The influence of rock fabric on the orientation of hydraulic fractures has been investigated for three wells in the Wattenberg field of Colorado. Oriented cores from the Sussex Formation of a depth of about 1560 m are shown to have a strong fracture anisotropy as determined from point-load, Brazilian and triaxial compression tests. The primary orientation is 337°-348°T, and a secondary alignment from 045°-060°T. This mechanical anisotropy is correlated with the fabric of the rock. The primary fracture orientation is aligned with the maximum residual strain direction in the bedding plane as determined by overcoring. The secondary fracture direction apparently is controlled by the preferred orientation of the dimensional axis of the elastic grains of the rock. All other fabric elements are either random in their orientation or not manifested. The fracture anisotropy is compared with field determination of the orientation of the induced hydraulic fractures. Their close coincidence is noticeable. These data are discussed with respect to previous work on the Muddy J Formation, at a depth of about 2400 m in the same area and differences are ascribed to local variations in the geologic structures producing the residual strain (Logan and Teufel, 1978a).
The presence of a strong fracture anisotropy argues against the traditional assumption of hydraulic fracturing methods that the rocks are isotropic. Either the prediction of the orientation of hydraulic fractures, or the determination of in situ stress states requires a consideration of properties of the rocks in addition to other conditions.
INTRODUCTION
The orientation of hydraulic fractures continues to remain enigmatic and a question for investigation in two areas: (1) the measurement of in situ stress, and (2) the production of hydrocarbon and other energy sources.
The measurement of in situ stress at depth is best approached through the use of hydraulic fractures, the basic assumptions of which are well defined by Hubbert and Willis (1957). Inherent in the use of this method is the premise that the induced fracture will develop only in response to the state of stress. That is, that the rocks are isotropic, homogeneous and continuous, and that their fabric or other characteristics do not influence the fracture orientation. It is also tacitly assumed that characteristics of the well bore, or the initiation and propagation processes do not affect the fracture orientation. Previous studies of five wells within the Wattenberg gas field of Colorado have clearly demonstrated the correlation of a dominant rock fabric element?residual strain?and the orientation of the induced fracture (Logan and Teufel, 1978a). Additional studies of other ares have suggested the importance of the character of the rocks when interpreting the condition of in situ stress (e.g., Daneshy, 1975; Engelder, 1982).
In addition to the basic assumptions used in interpreting the data, significant operational problems persist in determining the orientation of the induced fracture (and thus inferring the orientation of the stress field) from well-bore information. Workers have used a variety of technique to detect the fracture on the side walls of the well?impression packers, mechanical calipers, sonic logs, movie cameras, and televiewer cameras.
