Mathematical and Information Technologies, MIT-2016 β€” Mathematical modeling Multiscale Modeling of Strength Properties of Dispersion-Reinforced Ceramic Composite Materials Yuriy Dimitrienko, Yulia Zakharova, and Sergey Sborschikov Computational Mathematics and Mathematical Physics Department, Bauman Moscow State Technical University 2 Baumanskaya Str., 5, Moscow, 105005, Russia {dimit.bmstu@gmail.com,shpakovayuliya@bmstu.ru,servasbor@gmail.com} http://www.bmstu.ru/~fn11/english/echif.htm Abstract. In this work a three-level model of ceramic composites mate- rials based on a reaction bonded silicon carbide is developed. Numerical solution is based on the method of multiscale homogenization along with the finite element method. As a result a series of local problems on the periodical cells of 3 structure levels are solved. The calculations of stress concentration tensors in matrixes and weighing materials are presented. New criteria for matrixes and weighing materials is used to calculate the strength properties in multiaxis stressed condition. This criteria in- cludes essential differences (more than an order of magnitude) of ceramic properties under straining and compression. The model which includes scale effect of strength of ceramic composite materials is proposal. The computational research of sequential micro-destruction processes of ce- ramic composite until complete destruction is done. The results show that changing of concentration of larger fractions is less significant then content of smaller fractions in the presence of polydisperse structure in ceramics. Keywords: ceramic composites, reaction bonded SiC, microdestruc- tion, numerical simulation, finite-element method, multiscale homoge- nization method, strength criterion, stress concentration tensor, scale strength effect 1 Introduction Composite materials based on reaction bonded silicon carbide matrix (RBSiC) and SiC disperse filler are perspective materials for creation of shockproof pro- tecting systems because of their high strength, stiffness, destruction energy and relatively low cost. However characteristics of this materials significantly depend on manufacturing technological processes and on receipt of composite compo- nents. In addition during the hardening details can give strong shrink, giving significant residual stresses. This residual stresses can give deformation and even breakdown in the final product. To select the optimal content of ceramic com- posite components of SiC system and to calculate strength properties of such 277 Mathematical and Information Technologies, MIT-2016 β€” Mathematical modeling materials it is demanded to develop special mathematic model, which can fore- cast strength properties of composite materials including variation of content, form and disperse filling. This model also should take into account locked-up stresses, appearing in ceramic composite during agglomeration of particles. Existing analytical and numerical models of composite materials, armed with particles, allow to forecast elastic properties with certain precision, however nu- merical calculation of strength properties is essentially more complex problem, because it is necessary to build the appropriate model of microcrack emission in heterogeneous structure. Attempts to build such models using simple concen- tration of finite-element mesh were not successful, because of dramatic increase of non-physical singularity effect of calculation. Widely known commercial soft- ware not always allow to get adequate