Summary. This paper describesperforation/fracture tests performed perforation/fracture tests performed in large sandstone blocks in a triaxialstress cell to determine perforatinggeometry and perforating-fractureprocedures for optimal fracture procedures for optimal fracture initiation. Four-shot, 1800 phasedperforating guns, steel casing, and perforating guns, steel casing, and oilfield cement were used. In oneseries of experiments, the casing wascemented and cured while undertriaxial stress. Most tests were madewith pore pressure, and vertical andhorizontal wells were simulated. Ingeneral, the tests showed that(1) fractures initiate either at the baseof perforations or at the intersectionof the plane normal to the minimumhorizontal stress that passes throughthe axis of the wellbore and thewellbore surface and (2) fracture initiationdepended on perforation orientationwith respect to the plane normal tothe minimum horizontal stress andthe properties and injection rate ofthe fracture fluid. All fracturesreoriented into the plane normal to theminimum horizontal stress within onewellbore diameter; although multiplefractures were initiated, only primarysingle fractures propagated beyondone wellbore diameter. Introduction Laboratory simulation of fracturing throughcased and perforated wellbores generally hasbeen performed with one or more of thefollowing limiting conditions:scaled tests(rate effects ignored),artificial rock(cement, plaster of Paris, or hydrostone),isolation of the perforations from thewellbore (no pressurization between thescaled casing and the rock),noporoelastic effects (impermeable rock or poroelastic effects (impermeable rock or high-viscosity fracturing fluid),artificially scaled perforations(no perforation damage), andno pore pressure (impermeable ordry rock). Extrapolation of laboratory results todownhole situations must proceed withcaution whenever any of these conditions arepart of the laboratory experiment. For part of the laboratory experiment. For example, Condition 3 is unrealistic downholeand Condition 5 may apply only in specialsituations of high underbalance perforating. Conditions 4 and 6, however, may simulatea well with extensive wellbore andperforation damage. To the best of our knowledge, perforation damage. To the best of our knowledge, Warpinski's mineback experiments, inwhich annulus fractures were observed, were the only experiments conducted in theabsence of these conditions. The experiments discussed in this paperwere planned to evaluate the effect ofperformations on fracture initiation. Because of performations on fracture initiation. Because of the limiting-condition issues discussedabove, fill-scale experiments wereconducted in sandstone rock in a large triaxial stressframe that simulated downhole conditions. Test Fixture The fracture-initiation experiments wereperformed in 27×27×32-in. sandstone performed in 27×27×32-in. sandstone blocks in Terra Tek's 8, psitriaxial stress frame. The circular frame andits top and bottom platens surround the rockand provide reactions for three independentpairs of flat jacks. Flat-jack efficiency was pairs of flat jacks. Flat-jack efficiency was measured at 91 %; all applied stress datause this correction. Access to the centralwellbore is provided through the top loadingplate. Pore pressure was applied by placing plate. Pore pressure was applied by placing the rock in a stainless-steel can with rubberseals on the top and bottom faces. The canallowed the placement of about 0.25 in. ofbauxite beads around the four faces. Theborehole was cored to 4% in., and 3-in. OD × 2 1/2 -in. ID) steel tubing was cementedin place. Through-tubing, 1 11/16-in. orin., 4-shot/ft (SPF), 180 degs phased guns withthree or four shots were used. Table 1 givesthe rock properties. Experimental Procedure Three sets of tests were conducted (Table2). Experimental techniques were enhancedwith each additional set. Rock saturation andpore pressure were added to Set 2. An in-situ pore pressure were added to Set 2. An in-situ pore pressure gauge, a large 2.38-gal pore pressure gauge, a large 2.38-gal intensifier, and casing cemented and cured understress were added to Set 3. The general test procedure was as follows.1. Vacuum saturate the rocks with 3%brine, flowing brine from the uncasedwellbore to the rock sides (Sets 2 and 3 only).2. Cement casing into rock and allow tocure (Sets 1 and 2 only).3. Place rock into test frame. For Set 3, casing was cemented in place before theframe was closed and cured while understress.4. Place perforating gun in wellbore.5. Close frame, apply desired flat-jackand pore pressures; the wellbore is ventedto atmosphere.6. Fire gun. For Set 3, the casing cementcured for a minimum of 24 hours before thegun was fired.7. Flow brine at 2,000 psi from the beadpack through perforations; measure flow pack through perforations; measure flow rate8. Perform various prefractureprocedures depending on test set. procedures depending on test set. 9. Fracture rock with red dye used infracture fluids.10. Remove, cut open, and examine rock. Experimental Results Set 1, Torrey Buff Sandstone. Equipmentfailure prevented on of the specimensin this set. After perforation, the wellborewas flushed with brine, and on Tests 2 and4, brine was injected at low pressure throughthe perforations to saturate the rock locallynear the wellbore. JPT P. 608

Treating pressures from refracturing treatments and simulations of horizontal stress changes due to production suggest that the initiation of the refracture plane will be normal to that of the initial treatment. This possibility makes more wells candidates for refracturing.

