Several enhanced oil recovery (EOR) methods have been designed and developed in the past decades to maintain economic production from mature reservoirs with declining production rates. This chapter discuss mitigation of poor sweep efficiency in layered or naturally fractured reservoirs. EOR methods designed for such reservoirs all aim to reduce flow through highly conductive pathways and delay early breakthrough in production wells. Two approaches within this EOR class, injection of foam and polymer, specifically aim to improve the mobility ratio between the injected EOR fluid and the reservoir crude oil. Reduction in fracture conductivity may be achieved by adding a crosslinking agent to a polymer solution to create polymer gel. This may also be combined with water or chemical chasefloods (e.g. foam) for integrated enhanced oil recovery (iEOR). Polymer gel and foam mobility control for use in fractured reservoirs are discussed in this chapter, and new knowledge from experimental work is presented. The experiments emphasized visualization and in situ imaging techniques: CT, MRI and PET. New insight to dynamic behaviour and local variations in fluid saturations during injections was achieved through the use of complementary visualization techniques.
- polymer gel
- fractured reservoirs
- in situ imaging
Polymer gel as a conformance control technique is frequently applied in fractured reservoirs to reduce fracture conductivity close to production wells, but may also be applied to enhance oil recovery when used in the injection well to divert subsequent fluids injected from the fracture network into unswept matrix blocks. Fluid injection following a gel injection is here referred to as a chaseflood, and constitutes a range of EOR fluids depending on the reservoir and availability. Waterflooding was used as the standard chaseflood in this work, and the reader may assume that water is injected if not otherwise specified.
Generally, polymer gels are mixed in fresh water before injection to the field to achieve long‐term stability of the solution. In this work, polymer gels are made from saline brines and offer a significant advantage compared with fresh water gels because less interaction is expected between the gel and reservoir fluids when the gel solvent (most commonly water) is comparable to the formation brine, and gel blocking performance is significantly improved when combined with low‐salinity waterfloods in integrated EOR—iEOR, a term used here to describe the process where two or more proven EOR techniques are combined or performed in a smart sequence. In heterogeneous and fractured reservoirs, the iEOR approach is particularly effective when polymer gels are combined with mobility control techniques such as foam. The conformance control offered by polymer gels reduces flow in the largest fractures, whereas foam reduces gas mobility and increases the volumetric sweep efficiency of the reservoir resulting in incremental oil recovery. Gas mobility is reduced when bubbles are formed to make the gas phase discontinuous and separated by thin liquid films called lamellae. Foam application is particularly suited for heterogeneous reservoirs and is a field scale proven technique for EOR at both immiscible and miscible conditions and for aquifer remediation.
To successfully apply polymer gels or foams their behaviour and flow properties in porous media must be accurately predicted on a variety of scales. New insight to dynamic behaviour and local variations in fluid saturations during injections may be achieved through the use of complementary visualization techniques. Three imaging techniques are reviewed in this chapter: X‐ray computed tomography (CT), magnetic resonance imaging (MRI) and positron emission tomography (PET). The imaging techniques are used separately or combined to closely monitor flow processes and improve the understanding of EOR techniques and their underlying principles on core plug scale, and new results are discussed in this chapter.
2. Polymer gel placement in fractured systems
From a reservoir engineering point of view, achieving high sweep efficiency in the reservoir to maximize oil recovery represents a challenge when fractures are present. The fractures generally have permeabilities that exceed the rock matrix by several orders of magnitude and therefore represent thief zones for the injected EOR fluid. Rather than displacing oil from the oil‐saturated rock matrix, the injected fluids channel through the fracture network towards the producer, resulting in low sweep efficiency and excess production of water or injected chemicals.
