A hybrid method for generating heterogeneous digital models for granites using a conditional filtering process and multiple level Weibull distribution functions

Abstract

AbstractIn this study, combining geometric and statistical modelling techniques together, a novel hybrid heterogeneity digital model generation method for granites is developed. It is motivated by the complexity of the digital core method and the limitations of the traditional random Weibull distributed heterogeneous method in reflecting the mechanical behaviour influenced by the spatial distribution of mineral particles. The initial mineral proportion is acquired from the feature extraction of the input image. Then a conditional filtering process is proposed in order to acquire a reasonable mineral particle distribution. Finally, the multiple-level Weibull distribution function is applied to describe the heterogeneity of different minerals. The new method has been validated in particle distribution configuration and mechanical reflection using a simulation software. In contrast to the previous methods, the multiple heterogeneity material assignment results obtained from the conditional filtering process can better reflect the actual mineral particle distribution. Moreover, reflection of failure mode behaviour is greatly improved by the proposed method compared to the traditional single-level heterogeneity model. The current data emphasize the independent and significant influence of each mineral heterogeneity, which can only be investigated using the traditional method, which ignores mineral differences. The present study lays the groundwork for future research into the influences of mineral mechanical properties and mineral particle spatial distribution on macro mechanical behaviour.