| Geo-materials are multi-phase structural systems,which are composed of particles,pores and cracks.Their physical and mechanical properties are significantly different at different scales,showing multi-level structural characteristics.Based on the obtained meso-structure information,reconstructing the model of geotechnical materials and establishing the relationship between meso-structure and macroscopic properties of them is an efficient research method to invesgitate the mechanical properties of geotechnical materials and reveal the failure mechanisms of rock and soil mass at multiple scales.The results can be used as a theoretical guidance for the optimization of design and construction of geotechnical engineering.However,in the modeling of meso-structure of geotechnical materials,the particle shapes are often represented by idealized geometric shape.The overlap detection algorithm and particle packing method are only suitable for some irregularly shaped particles.Based on the theory of computational geometry,this thesis presents a systematic method to reconstruct meso-structure of geotechnical materials.Besides,the proposed method is applied in the study of numerical simulations on soil-rock mixtur slope stability.The main research contents of this thesis are as follows:(1)Based on random angular bend method,a novel method to generate two-dimensional random particle model for the numerical simulation of geotechnical materials are proposed.The relationship between input parameters and shape descriptors are obtained.Finally,a soil-rock mixture with predetermined rock shape is bulit to show the feasibility of proposed algorithm for the numerical modeling of realistic geotechnical materials.The results suggest that the proposed method has a good performance in the modeling of geo-materials(especially for nonstar-like particle reconstruction based on image information).(2)Based on the simulation method of two-dimensional irregularly shaped particle,a method to generate three-dimensional particle with irregular shape is proposed.The relationship between input parameter and shape descriptor is established by using the curve fitting method.The results suggest that four representative parameters could quantitatively control the shape feature of generated polygons by control three shape descriptors.Besides,these parameters have definite physical meaning,which makes the algorithm easier to understand.(3)A particle model base system of geotechnical materials based on graphical interface is developed.The system includes human machine interface,particle generation and import module,particle analysis and management module and particle output module.The system promotes the implication of proposed particle generation method.(4)By calculating the minimum distance between concave or convex particles,an overlap detection algorithm is developed to address the difficulties associated with spatial allocation of irregularly shaped particles.Based on gravity deposition method and particle expansion method,two modified particle packing methods are proposed.High or low packing density samples can be generated quickly by using proposed particle packing methods.(5)The proposed generation method of irregularly shaped particles,overlap detection method and particle packing method are applied in the numerical simulation of soil-rock mixture slope.The development of crack and strain after excavation were obtained and used to evaluate the deformation characteristics.Finally,the effects of shape and content of rock block on the stability of slope are investigated.The results suggest that the proposed method have a good performance in the modeling of geotechnical materials.Bedsides,the increase of rock block content will affect the slope stability and change the failure mode of soil-rock mixture slope.With the increase of elongation index and roundness index of rock block,the failure mode of soil-rock mixture slope with high rock content will gradually change from local shallow sliding to deep failure. |
PDF Full Text Request