Polymer flooding relative permeability

The circular and square points on Fig. 13 show the numerically calculated oil recovery curves for the two polymer floods with adsorption factor b = 0.13 and 0, respectively. A polymer water viscosity of 3 cp and polymer water relative permeability curve of 1/3, shown in Fig. 2, were used in the calculations. The linear portions of these two calculated recovery curves reflect the formation, high rate of travel, and production of the constant-saturation connate water bank (Fig. 4) ahead of the polymer front. Comparison with the experimental recovery curves shows that the numerical model predicts a considerably lower rate of production of the additional oil obtained by the polymer injection. The reason for this discrepancy is that the highly mobile connate water bank did not form in the experimental floods. The polymer water incompletely displaced and mixed with this connate water. 

Figure 5.54 shows an example of relative permeability curves in an oil-wet rock. The water relative permeability curve after polymer contact, k p, was parallel but signihcantly lower than the water relative permeability curve before polymer flood, kj i. k, with S increasing and kj 2 with Sw decreasing were different owing to hysteresis. The residual oil saturation decreased in the polymer/oil test as the k 3 shifted toward higher water saturation, as shown... 

According to the previons discussion, water relative permeability, k , in polymer flooding is rednced, whereas oil relative permeability, k, is little changed. There are several reasons as summarized here ... 

For oil and water relative permeability curves after polymer injection, Huang and Yn (2002), and Chen and Cheng (2002) reported their observations, which were similar to residnal permeability reduction after polymer flooding. Compared with the relative permeability curves before polymer flooding, the relative permeability curves had the following three characteristics (1) k was reduced at the same water saturation, and corresponding to the same k, , water saturation was larger (2) immobile water saturation was increased and (3) residual oil saturation was reduced. It is believed this result was caused by polymer adsorption, which made a rock surface more water-wet. 

This section discusses viscoelastic effect observed from core floods, its effect on relative permeabilities, and its effect on polymer flooding. 

Figure 6.19 shows a set of relative permeability curves for waterflooding and polymer flooding. The following observations can be made ... 

At the same water saturation, the permeability to polymer was significantly lower than that to water. However, the oil relative permeabilities in polymer flooding and waterflooding are not very different. 

Similar relative permeability measurements were reported earlier in the literature. However, the results were not consistent regarding whether polymer flood residual oil saturation is lower than waterflood. 

In ASP flooding, alkaline, surfactant, and polymer have different effects on relative permeabilities. Table 13.2 shows our attempt to summarize these effects compared with waterflood. From Table 13.2, we can see that the effect of alkaline flood in terms of emulsification is similar to the polymer effect, whereas its effect in terms of IFT is similar to the surfactant effect. Less rigorously, we may say that only polymer reduces k, and only surfactant reduces IFT. In ASP flooding, the viscosity of the aqueous phase that contains the polymer is multiplied by the polymer permeability reduction factor in polymer flooding and the residual permeability reduction factor in postpolymer water-flooding to consider the polymer-reduced k effect. Then we can use the k curves (water, oil, and microemulsion) from surfactant flooding or alkaline-surfactant flooding experiments without polymer. 

Ye and Peng (1995) measured ASP solution/oil relative permeabilities based on the preceding principle. They first conducted a core flood test using an ASP solution and calculated the Darcy viscosity for the solution, which included the polymer permeability reduction factor. Then they conducted ASP/... 

Alcohol-free chemical floods using an equimolar blend of an olefin sulfonate and a petroleum sulfonate were reported to give a final oU recovery of 94% with a 13% of PV slug size using 3 vol.% surfactant concentration. When the slug size was reduced to 3% of PV, the oil recovery was still 80% [J7]. The mobility was controlled by adding polymer so the minimum slug viseosity, Hs, was at least equal to the reciproeal value of the water mobility at residual oil saturation, Sor Rw is viseosity of water and fcrw is relative permeability of water, i.e. ... 

Another issue that has only been addressed in a few studies is the effect that polymer adsorption has on the relative permeability of the aqueous and oleic phases that subsequently flood a core. In conventional polymer flooding, this is not a very important consideration since the process usually occurs in one particular saturation direction for example, if the formation is strongly water-wet then the oil displacement by water or polymer solution is an imbibition process. In such a case, the oil would not normally flow at a high saturation in a polymer-flooded zone, although such behaviour is conceivable (but unlikely) in certain polymer oil displacements in layered systems (see... 

