SPE Members
Abstract
Optimization of fracture length and well spacing can be critical to the economics of the exploitation of natural gas resources. If the fracture length is small compared to well spacing, then fracture azimuth will not affect interference between wells; many fields will be drilled to dense spacings and therefore require knowledge of the fracture azimuth to optimize the number of wells, the placement of the wells, and the length of the fractures. This is contrary to historical approaches to the optimization of ultimate recovery and net present value from tight gas fields, which are based on the idealization of homogenous, isotropic reservoirs and which neglect interwell interference. This paper reviews current technology for determining hydraulic fracture azimuth, and presents a solution for uniform flux fractures which accounts for fracture azimuth. It is shown that the uniform flux fracture model fails when interference between wells is significant. Numerical simulation of high conductivity fractures confirms that knowledge of fracture azimuth will be important for cases for which the ratio of interwell distance to fracture length is less than 2.0. Permeability anisotropies increase the importance of knowing fracture azimuth. A general procedure for optimization of fracture azimuth and well spacing is presented.
OVERVIEW Increased well density is typically required when fracture length is limited for technical, economic, or regulatory reasons,
the costs of hydraulic fracture treatments are a significant fraction of well costs, or well performance indicates substantially smaller fracture lengths than are indicated by design fracture calculations.
In these cases, knowledge of fracture azimuth becomes increasingly important for the optimal location of infill wells. How sensitive is optimization to the accuracy of the azimuth determination? The major works discussing the effect of fracture azimuth on tight gas production were Smith 1 who discussed the Wattenberg Gas Field and Lacy who reviewed hydraulic fracture azimuth detection methods. They argued that as fracture length grew large compared to the interwell distance, the effect of fracture azimuth increased. This case had been proven unambiguously for the case of waterflooding and fluid displacement. Smith quantified the effects based on a descriptive argument he described as 'somewhat qualitative' in nature. In this paper, a solution to this problem has been accomplished and illustrates that for most homogeneous, single-phase, depletion cases of practical interest, the effect of fracture azimuth is small. When permeability anisotropy is large and for infill drilling cases, knowledge of fracture azimuth is quite important. Figures 1-2 illustrate the concepts involved. It is clear that transient flow around an hydraulic (vertical) fracture is approximately elliptical at early times. Both of these figures show six wells equally spaced on the corners of squares. Fractures are shown which extend 40% of the distance from the well to the next square. For the 0deg. case, considerable overlap of the drainage patterns is obvious. For the 45deg. case, interference is expected to be much less. Although this representation is qualitatively useful, the actual drainage patterns which result from the interference of such wells are different from this simple approximation and complicates interpretation.
DETERMINING FRACTURE AZIMUTH
Although the technology is far from mature, numerous techniques have been developed for determining fracture azimuth. It is important to consider the accuracy and cost of these methods as well as their reliability. For example, a method which can only be used in an open-hole section may require a specially designed well. Several different types of techniques may be required if very accurate estimates are desired. There is no current methodology to assess the value of a given level of accuracy. After brief discussions of these methods, some example applications will be reviewed. of these methods, some example applications will be reviewed The two most common types of seismic body waves are compressional waves and shear waves. The compressional or p-waves are generally used in surface seismic data. The p-wave motion is in the direction of propagation and consequently is sensitive to the acoustic impedance of the material through which the waves are propagated.
P. 567^