Characterization of a fractured rock mass using Geological Strength Index (GSI): A Discrete Fracture Network (DFN) and Machine learning (ML) approach

The rock mass is strongly influenced by the presence of discontinuities and their role is also strongly regarded in rock mass characterization. Different traditional methods were developed for accessing the rock mass condition for safely designing engineering projects such as slopes, tunnels, foundations, etc. The progress in computational techniques has led to a significant understanding of rock mass related problems. Among them, the discrete fracture network (DFN) technique based on statistical distribution gains significant importance in examining the rock mass. The applicability of remote sensing techniques such as photogrammetry has made it easy to collect the essential data, which otherwise was difficult to acquire using scanline survey or window mapping. The study aims application of DFN in estimating block volume distribution and Rock Quality Designation (RQD) for finding the Geological strength index (GSI) of the rock mass. The results also compare the aggregate and disaggregate DFN with GSI estimated using traditional methods in the field. Along with the estimation of GSI using the existing chart method, the work also proposed the applicability of machine learning (ML) in predicting the GSI value. It is easy and handy to use a chart but becomes time-consuming when dealing with a larger dataset. We have developed a ML inbuilt python-based GUI tool to estimate the GSI value from block volume and joint condition parameters quickly.