Computational Particle Mechanics最新文献

The presented surface indentation model is one step towards building a DEM model for wheel–rail sanding. In railways, so-called low-adhesion conditions can cause problems in traction and braking, and sanding is used to overcome this problem. Sand grains are blasted towards wheel–rail contact, fracture repeatedly as they enter the nip and are drawn into the contact and then increase adhesion. Research on this topic has mostly been experimental, but focussed on adhesion enhancement measurement. Thus, physical mechanisms increasing the adhesion are not well understood. Previous works involved experiments and DEM modelling of single sand grain crushing tests under realistic wheel–rail contact pressures of 900 MPa, focusing on sand fragment spread and formation of clusters of solidified fragments. In the experiments, indents in the compressing steel plates were also observed, which are also observed on wheel and rail surfaces in railway operation. These are now modelled by adapting an existing surface indentation model from literature to the case of surface indentations caused by granular materials. Two test cases are studied, and experimental spherical indentation tests for model parametrisation are presented. In a proof of concept, the mentioned single sand grain crushing tests under 900 MPa pressure are simulated including the surface indentation model. This work contributes to DEM modelling of wheel–rail sanding, which is believed to be a good approach to deepen the understanding of adhesion increasing mechanisms under sanded conditions.

This paper presents a constrained particle dynamics (CPD) framework with explicit and implicit algorithms for simulating differential–algebraic equations of motion for systems of particles under constraints. Addressing limitations in existing techniques such as position-based dynamics (PBD), commonly used in computer graphics but prone to inaccuracies, the CPD approach utilizes Hamilton’s variational principle and Lagrange multipliers to ensure accurate constraint enforcement. The explicit CPD (xCPD) algorithm employs a central difference scheme, enhancing efficiency by advancing the system under external forces and applying a correction term for constraints. The implicit CPD (iCPD) algorithm uses the Trapezoidal rule, solving a saddle point problem that integrates dynamic and constraint equations, offering robustness for larger time steps. The effectiveness of the CPD algorithms is demonstrated through mathematical analysis and numerical comparisons of benchmark problems. Results indicate that CPD algorithms achieve higher accuracy and superior energy conservation properties compared to PBD, exhibiting second-order convergence rates; whereas, PBD shows only first-order convergence.

Cavitation erosion is a pervasive issue in hydraulic machinery and ocean engineering, characterized by the collapse of bubbles, micro-jetting, and impact erosion, all exhibiting strong transient, microscale, and fluid–solid coupling features. Understanding these phenomena is essential for elucidating the mechanisms behind erosion and for developing strategies to prevent wear damage. Recognizing the limitations of conventional numerical methods, this study employs the smoothed particle hydrodynamics (SPH) method to develop a fluid–solid coupling model that simulates cavitation erosion at the bubble scale. The Lagrangian and mesh-free nature of SPH make it well-suited for tracking the transient processes of asymmetric bubble collapse, jet formation, and the subsequent impact on elastic–plastic materials. A comprehensive fluid–solid coupling SPH model is constructed, encompassing bubbles, surrounding liquids, and elastic–plastic materials. This model includes a compressible multiphase SPH approach for simulating the interaction between highly compressible bubbles and liquids. To address gas phase over-compression during bubble collapse, a modified particle regeneration technique (PRT) is introduced, allowing for automatic adjustment of particle resolution in the gas domain as it expands or compresses. For the solid simulation, an elasto-plastic constitutive model and a failure model are integrated into the SPH framework to describe material deformation and failure due to microjet impacts. These enhancements enable the simulation of the entire cavitation erosion process within a unified, mesh-free context. The SPH model is validated through simulating bubble collapse and jetting induced by shock waves. It is then applied to investigate the dynamics of cavitation erosion near both rigid and elastic–plastic materials, providing quantitative analysis of the erosion process. The outcomes of this research contribute significantly to our understanding of cavitation erosion mechanisms and offer a robust computational tool for predicting and mitigating erosion damage in related engineering applications.

