Altered-str~ss fracturing is a concept whereby a hydraulic fracture in one well is reoriented by another hydraulic fracture in a nearby location. A field test was conducted in which stress changes of 250 to 300 psi [1 .7 to 2.1 MPa] were measured in an offset well120 ft [37m] away during relatively small minifractures in a production well. Results show that altered-stress frqcturing is possible at this site and others. FRACTURE DIAGNOSTIC WELL 0 MWX-3 ADDITIONAL STRESS GENERATED BY FRACTURE WILL LIKELY TURN IN ONE OR BOTH DIRECTIONS FIRSTFL__~~-EFFECTIVE DISTANCE-f -, FOR STRESS DECAY 90 It VIRGIN" t:

A hydraulic fracture stimulation conducted during 1983-84 in nonmarine, deltaic. Mesaverde strata at 7,100 ft [2164 m] was cored in a deviated well in 1990. The observed fracture consists of two fracture intervals, both containing multiple fracture strands (30 and 8, respectively), abundant gel residue on the fracture surfaces, and many other complexities.3.7 x 10 6 psi [25 GPa] and a Poisson's ratio of 0.22. Few natural fractures 42 ,43 were found in the vertical core taken from the paludal interval (numerous fractures were found in the deviated SHCT core), but nearly all natural fractures found at MWX trended nearly east/west, indicating a highly anisotropic permeability system.Stress tests throughout the paludal interval yielded relatively high stress contrasts between the sandstones and the nonreservoir lithologies. 41 ,44 The azimuth of the maximum horizontal stress was found to be about 65° north of west at this depth. 44Zones 3 and 4 were perforated with two 14-g jet shots per foot from 7,120 to 7,144 ft [2170 to 2177 m] in Zone 3 and from 7,076 to 7, 100ft [2157 to 2164 m] in Zone 4. The 96 perforations were then broken down and balled out with 60 bbl [9.5 m 3 ] of 2 % KCI water.Extensive well testing in paludal Zones 3 and 4 gave an effective reservoir permeability of 36 Itd. 40 ,45,46 No interference was observed between wells during a 7-day drawdown at 200 to 250 Mcf/D [5565 to 7080 m 3 /d] or during a 7-day buildup. Calibration Injections. A step-ratelflowback test and a pumpin/flowback test 40 ,41 conducted in the two sandstones yielded a closure stress of 5,900 psi [40.7 MPa]. Each injection was about 120 bbl [19 m 3 ] of 2 % KCI, with the peak pressure reaching 600 psi [4.1 MPa] above closure. These pump tests were conducted in Dec. 1983. Immediately after the flowback tests, two minifractures 40 ,41 were pumped into Zones 3 and 4 to determine fracture parameters and to observe growth behavior in these lenticular sands. In Minifracture 1, 15,000 gal [57 m 3 ] of 30-lbm [14-kg] linear gel was pumped at 10 bbl/min [1.6 m3/min]. In Minifracture 2, 30,000 gal [114 m 3 ] of 60-lbm [27-kg] linear gel was pumped at 10 bbl/min [1.6 m 3 /min].

