An Analysis of Trace Information of Different-shaped Fracture Networks Having a Same Fracture Intensity (P32)

Discrete fracture network (DFN) simulation is the basis of studying the properties of a rock mass. An important premise of constructing three-dimensional (3-D) DFNs is to ensure that the fracture intensity ( P 32 ) of the simulated rock mass equals to that of the real one. In fact, the natural fracture shapes are various, and we can only obtain one-dimensional (1-D) and two-dimensional (2-D) measurements of a rock mass and usually infer 3-D DFNs based on 1-D and 2-D measurements. The following question would be induced: whether the DFNs with the same P 32 and different shapes have similar 2-D trace information. Hence, a series of hypothetical 3-D elliptical DFNs with different values of long-short axis length ratio k e are constructed, and disc models are used to represent them. The algorithms of obtaining traces on sampling windows concerning circular and elliptical fractures are developed. The results show that as the values of k e increase, the error rate of the trace numbers and lengths between the elliptical fractures and simulated disc models on the same sampling windows increases. Moreover, because of the significant impact of fracture shape, the Accuracy representation Index is proposed to better evaluate which model is more appropriate for constructing DFNs.