Summary Although even a perfunctory survey of the literature suggests that considerable information is available on the response of finite-conductivity fractures in single-layer systems, the influence of the settling of propping agents and the effect of fracture height on the well response need to be examined. These topics are examined in this paper. We suggest methods to analyze well performance when the fracture paper. We suggest methods to analyze well performance when the fracture conductivity is a function of fracture height and fracture length. The performance of wells with fracture height greater than the formation performance of wells with fracture height greater than the formation thickness is documented. The consequences of being unable to contain the fracture within the pay zone are also examined. Although incidental to this study, we found that solutions presented by various authors are not in agreement for all time ranges. In this paper, we discuss a systematic procedure to obtain a grid (mesh) so that paper, we discuss a systematic procedure to obtain a grid (mesh) so that accurate results are obtained by a finite-difference model. This procedure can be used for both two-dimensional (2D) and three-dimensional (3D) problems. problems. Introduction This paper examines the performance of wells intercepting finite-conductivity vertical fractures. Although much work has been presented in this area of pressure analysis, several aspects of well behavior have yet to be examined. We examine some of these topics. In this work we examine the influence of vertical variations in fracture conductivity on well performance. Concerns regarding the effect of the settling of propping agents on well productivity addressed in this paper complement our work on the influence of lateral variations in fracture conductivity. We also examine situations where the fracture extends below and/or above the productive interval. We consider two possibilities:the fracture length is assumed to be fixed and the fracture height is variable (volume of fracture treatment is variable); andthe fracture volume is assumed to be fixed (the product of the fracture half-length and fracture height is assumed to be constant). The latter case is of interest if one is not able to contain the fracture within the pay zone of interest. In the former case, we show that this approach provides a means to increase the effective fracture conductivity. These topics have not previously been examined in the literature. Verification of the finite-difference model used in this study consumed a significant portion of the time spent in this study. Although some works have reported problems in obtaining accurate solutions, no guidelines for choosing a grid (mesh) for this problem are available. We give empirical guidelines for systematically choosing a grid to obtain accurate solutions. We believe that these guidelines will significantly reduce the time spent by researchers in developing their own models and consider it an important contribution. We also present methods to modify grids developed for a given set of conditions if the fracture and/or reservoir dimensions are changed.

The high near-wellbore pressure drop which has frequently been reported in fracture treatments is indicative of ineffective communication between the wellbore and the fracture. Although numerous observations of such effects have been published, few attempts have been made to understand them. This paper describes three possible mechanisms of near- wellbore effects: perforation phasing misalignment-induced rock pinching, perforation pressure drop, and fracture reorientation (deviation tortuosity) - and their implementation in a numerical fracture simulator. Typical signatures of all three effects in fracture treatment records are shown, and a method proposed by which to distinguish them. Perforation phasing misalignment has been identified as a cause of near-wellbore restriction. Because the fracture does not always initiate at the perforation, the fluid must communicate with the fracture through a narrow channel (micro-annulus) around the casing. The paper describes this pinching effect and shows that it is related to the contact stress between the cement and the formation. Perforation pressure drop and deviation tortuosity, which have previously been proposed by other authors, have also been modeled. They have been incorporated in the simulator, together with the phasing misalignment effect, to allow investigation of the differences in response between the different mechanisms. Results of simulating the effects of erosion of near-well effects on treating pressure are also shown. Introduction Over the past five years high near-wellbore pressure drop has often been observed in fracture treatments, and attempts have been made to understand the effect of near-wellbore conditions on the placement of a hydraulic fracturing job and to develop new methods to prevent unplanned screenouts. Near-wellbore pressure drops have been attributed to phenomena such as wellbore communication (perforations), tortuosity (fracture turning and twisting), and multiple fractures. These geometries were identified as being detrimental to the success of a fracturing treatment, because of the increase in net pressure and the increased likelihood of unplanned screenouts. Fracture geometry around a wellbore Several researchers have investigated critical mechanisms related to fracture initiation from vertical and deviated wells. P. 569

A substantial proportion of hydraulic fracture treatments are performed in formation and well completion settings conducive to the extension of multiple fractures, a process not adequately described by existing hydraulic fracture simulators. This paper introduces a procedure for the simulation of multilayer fracture treatments. The approach uses an analytical PKN fracture model to describe the behavior of each fracture and couples the behaviors using a set of constraints describing conservation of volume and continuity of pressure. The performance of the simulator is demonstrated by application to several examples.