2.1. Gel or gelant injection
Polymer gel is usually injected to the field as a gelant solution—a mixture containing all chemical components, but not yet crosslinked to form a rigid and highly viscous polymer gel. Gelant in many ways resembles the standard polymer solution, but transforms into a gel depending on temperature conditions and time—the time it takes for a gelant to transform into a rigid gel at a given temperature is known as the gelation time. Gelant may flow both in the porous matrix and the fractures as a result of the lower viscosity and smaller particles relative to gel. When formed, long polymer chains in the gel structure prevent the gel from flowing through the small pores of the porous media, thus gel is contained to fractures during injection. In field applications, gelant is generally pumped from the surface and enters the reservoir either as a gelant or as a weak gel, depending on reservoir temperature and pumping time. The placement process must be carefully designed because the two fluids deposit differently in the fracture volume, which influences the ability of the gel to reduce fracture conductivity and impact its performance during conformance control operations. Gel blocking capacity as a function of gel state (gelant or gel) has therefore been thoroughly investigated at core plug scale [1–7].
2.2. The leakoff process
Although formed gel cannot flow through porous media due to its structure, the gel solvent (a range of solvents may be used, here we discuss water only) may leave the gel and progress through the pore network if the viscous forces become high enough. This fluid transport, called leakoff, increases the concentration of the gel in the fracture, resulting in a more rigid gel structure with a higher resistance to pressure during chasefloods. When the gel loses water due to leakoff its mobility decreases compared to the continuously injected (”fresh”) gel; fresh gel propagates through internal conduits within the concentrated gel called wormholes [3, 8–10]. The transport of water from the gel depends mainly on gel strength and differential pressures across the gel slug, in addition to rock wettability and two‐phase flow functions, therein capillary forces. Reduction of flow in fractures or high permeability zones after placement of polymer gels have been reported [2, 10–15].
2.3. Gel mechanical strength‐dependency on gel state during placement
The mechanical strength of a gel in a fracture is an indication of its blocking ability and is often determined in the laboratory by measuring the gel rupture pressure—the pressure at which the gel in the fracture breaks apart and permits fluids to pass through it. If the injection pressure exceeds, the rupture pressure of the gel during a chaseflood, the gel‐filled fracture will be partially re‐opened to flow and fluid channelling through the blocked fracture will start again [4, 5, 7, 16]. Measurements of rupture pressures in open fractures with defined geometries have shown that gel mechanical strength after placement is dependent on the gel state during placement . Rupture pressures measured after formed gel placement with leakoff were often to be consistent and predictable. It is assumed that water mobilizes gel from the wormholes during chasefloods and is thereafter contained in the narrow flow channels constituting the wormholes. Because of this, efficient gel blocking may also be achieved after gel rupture and significant water throughput. Rupture pressures measured after gelant injection and subsequent in situ gel formation, a process referred to as gelation, were comparable to measurements during gel injection, but with more inconsistency. It was suggested that gelation did not always occur , because gelant may experience compositional changes that interfere with gelation upon contact with the reservoir rock and fluids [5, 17], e.g. cross‐linker diffusion into the surrounding matrix that may lead to gelation failure . Although initially comparable, the blocking capacity over time during chasefloods was consistently higher for pre‐formed gels compared with gels with in situ gelation. This was explained by the presence of wormholes that collapsed and re‐opened depending on the injection pressure.
2.4. Gel volume alterations
The blocking capacity of a gel placed in a fracture is also determined by the ability of the gel to maintain its volume relative to the fracture volume. After placement, the gel may undergo processes that change its volume and impact the success of a chaseflood. Several processes, such as syneresis [18, 19], dehydration [20–23] and gel swelling [24, 25], have been studied experimentally using bulk gel, core plugs and micromodels. During syneresis and dehydration, the gel volume is reduced due to loss of water caused by chemical reactions over time, changes in pressure gradients or matrix capillary forces adjacent to the gel network. Other conditions, such as temperature, solvent composition, ionic strength and external electric field, are the main causes of gel swelling, which impact the gel volume without the loss of water by influencing the osmotic pressure balance between the gel network and its surroundings . The behaviour of a polymer gel during EOR chasefloods will, in summary, depend on both the properties of the gel itself and on the external conditions in the reservoir.
3. Foam flooding of fractured systems
EOR techniques in fractured reservoirs rely on maximizing the contact area between the injected EOR fluid and the rock surface, and therefore, the EOR fluid must transport effectively within the fracture network . Most fracture networks, however, are heterogeneous themselves, with varying fracture aperture, orientation and connectivity.
