Optimizing parameter combinations for clump models enabled by the bubble packing algorithm: Insights from geometrical and morphological approximation of typical geotechnical particles

The bubble packing algorithm has been widely used in previous DEM studies to generate clump models that approximate concave particle shapes. However, the selection of optimal parameters—specifically, the ratio ${\rho }_r$ and distance $\Phi$—is often subjective or based solely on volume ratios. This study offers a more comprehensive quantitative analysis by evaluating both the geometrical and morphological accuracy of clump assemblies generated using various combinations of ${\rho }_r$ and $\Phi$. The analysis focuses on typical geotechnical particles, including Fujian sand, ballast, and cobble particles. The study establishes correlations for the approximation degrees of each geometrical and morphological parameter and the average number of sub-spheres per clump in relation to ${\rho }_r$ and $\Phi$. Using a 5 % threshold for approximation accuracy, the optimal parameter combination of ${\rho }_r$ and $\Phi$ is identified and verified for each type of particles. The findings of this paper can provide a useful reference for future studies that utilize the bubble packing algorithm to generate clump models for approximating geotechnical particle shapes.