A comprehensive dissimilarity method of modeling accuracy evaluation for discontinuity disc models based on the sampling window

摘要

Discontinuity disc models are widely used to construct discrete fracture networks. It is important to evaluate the modeling accuracy and select the best network, which is most similar to the real rock mass, from multiple simulated networks. This study proposed a comprehensive dissimilarity method for modeling accuracy evaluation. The core of this method is the proposed index, comprehensive dissimilarity (CD), between the simulated model and the real rock mass, which is the average of the shape dissimilarity (SD), location dissimilarity (LD), and size dissimilarity (ZD). Some 3-D discontinuity networks were analyzed to validate the proposed comprehensive dissimilarity method. The results show that the method proposed in this paper is validated to quantitatively test the difference between the simulated models and real rock mass when multiple factors change at the same time. Finally, a rock slope at a hydropower station was used as a case study to illustrate the utility of the proposed method. The index CD comprehensively incorporates the orientation, location, size, and density of discontinuities and can be calculated by a computer code, the procedure of which is efficient and the results of which are objective. Therefore, the proposed method for evaluating the accuracy of models is worth popularizing.