3.1. Gas mobility reduction
Heterogeneous fracture networks may result in poor sweep of the injected EOR fluid, and mobility control should be implemented to maximize contact. Two‐phase flow within rough‐walled fractures has previously been investigated using Newtonian fluids [28–30] and non‐Newtonian polymer solutions [31, 32]. Foam has the potential to increase oil recovery by better areal sweep, better vertical sweep (less gravity override), less viscous fingering and diversion of gas from higher permeable or previously swept layers . Diversion of gas into lower permeable layers using foams may be important for fractured systems, where a very large permeability contrast exists and cross‐flow between the zones occur .
3.2. Foam generation in fractures
Foam is generated in porous media by snap‐off, lamella division or leave‐behind mechanisms. Foam generation is affected by porous medium topological properties, wetting content, capillary suction pressure, surfactant formulation, gas fractional flow and flow rate. Snap‐off and lamella division create “strong” foam in terms of separate gas bubbles above a critical capillary number , whereas leave‐behind forms lamellae oriented parallel to the flow direction with only moderate effect on flow resistance, and the injected gas remains as a continuous phase. In porous media with permeability discontinuities, snap‐off is an important mechanism for foam generation [36–38]. Foam generation within a fracture network was observed in a replica of a rough‐walled fracture with variable aperture . Foam was found to be generated by snap‐off and the bubble size was four times larger during in situ generation in the fractured systems as a result of fewer snap‐off sites in the fracture compared to the Berea sandstone used for bubble pre‐generation. The resulting foam texture and the relationship of curvature with foam viscosity determined the flow resistance of the system. Similar snap‐offs in idealized fractures at the transition from narrow to large aperture was reported , where the flow resistance increased with greater gas fractional flow and larger aperture. Similarly, flow in rough‐walled fractures was previously characterized [39, 41], and may be favourable for mobility in fracture networks of varying aperture .
4. Experimental procedures
In this work, experiments providing insight to polymer gel behaviour during and after placement were performed using cylindrical outcrop core plugs with a 1 mm longitudinal fracture and aged to varying wetting states. In situ imaging by CT, MRI and PET was used to quantify and explain local saturation behaviour during polymer gel placement and chasefloods. Foam was investigated using a fractured marble tile model encased in Plexiglas that enabled us to visualize and describe foam flow in fractures (the marble itself is impermeable) at different gas fractions.
4.1. Establishing stable wettability preferences by dynamic ageing
Stable, predictable and uniformly distributed wettability preference is essential during core analysis to achieve reliable results [43–45]. The simplest wettability preference to achieve is the strongly water‐wet condition, where the contact angle between oil and water on a smooth surface is between 0° and 30° . All clean outcrop rocks are strongly water‐wet before exposure to crude oil, and a strongly water‐wet wettability preference may be achieved and maintained using mineral oils . However, this wettability preference does not reflect the preference of most reservoirs , and it is therefore needed to change the wettability of outcrop cores to better reflect reservoir conditions. Because core plugs extracted from the subsurface are exposed to drilling muds and large changes in pressure and temperature conditions during extraction, wettability alteration is also needed for reservoir cores . Based on a decade of wettability alteration studies in chalk [50–54], a dynamic wettability alteration method  is outlined below
Cylindrical core plugs are drilled from outcrop rock blocks and cut to length. The core plugs are thereafter washed in tap water and dried at 60°C for no less than a week. The core plugs are thereafter saturated by brine, and porosity, and permeability measured.
The crude oil is prepared for ageing by being heated to 90°C and thereafter injected through a chalk core at a constant injection rate to remove impurities. Before use, the chalk core filter is flooded with crude oil to low water saturation and aged by flushing crude oil through the core at 1 cm3/h for 96 h, reversing the flow direction midways. The crude oil used for ageing the filter core plug is not re‐used in the wettability alteration study. Due to its aged state, it was assumed that the adsorption of active wettability alteration components on the filter was significantly reduced. After filtration, the crude oil is stored in closed containers at 90°C with minimal interference and without air exposure until used to age the cores.
