Abstract

Objective To identify key factors influencing the angle of repose of gravel soil by analyzing the impact of Johnson-Kendall-Roberts （JKR） surface energy， restitution coefficient， static friction coefficient， and rolling friction coefficient， aiming to achieve a precise prediction of the angle.

Methods The Generic EDEM Material Model （GEMM） database was used to obtain key parameters for the EDEM model. A Box-Behnken central composite design was employed in the experimental scheme. Discrete element simulations of the gravel soil’s angle of repose were conducted using the Hertz-Mindlin with JKR contact model. MATLAB functions were utilized to extract the boundary particle coordinates from the pile profile and fit the slope profile.

Results and Discussion The optimal parameter combination for simulated gravel soil particles was found to be： JKR surface energy of 0. 05， rolling friction coefficient of 0. 1， static friction coefficient of 0. 39， and restitution coefficient of 0. 45. With this combination， the EDEM simulation yielded a gravel soil angle of repose of 35. 41°， with an average angular error of only 0. 854% compared to the measured value of 35. 11°. No significant difference was observed in the particle pile morphology.

Conclusion To obtain the key parameters of the simulation of gravel soil particles， a combination of physical experiment and simulation was employed， and the GEMM database was innovatively introduced， which significantly narrowed the range of key parameters. During the post-processing stage of the stacking angle of the materials， MATLAB image processing technology was used to fit the outer contour line of the stacked slope， thereby obtaining the material’s angle of repose. Response surface analysis was applied to analyze the significance of DEM parameters affecting the angle of repose of gravel soil. The impact of different contact parameters between particles on the angle of repose varied. The JKR surface energy and rolling friction coefficient between sand and gravel particles were significant factors affecting the tangent value of the stacking angle， and their influence on the simulation accuracy was particularly significant. When setting contact parameters between particles， the values of rolling friction coefficient and JKR surface energy should be prioritized. When the JKR surface energy was 0. 05 J/m²， rolling friction coefficient was 0. 1， static friction coefficient was 0. 39， and restitution coefficient was 0. 45， the angle of repose of the particles was 35. 41°， with a relative error of only 0. 854% compared to the true angle of repose. This combination value was the optimal parameter configuration for gravel soil in DEM software.

Keywords：discrete element； gravel soil particle； particle reconstruction； contact model； numerical simulation