Three-dimensional simulations of large-scale long run-out landslides with a GPU-accelerated elasto-plastic SPH model

Numerical Simulation of the Large‐Scale Huangtian (China) Landslide‐Generated Impulse Waves by a GPU‐Accelerated Three‐Dimensional Soil‒Water Coupled SPH Model

“…Computation time for these four cases ranged from 0.4 to 23.2 hr. The final states of the four simulations are plotted in Figure 6, where the color of the water body represents water elevation, and the color of the landslide mass indicates accumulated plastic strain (Adps) (Huang et al, 2022), where…”

“…Computation time for these four cases ranged from 0.4 to 23.2 hr. The final states of the four simulations are plotted in Figure 6, where the color of the water body represents water elevation, and the color of the landslide mass indicates accumulated plastic strain (Adps) (Huang et al., 2022), where $$\mathrm{A}\mathrm{d}\mathrm{p}\mathrm{s}=\int \nolimits_{0}^{t}\sqrt{\frac{2}{3}{\dot{\varepsilon }}^{p}:{\dot{\varepsilon }}^{p}}dt$$. From Figure 6, it can be seen that a small wave near to the opposite bank is captured from dp = 5 to 3 m, while the other water areas have a similar color map.…”

“…Finally, the stress rate of particle p was computed in parallel for each thread. The more detailed program structure can be found in our previous contribution (Huang et al., 2022), in which the GPU‐accelerated elastoplastic soil model was developed for large‐scale landslides.…”

“…(2021) adopted the Drucker–Prager constitutive model with a non‐associated flow rule in SPH formulations to model LGIWs within a continuum framework. This constitutive model has been used to simulate granular failure and flow problems (An et al., 2016; Bui et al., 2008; W. Chen & Qiu, 2012; Huang et al., 2022), and it can successfully replicate the elastoplastic behavior of particle flow. To reproduce the occurrence process of a LGIW disaster chain, the elastoplastic constitutive model can better describe the slope stability, large deformation, and granular flow, compared with other model types (Bui et al., 2011; Peng et al., 2021; Zhang et al., 2018; Zhu et al., 2018).…”

“…Huang et al. (2022) pointed out that the number of particles in many studies is relatively small, limiting their applicability to large‐scale geological problems. Recently, graphics processing unit (GPU) technology has been shown to greatly improve computing efficiency in many studies in the field of SPH simulation (J. Y. Chen et al., 2020; Crespo et al., 2015; Zhan et al., 2019).…”