The fully water saturated core plugs are oilflooded to irreducible water saturation by crude oil injection at high differential pressures (2 bar/cm) at 90°C. When Swi is reached, the injection rate is reduced to an ageing flow rate of 3 cm3/h or 1.5 cm3/h, for 2″ and 1.5″ diameter cores, respectively. Crude oil is continuously injected throughout the entire ageing process, and the direction of flow reversed midway. The ageing time is determined by the crude oil/rock/brine system and the desired wettability preference. After ageing, the crude oil is miscibly displaced from the core with 5PV of decaline (decahydronapthalene) followed by 5PV mineral oil (e.g. decane), all at 90°C.
Wettability is measured using the Amott‐Harvey method: spontaneous and forced imbibition of water, followed by spontaneous and forced imbibition of the preferred mineral oil: using mineral oil to represent the oil phase ensures that the wettability of the core plugs is not further altered for the duration of experiments.
Experiments to investigate enhanced oil recovery (EOR) potential are commonly started at the irreducible water saturation, Swi.
4.2. Imaging techniques applied in porous media
Several imaging techniques are currently applied to study flow in porous media, and most techniques originate from medical imaging. Here, we briefly review three imaging techniques used to study mobility control for EOR in fractured core plug systems.
4.2.1. Positron emission tomography (PET)
PET is based on positron‐emitting radionuclides where a positron is emitted from the nucleus accompanied by an electron to balance atomic charge. Radioactivity is a spontaneous nuclear phenomenon insensitive to temperature and pressure . The positron loses kinetic energy by interactions with the surroundings, and at near‐zero momentum the positron combines with an electron and annihilates. The use of PET to study flow in porous media was recently reviewed and compared with CT . Spatial fluid saturation is calculated based on the registered activity of the labelled phase, which in this work is water using the fluorine radioisotope 18F.
4.2.2. X‐ray computed tomography (CT)
CT measures the X‐ray attenuation through gradual loss in flux intensity through the medium, and produces a time‐averaged density distribution image of the rock [58, 59]. CT imaging relies on sufficient density contrasts which may often be fulfilled in low‐porosity media. Because attenuation depends on both the electron density and the effective atomic number, the CT number will vary with the energy of the rays. Spatial fluid saturation is derived with sufficient fluid density difference.
4.2.3. Magnetic resonance imaging (MRI)
MRI provides high spatial resolution imaging and fast data acquisition necessary to study flow in porous media. When replacing hydrogen in the aqueous phase (H2O) with deuterium (D2O), the magnetic signal in the water phase is removed and one explicit fluid phase (oil) is imaged in an oil‐water system. MRI has successfully been used to characterize flow behaviour and production mechanisms in fractured chalk and their dependency on wettability, injection rate, fracture aperture and configurations [56, 60].
Figure 1 shows the different views applied during in situ imaging of cylindrical core plugs in this study. The sagittal view was often used to capture the flow dynamics inside the open fracture—which was vertically aligned in all experiments.
5. Recent advances in laboratory evaluation of polymer gel and foam for EOR
5.1. In situ imaging of polymer gel transport in fractured core plugs
Complementary in situ imaging techniques have been used to investigate different aspects of polymer gel treatments described in the following sections. For most experiments, artificially fractured core plugs with open, longitudinal fractures have been used and the matrix has been saturated by two immiscible phases: oil and water, to mimic the environment in an oil reservoir. The polymer gel used for all experiments was 0.5% HPAM (∼5 million daltons molecular weight) cross‐linked by 0.04% (417 ppm) Cr(III)‐acetate and aged at 41°C for 24 h (five times the gelation time). After ageing, the gel was cooled to ambient temperature of ∼23°C.
