Volume 113
您当前的位置:首页 > 期刊文章 > 过刊浏览 > Volumes 108-119 (2025) > Volume 113
A 3D lattice Boltzmann method with μ(I)-rheology for simulating large-scale granular flows
Yiyang Shen, Liuchao Qiu *, Tangjing Yuan
College of Water Resources & Civil Engineering, China Agricultural University, Beijing, 100083, China
10.1016/j.partic.2026.03.027
Volume 113, June 2026, Pages 199-209
Received 13 January 2026, Revised 6 March 2026, Accepted 19 March 2026, Available online 1 April 2026, Version of Record 6 April 2026.
E-mail: qiuliuchao@cau.edu.cn

Highlights

• A GPU-accelerated 3D lattice Boltzmann method for large-scale landslide simulation is proposed.

• Multistate constitutive behavior of landslide materials is captured by μ(I)-rheology model.

• A granular collapse experiment with a step beneath the column was performed.

• The proposed method was used in reconstruction of the 2014 Hongshiyan landslide evolution.


Abstract

The simulation of granular flows is highly challenging due to the involvement of large deformations and complex flow dynamics. In this paper, we developed a 3D numerical method to simulate the evolution processes of granular flows. The proposed approach employs the lattice Boltzmann method (LBM) as the solver, incorporates the μ(I)-rheology model for characterizing the constitutive response of granular flows, and uses a single-phase volume of fluid (VOF) scheme for tracking its free-surface. To enhance computational efficiency for large-scale simulations, an OpenCL-based GPU parallelization technique is used to accelerate computations. After validation via a granular collapse experiment, the model is applied to simulate the large-scale granular flow of the 2014 Hongshiyan landslide. Numerical results agree well with field observations, accurately reproducing the runout distance and deposit extent, which demonstrates the model's capability for simulating large-scale granular flows in complex 3D terrain.

Graphical abstract
Keywords
Granular flows; Lattice Boltzmann method; μ(I)-rheology model; Free-surface flow; Numerical modeling