Static Proppant-Settling Characteristics of Non-Newtonian Fracturing Fluids in a Large-Scale Test Model

Abstract

Summary Large-scale testing of the settling behavior of proppants in fracturingfluids was conducted with a slot configuration to model realistically theconditions observed in a hydraulic fracture. The test apparatus consists of a1/2 × 8-in. [1.3 × 20.3-cm] rectangular slot 14 1/2 ft [4.4 m] high, faced with Plexiglas (TM) and equipped with pressure taps at 1-ft [0.3-m] intervals. Thisconfiguration allows both pressure taps at 1-ft [0.3-m] intervals. Thisconfiguration allows both qualitative visual observations and quantitativedensity measurements for calculation of proppant concentrations and settlingvelocities. We examined uncrosslinked hydroxypropyl guar (HPG) andhydroxyethylcellulose (HEC) fluids, as well as crosslinked guar, HPG, andcarboxymethyl HPG (CMHPG) systems. Sand loadings of 2 to 15 lbm/gal [240 to1797 kg/m3] (8 to 40 vol% of solids) were tested. Experimental results werecompared with the predictions of existing particle-settling models for a40-lbm/1,000-gal predictions of existing particle-settling models for a40-lbm/1,000-gal [4.8-kg/m3] HPG fluid system. Introduction The problem of particle or proppant settling in hydraulic fracturingoperations is of great importance. Knowledge of proppant-settling velocity isnecessary for prediction of the proppant-settling velocity is necessary forprediction of the final proppant distribution in the fracture. A thoroughunderstanding of proppant transport aids in designing fracturing treatments toobtain more effective proppant distribution. During a hydraulic fracturingprocess, some fluid is lost to the formation, thus increasing the proppantconcentration of the slurry pumped. Proppant settling in slurries is a verycomplex phenomenon compared with single-particle settling in a fluid.phenomenon compared with single-particle settling in a fluid. Single-particlesettling in a non-Newtonian fluid in the fracture has been studied extensively. In contrast, very little literature exists that explains the complexhindered-settling effects of particles in non-Newtonian fracturing gels. Without adequate particles in non-Newtonian fracturing gels. Without adequatetheory for description of this phenomenon, most fracturing-design simulatorstoday use correlations for single-particle-settling velocities in fracturinggels, and from the theory of Newtonian suspensions or slurries, correct thesevelocities to account for the hindered settling effects. For example, from hisexperimental work with water and 35-cp [35-mPa] oil, Novotny suggested using Richardson and Zaki's Newtonian-fluid correlations. Novotny demonstrated fromlimited experimental data that the correction for Newtonian-fluid hindranceeffects was also reasonably good for non-Newtonian fracturing gels. Novotnyalso reported that in some cases proppants agglomerated or clustered duringmeasurements of settling velocities of proppant-laden slurries flowing throughfractures. Kirkby and Rockefeller conducted static sedimentation experiments toevaluate the settling behavior of nonflowing slurries with thedifferential-pressure technique for measuring proppant concentration andsettling velocity in concentrated slurries. A proppant clustering phenomenonoccurs in all fracturing gels at proppant clustering phenomenon occurs in allfracturing gels at typical proppant concentrations, resulting in average staticslurry-settling velocities that are considerably greater than those of singleparticles. Dunand and Soucemarianadin used viscometric measurements to studysedimentation of single particles and suspensions in static HPG fracturingfluids. They showed that the viscosity function of these fluids can bepresented by the Ellis model and that theoretical predictions andexperimentally measured velocities of single particles agree well. Using gammaray attenuation measurements, Dunand and Soucemarianadin studied the settlingbehavior of concentrated proppant suspensions. They claimed that, in fluidswith identical single particle-settling velocities, the average settling rateof a concentrated suspension in a static non-Newtonian fluid is two to threetimes higher than in a corresponding Newtonian fluid. These two studies triedto answer some of the questions related to proppant settling in fracturing gelsin a multiparticle environment. Despite the contributions of these and otherauthors, however, further investigation is needed to understand this complexphenomenon fully. Therefore, an experimental study was undertaketo phenomenonfully. Therefore, an experimental study was undertaketo explain the complexphenomenon of particle settling in fracturing slurries. The objectives of thisinvestigation are two-fold:to design and build a large-scale testingapparatus for simulating the behavior of proppant-laden fluids in a fracture andto report observations of proppant settling in the presence of multipleparticles in controlled test conditions. Preliminary findings of particles incontrolled test conditions. Preliminary findings of sedimentation experimentswith various fluid systems also are reported. Theory When a particle settles in a given fluid and its fall is not affected by thecontainer boundaries or by the presence of other surrounding particles, theprocess is referred to as "free settling." When the particles, theprocess is referred to as "free settling." When the particles are neareach other, the motion of a particle is impeded particles are near each other, the motion of a particle is impeded by other particles and the process iscalled "hindered settling." Steinour proposed the following equationfrom his experimental study on the sedimentation of Newtonian suspensions: vH/vt = =, ..........................(1) where a = 1.82. Ramakrishna and Rao have used this equation extensively. Since Steinour's pioneering work, many investigators have derived the functionf() in Eq. 1. In the absence of wall effects, Richardson and Zaki proposed thefollowing empirical equation for N Rep less than 0.2: vH/vt = 4.65.........................................(2) They also give equations for other NRep ranges. Barnea and Mizrahi developeda generalized correlation: vH/vt =, ................................... (3) where f()={1+(1−)1/3 exp[5(1-)/3]}-1 ................. (4) Barnea and Mizrahi tested this correlation, which holds over the entirerange of applicability of the drag-coefficient curve, to NRep =3,500. Slatteryshowed theoretically that for creeping flow past a sphere in a non-Newtonianpower-law fluid, CD = 24X(n')/NRep,......................................(5) where X(n') is a function of power-law flow-behavior index, n'. Thus, X(n')can be considered a "correction factor" to the Newtonian equation. Using the definitions of CD and NRep, we can reduce Eq. 5 to v = gd (p-P)/18K'X(n'),..........................(6) Assuming that Stokes' law (Eq. 6) is modified by a factor f() when otherparticles are present, then v = [gd (Pp-Pf) /18K'X(n')]f ...................(7) SPEPE P. 305

Quantifying Proppant Transport in Thin Fluids - Theory and Experiments

“…However, strong, smooth particles are more likely to bounce off each other, separating and settling as individual particles. McMechan and Shah (1991) found that hindered settling is only important at proppant concentrations exceeding 10 lb/gal but that clustered settling can be important at much lower concentrations. The proppant concentration in most slickwater treatments is well below 10 lb/gal, even after leakoff, suggesting that clustered settling is the more important mechanism affecting proppant transport in suspension.…”

Fracturing Fluids

“…The rate of fall for proppant is normally calculated using Stoke's Law which can be written as: Stokes's Law is generally not valid for Reynolds numbers much in excess of unity 15 or for hindered settling due to proppant clustering in static fluids 16 . For crosslinked fluid the actual fall rate may be much less than Stokes Law.…”

Bed load proppant transport during slickwater hydraulic fracturing: Insights from comparisons between published laboratory data and correlations for sediment and pipeline slurry transport

“…Clustered settling can accelerate settling velocity at volume fractions below 0.1, but only in quiescent or slowly moving fluid (Kirby and Rockefeller, 1985;McMechan and Shah, 1991;Brannon et al, 2005;Liu and Sharma, 2005). Conversely, hindered settling can slow the rate of settling at high concentrations (McMechan and Shah, 1991). The horizontal superficial velocity can be defined as Vh.…”