5.1.1. Gel shrinkage from spontaneous imbibition
Gel placement in fractured reservoirs and its behaviour during chasefloods have been widely studied [4, 15, 61], but gel interaction with the porous rock in which the fractures are embedded has been less discussed. The exchange of fluids between the gel and adjacent porous rock may be substantial, even without a viscous pressure gradient across the system, if spontaneous capillary imbibition of water occurs . Spontaneous imbibition impact leakoff during gel placement and gel stability after placement, and the loss of water from the gel is influenced by relative permeability, capillary pressure and the wettability of the porous medium. A severe reduction in gel volume (up to 99%) was documented due to spontaneous imbibition into porous media from adjacent, non‐flowing bulk gel , suggesting that gel treatments lose efficiency over time where a potential for spontaneous imbibition exists, because gel‐treated fractures may be opened to flow. Polymer gel is often used to reduce water cut and is therefore placed in a well or zone of the reservoir that is watered‐out, where spontaneous imbibition is reduced. During EOR operations, however, the local fluid saturations may change considerably, i.e. if an oil bank is generated, and spontaneous imbibition may again have impact on gel mobility control.
5.1.2. Leakoff dependency on wettability
In a strongly water‐wet rock, a positive capillary force exists to attract water from the gel into the rock matrix. In an oil‐wet rock, however, water will be repelled and must overcome a capillary threshold pressure to invade the pores. These two wettability conditions were directly compared using complementary imaging techniques during mature gel injection into open fractures. Cylindrical core plugs with longitudinal fractures of 1 mm aperture were used, and gel was injected with a constant volumetric injection rate into the fracture at ambient conditions.
In strongly water‐wet, low‐permeable chalk rock samples, where a strong positive capillary pressure exists, residual oil saturations were reached during gel placement due to water leakoff . The chalk core was placed in a Bruker 4.7T MRI, and a spin‐echo based sequence was used to produce images during D2O‐gel injection. MRI settings must be tuned for each experiments, and depend strongly on the core material used, the fluid saturation and type of experimental setup (e.g. if a core holder or epoxy layers were used as confinement vessel). Finding the right settings for each experiment requires expert knowledge of the imaging apparatus and is not the focus of this chapter. Readers are referred to . Figure 2 shows MRI images of water leakoff during gel placement in a low‐permeable, strongly water‐wet chalk core (permeability, K <10 mD, porosity, ϕ = 45%, d = 4.96 cm, L = 7.6 cm). The rock matrices adjacent to the fracture were initially fully saturated with mineral oil (n‐decane) and appear light grey in the images. Bedding planes in the rock sample may be observed as darker features that are constant in the images. A strong positive capillary pressure exists due to the small pore sizes of this rock . Water leaves the gel during leakoff and reduces the oil saturation in the rock matrix adjacent to the fracture first (darker grey regions due to loss of magnetic signal from the oil).
The water leakoff from the gel in the fracture to the oil‐saturated matrix was symmetric around the fracture and occurred by combined spontaneous and forced imbibition from the differential pressure generated. This combined displacement resulted in residual oil saturation in the matrix after gel placement, corresponding to 64% OOIP oil recovery during gel injection for the chalk core used. Although significant oil was produced in experiments at core plug scale, gel placement will not in itself contribute significantly to oil production at field scale because of the small fracture volume (1–2%) relative to matrix volume in most reservoirs. Hence, chasefloods are important to recover additional oil after a gel placement. The concept of gel placement is, however, important to study because it greatly affects the success of chaseflood oil recovery.
In oil‐wet and low permeable rocks, a threshold pressure must be overcome to achieve water leakoff. Local water leakoff during mature gel placement in a fractured, oil‐wet carbonate core plug (K = 11.7 mD, ϕ = 17.6%, d = 3.8 cm, L = 10.1 cm) was investigated by CT . The CT was a Picker PQ‐6000 with 120 kVP tube voltage. Scans were produced with 4 s scanning time, and the slice thickness was 2 mm. Using a low injection rate of 6 cm3/h during gel injection reduced the differential pressure and no leakoff was observed. Gel progressed through the fracture without losing water and increasing its concentration—thus, rock saturation remained constant at the irreducible water saturation (Figure 3, top and middle). In fully water saturated rock, where capillary forces are not present, the differential pressure limit to achieve leakoff corresponds to the pressure needed to extrude gel through the fracture . Improved sweep efficiency was observed during the subsequent chaseflood when the injected water displaced oil from the rock matrix, predominantly in the inlet part of the core (see Figure 3).