results of microdestruction modeling of composite. Nowadays the great attention is paid to the development of numerical fi- nite element methods of microstress modeling in composites [4, 18, 21]. One of the most efficient method for calculation of microstress in composites is the method of asymptotic averaging (MAA) (or homogenization method) [2, 3, 5, 23, 25], which ensures the high accuracy of calculation of microdestruction in math- ematical terms. Possible errors of calculation by this method can be related only to errors of its numerical application as well as to inaccurate specifications of the component characteristics and the microstructure geometry. In [1, 6–8, 13] algo- rithms of finite element solution of the so-called local problems in periodicity cells, which appear when MAA is used, were developed. This work continues the development cycle [14, 16, 17, 20] of creating models and numerical methods for modelling microdestruction processes in composite materials. The new 3-level strength model of ceramic composite is presented. This model is based on RBSiC and allows to describe the effect of strengthening of composite material during the changing particles content of SiC including the production technology of its manufacturing. 2 Microstructure of the reaction bonded silicon carbide composite A composite based on reaction bonded silicon carbide consists of a filler and a silicon carbide matrix. The filler is powder of silicon carbide of the different fractions. The silicon carbide matrix is synthesized by chemical reaction of liq- uid silicon, carbon and solid carbon, which is produced during the pyrolysis of phenol-formaldehyde resin [22, 24]. The filler is, as a rule, fission fragments which have random character and big difference in fractions. Generally it can be iden- tified large fractions of the size of 20-100 microns and small fractions of 1-10 microns. Photographies of real microstructure of RBSiC are shown in Fig. 1. We consider a model of reaction bonded silicon carbide composite material which has three structure levels [5, 15, 6, 19, 26] (Fig 2). The first level is formed by the periodicity cells 1 (PC1) consisting of a filler of a coarse fraction and a matrix π‘š1 . On the second level the matrix π‘š1 is formed by the periodicity 278 Mathematical and Information Technologies, MIT-2016 β€” Mathematical modeling a) b) Fig. 1. The microstructure of the material with the original grains of silicon carbide of 28 microns. cells 2 (PC2), each of them consists of a filler of a fine-grained fraction and a reaction bonded silicon carbide matrix π‘š2 . The matrix π‘š2 has defects, for example, high concentration of dissolved, but unreacted component C and Si, microcracks due to technological stresses and mainly pores. So, we introduce the third structural level formed by the periodicity cells 3 (PC3). Each periodicity cell of type 3 is formed by a zero defect silicon carbide matrix π‘š3 and a defect. Fig. 2. The three level structure of silicon carbide ceramics. 3 Mathematical formulation of local problems All structural levels may be considered as independent according to the method of multiscale homogenization [17]. At first we compute the effective elastic and strength properties of the third level, then we calculate the effective characteris- tics of the second level, considering the composite matrix π‘š2 as a homogeneous 279 Mathematical and Information Technologies, MIT-2016 β€” Mathematical modeling material with effective characteristics of the third level, and then we calculate the characteristics of the first level. Consider the solution of local problems for the periodicity cell of the second level having the volume π‘‰πœ‰ . It includes the matrix π‘š2 and fine-disperse filler. We believe that the PC2 has three-axial symmetry, therefore instead of a full volume of PC2 π‘‰πœ‰ we can consider its 1/8th part of volume π‘‰Λœπœ‰ . This volume π‘‰Λœπœ‰ consist of the 𝑁 components: 𝑁 βˆ’ 1 pieces of fine-disperse particulate of filler of the volume π‘‰Λœπœ‰π›Ό , 𝛼 = 1...