Numerical Analysis About Effects Of Microstructure On Mechanical Properties Of Particle Reinforced Composites

Particle reinforced metal matrix composites(MMCs) have been widely applied in areas like aeronautics, electronics, automobile and construction industry due to their advantages of high strength, high elastic modulus, high abrasion resistance,good thermal and electrical conductivity. While due to low elongation ratio and poor fracture toughness of such material, advanced research has been focused on the mechanism which relates microstructrue and macro material deformation. Microscopic features such as particle morphology, size, distribution, volume fraction, and interface property will all cast influence on macro material performance. Till now the mechanism remains unclear. Numerical simulation has been applied to the study of it. Three-dimensional(3D) multiple particles finite element(FE) model has shown the effect of particle morphology, distribution and interface damage on material deformation, both respectively and concurrently. However the influence of strain gradient is ignored which makes it incapable of reflecting the size effect of the composites under the scale of micrometer.To overcome such difficulties, in present work 3D periodic FE model has been developed for the MMC where extended strain gradient theory has been implemented to study the effect of particle size on strain gradient in matrix. A cohesive model has also been developed to analyze the reduce effect caused by interface damage. First the stress-strain relationship in incremental form is derived based on extended strain gradient theory. Then the user subroutines UMAT and URDFIL which are integrated in Abaqus are developed where plastic strain gradient is calculated by average nodal plastic strain. The method is validated by comparison between experimental results and simulation results.Based on the validated method various FE models have been developed features different particle size, morphology, interface strength and distribution. The benchmark model has been modeled as Aluminum matrix reinforced with Si C particles. The study focuses on the influence of microstructure on the material behavior including elasto-plastic behavior under uniaxial tension, particle/matrix load allocation, local distribution of stress/strain field, initiation and evolution of interface damage. Furthermore this work also carries out an insight into the magnitude and distribution of residual thermal stress and the subsequent material behavior caused by different particle sizes and shapes. The results are summarized as follows:(1) Flow stress, tensile strength and average strain are higher with stronger interface when particle shape and size remain unchanged. For composites with same interface strength and particle morphology, the smaller the particle, the higher the flow stress, tensile strength and average strain. For composites with same interface strength and particle size, composites reinforced with spherical particles show higher uniform strain. Flow stress is influenced by particle morphology as well as interface strength. When interface strength is low, the flow stress of composites reinforced with spherical particles is almost the same as the case where the materials are reinforced with cubic particles. For composites with strong interface, the flow stress is higher in cubic-particle-reinforced composites than their spherical counterparts. In sum, the flow stress achieves the highest value in the smaller cubic particle reinforced composite with strong interface, then that composite has the best strength property. The uniform strain achieves the highest value in the smaller spherical particle reinforced composite with strong interface, then that composite has the best ductility.(2) The comparison between 3D single-particle model and multi-particle model has shown that for spherical-particle-reinforced composites the flow stress and uniform strain are almost the same between results derived from both types of models, which makes it sensible to evaluate the property of the composites by single-particle model in order to achieve computational cost reduction. For cubic-particle-reinforced composites, the flow stress derived from single-particle model is lower than that from the multi-particle model. The main reason behind it is the assumption applied in the single-particle model that all interface damage are activated simultaneously which overestimates the average damage in the early stage of loading. Thus the average stress is lower than that of the multi-particle model from the very beginning. Due to this deficiency it is more reliable to simulate the behavior of the cubic-particle-reinforced composites with multi-particle model.(3) Flow stress is nearly identical to each other when the particles are distributed by clusters, layers or randomly. While uniform strain is sensitive to particle distribution pattern. For models where particles are distributed by layers, uniform strain is lower than the other two cases. Because the layer is perpendicular to the loading direction, interface damage is easier to develop. The material strength is not reduced in models with particles distributed by clusters which contradicts with the experimental results. This inconsistency is caused by model simplification where possible material defects such as voids caused by clusters of particles are ignored. The results show that the clusters of particles will not reduce material performance by their own.(4) When composites are cooled down from high temperature to room temperature, the magnitude of average residual thermal stress is higher when particles are larger. In cubic-particle-reinforced composite the value is higher than its spherical counterparts. Residual thermal stress(plastic strain) has a strong effect on the elastic modulus and micro-yielding behavior of the composites, rather than on the flow stress. The general rule is residual thermal stress will reduce micro-yielding stress, increase the flow stress a little,and residual thermal plastic strain will reduce the elastic modulus. Interface damage is hardly effected by residual thermal stress where the uniform strain of composites are close to each other in models with and without residual thermal stress.The results summarized above serve as valuable reference for practical composites design optimization.