acts as the final line of defence, removing a fine mist of droplets with diameters of around 10 mm or less. For effective operation, it is critical that the demister is not blocked by much larger oil particles, which, given enough time, should fall onto the surface of the liquid layer due to the influence of gravity. The separator therefore needs to be long enough, upstream of the demister, to ensure that the
gas flow has sufficient residence time to allow these larger particles to fall into the liquid layer of hydrocarbon at the bottom of the tank.

The simulation results reported in Figure 3 show that the majority of 65 mm and 35 mm particles hit either the vessel wall or the liquid surface a short distance after entering the separator. By contrast, many of the small 5 mm particles are carried with the gas flow until it passes through the vane pack, at which stage they are removed. The simulation predicted an overall trapping efficiency of 90%, with almost a 100% of particles of diameter 40 mm or higher removed by the separator. The separator manufacturers were able to significantly reduce the length of the separator, after establishing with the aid of further simulation, that since larger particles were hitting the walls of the liquid surface soon after entering the separator, much of the length upstream of the demister was unnecessary.