𝑁 βˆ’ 1, and binding matrix π‘š2 (𝛼 = 𝑁 ). For calculating microstresses in PC2 by homogenization method [15, 11, 12] we formulate a series of the so-called local problems πΏπ‘π‘ž of the elasticity theory on the 1/8th part of the periodicity cell ⎧ βŽͺ βŽͺ πœŽπ‘–π‘—(π‘π‘ž)/𝑗 = 0, π‘‰Λœπœ‰ ⎨ πœŽπ‘–π‘—(π‘π‘ž) = 𝐢(οΈ€π‘–π‘—π‘˜π‘™ (πœ‰π‘  , 𝑧)(πœ€π‘˜π‘™(π‘π‘ž) βˆ’ )οΈ€π›Όπ‘˜π‘™ (πœƒ βˆ’ πœƒ* )), π‘‰Λœπœ‰ βˆͺ 𝛴𝑠′ βˆͺ 𝛴𝑠 (1) βŽͺ βŽͺ πœ€ = 21 π‘ˆπ‘–(π‘π‘ž)/𝑗 + π‘ˆπ‘—(π‘π‘ž)/𝑖 , π‘‰Λœπœ‰ , ⎩ 𝑖𝑗(π‘π‘ž) Λœπœ‰π›Όπ‘ , [π‘ˆπ‘–(π‘π‘ž) ] = 0, [πœŽπ‘–π‘—(π‘π‘ž) ]𝑛𝑗 = 0, 𝛴 where 𝑝 and π‘ž are the indexes of the local problems changing from 1 to 3 (there are a total of nine different problems πΏπ‘π‘ž ); π‘ˆπ‘–(π‘π‘ž) (πœ‰π‘  ) are the components of the displacement vectors (the unknown functions) in the problem πΏπ‘π‘ž ; πœŽπ‘–π‘—(π‘π‘ž) , πœ€π‘–π‘—(π‘π‘ž) are the components of the stress and deformation tensors in π‘‰Λœπœ‰ ; πœ‰π‘  are the local Cartesian coordinates in the 1/8th PC; /𝑖 = πœ•/πœ•πœ‰π‘– are the derivatives of the local [οΈ€ ]οΈ€ coordinates; π‘ˆπ‘–(π‘π‘ž) are the jumps of functions at the interface 𝛴 Λœπœ‰π›Όπ‘ of the cell components; πΆπ‘–π‘—π‘˜π‘™ (πœ‰π‘  , 𝑧) are the components of the tensors of the elasticity mod- uli of the composite structural components of PC2 (they are described by the dependencies of the coordinates πœ‰π‘  ); 𝑧 is the parameter of the component dam- ageability; π›Όπ‘˜π‘™ (πœƒ) are the components of the tensor of thermal expansion, which depend on the temperature; πœƒ is the current temperature; πœƒ* (πœ‰π‘  ) is the sintering temperature of the ceramic particles, depending on the local coordinates. System (1) is supplemented by the special boundary conditions at the surfaces 𝛴𝑠′ = {πœ‰π‘  = 0.5} of the 1/8th part of PC β€² πœ€π‘π‘ 𝛿𝑖𝑝 , 𝑆𝑗(𝑝𝑝) = 0, π‘†π‘˜(𝑝𝑝) = 0, 𝑖 ΜΈ= 𝑗 ΜΈ= π‘˜ ΜΈ= 𝑖, π‘Žπ‘‘ 𝛴𝑖 : π‘ˆπ‘–(𝑝𝑝) = 1/2Β― β€² π‘Žπ‘‘ 𝛴𝑗 : π‘ˆπ‘–(π‘π‘ž) = (1/4)Β―πœ€π‘–π‘ 𝛿𝑖𝑝 , 𝑆𝑗(π‘π‘ž) = 0, π‘ˆπ‘˜(π‘π‘ž) = 0, 𝑖, 𝑗 = {𝑝, π‘ž}, (2) β€² π‘Žπ‘‘ π›΄π‘˜ : 𝑆𝑖(π‘π‘ž) = 0, 𝑆𝑗(π‘π‘ž) = 0, π‘ˆπ‘˜(π‘π‘ž) = 0, 𝑖 ΜΈ= 𝑗 ΜΈ= π‘˜ ΜΈ= 𝑖, where πœ€Β―π‘π‘ž are the components of the averaged deformation tensor for PC, 𝑆𝑖(π‘π‘ž) ≑ πœŽπ‘–π‘™(π‘π‘ž) 𝑛𝑙 are the vectors of forces. The boundary conditions at the symmetry planes 𝛴𝑠 = {πœ‰π‘  = 0} are similar to relations (2), where we assume πœ€Β―π‘π‘ž = 0. 4 Effective elastic characteristics of the periodicity cells of the second level structure Using the numerical solution of problems πΏπ‘π‘ž (1), (2) we find the fields of dis- placements π‘ˆπ‘–(π‘π‘ž) and stresses 𝜎 ¯𝑖𝑗(π‘π‘ž) (πœ‰π‘  ) in the PC2 at given values of average deformations πœ€Β―π‘˜π‘™ . 