5.2. Visualization of EOR chaseflood behaviour
During gel placement in initially oil‐filled fractured systems, it is assumed that the oil residing in the fracture is produced at gel breakthrough . This assumption was based on the high differential pressure observed during chasefloods, comparable to conventional experiments using 100% water saturated cores.
5.2.1. Oil trapped in gel
Upon careful inspection of Figure 2, it is possible to observe oil remaining in the fracture (bright sections) during gel flooding, and also after significant gel throughput. A clear rupture pressure during water chaseflooding suggests that the oil did not impact the blocking capacity of the gel. By employing the MRI further to image the chaseflood (Figure 4), we determined that the water flow paths were narrow and presumably through wormholes in the dehydrated gel, and that the oil remaining in the fracture was trapped within the gel structure (light grey areas in Figure 4). Figure 4 shows the water flow paths through the gel in the fracture (bright white). A thresholded image (Figure 4B) shows the wormhole flow pattern in more detail.
Wormhole size is known to change with injection rate due to the gel elasticity. Although the gel pressure resistance is stable at low injection rates (below 600 cm3/h, corresponding to a maximum injection flux of 1100–1600 cm/h when all flow of water is through the fracture), the gel may erode in the vicinity of wormholes at higher injection rates over time (significant water throughput). In experiments where the purpose is to measure wormhole size, it is therefore advisable to use high flow rates within relatively short time intervals. Without imaging, the average wormhole size can be calculated indirectly by measuring the injection rate and average pressure drop across the core plug. With in situ imaging, such as applying PET, the wormhole size may be measured directly, and the wormhole dynamics with changes in injection rate directly observed. Wormholes within a gel‐filled fracture were observed and quantified using PET, as shown in Figure 5. PET/CT sequences were acquired on a CT 80 W Nanoscan PC imager, featuring spatial resolutions of 800 and 30 μm of the respective PET‐ and CT‐detector systems. The core used was of Edwards limestone material  and was fully saturated by brine with a composition equal to the gel solvent. The PET signal was not corrected for attenuation in the analysis because the low CT tube energy (70 kVP) did not penetrate the core to produce spatial core plug density (needed for attenuation correction). The CT was thus only used for core plug positioning. In Figure 5, radioactively labelled water (18FDG) was injected into the gel‐filled fracture with a constant injection rate of 60 cm3/h. When the pressure drop had stabilized across the core (at 1.6 kPa/cm for this injection rate), the wormhole flow pattern, visualized by PET, remained unchanged and covered 22.3% of the fracture volume. Dehydrated gel (no signal) covered the remaining fracture volume. In situ imaging by PET showed locally large variations in the wormhole width, from very narrow (almost invisible on images) in some sections of the fracture to spanning large parts of the fracture height in other sections.
Waterflooding after polymer gel placement may be efficient, and even if the gel has ruptured it provides significantly reduced fracture conductivity compared to open fractures. In situ imaging of core floods (Figure 3) have, however, shown that displacement fronts may be dispersed during water chasefloods. Enhanced oil recovery may be achieved by stabilizing the displacement front, e.g. using foam . The gel treatment may also be strengthened after placement by injecting a water phase with lower salinity relative to the water in the gel . Both injection pressures and matrix production rates during chasefloods increased with reduced salinity compared with chasefloods where the salinity was matched with the gel , presumably due to swelling of the polymer gel network. The improvement of the polymer gel treatment was reversible, and gel blocking efficiency immediately decreased during chasefloods with matched salinity. This shows that is possible to control the success of a chaseflood by influencing the gel properties. Combining polymer gel placement in fractures with low‐salinity chasefloods is therefore a promising approach in iEOR. The new findings from PET imaging (Figure 5) may be helpful in explaining the mechanism for improved conformance control during low‐salinity chase waterfloods and shed light on wormhole morphology when injected water salinity is varied.
