Barium and Calcium Sulfate Precipitation and Migration Inside Sandpacks

Abstract

Summary This paper is an experimental and theoretical study of permeabilityreduction of sandpacks caused by solids generation permeability reduction ofsandpacks caused by solids generation and migration. Two solids, calcium andbarium sulfate crystals, were injected into and generated within the sandpacksby chemical reaction. We found that we could simulate the effluentconcentrations, including the produced solids, with a fractional flow model forsolids migration. Introduction Mineral precipitates, like barium and calcium sulfate, are common oilfieldscales that can cause a significant reduction in permeability by plugging porethroats of reservoir rocks. Precipitates form because of the pore throats ofreservoir rocks. Precipitates form because of the interaction of injected waterand other chemicals with the subsurface rock. Mechanisms by which a precipitate reduces permeability include solidsdepositing on the pore walls because of attractive forces between the particlesand the surface of the pore, a single particle blocking a pore particles andthe surface of the pore, a single particle blocking a pore throat, and severalparticles bridging across a pore throat. The characteristics of the precipitateinfluence the extent of formation damage. Such conditions as a large degree ofsuper-saturation, the presence of impurities, a change in temperature, and therate of mixing presence of impurities, a change in temperature, and the rate ofmixing control the quantity and morphology of the precipitating crystals. This paper describes linear flooding experiments in sandpacks whereprecipitation, dissolution, and migration of barium and calcium sulfateprecipitation, dissolution, and migration of barium and calcium sulfateoccurred. Type A experiments investigate the effects that precipitates have onthe permeability reduction of unconsolidated sandpacks. The initialpermeabilities of the sandpacks were adjusted with a mixture of silica flourand sand. In Type B experiments, we measured the ion and solid wave producedwhile the precipitation/dissolution reactions were occurring in the sandpacksand interpreted the results with a geochemical flow model. The solids migratinginside the sandpack were traced with an X-ray computerized tomography (CT)scanner. The objective of this work is to investigate permeability reductionmechanisms and the extent to which they can be quantified. Previous Work Previous Work The formation of a precipitate starts withnucleation, a process in which attracted atoms join to form submicron nuclei. When impurities are present in the fluid, the energy required to form thesenuclei is smaller than in a pure solution. The impurities act as nucleationsites, and this process, pure solution. The impurities act as nucleation sites, and this process, called heterogeneous nucleation, occurs commonly insubsurface reservoirs where fluids contain mineral particles in suspension. Laboratory Studies. Several experiments studied the extent of permeabilitydamage caused by flowing precipitates in cores and sandpacks. Read and Ringentested incompatible waters from the North Sea that produced barium andstrontium sulfate precipitates. They reported that 15- to 20- m crystalsblocked the pore throats inside glass-bead packs and alumina cores by sizeexclusion and bridging mechanisms. Crystals often grew perpendicular to thepore walls, and some crystal aggregates also had the perpendicular to the porewalls, and some crystal aggregates also had the form of "books" androsettes. Todd and Yuan also conducted laboratory investigations using North Sea reservoir brines that produced barium and strontium sulfate scales. Most ofthe reduction in core permeability was caused by crystals depositing along andgrowing perpendicular to the pore surface. They observed that doubling thesupersaturation ratio of both barium and strontium sulfate produced an increasein the quantity of scale formed inside the pores and a change in the morphologyof the crystals. Both changes increase the rate of permeability decline. Thebasic structures of crystal formed under dynamic flow conditions and in astatic solution are similar, but some faces of the crystal disappear and othersbecome more plate-like when subjected to shear forces. Results of Cusack etal.'s injection tests on cemented glass-bead packs showed a 95.5 % reduction inpermeability because of the bridging of 2- to 10- m calcium carbonate crystalsthat blocked 33- m pore throats. Potter and Dibble showed how iron oxyhydroxidecolloid particles block the pores of a quartz sand. They suggested twomechanisms by which the plugging of a quartz sand occurs: flocculation/coagulation of the iron oxyhydroxide phase producing a filter cake near theinlet of the sandpack and colloid-quartz surface interaction by which theparticles attach to the quartz surface. Theoretical Studies. In 1984, a geochemical flow model was developed forboth mineral dissolution and precipitation processes during convectivetransport. The model was based on the assumptions of local thermodynamicequilibrium and negligible dispersion. The model did not allow for solidmigration, an assumption that was severely limiting because experimentalstudies demonstrate that precipitated solids can migrate with the fluid. In 1988, a theory of solids fractional flow was proposed in an attempt tocharacterize the local permeability reduction by incorporating the dominantpore-plugging mechanisms. A fractional flow function for each migratingmineral, coupled with the mass conservation of each element and the assumptionof local equilibrium between the solid and aqueous species, provided a completedescription of reaction during flow through permeable provided a completedescription of reaction during flow through permeable media. Experimental Setup The sand was packed in a 9-in. [22.86-cm] -long aluminum tube with a 1-in.[2.54-cm] I.D. The tube had four pressure taps along its length to measurepressure drops. In Type A experiments, one of the solutions was injectedpressure drops. In Type A experiments, one of the solutions was injectedthrough the three injection ports shown in Fig. 1. The second solution wasinjected through a port in the aluminum cap at the entrance end of the tube. In Type B experiments, the fluids were injected only through the port in the capat the entrance end. Caps at each end of the tube had a port in the cap at theentrance end. Caps at each end of the tube had a fiberglass cloth thatprevented sand grains from escaping. Fig. 2 shows the experimental apparatusmounted inside the CT-scanning apparatus. The Ottawa sand had a mean grain size of 0.095 mm and a sorting value of0.48, indicating that it was a well-sorted sand (Fig. 3). Silica flour withless than 0.045-mm grain size was mixed with Ottawa sand to produce sandpackswith less permeability than the sand alone. Fig. 3 shows the grain-sizedistribution for the 10 wt % silica flow mixture. Only the larger grain sizesare affected (recall that the Wentworth scale is -log2 of the grain diameter inmillimeters). The sorting coefficient has increased from 0.48 to 0.67. Weprepared the sandpacks to be free of clays and fines to study precipitation, dissolution, and migration of specific scaling crystals. The chemical reactants used were barium chloride dehydrate, calcium chloridedihydrate, anhydrous sodium sulfate, and calcium sulfate anhydrate. SPEFE P. 79