In this paper, we focus on three issues related to applications of material point methods (MPMs) to objects with complex geometries. They are material point generation, compatibility of material points with a mesh, and sensitivity to mesh orientation. An efficient method of generating material points from a stereolithography (STL) file is introduced. This material point generation method is independent of the mesh used in MPM calculations. The compatibility between the material points and the mesh is then studied. We also show that the original MPM and the dual domain material point (DDMP) method are sensitive to mesh orientation. These issues are related to the calculation of the internal force and are concerns of the MPMs. They become more prominent when MPMs are applied to complex geometries. Our numerical results show that the recently developed local stress difference (LSD) algorithm (Perez et al. in J Comp Phys 498:112681, 2024) can be used to effectively address them.

The Discrete element method (DEM) is a robust numerical tool for simulating crack propagation and wear in granular materials. However, the computational cost associated with DEM hinders its applicability to large domains. To address this limitation, we employ DEM to model regions experiencing crack propagation and wear, and utilize the finite element method (FEM) to model regions experiencing small deformation, thus reducing the computational burden. The two domains are linked using a FEM–DEM coupling, which considers an overlapping region where the deformation of the two domains is reconciled. We employ a “strong coupling” formulation, in which each DEM particle in the overlapping region is constrained to an equivalent position obtained by nodal interpolation in the finite element. While the coupling method has been proved capable of handling propagation of small-amplitude waves between domains, we examine in this paper its accuracy to efficiently model for material failure events. We investigate two cases of material failure in the DEM region: the first one involves mode I crack propagation, and the second one focuses on rough surfaces’ shearing leading to debris creation. For each, we consider several DEM domain sizes, representing different distances between the coupling region and the DEM undergoing inelasticity and fracture. The accuracy of the coupling approach is evaluated by comparing it with a pure DEM simulation, and the results demonstrate its effectiveness in accurately capturing the behavior of the pure DEM, regardless of the placement of the coupling region.

A three-dimensional grain-based model based on the discrete element method is proposed for reconstructing the filling and grouping of minerals in granite, then a batch of numerical disc specimens with different grain sizes R_G are subjected to the Brazilian splitting test. In addition, the force chain networks in the numerical samples are subjected to multi-level classification and quantitative analysis, and the grain size effect on the tensile mechanical behavior of granite is discussed from the perspective of force chain networks. The results show that the mechanical properties and micro-cracking behavior of fine- and coarse-grained samples obtained experimentally and from simulation are consistent, including the load–displacement curve, the peak load, the failure displacement, and the proportion of intergranular/transgranular cracks. Therefore, the reliability of the model is verified. As R_G increases, the number of intragranular contacts increases, while the number of intergranular contacts decreases. The bearing capacity and deformation resistance of the samples increase. As R_G increases, both the number and sum of force chains for intragranular structures increase gradually, while these two parameters for intergranular structures decrease; meanwhile, the average values for intragranular and intergranular structures increase with increasing R_G. As R_G continues to increase, the number of contacts within mineral grains capable of withstanding external loads increases, forming a robust force chain network to bear external loads. It becomes challenging for a low-level load to break the contacts within the mineral, leading to an increase in the sample’s load-bearing capacity.

Ejector deep hole drilling achieves high-quality boreholes in production processes. High feed rates are applied to ensure a high productivity level, requiring reliable chip removal from the cutting zone for a stable process. Therefore, a constant metalworking fluid flow under high volume flow rates or high pressure is required. Experimental results show a vortex formation at the outer cutting edge. This vortex can lead to delayed chip removal from the cutting zone, and ultimately, it can lead to chip clogging and result in drill breakage due to increased torque. This paper investigates modified drill head designs using the smoothed particle hydrodynamics method. The investigated modifications include various designs of the chip mouth covering. Besides graphical analysis based on flow visualizations, flow meters are placed at the tool’s head to evaluate the impact of the modifications on the flow rate and possible increased resistance and relocation of the fluid flow from the outer cutting edge to other parts of the tool. The simulation results for the reference design show the experimentally observed vortex formation, validating the simulation model. By adding the tool’s rotation in the SPH simulation, which is not included in the experiments for observation reasons, the vortex formation is positively influenced. In addition, some designs show promising results to further mitigate the vortex formation while maintaining a sufficient fluid flow around the cutting edges.