5.3. Fractured tile model to investigate 2D foam flow in fractured systems
An experimental setup was built to visualize flow and to study mobility control by foam generation in fracture networks. A white marble tile was fractured using a ball‐peen hammer and then carefully reassembled within a frame to create a fracture network with rough, calcite fracture surfaces. The (effective) porosity was assumed to be zero in the marble matrix blocks, and flow occurred through the fracture network only. The fracture network was characterized by injecting brine and gas separately into the model. Pure nitrogen (N2) gas injection provided a baseline to which foam injection could be compared. Figure 6 shows the marble tile model filled with red‐dyed brine.
Aqueous surfactant solution (1 wt% alpha olefin sulfonate (AOS) surfactant in 5 wt% NaCl brine) and N2‐gas were co‐injected into the fracture network at ambient conditions to form foam, using a range of gas fractions, to investigate the impact of the gas fraction on sweep efficiency. The fracture network was initially filled with aqueous solution, and the total flow rate for the co‐injections was 60 cm3/h.
5.3.1. Sweep efficiency
Areal sweep efficiency with changes in gas fraction was evaluated using the fracture network: a high‐resolution camera was used to capture images at short time intervals throughput foam co‐injection. From the images, total sweep was evaluated and it was possible to map flow patterns and calculate dynamic sweep efficiency as function of foam volume injected. Images could capture drainage of liquid from each fracture element, including vugs and narrow fractures. Figure 7 compares total areal sweep efficiency for three N2‐gas fractions during foam co‐injection: fg = 0.60, fg = 0.80 and fg = 0.90.
Figure 7 shows the development in sweep efficiency within the fracture network during co‐injection with three different gas fractions. The sweep efficiency is compared at three time steps: after three fracture elements are filled; after five fracture elements and after seven fracture elements are filled. Only foam in the larger vugs could be visualized directly, but foam was also generated in, and flowed through, the smaller apertures. In general, we see that the sweep efficiency and spread of foam was higher with the increase of gas fractional flow. At higher gas fractions, breakthrough of gas was delayed, but in all cases delayed compared with pure gas injection. This demonstrated that foam reduced gas mobility and was formed within the fracture network itself.
Unlike the foam bubble shape in a porous medium, where bubble shape is governed by pore configurations [35, 67], the bubbles in fractures are deformed according to interfacial tension, and the gas‐liquid interfacial curvature varies according to gas fractional flow . Foam coalescence in a porous medium is controlled by the ‘limiting‐capillary‐pressure’  at moderate to high gas fractional flow. The limiting capillary pressure is determined by the pore size, surfactant formulation, local fluid saturation and the interstitial velocity. The capillary pressure in a fracture is much lower (generally set to zero in most simulations—although strictly not correct) at a given aqueous‐phase saturation in comparison to permeable matrix. Capillary suction coalescence of foam lamellae appeared to become significant when the local aqueous phase saturation became very low. With increased gas fraction, the gas bubbles increased in size, on average. Large gas bubbles separated by thin lamellae occupied the majority of the fracture volume, especially at 0.95 gas fraction. This effect may be attributed to thinning of lamella at high gas fractional flow , resulting in foam coalescence and merging of smaller bubbles. Fewer and coarser bubbles related to the highest gas fraction are likely caused by thinning of lamellae (a larger fraction of the liquid is located in the Plateau border), followed by film rupture due to higher capillary‐suction pressure compared to lower gas fractional flow. The observations are consistent with earlier results, where larger bubbles were observed with increased gas fractional flow in a fracture network [39, 70].
