Formation damage

Fines Migration in Fractured Wells: Integrating Modeling With Field and Laboratory Data

Production and drawdown data from 10 subsea deepwater fractured wells have been modeled with an analytical model for unsteady-state flow with fines migration.

jpt-2015-02-fig1finesmigration1.jpg
Fig. 1—Simulation results for Well 2 with an analytical model for well productivity with fines migration. Blue stars correspond to well data. The model match is shown in hollow black circles.

Production and drawdown data from 10 subsea deepwater fractured wells have been modeled with an analytical model for unsteady-state flow with fines migration. The simulation results and the field data indicated a good match, within 5%. This paper describes the methodology used to integrate the modeling predictions with field and laboratory data to identify probable causes for increasing skins and declining productivity-index (PI) values.

Introduction

Fines migration is a complex phenomenon that can challenge the economic viability of a project because of well-­productivity decline, lower-than-expected hydrocarbon recoveries per well, large capital expenditures to drill and complete additional wells, and high operating costs from suboptimal facility designs. Excessive fines production may also result in equipment erosion and corrosion, formation of hard-to-break emulsions, and plugging of flowlines and surface facilities, all leading to potential hazardous situations.

This paper describes a multidisciplinary approach in which fines-migration modeling has been integrated with field and laboratory data to ascertain whether fines migration may be associated with rapidly increasing skins and declining PI values observed in a subset of deepwater fractured wells. Laboratory studies exhibit fines release and migration during coreflooding and stress testing; the field well-productivity data are well-matched with the mathematical modeling.

Fines-Migration Modeling

The traditional mathematical model for fines migration assumes release intensity to be proportional to differences between the current and critical values of velocity, salinity, pH, and stress. The shortcomings of the classical model for fines migration have been discussed in the literature. Another approach to fines-mobilization modeling is use of the maximum-retention function. This approach was adopted in the current work. Steady-state and quasisteady-state regimes of flow toward wells are described by analytical modeling. In the current work, the analytical model was developed for unsteady-state fines lifting, migration, and straining in fractured wells.

Treatment of Well Data

Production and drawdown data from 10 subsea deepwater fractured wells were modeled with an analytical model for unsteady-state flow with fines migration. The simulation results and the field data showed a good match, within 5%. Fig. 1 above, Figs. 2 and 3 show the time dependency of rate and drawdown for Wells 2, 4, and 6, respectively. These wells have been selected for discussion on the basis of their completion and production history.

jpt-2015-02-fig2finesmigration1.jpg
Fig. 2—Simulation results for Well 4 with an analytical model for well productivity with fines migration. Blue stars correspond to well data. The model match is shown in hollow black circles.

 

jpt-2015-02-fig3finesmigration1.jpg
Fig. 3—Simulation results for Well 6 with an analytical model for well productivity with fines migration. Blue stars correspond to well data. The model match is shown in hollow black circles.

 

Well 2 is the best producer in the field with zero water cut and negative skin. The frac-pack completion was executed with precision, achieving a hard tip screenout (TSO) with more than 600‑psi net pressure increase after TSO. As shown in Fig. 1, there is a good match between the model and the well data (R2=0.996). Moreover, it is important to note that the well drawdown (Fig. 1b) has a small impact on the flow rate under current conditions (Fig. 1a). This signature response was also observed for Wells 1, 9, and 10. The good quality of the completion (i.e., wider, more-densely-packed fractures, and full annular packs) may be the reason that this set of wells is able to withstand higher drawdowns and the effects of reservoir depletion.

As seen in Fig. 2, Well 4 experienced a rapid decline in productivity since the start of production. Well 6 also undergoes a fairly rapid drop in rate when the dimensionless time T equals 0.02 (Fig. 3a), which coincides with a decision to increase the well drawdown to arrest production decline (notice the change in slope in Fig. 3b at T=0.02).

As illustrated in Figs. 1 through 3, the fines-migration model discussed here adequately captures the well data, except for a relatively small time interval at the end of the production period where the modeling slightly underestimates the well drawdown. This can be explained by the reduced flow rate and the time dependency of fines transport from the reservoir to the wellbore. Although the analytical model accounts for rate decrease in the flow equation, it calculates the initial areal distribution of released fines from an average well rate; therefore, those particles lifted by higher rates early in the production history effectively arrive at the near-wellbore region later than predicted, causing an additional pressure drawdown over time.

Given the good match achieved between the well data and the model output, the analytical model can be used for long-term prediction of well productivity on the basis of the matched production history. Moreover, a 2D numerical model may be applied to history match the well data on the basis of the assumption of radial axisymmetric flow because the fracture half-lengths (18.0 to 29.4 m) are significantly smaller than the drainage radius (1500 m). Henceforth, the flow toward the fracture can be assumed radial, with the effective well radius equal to half of the well half-length. With this premise, the radius of the fines-mobilization zone was estimated to be in the range of 314 to 578 m, which also exceeds the fracture half-length and validates the assumption of axisymmetric flow. Another assumption of the analytical model is the condition of oil incompressibility in the drainage zone, which is reasonable given the negligible effect of oil compressibility on fines migration.

For a discussion of the sensitivity analysis carried out to determine the impact of the model matching parameters, please see the complete paper.