Summary Vertical distribution measurements of the minimumprincipal in-situ stress in the lower Mesaverde group principal in-situ stress in the lower Mesaverde group (7,300- to 8,100-ft [2225- to 2470-m] depth) at the U.S. DOE'sMultiwell Experiment (MWX) site have been made by conducting small-volume, hydraulic-fracture stress teststhrough perforations. Accurate, reproducible results wereobtained by conducting repeated injections in each zoneof interest with a specially designed pump system, modified high-resolution electronic equipment, and adownhole shut-off tool with a bottomhole pressure (BHP)transducer. Stress tests were conducted in marines and stones and shales as well as in coal, mudstone, andsandstone in a paludal depositional environment; these testsprovide a detailed stress distribution in this region. provide a detailed stress distribution in this region. The stress magnitudes were found to depend on lithology. Marine shales above and below the blanket sands have large horizontal stresses that are nearlylithostatic, with a fracture gradient greater than 1.0 psi/ft[23 kPa/m]. This indicates that these rocks do not behaveelastically and processes such as creep and possibly fracturing are the dominant mechanisms controlling the stressstate. Sandstones and siltstones have much lower stresses. with a fracture gradient of 0.85 to 0.9 psi/ft [19 to 20kPa/m]. Containment of hydraulic fractures would beexpected under these conditions. Only three data points wereobtained from the paludal interval; no significant stressdifferences were observed in the different lithologies. Introduction The vertical distribution of the minimum principalhorizontal in-situ stress, has a significant influenceon hydraulic fracture geometry. Perkins and Kern notedits importance with respect to fracture height, and Simonson et al demonstrated how to calculate fracture heightin a nonuniform, but symmetric, stress field. Laboratory and mineback experiments have provedthe effect of differences on fracture height, but, as yet, field experiments have not yielded conclusive results. This results primarily from the lack of detailedin-situ stress data and viable fracture height measurement techniques. In addition, few in-situ stress measurementshave been obtained in intervals where core is availableso that stress/rock-property correlations can be attempted. The present work on in-situ stress measurements is partof DOE's MWX program, which is being conducted inthe Piceance basin near Rifle, CO. In-situ stressmeasurements currently are planned throughout the entire4,000 ft [1200 m] of Mesaverde rocks encountered at thislocation, with particular emphasis on obtaining detailedstress measurements around formations to be stimulated. Over 4,000 ft [1200 m] of core have been obtained from three closely spaced wells (130 to 180 ft [40 to 65 m])so an abundance of core data is available. Complete conventional log suites, as well as various advanced and experimental logs including the long-space sonic logs, havebeen run. This paper presents the results of the initial series of in-situ stress tests, which were conducted at thebase of the Mesaverde in marine sandstones and shales and at various horizons in a paludal zone. These data willbe used to design hydraulic fracture treatments and aidin the analysis of postfracture performance. In-Situ Stress Measurements At present, the only reliable method of obtaining distribution is by measurement with small-volume hydraulic fractures. Two techniques are currently in use. The step-rate/flowback procedure, pioneered by Nolteand Smith, yields a reliable, reproducible estimate ofand typically is conducted in an interval soon tobe stimulated. These tests have been called "minifraes"because they use small volumes of fluid (500 to 10,000gal [2.0 to 40 M ]) compared with conventional fracture treatments. The technique used in this study also is calleda "minifrac, "although it uses much smaller pumped volumes (1 to 250 gal [0.004 to 1.0 m ]). Minifracsusually are conducted by injecting a small volume intothe formation, shutting in, and measuring the instantaneous shut-in pressure (ISIP). For openhole tests, several authors have discussed the technique'sdetails, and it is clear that when adequately conducted, the test yields an accurate, reproducible estimate ofand a somewhat less reliable estimate of the maximum horizontal in-situ stress,. For most oil and gas applications, however, it isimpossible or impractical to conduct these tests in anopenhole environment. Problems with hole stability, gaspressure, cementing, and cost factors usually require that pressure, cementing, and cost factors usually require that the tests be conducted in cased holes through perforations. This causes additional complications because of casing, cement annulus, explosive perforation damage, andrandom perforation orientation effects. JPT p. 527

Six hydraulic-fracture injections into a fluvial sandstone at a depth of 4500 ft were monitored with multi-level triaxial seismic receivers in two wells, resulting in maps of the growth and final geometry of each fracture based upon microseismic activity, These diagnostic images show that the hydraulic fractures are highly contained for smaller-volume KC1-water injections, but height growth is significant for the largervolume, higher-rate, higher-viscosity treatments. Fracture lengths for most injections are similar. Final results are also compared with fracture models.

Experiments at the GRI/DOE M-Site have shown that hydraulic fractures have a considerable degree of complexity that is difficult to account for using current understanding. Several of the key features are reviewed and an attempt is made to examine them relative to fracture mechanics, elasticity, and fluid-flow mechanisms. Key features include unexpected fracture containment, multiple fracture strands, secondary and T-shaped fracturing, large pressure drops down the fracture, significant differences in microseismically imaged geometry as a result of fluid-system changes, unexpectedly large residual width/deformation measurements after unpropped fracturing, large discrepancies in fracture area (and, thus, volume) between imaging and modeling results, evidence of complexities in proppant transport, and measurement of a clearly defined fracture closure pressure using a pressure independent technique. P. 395