5.4. Combined polymer gel and foam chasefloods for iEOR
Using a low‐field 8 MHz MRI (Oxford Instruments Maran DRX), we experimentally investigated an iEOR approach where the first step was polymer gel placement to reduce fracture conductivity and the second step was CO2‐foam injection to improve sweep efficiency compared to water chasefloods . The experiments were performed at ambient conditions, thus the CO2 was not miscible with the oil (n‐decane). Pre‐generated foam was injected into the core after gel placement, using a gas fraction of 0.8 and a surfactant solution containing 1 wt% AOS surfactant. The development in oil saturation during all experimental steps was compared to oil recovery during foam injection without mobility control, using magnetic resonance imaging. Two fractured carbonate core plugs were used (K ≈ 30 mD, ϕ = 23–24%, d = 3.8 cm, L = 5.8 cm), with a measured Amott‐Harvey wettability index after dynamic ageing of IAH = −0.2, which represents weakly oil‐wet conditions. Waterfloods were not efficient due to the oil‐wet conditions and only recovered 10–20% of the oil in place. In situ imaging showed that oil was primarily displaced from the inlet end of the core plugs, where water was injected directly into the matrix through specially designed end pieces . Oil was displaced into the fracture, and no further oil recovery from the core matrix was detected after water breakthrough. Polymer gel placement was performed at 100 cm3/h and yielded 20% oil recovery that was caused by the large fracture volume compared to the core pore volume at high injection pressure and low matrix threshold pressure due the weak oil‐wet conditions. Chaseflooding by immiscible CO2‐foam was performed to mitigate fluid dispersion around the displacement front. Figure 8 shows the in situ development of the oil saturation and displacement fronts during foam flooding in the twin cores, with and without polymer gel conformance control present in the fracture. Oil production was prevailing in the near‐fracture area by direct foam injection (Figure 8A). Haugen et al.  showed, using CT imaging of oil‐wet core plug systems, that oil was viscously displaced by surfactant solution, forced into the matrix by the increased pressure gradient provided by the foam. In Figure 8B, polymer gel was placed in the fracture prior to CO2‐foam flooding, and both foam constituents (gas and surfactant solution) entered the matrix during the foam flood. Immiscible foam injection after conformance control by polymer gel provided 28% OOIP (original oil in place) additional oil recovery during 12.7 PV foam injected. In comparison, recovery factors of 14% OOIP (Figure 8A) to 20% OOIP were obtained during injection of more than 30 PV foam in comparable core plugs. Improved sweep efficiency was visually observed by MRI during foam chasefloods (Figure 8); in which more stable displacement fronts and less tendency to viscous fingering were observed when compared to e.g. chase waterflooding. The combination of two established EOR methods to improve conformance in both the fracture and matrix has important implications for future work on the development of efficient integrated EOR schemes.
In situ imaging provides new insight and corroborates results from core analysis applying material balance or gravimetrical measurements, and several different imaging techniques may be used to closely investigate local changes in saturation, fluid flow and oil recovery mechanisms during EOR experiments. Using in situ imaging techniques in this work, we found that:
Water leakoff during polymer gel propagation through a fractured, oil‐saturated core plug could be captured using magnetic resonance imaging (MRI). Water leakoff was symmetric around the fracture and occurred by combined spontaneous and forced imbibition from the differential pressure generated by the gel injection.
Residual oil within the gel structure was observed with MRI after gel injection into a fractured, oil‐saturated core plug. The remaining oil did not impact the flow paths during water chasefloods, and it is likely that only a thin oil layer partly coated some region of the gel.
Water flow paths known as wormholes were observed during chase waterfloods applying positron emission tomography (PET). We found large local variations in the wormhole shape and width that have not previously been observed during waterfloods. PET is able to catch rapid changes in fluid saturation during dynamic floods and may thus give valuable information about changes in wormhole morphology as a function of flow rate and pressure.
In situ imaging by CT and MRI were used to investigate the shape of displacement fronts during chasefloods when polymer gel had been placed in a fracture. Imaging showed a dispersed displacement front during waterflooding of an oil‐wet core plug. CO2‐foam injection stabilized the displacement front and improved sweep efficiency.
Advanced medical imaging is not always required or available, and important insight may be gained from relatively simple experimental setups, such as the study foam generation in fractures where Plexiglas models provide the visualization needed. We found that foam was generated in a fracture model during co‐injection of surfactant and N2‐gas. A higher fractional flow of gas increased the sweep efficiency and spread of foam in vugs and fractures.
The crux of core plug scale experiments is proper design and planning of the experimental schedule using suitable equipment and imaging tools. When the core plug system is positioned within the imaging apparatus it cannot be moved. Hence, Testing the experimental setup without imaging, using simple material balance, is a good idea.