Inductive Detection of Fractures in Geothermal Systems - Petrophysical Aspects and Geophysical Implications

A reliable electromagnetic (EM) means of detecting and delineating geothermal fracture systems would facilitate geothermal reservoir development. To this end, there have been rapid advances in our geophysical capability to gather extensive electromagnetic data sets and to generate elaborate numerical simulations mapping electrical conductivity distributions to electromagnetic field responses. However, the nonuniqueness of electromagnetic interpretations requires that formal interpretations of EM field data be conditioned by geological constraints. Indeed, the electrical conductivity expression of a fault is far from intuitive and must be based on geological investigations. FLUID TRANSPORT IN FRACTURED RESERVOIRS: IMPETUS FOR INDUCTIVE LOGGING Fractures influence the migration and accumulation of groundwater, geothermal and ore bearing fluids, and hydrocarbons. Estimating the permeability of individual fractures and fracture arrays is an important problem with application to a wide variety of reservoirs in diverse environments (Price, 1966; Stearns and Friedman, 1972; Bruhn et al., 1997; Brown and Bruhn, 1998). The importance of fracture permeability to fluid production and injection in commercial reservoirs is well documented and in many reservoirs fracture permeability dominates over intergranular permeability (Hunt, 1996; Tsang and Neretnieks, 1998). This is especially the case in fine-grained sedimentary reservoirs and in geothermal reservoirs. Considerable effort has been expended in the past several decades to characterize natural fracture patterns by field mapping and analysis of drill core (e.g. Price, 1966; Stearn and Friedman, 1972; Kulander et al., 1979), in characterizing fluid transport properties of both discrete fractures and fracture networks (Long et al., 1982; Purak-Nolte et al., 1988; Brown, 1989; Oda et al., 1997; Brown and Bruhn, 1998; Tsang and Neretneiks, 1998), and obtaining better insitu measurements of fracturing and fluid flow using down-hole logging tools (e.g. Barton et al., 1997). In spite of this effort, predicting the occurrence, geometry, and fluid transport properties of fractures in reservoirs remains a difficult task, which is hindered by our inability to measure fracture heterogeneity and anisotropy insitu, by the tendency for flow to become concentrated or 'channelized ' within only a fraction of the fracture network, and by the response of permeability to stress changes and mineral precipitation during fluid production. Determining the geometry and fluid transport properties of fractures in the vicinity of a well bore is fundamental to understanding fluid production and well bore stability. Several existing logging tools provide useful, but limited information. Acoustic logs allow one to infer the presence of fractures, but provide little information on fracture geometry and transport properties. Bore hole televiewer and microscanner logs provide measurements of fracture orientation, and rough estimates of fracture filling and aperture, for those fractures that intersect the well bore. This latter information can be used to construct statistical models of fracture network geometry based on fracture orientation and frequency, but lack robustness because fracture size and connectivity must be modeled in an ad hoc fashion (Oda et al., 1997). Furthermore, recent work has demonstrated that the heterogeneity and anisotropy of fracturing and fluid transport properties are not the same in a number of cases. For example, Tsang and Neretnieks (1998) summarized evidence for flow channeling as a common phenomenon in heterogeneous fractured rocks and Barton et al. (1997) demonstrated that fluid production is concentrated along fractures most favorably oriented for Coulomb fractional (shear) failure rather than ubiquitously within a complex network. These and similar results from other regions are motivating development of new fracture permeability algorithms to account for the effects of stress on fracture closure (e.g. Bruhn et al., 1997; Brown and Bruhn, 1998), and also provide new impetus to develop logging methods that sample a volume of rock around the well bore in order to more directly characterize fluid conducting fractures. The purpose of this paper is to examine inductive means of characterizing fracture zones in particular, emphasizing the definition of an electrical conductivity structure of such a zone. CONDUCTIVITY MODELS FOR FRACTURED ROCK The morphological complexity, which can be encountered in fracture zones, is illustrated in several figures. Figure 1a shows a cross section of a fault zone. Although the zone as a dimensional scale of meters, the individual fractures are much smaller, and form a network. 1 m P r i m a r y S h e a r F r a c t u r e E x t e n s i o n F r a c t u r e S e c o n d a r y ( W e s t ) S h e a r F r a c t u r e Figure 1a: Fault zone cross section. Rather than "channels". Primary shear fractures have a given orientation, while secondary shears have a second orientation, and extension fractures have a third orientation. This pattern can be fractal and persist to different dimensional scales. D u s c h e s n e F a u l t O u t c r o p # 1 W i d t h = 1 . 1 6 m H e i g h t = 1 . 6 7 Figure 1b: Bedding surface fracture pattern. Figure 1b illustrates a bedding surface fracture pattern taken from the Duchesne Fault. Again, the fracture patterns are complicated and highly anisotropic. Fractures are also associated with folds. As Figure 2 demonstrates, these fractures again form networks with multiple orientations.