When metals and alloys are exposed to ultrasonic vibrations (UV), a softening behavior occurs, caused by the phenomenon of acousto-plasticity. To obtain accurate results in a deformation analysis, this phenomenon must be included in the formulation of the constitutive material model. In this work, an acoustic-plastic model is proposed to capture the effects of ultrasonic vibrations during machining. The desired effect is to modify the chip morphology to reduce the magnitude of the cutting forces and thus reduce the energy consumption of the process. The study focuses on the modeling of ultrasonic vibration-assisted micromachining (VAMM). The particle finite element method is used and extended to perform a thermo-mechanical analysis capable of capturing the responses of conventional micromachining (CMM) and VAMM operations of 32 HRC stainless steel. The cutting speed and UV parameters, including amplitude and frequency, are integrated into the Johnson–Cook constitutive model to account for the effects of acoustic softening on the machining characteristics. The results show that the influence of UV on microcutting leads to thinner chips and lower cutting force. In the VAMM operations, an average reduction in cutting forces of 20% is achieved at five different cutting speeds. In addition, the contact length between the tool and chip decreases at different cutting speeds from 29% to a maximum of 44%. Furthermore, the thermal analysis results show that there is a negligible temperature change during the CMM and VAMM simulations, indicating that the study of the machining process can focus exclusively on its mechanical aspects when performed at the microscale. The predicted average chip thickness and effective shear angle of the workpiece material are in strong agreement with the experimental results, emphasizing the importance of considering acoustic softening in VAMM studies.

This paper presents a novel method for smoothed particle hydrodynamics (SPH) with thin-walled rigid structures. Inspired by the direct forcing immersed boundary method, this method employs a moving least square method for the velocity interpolation instead of the linear interpolation. It reduces oscillations due to changing relative positions between fluid grids and structures. It also simplifies thin-walled rigid structure simulations by eliminating the need for multiple layers of boundary particles, and improves computational accuracy and stability in three-dimensional scenarios. Results of the impulsively started plate test demonstrate that the proposed method obtains smooth velocity and pressure, as well as a good match to the references results of the vortex wake development. Results of the flow past cylinder test show that the proposed method avoids mutual interference on both side of the boundary, while accurately calculating the forces acting on structure. By comparing to linear least square direct forcing scheme and the diffusive direction scheme, advantages of lower oscillation and higher accuracy are proven. Results of flow past a sphere further indicate the stability of the proposed method for three-dimensional simulations.

The material point method (MPM) is computationally costly and highly parallelisable. With the plateauing of Moore’s law and recent advances in parallel computing, scientists without formal programming training might face challenges in developing fast scientific codes for their research. Parallel programming is intrinsically different to serial programming and may seem daunting to certain scientists, in particular for GPUs. However, recent developments in GPU application programming interfaces (APIs) have made it easier than ever to port codes to GPU. This paper explains how we ported our modular C++ MPM code Karamelo to GPU without using low-level hardware APIs like CUDA or OpenCL. We aimed to develop a code that has abstracted parallelism and is therefore hardware agnostic. We first present an investigation of a variety of GPU APIs, comparing ease of use, hardware support and performance in an MPM context. Then, the porting process of Karamelo to the Kokkos ecosystem is detailed, discussing key design patterns and challenges. Finally, our parallel C++ code running on GPU is shown to be up to 85 times faster than on CPU. Since Kokkos also supports Python and Fortran, the principles presented therein can also be applied to codes written in these languages.