Before discussing the issues concerning the polymer experimental procedures, it is necessary to establish the conditions under which the more traditional field core data have been gathered (i.e. core permeabilities, relative permeabilities, etc.). Central to such consideration is the matter of core wettability and how the core has been conditioned or restored for the relative permeability experiments and, therefore, for the polymer flooding experiments. This very important matter will not be considered here, but it will be assumed that the wettability and conditioning of the reservoir core have been satisfactorily achieved. This is important for polymer properties, since the adsorption is thought to be greater in water-wet cores than in oil-wet systems. In the discussion below, it will be assumed in all cases that experiments in porous media use correctly conditioned field cores at residual oil (unless otherwise stated). The oil will be the (dead) field oil, and conditions of reservoir temperature, but not necessarily pressure, will be established in the core. 

Measurement of transport parameters The main measurement of interest under this heading is of the excluded/inaccessible pore volume (IPV) of polymer relative to tracer as parameterised by the core permeability. If this quantity is known, then it should be included in the simulation studies since it may have some effect on the relative breakthrough times of polymer and tracer. However, it has been found that the IPV effect is usually dominated by the frontal retardation of the polymer as a result of adsorption/retention, and it is not generally of major importance in the assessment of the outcome of the polymer flood. Other measurements, such as of polymer dispersion coefficient and viscous fingering parameters, are primarily of importance for interpreting detailed core flood experiments since they do not scale in a simple way to the field and cannot therefore be used directly in the polymer field-scale simulations. 

If the polymer application is for viscosity control, then the simple model may be a 2-D areal or simple layered five-spot pattern the well locations will be known, and any faults that are known should be included using transmissibility modifiers. In this areal system, the key features in the calculation are that the oil viscosity is correct and reasonable estimates of the reservoir relative permeabilities are available. For a heterogeneity control flood, a multilayer cross-section similar to the eight-layer system described in Chapter 8 may be selected. Here, the most important factor is to get the layering structure and layer permeabilities correct, especially the permeability contrast between the high-permeability layer(s) and the position of this layer. The value of the kjk ratio may also be very important in this type of polymer flood, but it may be sufficient to know that it is > 0.05 for example (see Chapter 8). In such an application, the mobility ratio may be close to... 

Example 5.7—Polymer Flood in a Linear Reservoir Originally at Intensdtial (Immobile) Water Saturation. The potential of using polymer-augmented waterflooding to increase oil recovery from a uniform reservoir must be evaluated. For the purposes of this example, consider a linear reservoir segment that is 500 ft wide and 20 ft Ihiek. Production and injection wells are 1,000 ft apart. Properties of the reservoir rock and fluids, summarized in Table 5.19, are identical to those used in Example 3.1. Injection rate is constant at 200 B/D. Relative permeability relationships are =0.8(1.(5.66)... 

Consider the effect of adding sufficient polymer to the injected water so that the apparent viscosity of the polymer solution is 4 cp. Assume that no reduction in water relative permeability is caused by the polymer. The mobility ratio, Mg, would be expected to decrease by a factor of four to about 0.54, which is clearly favorable. In Example 3.5, the displacement performance of a polymer flood was estimated for a polymer solution containing 300 ppm polymer with an apparent viscosity of 4.0 cp. Retention is 17.5 /tg/g at 300 ppm, so that Dp =0.424. Effective inaccessible PV for this system is estimated to be 0.25. Thus, —< g+Z>p =0.174. 

A reservoir consists of two layers separated by a permeability barrier so that there is no flow between layers except at the wellbores. The permeability of the upper layer and lower layers is 1,400 and 33 md, respectively. Table 5.74 gives other properties of the reservoir. A determination of the potential of a polymer flood for this reservoir is desired assuming that displacement is linear. The wells are on 10-acre spacing so that the distance between rows of injection and production wells is 660 ft. Polymer is retained by the reservoir rock Table 5.74 gives the values of polymer retention at the injected concentration of 250 ppm. Retention is assumed to be irreversible, and the density of the rock matrix is taken as 2.65 g/cm. The eflfec-tive IPV is 0.25. Relative permeability relationships given by Eq. 3.197 and 3.198 are assumed to be applicable for both layers. Use constants in Table 3.20 for Layers 1. a , = — 3 for Layer 2. 

