Affiliation:
1. Pinnacle Technologies Inc.
2. Chevron Texaco Corporation
Abstract
Abstract
Water injection and reservoir fluid production result in poroelastic stress changes that can dramatically alter the created fracture geometry on infill wells. This basic conclusion is not new. It has been documented in many different environments, and is supported by theoretical modeling. However, this paper presents for the first time a large data-set of 76 fracture treatments in a concentrated area that not only shows stress reorientation, but also shows how fracture reorientation depends critically on the pattern of injectors and producers and their interaction. This knowledge can be used to improve recovery in water injection projects that depend on closely spaced fractured wells.
Data is presented from 76 fracture stages in 16 infill wells completed within a one-year period in Chevron's Lost Hills diatomite waterflood. Surface tiltmeter fracture mapping determined the fracture orientation (azimuth and dip) of the induced fractures, consisting of a vertical fracture component and a secondary sub-horizontal fracture component. This dataset clearly shows that the interaction between injectors (located in a line along preferred fracture azimuth) and producers results in a large-scale stress perturbation, producing a "room-and-pillar" stress structure. Both the fracture azimuth and the degree of secondary sub-horizontal fracturing are controlled by the location of infill wells with respect to nearby injector wells. Infill wells "inline" with injector wells yield fractures that grow close to the initial preferred fracture orientation. In contrast, infill wells that are "offset" from the line of injector wells yield highly variable fracture azimuths (often rotating towards injector wells) and greatly increased secondary sub-horizontal fracturing - both of which raise the risk of "short-circuiting" waterflood sweep. The data is presented and phenomenologically explained. A geomechanical model explains the observed stress changes, allowing predictive modeling of various infill-drilling scenarios in waterfloods to optimize recovery.
For a number of project treatments, downhole tiltmeter fracture mapping was also used to evaluate fracture dimensions, and fracture pressure analysis was performed to link tilt observations with treatment pressure behavior. Poroelastic effects have resulted in increased stress magnitudes in some layers and decreased stress in others, impacting the created fracture dimensions (height, length, and width). A summary of the mapped fracture dimensions and pressure analysis findings is presented together with a brief discussion of the poroelastic impacts on fracture geometry.
Field Setting
The Lost Hills field1,2 is an asymmetric anticline about 1 mile wide and 12 miles long, located approximately 45 miles northwest of Bakersfield, California. The structure is oriented NW-SE, which is nearly parallel to the San Andreas fault (located ~15 miles to the West). Although there is some production from other shallower and deeper intervals, the main reservoir rock is the late-Miocene Belridge diatomite, which is 600–1200 feet thick and located in the depth range of 1000–3000 feet.
The Belridge diatomite is primarily composed of siliceous diatom shells, the skeletal remains of single-cell shallow marine algae-like plants. As shown in Table 1, the rock is distinguished by its very high porosity (as high as 65%), low average permeability (~0.1 mD), and very low Young's modulus (average 100,000 psi). Lithology, and thus permeability and porosity, varies significantly across the diatomite column due to variable clay, silt and sand content, and the transition from amorphous Opal A quartz phase to Opal CT. The resulting lithological layers are well defined and are consistently present across the field. The large vertical variation in reservoir properties, combined with production and waterflood operations, has resulted in large variations in pore pressure, both vertically across the diatomite column and laterally from well-to-well.
Cited by
11 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献