China
Africa
Explore chapters and articles related to this topic
An iterative sequential Monte Carlo filter for Bayesian calibration of DEM models
Published in António S. Cardoso, José L. Borges, Pedro A. Costa, António T. Gomes, José C. Marques, Castorina S. Vieira, Numerical Methods in Geotechnical Engineering IX, 2018
H. Cheng, S. Luding, V. Magnanimo, T. Shuku, K. Thoeni, P. Tempone
The Discrete Element Method (DEM) captures the collective behavior of a granular material by tracking the kinematics of the constituent grains (Cundall and Strack 1979). By just a few micro-mechanical parameters, DEM can provide comprehensive cross-scale insights (Cheng et al. 2016, Cheng et al. 2017) that are difficult to obtain in either state-of-the-art experiments or sophisticated continuum models. Nevertheless, fast and automated parameter estimation is still lacking for DEM models of granular materials against the time—or history-dependent macroscopic behavior measured in experiments. A successful calibration method should provide two key ingredients for the DEM models, namely, particle configuration and micromechanical parameters, conditioned to the granular material and experimentally measured macroscopic behavior. Each takes significant computational effort. This work aims to bring together the two ingredients for DEM modeling of granular materials within an iterative Bayesian framework.
An iterative sequential Monte Carlo filter for Bayesian calibration of DEM models
Published in António S. Cardoso, José L. Borges, Pedro A. Costa, António T. Gomes, José C. Marques, Castorina S. Vieira, Numerical Methods in Geotechnical Engineering IX, 2018
H. Cheng, S. Luding, V. Magnanimo, T. Shuku, K. Thoeni, P. Tempone
The Discrete Element Method (DEM) captures the collective behavior of a granular material by tracking the kinematics of the constituent grains (Cundall and Strack 1979). By just a few micro-mechanical parameters, DEM can provide comprehensive cross-scale insights (Cheng et al. 2016, Cheng et al. 2017) that are difficult to obtain in either state-of-the-art experiments or sophisticated continuum models. Nevertheless, fast and automated parameter estimation is still lacking for DEM models of granular materials against the time—or history-dependent macroscopic behavior measured in experiments. A successful calibration method should provide two key ingredients for the DEM models, namely, particle configuration and micromechanical parameters, conditioned to the granular material and experimentally measured macroscopic behavior. Each takes significant computational effort. This work aims to bring together the two ingredients for DEM modeling of granular materials within an iterative Bayesian framework.
Measuring stiffness of soils in situ
Published in Fusao Oka, Akira Murakami, Ryosuke Uzuoka, Sayuri Kimoto, Computer Methods and Recent Advances in Geomechanics, 2014
Fusao Oka, Akira Murakami, Ryosuke Uzuoka, Sayuri Kimoto
The Discrete Element Method (DEM) is an accepted method of simulating granular materials, and of investigating the mechanical behavior in the macro level as well as micro responses. For example, Thornton (2000) revealed that the effect of intermediate principle stress under a constant mean stress condition, supported Lade & Duncan's (1975) failure mode. Sazzad et al. (2011) used DEM to simulate the macro mechanical responses and explore the micro characteristics of granular materials under constant b values. Barreto & O'Sullivan (2012) also found that the b value had an influence on the coefficient of interparticle friction μ by quantitatively using DEM simulation data compared with experimental data. This study will examine how granular materials react to more generalized stresses that have continuously varying b values. Although DEM has been used to simulate different b values, the behavior of sand has not yet been studied under a number of complex general stress conditions.
Discrete element numerical analysis for bearing characteristics of coral sand foundation considering particle breakage
Published in Marine Georesources & Geotechnology, 2023
Fenghui Hu, Xiangwei Fang, Chunni Shen, Zhihua Yao, Ganggang Zhou, Zhiqiang Wang
There are two main methods for simulating particle breakage using the discrete element method: the fragment replacement method and bonded particle method. The former involves replacing the original particles with particles smaller than the original size when broken (Åström and Herrmann 1998; Lobo-Guerrero and Vallejo 2005; Ben-Nun and Einav 2010; Ciantia et al. 2015). The latter involves bonding multiple small particles to form a large crushable particle, which breaks when the bond breaks (Bolton, Nakata, and Cheng 2008; Wang and Yan 2013; Ge, Wang, and Zhou 2019; Cil, Sohn, and Buscarnera 2020). The bonded particle method is more accurate than the fragment replacement method for simulating the interaction between particles; however, its calculation efficiency is lower. Therefore, the bond particle method is mainly used for a few three-dimensional particles crushing under loading (Wu and Wang 2020) and two-dimensional particles breakage with different shapes under shear stress (Shi et al. 2015), while the fragment replacement method is mainly used for two-dimensional and three-dimensional particles breakage during simulation tests or construction (Yang, Xu, and Liu 2015; Zhang et al. 2017; He et al. 2021).
Using discrete simulations of compaction and sintering to predict final part geometry
Published in Powder Metallurgy, 2023
Gilmar Nogueira, Thierry Gervais, Véronique Peres, Estelle Marc, Christophe L. Martin
The discrete element method (DEM) is an alternative simulation method that can be used to take into account powders’ granular microstructure directly. Initially introduced by Cundall & Strack [4] for geomaterials, the DEM consists of simulating a granular medium by taking into account each particle, generally represented by a sphere (Figure 1(a)). Contact laws with a normal and a tangential component define the interaction between the particles that indent each other (Figure 1(b)). The particle’s motion is calculated to satisfy force and moment equilibrium. This method has been used mainly to simulate metallic [5–7], ceramic [8] and composite powders [9,10]. With DEM, it is possible to simulate the rearrangement and the plastic deformations during cold isostatic and closed die compaction of powders [6,11], as well as the fracture of agglomerates [12]. DEM has also been used to model sintering, taking into account diffusion as well as coalescence [13]. Thus, the DEM is a powerful method to simulate the entire process from compaction to sintering.
Experiments and simulation of torque in Anton Paar powder cell
Published in Particulate Science and Technology, 2018
Hamid Salehi, Daniele Sofia, Denis Schütz, Diego Barletta, Massimo Poletto
In comparison with experimental works, in numerical simulations, it is possible to observe process dynamics in given time and spatial coordinates (Müller and Tomas 2014). Discrete element method (DEM) is a powerful tool to predict the behavior of granular materials in industrial operations. It is acknowledged that shape (Lu, Third, and Müller 2015) and size (Windows-Yule, Tunuguntla, and Parker 2016) of particles are important input parameters in DEM. Nevertheless, the main difficulty for the application of DEM is the estimation of correct and realistic contact law parameters of bulk materials such as shear modulus, Poisson’s ratio, coefficient of restitution, coefficient of sliding friction, and coefficient of rolling friction to be used in DEM calculation (Coetzee 2017). There are two ways to attain these mechanical properties. In one case, the particle properties can be directly measured, especially if particles larger than 1 mm are used (Bharadwaj, Ketterhagen, and Hancock 2010; Kubík and Kažimírová 2015; Stasiak et al. 2015; Pasha et al. 2016). The other approach is the model calibration, which means changing the parameter values until a satisfactory match has been acquired between the simulation and experimental results. This approach is particularly useful for nonspherical and small particles. In these cases, the model parameters have to be calibrated because they have to account also for the shape differences between simulated and experimental particles. In fact, in this case, the direct measure of the parameter is difficult or the parameter itself, like the rolling friction, does not have a physical analogue (Rackl and Hanley 2017).