Laboratory Assessment of Fines Production

Laboratory experiments were conducted to assess the potential for fines production and permeability impairment caused by reservoir depletion and near-wellbore stresses at various flow conditions. It is important to note that conventional critical-velocity tests performed on core material before field development did not show evidence of fines migration. Those results were supported by X-ray-diffraction data and thin-section images of the core material, showing relatively low clay content and good cementation by tight grain-to-grain contacts.

One limitation of conventional critical-velocity tests is that they do not capture flow conditions in fractured wells adequately, in particular the effect of multiple operational shutdowns/restarts and pressure-buildup tests, which can induce stress cycling and frictional forces capable of breaking and crushing in the formation and the proppant pack, and releasing fine particulate material therefrom. Changes in reservoir stresses with depletion can intensify this effect further, increasing the potential for fines production, skin increase, and productivity loss through the life of the well. To address this gap, alternative methods to test fines migration in the laboratory were considered in this study.

The first set of tests involved single-phase-flow experiments with surging steps to simulate the effect of multiple shut-ins and restarts of production. A two-phase coreflow experiment was also conducted with coinjection of synthetic formation brine and treated kerosene, starting with kerosene at irreducible water saturation (Swi), and progressively increasing the water-cut ratio (oil/water ratios of 90:10, 75:25, 50:50) to 100% synthetic formation brine. For a discussion of the procedure for conducting single-phase and two-phase coreflow experiments, please see the complete paper.

Permeability to brine vs. cumulative injected fluid in pore volumes (PV) is plotted for each sample in Fig. 4. As seen in the figure, approximately 50% of the initial permeability for each sample is lost after 30,000 PV of cumulative injection, regardless of flow rate.

jpt-2015-02-fig4finesmigration1.jpg
Fig. 4—Single-phase coreflow tests with surging and constant flow rate. Permeability to liquid vs. cumulative injected fluid in PV: (a) 10 mL/min; (b) 100 mL/min. k=permeability; kair=air permeability; kinitial=initial permeability.

 

Unlike conventional critical-velocity tests, in which fines production and permeability impairment are correlated to a critical velocity for mobilization of fines, the results in Fig. 4 suggest that the amount of damage may be correlated to a critical cumulative volume of fluid injected through the sample, in this case 30,000 PV.

A possible reason that a number of conventional critical-velocity tests do not show evidence of fines migration, despite field evidence in support of it, is that these tests often flow a limited volume of fluid through the sample (e.g., a few hundred PV).

To assess the impact of multiphase flow and increasing water cuts on fines production in the vicinity of the fracture face and the annular pack—a region subjected to large volumes of fluid flow—an extended two-phase coreflow experiment was conducted by coinjecting synthetic formation brine and treated kerosene, starting with kerosene at Swi and progressively increasing the water cut (oil/water-injection ratio=90:10, 75:25, 50:50) until synthetic formation brine at irreducible oil saturation (Sor) was reached, as illustrated in Fig. 5. Care was taken to remove drilling-fluid filtrate from the sample while preserving the native wettability of the sample.

jpt-2015-02-fig5finesmigration1.jpg
Fig. 5—Relative permeability vs. cumulative fluid injected in PV for steady-state, native-state sample (extended two-phase coreflow test). kair=air permeability; ko=oil permeability; kw=water permeability.

 

Steady-state relative permeability vs. water-saturation data for an extracted sample, and a wettability-restored sample, are shown in Fig. 6, respectively.

jpt-2015-02-fig6finesmigration1.jpg
Fig. 6—Steady-state relative permeability vs. water saturation for (a) extracted sample and (b) wettability-restored sample.

 

Direct comparison of Figs. 5 and 6 reveals a significantly-more-damaged sample from the two-phase extended-flow test (Fig. 5). For instance, the relative permeabilities to oil (Ko) at 75:25 oil/water-injection ratio are 0.3 for the extracted sample in Fig. 6a and 0.4 for the wettability-restored sample in Fig. 6b, but only 0.1 for the sample subjected to extended flow (Fig. 5). At 50:50 injection ratios, Ko ranges from 0.2 to 0.25 for the extracted and the restored samples, respectively, but it drops to 0.035 for the sample in Fig. 5. The permeability to brine at Sor ranged between 0.25 and 0.45 for the extracted and the wettability-restored samples (Figs. 6a and 6b, respectively), but it was less than 0.05 in Fig. 5—an order of magnitude less than the samples tested by conventional relative permeability experiments. From these results, it is evident that the sample subjected to extended flow (greater than 30,000 PV) experienced some sort of mechanical damage beyond the damage normally associated with relative permeability changes. As in the single-phase coreflow tests (Fig. 4), the cumulative fluid injected through the sample appears to be a critical variable when it comes to assessing fines-production potential and permeability impairment in the laboratory. For a discussion of triaxial-stress tests, please see the complete paper.

The integrated approach to evaluating fines production discussed in this study offers an advantage over conventional critical-velocity experiments in that it evaluates the impact of rock stresses and various flow conditions, including extended flow periods, thus addressing potential fines migration from both native (i.e., movable silty/clayey particles present within the rock fabric) and nonindigenous particulates generated by the crushing and shearing of proppant and formation-sand grains.

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 165108, “Fines Migration in Fractured Wells: Integrating Modeling With Field and Laboratory Data,” by M. Marquez, W. Williams, and M. Knobles, Chevron, and P. Bedrikovetsky, SPE, and Z. You, SPE, University of Adelaide, prepared for the 2013 SPE European Formation Damage Conference and Exhibition, Noordwijk, The Netherlands, 5–7 June. The paper has not been peer reviewed.