Simulation Of Heterogeneous Material Impact Properties Of Three-dimensional Material Point Method

Heterogeneous material widely used in industrial production and in our daily life. Previous studies the impact of dynamic characteristics of heterogeneous materials are mostly dependent on the experimental, theoretical studies also limited to a simple model of multiphase flow. Resulting a result of the dynamics behind the results of direct measurements, thermodynamic processes, as well as some fast process can not be measured directly understanding is not enough. The best way to fill the vacancy for the numerical simulation.Inhomogeneous materials, numerical simulation of large deformation problem is always difficult. Molecular dynamics simulations can reveal holes grow, some of the basic mechanisms of evolution of dislocation, but time and space scales which can not be covered by comparing with the experimental results. When using the macro-simulation approach to heterogeneous material impact dynamics, the traditional method (either Eulerian or Lagrangian method method) have encountered some difficulties not easy to overcome:Eulerian method is not easy to trace the material boundary; Lagrangian method will generate a grid distortion. New material point method can be compared with the experimental results on time and space scales, and material point method is a method of combining the advantages of Eulerian and Lagrangian method, avoiding its disadvantages. However, due to the material point method of calculation model becomes more complex, artificial modeling efficiency can not use the fast-paced scientific work, and after the treatment was non-gridded data format, derivation, the more difficult analysis.Based on this, the material point method for pre-processing and post-processing problems, as well as the impact properties of heterogeneous materials to explore, the main innovation is divided into the following five points:First, pre-treatment. Machine learning based on the idea of artificial intelligence system that can replace the manual work for a series of decisions; based on digital image processing technology that enables accurate identification of light, texture, gray. Combines machine learning ideological thoughts and image digital processing, compilation intelligent, automated procedures. By calling QQ, CAJ reader, web browser, achieve reptile features to the mass of information on the Internet examples of capture and information as an example for machine learning, image by image processing module light gray transformed to for program textures, materials, chaining identification.Second, after the treatment. Numerical algorithms for post-processing, in order to be accurate derivation operation, the use of thought and image restoration algorithms, data interpolation and recovery, complete reproduction of the trends in data and dissipation details.Third, the single hole model. Use MPM3D program, analysis of the impact of the problem hole model to study the hole collapse and jet deformation characteristics, focusing on the various physical quantities (such as speed, temperature, pressure, divergence, curl, etc.) in three-dimensional space shown by the law.Fourth, multi-hole model. Use MPM3D program, multi-hole model to explore the collapse of the hole collapse characteristics; hole model for multi-impact characteristics of setting a control experiment, porosity analysis, the impact speed.Fifth, concrete mesoscopic defects. Concrete impact, under the force of the hole by the growth, evolution, defects result explore concrete impact for the dynamics of the concrete impact of macroscopic mechanical properties, looking for mathematical laws of different parameters and correlation characteristics.