This paper is a case study of the stimulation and testing of tight, lenticular sands in the paludal interval of the Mesaverde group in the Piceance basin at DOE's Multiwell Experiment (MWX) site in Colorado. Topics discussed include geologic data, stress test results, well testing, laboratory core studies, stimulation and stimulation analyses, and postfracture operations.

A multi-well test program designed to study the gas production mechanisms of the eastern Devonian shale reservoirs was completed. Two offset wells were drilled as observation wells in Meigs County, OH. This paper presents the engineering design of the tests, data acquired, analysis techniques, and results of the analysis. The results indicated a complete anisotropic, layered reservoir system which implies directional gas flow and orientation of natural fractures. This study has provided an insight into the production behavior of reservoirs. It will aid future development of shale gas by optimizing well spacing and understanding of the gas release mechanisms of the Devonian shales.

The Multi-Site Hydraulic Fracture Diagnostic Project was initiated to develop independent diagnostic technologies and methodologies that will result in increased accuracy in measuring hydraulic fracture dimensions. Through a series of field experiments, the project has provided key insights into the time-dependent growth and total dimensions of hydraulic fractures. Wire line-retrievable and cemented-in triaxial accelerometer arrays provide the basis for plan and profile maps of microseismic events associated with the hydraulic fracture. A vertical inclinometer array provides data on the earth's mechanical response to fracture opening and closure. A directionally drilled wellbore remotely intersected the hydraulic fracture interval, verified the accuracy of microseismically imaged hydraulic fracture azimuth, and provided ground-truth observations on the far-field character of hydraulic fractures. Although eleven far-field hydraulic fractures were observed, it cannot be definitively determined if they are the result of multiple injections. Collectively, data resulting from M-Site research indicate that hydraulic fracture wings are not always symmetrical, rapidly attain their full length extent with only limited fracture height growth, and that hydraulic fractures propagated only with fluid retain their width for a long time period. Although these results apply to the zones being tested at the M-Site, they should also transfer to other zones with comparable thickness and stress contrast. Introduction Project Objective and Initiation Logic. The jointly funded Gas Research Institute (GM) and U.S. Department of Energy (DOE) Multi-Site (M-Site) Project is focused on performing field-scale experiments and gathering high-quality, independent diagnostic data that will result in increased accuracy in measuring hydraulic fracture dimensions. The ultimate goal of the project is to develop the hardware and methodology necessary to establish a microseismic-based system that can be used as the foundation for a commercial hydraulic fracture mapping capability. Several inter-related technology- and economics-based reasons prompted GRI and DOE to initiate a project of this magnitude. Each organization has a long history of supporting research directed toward determining the size and shape of hydraulic fractures. These research efforts include the development of real-time models for history matching of fracture treatment pressures, the collection and analysis of rock and reservoir properties that control fracture growth (e.g., in-situ stresses and rock properties), the estimation of fracture length and conductivity based on production, and the development of various fracture diagnostics for azimuth and height measurement. However, throughout all of these efforts, no reliable tool or technique was developed for determining the total fracture extent, primarily because of the difficulty in estimating fracture length. Fracture modeling is currently the most common method for estimating hydraulic fracture extent. However, without a definitive method for imaging or mapping the hydraulic fracture extent there is no field-scale ground-truth measurement with which to compare fracture model results. Lacking an accurate measurement of fracture dimensions, modelers and stimulation treatment designers are likely to continue to have diverse opinions regarding the importance of various fracturing mechanisms (e.g., tip effects, complex fracturing and height growth) and controlling parameters (e.g., closure stress measurements and in-situ modulus). P. 315

The U.S. DOE's Multiwell Experiment (MWX) was a field laboratory aimed at improved characterization and gas production from low-permeability reservoirs typified by the Mesaverde Group in western Colorado. A broad spectrum of activities was conducted over 8 years at a site containing three closely spaced « 225 ft [ < 68 m]), deep (7,550 to 8,350 ft [2300 to 2550 m])wells. The results yielded insights and contributions into the technology of gas production from this resource.

Summary Gas-production characteristics of naturally fractured Devonian shale have been quantified through a three well interference field test by use of an established producing well and two offsets placed on the primary and secondary regional fracture trends relative to the producer. Three individual shale zones were evaluated simultaneously by buildup, drawdown, and pulse tests to investigate reservoir gas flow characteristics, natural fracture properties, and gas storage and release mechanisms. Test results show severe permeability anisotropy, indicating elliptical drainage pattern with an 8:1 axis ratio. Essentially all gas is stored in a sorbed state in the shale matrix and is transported toward the wells through the native fracture system. Introduction The Devonian shales of the Appalachian basin underlie approximately 68,000 sq miles [175 000 km] from New York to Tennessee. These massive shales range in thickness from a few feet at the basin margin outcrops to thousands of feet at the basin center, and are made up of both organically lean ("gray") and rich ("brown") intervals. The shales contain natural gas in considerable quantity, primarily held in solution in solid organic constituents (kerogen) of the shale matrix, which makes the gas resource truly "unconventional." Recent resource estimates for the Appalachian basin shales range from 585 to 2,500 Tcf [16.5 × 10(12) to 70.8 × 10(12) m3] of natural gas in place, but because of the extremely low matrix permeability of the shale, gas is often not economically recover-able by conventional industry practice. The approximately 12,000 wells drilled to date historically have recovered only about 3 Tcf [0.08 × 10(12) m3] over the last 50 years. The Appalachian basin shales are considered blanket formations because discrete members are correlative over wide geographic areas. However, even though stratigraphically continuous and gas-containing across their extent, the Devonian shales do not produce uniformly when drilled. Current commercial production is a function of connecting the well with the primarily vertical natural fracture systems present in the shale, which form gathering and transportation networks to move gas from the matrix to the wellbore. Historical production has been limited to discrete areas where natural fracture density was high enough to support development. The Eastern Gas Shales Project (EGSP) is a long-term R and D effort by the U.S. DOE to improve overall gas recovery from the shale and to stimulate development of the resource by the private sector. An important part of the R and D thrust has been the development of a technical data base on shale characteristics and production behavior. The offset well test (OWT) described in this paper is a field experiment conducted to improve understanding of the basic gas-production mechanism of this unconventional reservoir. The test project was designed to investigate both qualitatively and quantitatively the gas-production characteristics of the shale in an area of natural fracturing and commercial development, and used a series of reservoir interference tests to achieve the test objectives. Specifically, the experiment addressed the following objectives. 1. Investigate the tow mechanics of gas in shale matrix and fractures. 2. Determine fracture orientation and distribution. 3. Determine how gas is stored in and released from the shale. 4. Verify the existence of directional drainage patterns and their impact on production practice. A three-well pattern was developed. consisting of an existing gas well with a 22-year production history and two newly drilled holes offsetting the producer by 120 and 90 ft [36.6 and 27.4 m], respectively, on the major and secondary directional fracture trends predicted for the test area. The interference series was conducted by perturbing the shale reservoir in the base well and monitoring the effects in the offsets. Fig. 1 shows a schematic of the OWT layout and instrumentation used during the interference testing. Earlougher provides a thorough description of interference testing, as well as many references on the subject. Site Selection The established producer (control well) used in the OWT formed an important part of the experiment and was carefully selected. During the test planning phase, design criteria were developed to define well parameters required for the planned interference tests. Table 1 lists these requirements. JPT P. 291^