A micellar-polymer flood is to be designed for a reservoir which has been waterflooded to ROS. Viscosities of the oil and formation water at reservoir temperature are 3 and 0.9 cp, respectively. Eqs. 3.14 through 3.16 give relative permeability relationships, and Table 5.77 gives rock and fluid properties. Determine the mobility of the micellar-polymer solution required to obtain mobility control in the displacement. 

The water-oil mobility ratio based on average relative permeability data from similar Minnelusa reservoirs was in excess of 15. The use of polymers to alter this unfavorable mobility ratio to one nearer unity was suggested. Simulated pplymer flood performances using a model similar to that of Bonder, et al. indicated that the conversion of C-H 1 well to polymer Injection could be economic. Polymer entrainment by the porous medium was estimated to be 50 pounds per acre foot in the relatively clean Minnelusa sand. Formation brine had an equivalent sodium chloride concentration near 45, 000 mg/ liter with approximately 1, 500 mg/liter calcium ion present. However, the waterflood injection brine was particularly well suited for polyacrylamide, having less than 1, 000 mg/1 equivalent sodium chloride concentration and less than 12 mg/1 calcium ion in solution. It was anticipated that a fresh water buffer would need to be injected prior to commencing polymer injection. 

Fractional Flow Curves. Initially, fractional flow calculations were made to determine the amount of additional recovery expected from a polymer flood (Fig. 1). The calculations were based on relative permeability curves derived from clean-core experiments. Polymer flood recoveries were predicted for a range of polymer-to-oil viscosity ratios. For instance, at a ratio of 1, a wide separation in the two fractional flow curves (waterflood and polymer flood) was observed (Fig. 1). This indicated an excellent chance for increasing the displacement efficiency by substituting polymer solution for water as the displacing phase. 

The numerical model described here simulates linear or five-spot polymer floods in a single-sand reservoir or in a stratified reservoir consisting of several noncommunicating sands of varying thickness, permeability and porosity. Different relative permeability curves and initial water and gas saturations may be used for each different sand layer in the stratified case. 

The numerical model was run for the case of a single tube to simulate the linear waterflood and two polymer floods. The relative permeability curves shown in Fig. 2 are one of a number of sets measured for different mesh-size Nevada sands having permeabilities ranging down to 300 md. As stated above, these curves were provided by an oil company. No alterations of the 6,000-md curves were made to attempt to match die waterflood oil recovery curve. Fig. 13 shows diat the waterflood oil recovery curve calculated by the numerical model agrees well with the experimental data. This indicates that the given relative permeability curves are closely representative of the true curves for the 6,000-md sand. 

The numerical model was run to predict oil recovery from a five-spot polymer flood in a California viscous-oil reservoir. The relative permeability curves used corresponded to a 310-md Nevada sand representative of that reservoir. Oil viscosity was 34 cp, water viscosity 1 cp and polymer water viscosity was represented by 1 + 2C/Cq. Feed concentration was 268 ppm Kelzan M and an adsorption level corresponding to b = 0.13 was used. The relative permeability curve for polymer water was taken as 1/3 the normal water relative permeability curve. Initial connate water saturation was 26.6 percent. 

FIG. 7—WATERFLOOD AND POLYMER FLOOD WATER. OIL RELATIVE PERMEABILITIES. 

The viscosity of the live oil, water and polymer solution was 950 mPa-s, 1 mPa-s and 25 mPa-s, respectively. The relative permeability curves used to achieve the history match for the water-flood and subsequent polymer floods, are shown in Figure 3. Based on the shape of the relative permeability curves, the core is water wet, = 0.1. The water and polymer coreflood parameters were used to calibrate the field scale simulation model described in the next section. 

Fractional flow curves presented in Figure 3 were derived from relative permeability experiments and represent both waterflood and polymer flood data. Apparent fluid viscosity for the polymer flood curve is 22.0 cp as calculated using 750 ppm of American Cyanamid Cyanatrol 970. These curves show that the increased viscosity of the polymer solution displaces the fractional flow curve to higher water saturations at a given fractional flow of oil, illustrating the potential for additional oil recovery. Using a laboratory measured adsorption value of 0.058 lbs. polymer/bbl pore volume and a S. of 0.54 (water-oil ratio of 25) for the polymer curve, the fractional flow predictive method was used to calculate that an incremental 3.83 MBO (10.1% OOIP) could be recovered within the polymer flood area. 

---