A robust potential-based contact force solution approach for discontinuous deformation analysis of irregular convex polygonal block/particle systems

This paper presents a machine learning (ML)-enabled discrete element method (DEM) for the computational mechanics of irregular-shaped particles. MLenabled DEM, as with most conventional DEMs, encompasses four main steps in one typical calculation cycle, namely, (1) the detection and resolution of contacts, (2) the evaluation of contact behavior, (3) the calculation of particle motion, and (4) the updating of particle geometric descriptions. Unlike conventional DEMs, the proposed method constructs and employs neural networks to detect particle contacts and resolve contact geometric features. Neural networks take particle geometric descriptors as inputs and output the contact status and contact geometric features. Using two-dimensional elliptical particles as an example, the performance of the ML-enabled DEM is investigated through five numerical experiments and compared with analytical solutions or conventional DEM methods. A sixth numerical experiment involving irregular-shaped particles is also presented to showcase the potential and applicability of the proposed method for other particle shapes. ML-enabled DEM can accurately capture the trajectory and energy evolution of individual particles, the fabric characteristics of dense packing, and the mechanical behavior of packing under large loads, while demonstrating computational efficiency over conventional methods.

Section: Discussionmentioning

confidence: 99%

Machine‐learning‐enabled discrete element method: Contact detection and resolution of irregular‐shaped particles

Lai

Chen

Huang

2021

“…Potential-based penalty function methods are established to calculate the contact forces. [40][41][42] Because of the irregularity of polyhedra, algorithms for complete contact search and efficient contact identification are difficult to develop. The establishment of an effective contact detection theory is a very difficult task.…”

Section: Introductionmentioning

confidence: 99%

“…43 Examples of this strategy are the penalty method proposed by Cundall 3 and the potential-based penalty function methods. [40][41][42] These methods use an explicit approach. The other kind is based on the contact dynamics method.…”

Section: Introductionmentioning

confidence: 99%

“…Based on this contact theory, cover‐based contact detection methods 37–39 have been developed. Potential‐based penalty function methods are established to calculate the contact forces 40–42 . Because of the irregularity of polyhedra, algorithms for complete contact search and efficient contact identification are difficult to develop.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A new cube movement test for verification of simulations of contact processes of blocks of different size in geological hazards

Wang,

Feng,

Lahayne

et al. 2024

In many geological hazards, such as landslides, a large number of irregular blocks start moving. Their interaction on the way down renders prediction of disaster scopes difficult. To study this process and to provide a novel method for validation and calibration of numerical tools for its simulation, a cube movement test is designed. The goal of this research is to obtain patterns of movement of cubes, starting from different initial stacking arrangements. Cubes of four sizes are inserted into a hollow cylinder. Their distribution after lifting the cylinder is determined. Three categories of tests refer to three different strategies of filling the cubes into the cylinder. In order to simulate cube movement tests, a numerical tool is developed in the framework of the continuum–discontinuum element method (CDEM). The contact between the individual cubes is modeled by the contact‐pairs‐based algorithm. Both the contact state and type are detected by determining the half‐space relation between contact pairs. The final positions of the cubes are strongly related to their initial arrangement. The latter is different in every test, even if the same strategy is used to fill the cubes into the cylinder. It is found that at least 20 experiments/simulations are required to obtain statistically representative results. The new test provides valuable data for validation of numerical tools used for the simulation of mass movement processes. The proposed numerical method captures the complicated movements of blocks.

“…The nonsmooth changes of contact normal for polygon and polyhedron contact problems have been investigated, and several algorithms have been proposed to solve this issue within the explicit solution framework, such as the corner rounding scheme, 52,53 the potential particle method, 54,55 the energy conservation method, 56 and the potential function method. 4,[57][58][59][60][61] The iterative methods in conventional DDA program can solve the polygon contact problems while remaining the nonsmooth feature in contact normal change. [62][63][64][65] The indeterminacy in solving vertexvertex contact problems 66 and the algorithmic robustness problem 25 for contact analysis of quasi-parallel edges/faces pairs was also discussed.…”

Section: Introductionmentioning

confidence: 99%

A cover‐based contact detection approach for irregular convex polygons in discontinuous deformation analysis

Zhuang

Zheng

et al. 2020

Self Cite

Irregular polygon shapes (eg, with small edges or small angles) are usually encountered in the contact simulation of discrete block systems. Treatment of irregular polygons in contact detection process has critical effects on the robustness and efficiency of the discontinuous computation approach. The present work proposes a cover‐based strategy to detect and solve contacts of irregular convex polygons in a robust and efficient way. Contact constraints of two polygons are represented by vertex‐edge and edge‐vertex contact covers in 2D. Two loops, namely vertex‐edge loop and edge‐vertex loop, and two filter criteria, namely the entrance filter criterion and the distance filter criterion, are used to establish the potential contact cover list of two neighbor polygons. The initial active and closed contact covers are chosen based on block configuration at the beginning of the step and they are then updated in the open‐close iteration process using proposed criteria. This strategy is implemented in discontinuous deformation analysis. The robustness of the proposed cover‐based approach and the conventional type‐based approach in handling contact of irregular blocks is verified first. Then, the contact analysis efficiency of the cover‐based approach with different contact tolerances is evaluated. This cover‐based method can be extended to 3D case for efficient and robust contact analysis of irregular polyhedral blocks.