Calibration of Rockmass Failure Criteria for Mines Using Seismic Monitoring Data

Dmitriy Malovichko,A. Rigby

DOI: https://doi.org/10.56952/arma-2024-0309

2024-06-23

Abstract:A method is proposed for the calibration of a mine-scale numerical stress models using seismic monitoring data. The method is based on comparing the gridded observed and modelled seismic potency of slip-type seismic events according to a misfit/cost function. Expressions are derived for the gradients of this misfit function with respect to the parameters of Mohr-Coulomb failure criteria defining the intact and post-peak behaviour of the rockmass. Determining these gradients requires a single modelling simulation regardless of the number of model parameters; this affords a significant reduction in runtime compared to a simple finite-difference approximation, which requires a number of simulations equal to twice the number of model parameters. The gradient of the misfit function is used in an optimization algorithm (nonlinear conjugate gradient method) to update the failure criteria. The performance of the optimization is demonstrated using synthetic and real data. The proposed method makes it possible to promptly reconcile the numerical model of mines with recent seismic data, which is important for operational forecast of seismic and ground motion hazards for planned mining. An example of the effect of updating the stress model to the assessment of ground motion hazard is presented. Numerical stress models are used extensively in mine engineering practice. The models make it possible to assess the deformations and stresses in the rockmass, which have been disturbed by excavations (shafts, tunnels, stopes, caves, etc.). This helps to ensure that planned design of the excavations is safe. Modern numerical stress modelling algorithms are powerful. They can describe complex 3D geometry of mining and geology with high resolutions. Various constitutive relations are available to represent the behavior of the rockmass before and after the failure. One of the main challenges in modelling is the reliability and completeness of input information about rockmass properties. For example, the properties of rocks can be assessed for small samples (few centimeters size) extracted from boreholes and these need to be extrapolated to larger volumes (meters and tens of meters). Some parts of the rockmass may have sparse coverage by exploration drilling and, therefore, knowledge of their properties is approximate.

Engineering,Geology