Section V Fracture Analysis
Thirty-six sets of surface lineaments and fractures mapped from satellite images and aerial photos from parts of the Mid-continent and Colorado Plateau regions were collected, digitized, and statistically analyzed to obtain the probability distribution functions of natural fractures for characterizing naturally fractured reservoirs. The orientations and lengths of the surface linear features were calculated using the digitized coordinates of the two end points of each individual linear feature. The spacing data of the surface linear features within an individual set were obtained using a new analytical sampling technique that involves overlapping a set of uniform imaginary scanlines orthogonally on top of an individual fracture set and calculating the distance between two adjacent intersection points along each scanline. Statistical analyses were then performed to find the best-fit probability dis-tribution functions for the orientation, length, and spacing of each data set. Twenty-five hypothesized probability distribution functions were used to fit each data set. A chi-square goodness-of-fit value was considered the best-fit distribution.
The orientations of surface linear features were best-fitted by triangular, normal, or logistic distributions; the lengths were best-fitted by PearsonVI, PearsonV, lognormal2, or extreme-value distributions; and the spacing data were best-fitted by lognormal2, PearsonVI, or lognormal distributions. These probability functions can be used to stochastically characterize naturally fractured reservoirs.