Authors: Yoshitaka Wada, Takayoshi Matsumoto, Masanori Kikuchi

Abstract: We have proposed a cell based parallel computing system using a cellular automaton
model. The proposed system is designed by object-oriented technique to realize extensibility for parallel cellular automaton model. Because the system needs only information of model as cells which are represented by pixel and voxel mesh, model generation is very easy from any data formats. We implement several applications, which are 1-D life game, 2-D life game, 2-D diffusion, 2-D parallel diffusion, 2-D wave propagation, 3-D wave propagation and 3-D parallel wave propagation on the system. Status of cells is updated by adjacent status of cells. In other worth, behavior of system is completely specified in terms of a local relation. Cells can hold many properties to represent its status. Cellular automata system can calculate physical phenomena and discrete problem like a life game simultaneously. We present some results of a physical analysis and discrete analysis, and discuss effectiveness and parallel efficiency of proposed system.

571

Authors: Jing Zhang, Nai Hui Song, Xiao Peng Li, Zhao Hui Ren, Bang Chun Wen

Abstract: A two-dimensional cellular automaton (CA) model is developed to simulate damage and
fracture morphology evolution on the mesoscale in materials. The plastic convection of damage is
mapped onto the CA lattice, and initiation, propagation and coalescence of damage are simulated
with a local rule-based scheme of a probability cellular automaton. The model includes known
physical distinctions of fracture behavior between microcracks and microvoids, and they are
characterized by modifying the probability rule of the cellular automaton. The simulations provide
visual insight to understand how those physical processes dynamically progress and they affect the
damage evolution in materials. The modeling can be used to link micromechanical models to
continuum damage models.

1060

Authors: Szilvia Gyöngyösi, Anita Tóth, Péter Barkóczy

Abstract: The same property of the phase transformations driven by short range diffusion (recrystallization, allotropic transformation, grain coarsening) is that the movements of the grain or the phase boundaries take place by atomic jumps through the boundaries. The probability (frequency) of these jumps depends on only on the energy state of the closenear neighborhood of the atoms. In the operation of cellular automata Consequently, only the closenear neighborhood of the cells is taken into account in the operation of the cellular automaton. This similarity makes applicable the cellular automaton applicable to simulate the aforementioned phase transformation processes. A condition (rule) of the movement of grain and phase boundaries is introduced, which makes it possible to simulate all the all mentioned phase transformation by the same automatona.

405

Authors: Szilvia Gyöngyösi, Péter Barkóczy

Abstract: By applying the cellular automaton method the short-range diffusion processes in metals can be efficiently simulated. Several examples for the two-and three dimensional modeling of recrystallization and grain-coarsening are know at the literature. In some previous works, results have been performed concerning the two-dimensional, stochastic automatons of grain-coarsening, recrystallization and allotropic transformation. In order to use these simulations also in technological processes, it is necessary to scale the results reached by the simulation. The primary aspect of adapting the automaton in technological processes is the quick-operating simulation. The aim is to develop a most simplified, scalable cellular automaton by which scaling can be efficiently performed.

150

Authors: Szilvia Gyöngyösi, Peter Barkoczy

Abstract: Numerous literature [1,4,5] has reported on the effective use of cellular automaton method for the simulation of short-range diffusion. Using this model for the simulation of short-range diffusional phase transformations therefore is a resolved issue. It is proven that two- or three-dimensional automata can reflect the course of the abovementioned processes realistically. What our study demonstrates more than in the past [1] is that two-dimensional stochastic cellular automaton simulation already presented before has been simplified. This time our automaton operates in one dimension [2], which has consequently reduced running time, thus, made it possible to enhance the efficiency of the scaling of simulation. In our previous work the results of scaling of one-dimensional simulation of the recrystallization process [3] were demonstrated, in our current study fitting is performed for measurement results of grain coarsening using one-dimensional cellular automaton.

217