The production of oil from horizontal wells in thin rims sandwiched between gas and water is notorious for coning problems. There is a strong tendency for early gas or water breakthrough at the heel, especially if the pressure drop over the length of the well is in the same range as the drawdown. We present two conceptual solutions to counteract the negative effect of well bore pressure drop through the application of downhole measurement and control. One solution concerns inflow switching in a segmented well bore and allows for coning control after breakthrough has occurred. The other solution aims at preventing breakthrough as long as possible. This is achieved by flattening the drawdown profile along the well through controlling inflow at one or more points in an extended stinger. The feasibility of the solutions was demonstrated through numerical simulations over a range of reservoir and well bore parameters. Implementation would require further development of downhole water and gas detection capabilities. Introduction Thin oil rims are relatively thin oil columns, in the order of a few to tens of meters thick, sandwiched between water and gas layers. They often occur in reservoirs with lightly compacted sands having high porosities and high permeabilities of up to several Darcy, and they commonly contain light oils. These properties culminate in favorable reservoir flow conditions, reflected in a low drawdown required for production. However, these properties can also cause production problems, in particular water or gas coning leading to early water or gas breakthrough. Horizontal wells are an attractive solution to reduce the potential for coning because they require lower drawdown than vertical wells for the same production rates. However, a possible problem of horizontal wells is the pressure drop over the well bore caused by friction forces between the fluid and the well bore (Fig. 1). As a result the drawdown at the heel of the well becomes higher than the drawdown at the toe, which increases the tendency for water and gas coning at the heel and thus partially cancels the beneficial effect of the horizontal well. The reduced draw down near the toe of the well also lowers the effectiveness of increasing the well length. Pressure drop over horizontal oil wells has been modeled to various degrees of sophistication1–3. These studies show that, roughly speaking, well bore pressure drop over a horizontal well becomes a problem when it is in the same order of magnitude as the drawdown at the heel. The ratio between pressure drop and drawdown increases for reducing well diameter, increasing well length, and, importantly for many oil rims, increasing reservoir permeability and reducing oil viscosity. Stinger completions Passive stinger completion One of the solutions to the non-uniform drawdown problem is the use of an ‘extended stinger’ to shift the tubing inflow point from the heel of the well to somewhere near the middle4,5. This effectively replaces the horizontal well by two shorter ones (Fig. 2). However, such a ‘passive’ stinger has a number of practical disadvantages:Its dimensions are based on a fixed inflow profile along the well. However, the inflow profile may change over the life of the well due to reservoir pressure transients and due to the breakthrough of gas or water.It requires that the inflow profile along the well bore is known at the design stage. This is usually quite unrealistic because of unpredictable reservoir heterogeneities, in particular near-well bore permeability fluctuations. Passive stinger completion One of the solutions to the non-uniform drawdown problem is the use of an ‘extended stinger’ to shift the tubing inflow point from the heel of the well to somewhere near the middle4,5. This effectively replaces the horizontal well by two shorter ones (Fig. 2). However, such a ‘passive’ stinger has a number of practical disadvantages:Its dimensions are based on a fixed inflow profile along the well. However, the inflow profile may change over the life of the well due to reservoir pressure transients and due to the breakthrough of gas or water.It requires that the inflow profile along the well bore is known at the design stage. This is usually quite unrealistic because of unpredictable reservoir heterogeneities, in particular near-well bore permeability fluctuations.

The Nova subsea oilfield development is located in the Norwegian sector of the Northern North Sea and will be developed with 3 oil-producer / water-injector pairs. Two of the production wells will be open hole horizontal wells of moderate length (400 – 1000 m) completed with sand screens equipped with ICDs (inflow control devices) to facilitate the best possible clean-up and to optimize inflow distribution. The injection of filtered and treated seawater has been selected for reservoir pressure support. Despite the relatively low Barium content (<70 mg/L) in the formation water, sulphate scaling is expected to appear in the production wells once injection water breakthrough occurs. In order to maintain well productivity, periodic successfully placed scale squeezes are essential. As the horizontal wells are completed across multiple reservoir segments in the same layer, uncertainties in fault transmissibility and water injector connectivity may result in variation of pressure along the well, thereby complicating squeeze placement. Traditionally, variation in rate and fluid viscosity have been applied to improve placement of scale inhibitor in the higher-pressure layers connected to the water injector. Chemical vendor simulations indicated that for a pressure variation below 5 bar adequate placement could be achieved with non-Newtonian fluids, however for higher differential pressures significant portions of the well would go untreated. The Nova sand screens are equipped with inflow control devices designed to allow the best possible clean-up after drilling and optimize inflow during production by distribution of drawdown and limiting of annular flow. The idea of using the same principles for scale squeeze was difficult to prove due to limitations in vendor placement software and other industry standard modelling packages not being able to model the combination of ICDs and non-Newtonian fluids. Computational Fluid Dynamics (CFD) on the other hand can simulate various factors that could influence the placement of scale inhibitors along the well length. Factors such as Newtonian versus non-Newtonian scale inhibitors, different reservoir pressure profile, ICD nozzle sizes and different well completions. This is all possible because CFD is largely based on physics and rigorous geometry, which is a significant step forward compared to the industry-standard scale squeeze placement software (Byrne 2010, 2011, 2014). Application of a computational fluid dynamics well model has provided confidence that the combination of non-Newtonian fluids and production ICDs would allow scale squeeze placement along the whole reservoir interval despite significant pressure variations across it.

To develop scale management strategies and plans during field development planning, it is important to know the composition of formation water in the reservoir. Typically, formation water samples will be collected from appraisal wells and analysed for this purpose. However, when the wells are drilled with water-based mud, the samples are often contaminated with mud filtrate that has invaded the formation during drilling. By adding a tracer to the drilling mud and using a simple mass balance correction technique, it is possible to correct for the effects of contamination and obtain an estimate of the formation water composition. But, where reactions occur during invasion or within the sample after collection, this method of correction will generate an erroneous estimate of the composition. The errors will increase with the extent of reaction and degree of contamination.