Probabilistic predictions regarding key blocks using stochastic discrete fracture networks - example from a rock cavern in south-east Sweden

The natural variation of the fracture/joint network geometry and limited data access are the main sources of uncertainties in key block predictions. To make the uncertainties easily quantifiable the predicting variables should be presented as probability density functions. This paper illustrates the implementation of a stochastic discrete fracture network to predict the amount and size of key blocks around the rock cavern of the Central Storage Facility for Spent Nuclear Fuel (CLAB-2, Central Lager Anvant Bransle) in south-eastern Sweden. The data used in the study were effectively limited to fracture mapping in boreholes. The stochastic fracture model was generated with a FracMan discrete fracture simulator by adopting random fracture locations. Subsequently, the key block statistics along a simulated tunnel positioned inside the fracture model were generated. To illustrate the value of the predictions made, block statistics were undertaken for two different tunnel orientations. The methodology presented offers the potential to optimize the excavation design.