Visualization and quantification of pore structure in cement tailings waste rock composites using X-ray computed tomography and deep learning
Jiajian Li, Shuai Cao, Weidong Song, Lijuan Sun
Abstract
The aim of this study is to examine the pore structure of cement tailings waste rock composites in relation to the particle size and content of the tailings and waste rock. The pore structure was visualised and quantified (porosity, equivalent diameter , sphericity , orientation angle) by X-ray CT scanning, deep learning segmentation techniques , and 3D reconstruction . Finally, the relationship between porosity and elastic modulus was analysed. The results show that the particle size and content of tailings and waste rock and their interactions significantly regulate the nonlinear evolution of the cement tailings waste rock composites pore structure: increasing tailings particle size decreases the porosity 3D, induces local aggregation of pores and degrades the pore morphology. The high content (50 wt%) and large particle size (5–7 mm) of the waste rock resulted in an increase in the volume percentage of large and extra-large pores. The increase in waste rock content and particle size leads to an increase in pore aggregation in the cement coarse tailings waste rock composites, as well as a migration of the pore orientation θ angle to higher angles. The addition of 30 wt% waste rock led to a notable reduction in the porosity 3D of the cement fine tailings waste rock composites. The above changes stem from the synergistic effects of tailings-waste rock properties on slurry rheology , bleeding rate, and particle arrangement. This ultimately leads to a significant negative correlation between porosity 3D and elastic modulus . It is shown that the pore structure is a key factor in regulating the mechanical response of the material.