Generation and verification of three-dimensional network of fractured rock masses stochastic discontinuities based on digitalization

Abstract

Accuracy of seepage simulation in rock masses is directly influenced by related geometrical parameters of DFN (discrete fracture network) models. In this paper a well-planned investigation of rock mass discontinuities was conducted on well-exposed outcrops in Beishan (Gansu province), one of the main candidate sites for the Chinese high-level radioactive waste repository using digital techniques, e.g. global positioning system, close range digital photogrammetry and geographical information system. This allows for very efficient collection of large volumes of data, over 10,000 discontinuities were obtained at the study site (Jijicao block), based on statistical parameters two homogeneities were identified, and the dominant sets were also delineated, which suggested that two sets exist in homogeneity I, while three sets in homogeneity II. In homogeneity I, The optimal fracture diameter probability distributions of both sets were lognormal. Three-dimensional stochastic DFN models were established for both homogeneities based on stochastic process as well as statistical parameters. The reliability of the model was validated by both graphical and numerical methods. Results showed that: digital techniques used in this study can overcome the problem of sample size and precision restriction posed by traditional methods, which leads to a more effective application of DFN model, and are thus suitable for tasks such as detailed mapping of rock masses in large area. It is believed that this verified DFN model is more appropriate for subsequent calculations of mechanical properties and seepage of fractured rock masses.