The Value and Analysis of Core-Based Water-Quality Experiments as Related to Water Injection Schemes

Abstract

Summary This paper describes an experimental study of formation damage associated with low-concentration particle invasion in water injection schemes. Depth-of-invasion results are presented, the importance of core preparation is stressed, and the results from conventional cut-faced plugs are compared with those from the broken-faced plugs used in this study. Introduction The injection of water for reservoir pressure maintenance or secondary recovery purposes is important to oilfield development; therefore, the effect of water quality on well injectivities continues to receive attention. The main physical processes of formation damage in water injection wells can result from plugging by solids present in injection water, the entrainment and redeposition of in-situ fines, and the precipitation and deposition of scale in the flow constrictions of porous media. The degree of filtration -used to treat injection waters, including seawater, with very low particle concentrations varies within the industry. A common practice in coreflooding investigations, however, is to predict permeability impairment from laboratory cores prepared by cutting the core-plug ends with a saw. The obvious difficulty with cut-faced core is that the permeability characteristics of the inlet and outlet faces are altered. Numerous studies have been done on formation damage resulting from invading solids, particularly in relation to water injection where low-concentration micrometer and submicrometer solids are concerned. We discuss some of the more simplistic approaches from the literature in terms of the porous media (membrane filters and unconsolidated sandpacks), the duration of the tests, and the overall permeability measurements made in tests on nonstandard-length cores that give little or no information about the damage within the core. Further, because the permeability of the inlet face influences the formation of an external and internal filter cake, we question the validity of test results based on conventionally prepared cores. The influence of particle plugging on injectivity decline has been investigated by several researchers. Barkman and Davidson proposed a water-quality measurement that can be obtained directly from membrane or core-filtration data and can be used to calculate the rate of formation impairment in terms of the injection well's half-life. Anomalies were found, however, in their predicted data and in the actual behavior of the formation as a result of particle invasion. Their model also failed to project the depth of particle invasion into the formation. Donaldson et al. used 27-mm-long × 18-mm-diameter sandstone cores of three different ranges of pore size (10 to 30 mu m) through which slurries of 4- to 6-mu m quartz particles in 1 wt% brine were injected for up to 500 PV. Plugging was generally attributed to internal blockage of pores and cake buildup. In their statistical random-walk model, they defined a threshold pressure that corresponds to the critical velocity suggested by Maroudas. Donaldson et al.'s data relate to high solid concentrations; they did not investigate depth of invasion. The micromodels Muecke used to explain the behavior of fine-particle movement in a porous medium were made of a single-layer thickness of 200-mu.m commercial glass chips. His single-phase flow tests showed that the percentage of fines bridging the pores depends greatly on their concentration; i.e., increasing the concentration of fines increases their tendency to bridge. He also reported results for linear flow tests through unconsolidated sandpacks. Gruesbeck and Collins investigated the mechanism of entrainment and redeposition of naturally occurring fine particles in porous media. They used some permeability data from 300 PV of a glass-bead suspension injected into a sandpack. The amount of injected fines accumulated within the pack was presented for core experiments but no relation to depth of invasion was given. In tests of limited duration, Davidson used 31-mm-long × 25.4-mm-diameter artificial cores and 3-mu m silica particles to investigate constant flow rate. The superficial velocity corresponded to 12.5 to 50 m3/d m in a 203-mm-diameter wellbore. Todd et al. observed that the injectivity variations predicted by core studies were not observed in practice, particularly in North Sea water injection wells. They suggested that this lack of agreement could be caused by the existence of microfractures in the wellbore face that increase injectivities. Water-quality experimental results presented by Todd et al. gave permeability profiles obtained with a multiport pressure-tapped core holder that demonstrated that local permeability impairment was a function of particle size and rock properties. Depth-of-invasion results covered a range of particle sizes and core permeabilities. However, while general trends were observed, the data did not reveal consistent parameter variations. In these experiments, a conventional procedure was used to prepare the core-plug faces; i.e. the ends were cut with a core-trim saw. Todd et al. also constructed a network model with a built-in particle capture mechanism to predict permeability damage resulting from the depth of invasion of the solid particles. Simulation results, however, did not give an acceptable match to experimental data. Eylander proposed a new mathematical model to predict an injection well's half-life in relation to the quality of injection water from coreflood-filtration experiments. His method accounts for dynamic and plugging types of impairment. He observed that internal filter-cake formation initiates at the injection face and that the depth of impairment progresses inward. Eylander suggested further research to assess the factors controlling depth of impairment, which is an important factor in the prediction of injectivity decline rate. According to a literature survey conducted by Vetter et al., most past researchers have assumed that particles less than 2 mu m move away from the wellbore and reach locations in the reservoir where their impact on fluid flow is negligible. Their survey indicated that the mechanism of particle invasion and movement is not sufficiently known. Using a maximum permeability of 200 md in porous media, Vetter et al. concluded that, contrary to the commonly accepted philosophy, submicrometer particles can enter the reservoir rock and cause severe damage to relatively tight sandstones. They further concluded that particle penetration into the core material (depth of penetration) is related to the linear velocity of a given type, size, and concentration of particles. This paper presents results from laboratory investigations aimed atthe influence of core-plug preparation on particle invasion,the effects of particle penetration, flow rate, and particle concentration on formation damage; anddepth-of-invasion characteristics rather than overall permeability variation.