This paper presents an experimental approach on a segregation phenomena of two sizes mixed granular with rough surface occurred in rolling cylinder experimental device. And this paper presents DEM simulations of those results in order to analyze segregation mechanism. Two sizes gravels and two sizes glass beads are used, and two of them are combined as a grain sample. Several roughness conditions at bottom surface of the rolling cylinder as well as no-roughness are given as an experimental parameter. For spherical surface granular, the roughness of bottom surface describes the occurrence of segregation. For rough surface granular, an adequate roughness of the bottom is necessary to the segregation. DEM simulates the segregation occurrence or no-occurrence as same as experimental results, and gives a fundamental information on the segregation occurrence mechanism.
Detection of micro-cracks is very important to lengthen the lifetime of structures because micro-crack is a signal of initiation and growth of deterioration of materials and eventually degrades the mechanical performance of structures. However, no practically useful non-destructive testing device for large scale structures can capture the essential micro-crack for large structures in reality. In the meanwhile, nowadays, non-destructive testing simulation is also gaining attention. The term, non-destructive testing simulation, means a numerical testing for crack inspection applying computational mechanics. The present study proposes a method to determine micro-crack topology giving an equivalent homogenized macro-stiffness tensor to the prescribed anisotropic material stiffness tensor by applying an inverse-homogenization method. We demonstrate the accuracy of the evaluation of the proposed method in terms of a series of numerical simulations.
Cracks in desiccation crack phenomenon show a wide variety of geometric patterns. Our intuition tells that the major source of this wide variety is the coupling of water diffusion induced volume change, deformation, and fracture. In this paper, we propose a coupling model of i) change in water volume fraction due to desiccation, ii) deformation field corresponding to the inhomogeneous volume shrinkage, and iii) crack formation. Geometric patterns in desiccation crack phenomenon have been reproduced by weakly coupling the finite element analysis of water diffusion and analysis of deformation and failure by Particle Discretization Scheme Finite Element Method.
To optimize a shape of hopper which is widely used in various industries, we developed a particle-swarm optimization (PSO) method with discrete-element method (DEM) simulations of powder discharge: DEM-PSO. We determined reasonable PSO parameters (number of PSO agents and weight coefficient) for the hopper shape optimization problem by parametric studies: 10 for number of PSO agents and 0.5 for weight coefficient. An effective shape of hoppers was uniquely obtained from the DEM-PSO analysis, even though we used random and irregular shapes of hoppers in initial state of DEM-PSO analysis. Also, we confirmed that the optimized hopper shape discharged powder quickly compared with conventional conical-shaped hopper.
This study presents a new framework of the probabilistic tsunami hazard evaluation which enables us to use numerical simulation effectively. The basic idea of the framework had been proposed for the reliability analysis of geostructures, and the key issue of the framework is use of the response surface which obtained from the results of numerical simulations. In this study, the response surface of the tsunami height is obtained from the results of the tsunami simulation. In addition, uncertainties related to tsunami simulations are quantified. By using the framework, we can create an environment in which the Monte Carlo Simulation can be performed based on highly-developed numerical simulations. The proposed framework is finally applied to the probabilistic evaluation of the tsunami that was induced by the great east Japan earthquake. The probability density distributions of tsunami height at Sendai, Ishinomaki, and Kamaishi were evaluated with consideration of uncertainties of some fault parameters and a modeling error of the numerical simulation. The obtained results indicate that the framework is effective for the probabilistic tsunami hazard evaluation.
In the thick shell theory based on first-order shear deformation theory, the 1/(1 + ζ/Ri) term of strain-displacement relations is treated approximately. When the terms of strain-displacement relations and stress resultants are treated exactly, equilibrium of the moment about the normal to mid-surface is satisfied without using modified strain-displacement relations. The error of the displacement, stress and frequency caused by first- and second-approximation is investigated.