| Abstract |
In modeling studies of geothermal reservoirs, quantitative characterization of fracture network systems is essential and fractal geometry has been recently shown to be useful for quantification. Fractal geometry not only provides a characterization and a quantification of fracture systems but it can be used to generate representative patterns synthetically. For this generation process, however, reliable estimates of fractal properties of the fracture network are required. In this study, an extensive fractal analysis of fracture patterns collected from the outcrop of producing geothermal reservoirs in southwestern Turkey was performed and the preliminary results are presented. Data used throughout the study were collected at different scales, i.e., aerial (order of kilometers), outcrop (order of meters), rock (order of centimeters), and thin section (order of microns) photos. Outcrop photos were obtained on a field trip to outcrop locations of the production formations of different geothermal fields including the major ones, Kizildere and Germencik Fields. These patterns are generated from the photographic images through an image processing application. First, the box counting technique is applied to fracture patterns at four different scales. It is observed that the box counting dimensions change with the scale of the fracture traces and, as the scale increases, more complex networks are obtained resulting in an increase in fractal dimension. This is an important issue because knowing the scale limits and selection of scale, at which the maps are prepared, are critical points in data preparation (e.g., conductivity data) for simulation studies. Fracture properties that might be related to the conductivity of the fracture networks such as density, connectedness and length are also considered and the fractal dimensions of their distributions are measured by different methods. Box counting (to measure the ìbox dimensionî) and sand box techniques (to measure the ìmass dimensionî) are applied for these measurements on natural fracture patterns at mega scale (order of meters). It is observed that the distribution of all these properties mostly represents fractal characteristics but different methods may yield notably different fractal dimensions for the same fracture property. |