Development of numerical geomechanical models with different levels of detail using the example of the Angidrit underground mine of the Kayerkansky ore mine

Abstract

As mining conditions are getting more complex, numerical modelling is becoming one of the most promising areas to obtain data for developing efficient technological solutions. However, creation of high-quality numerical models is an extremely labour-intensive and knowledge consumptive task. Optimization of the numerical modelling process is currently highly demanded. The use of global numerical models with a high level of detail not only makes it possible to evaluate the stress-and-strain state of the rock mass over a large area, but also to qualitatively assess some local effects. Such models help to select more correctly the most hazardous areas for designing local calculation models. Thanks to the high level of detail of the simulated underground structure, the global numerical model can act as a "donor" of the initial geometry for local calculation models, and the exported stress tensor can be used as the boundary conditions, which will increase the accuracy of the local numerical simulation. This approach to numerical modelling can significantly improve the quality of numerical simulations. Although the detailed global numerical models can represent some local phenomena of the rock mass response, they should not be taken as a cure-all solution. The results obtained in a global numerical model are rather aggregated and in case of local tasks only indicate the presence of this or that phenomenon in a particular zone, but cannot describe it quantitatively. Therefore, the transition from global to local numerical models is a necessary part of the work. This paper provides an example of a complete cycle of creating a set of detailed global and local numerical models. The cycle includes all stages of development from optimization of the initial geometry to the step-by-step calculation and analysis of the obtained results. Thanks to the approach used, both qualitative and quantitative convergence was achieved with the results of in-situ observations.