280 Mathematical and Information Technologies, MIT-2016 β€” Mathematical modeling These fields are used to find the average values of stress: 3 βˆ‘οΈ 𝜎 ¯𝑖𝑗 =< πœŽπ‘–π‘— >= 𝜎 ¯𝑖𝑗(π‘π‘ž) , 𝑝,π‘ž where ∫︁ βŸ¨οΈ€ βŸ©οΈ€ 𝜎 ¯𝑖𝑗(π‘π‘ž) = πœŽπ‘–π‘—(π‘π‘ž) = πœŽπ‘–π‘—(π‘π‘ž) (πœ‰π‘  )π‘‘π‘‰πœ‰ . (3) π‘‰Λœπœ‰ Then the components of the tensor of effective elasticity moduli of the composite are calculated by the formulas 𝜎 ¯𝑖𝑗(π‘π‘ž) πΆΒ―π‘–π‘—π‘π‘ž = , (4) πœ€Β―π‘π‘ž where there is no summation over 𝑝 and π‘ž. After that we calculate the effective Β― π‘–π‘—π‘π‘ž , that is inverse to πΆΒ―π‘–π‘—π‘π‘ž , and technical elastic tensor of elastic compliances 𝛱 constants of the composite, such as effective Young moduli 𝐸𝛼 = 1/𝛱 Β― 𝛼𝛼𝛼𝛼 , Β― effective Poisson constants 𝑣𝛼𝛽 = βˆ’π›±π›Όπ›Όπ›½π›½ 𝐸𝛼 , and effective shear moduli 𝐺𝛼𝛽 = 𝐢¯𝛼𝛽𝛼𝛽 . (𝛼) The components of the tensor of stress concentrations π΅π‘–π‘—π‘˜π‘™ connect mi- (𝛼) βˆ‘οΈ€ 3 crostresses πœŽπ‘–π‘— (πœ‰π‘  ) = πœŽπ‘–π‘—(π‘π‘ž) (πœ‰π‘  ) in the matrix and the filler (the fine disperse 𝑝,π‘ž particles SiC) with average stresses 𝜎 Β―π‘˜π‘™ in the PC2 by the formulas (𝛼) (𝛼) πœŽπ‘˜π‘™ , πœ‰π‘  ∈ π‘‰Λœπœ‰π›Ό , 𝛼 = 1...𝑁. πœŽπ‘–π‘— (πœ‰π‘  ) = π΅π‘–π‘—π‘˜π‘™ (πœ‰π‘  )Β― (5) (𝛼) The components π΅π‘–π‘—π‘˜π‘™ in the matrix and the filler are calculated by the formulas (𝛼) Β― π‘π‘žπ‘˜π‘™ , πœ‰π‘  ∈ π‘‰Λœπœ‰π›Ό , 𝛼 = 1...𝑁. π΅π‘–π‘—π‘˜π‘™ (πœ‰π‘  ) = πœŽπ‘–π‘—(π‘π‘ž) (πœ‰π‘  )𝛱 (6) 5 Model of the strength properties of the components The strength criterion of ceramic materials should take into account the signif- icant differences in their properties in tension and compression. Therefore, we introduce a failure criterion of isotropic matrix π‘š2 and filler particles [19] based on Pisarenko-Lebedev criterion: (𝛼)2 πœŽπ‘’ 𝑧= (𝛼)2 (𝛼) , (7) 3πœŽπ‘† (1 + 𝐡 (𝛼) 𝑉 (πœŽβˆ’ )) (𝛼) βƒ’ βƒ’ (𝛼) (𝛼) (𝛼) where πœŽβˆ’ = 21 (βƒ’πœŽ (𝛼) βƒ’ βˆ’ 𝜎 (𝛼) ), 𝜎 (𝛼) = 𝜎11 + 𝜎22 + 𝜎33 are the invariants of (𝛼) the (οΈ‚ stress tensor )οΈ‚ in the matrix and fillers, πœŽπ‘’ are stress intensity [10], 𝐡 (𝛼) = (𝛼)2 𝜎𝐢 1 (𝛼) (𝛼) (𝛼) (𝛼)2 βˆ’1 (𝛼) is the constant, 𝜎𝐢 , πœŽπ‘‡ , πœŽπ‘† are the ultimate compression 3πœŽπ‘† 𝜎𝐢 281 Mathematical and Information Technologies, MIT-2016 β€” Mathematical modeling strength, ultimate tensile strength and ultimate shear strength. For ultimate √ strengths the following relationships should be taken into account: 𝜎𝐢 > 3πœŽπ‘† , (𝛼) 𝜎𝐢 > 0, πœŽπ‘† > 0. In (7) a continuous positive function of the 1st invariant 𝑉 (πœŽβˆ’ ) is introduced ⎧ ⎨ 0, 𝜎 (𝛼) > 0, (𝛼) (𝛼) (𝛼) 𝑉 (πœŽβˆ’ ) = βˆ’πœŽ , βˆ’πœŽπΆ < 𝜎 (𝛼) < 0, (8) ⎩ (𝛼) (𝛼) 𝜎𝐢 , 𝜎 (𝛼) < βˆ’πœŽπΆ . The failure criterion 𝑧, which is calculated by the formula (7), has the value 0 if (𝛼) the stress is absent in the composite. It is ranged within 0 < 𝑧(πœŽπ‘–π‘— ) ≀ 1 in the (𝛼) loaded condition if there is not damage. And it takes the values 𝑧(πœŽπ‘–π‘— (πœ‰π‘  )) β‰₯ 1, if the fracture initiation occurs at some point πœ‰π‘  . If the failure criterion reaches the value 𝑧 = 1, then we obtain strength surface of a component (𝛼)2 (𝛼) πœŽπ‘’(𝛼)2 = 3πœŽπ‘† (1 + 𝐡 (𝛼) 𝑉 (πœŽβˆ’ )). (9) In the tensile area 𝜎 (𝛼) > 0 the strength surface is the von Mises ellipsoid (𝛼)2 (𝛼)2 (𝛼) πœŽπ‘’ = 3πœŽπ‘† . In the compression area βˆ’πœŽπΆ < 𝜎 (𝛼) < 0 the tensile strength (𝛼) is increased. And in the ”supercompression” area 𝜎 (𝛼) < βˆ’πœŽπΆ the strength surface again is the von Mises ellipsoid, but with the modified tensile strength: (𝛼)2 (𝛼)2 πœŽπ‘’ = 𝜎𝐢 . (𝛼) If the condition 𝑧(πœŽπ‘–π‘— (πœ‰π‘  )) β‰₯ 1 is satisfied at the point πœ‰π‘  or in a certain area PC2, there is no complete destruction. This is partial destruction of PC2, hereinafter called microdestruction. Introduce the dependence of the components of the elastic modulus of the failure criterion for accounting microdestruction of components in the model: (𝛼) (𝛼) πΆπ‘–π‘—π‘˜π‘™ (πœ‰π‘  , 𝑧) = (1 βˆ’ β„Ž(𝑧(πœŽπ‘–π‘— (πœ‰π‘  ) βˆ’ 1))πΆπ‘–π‘—π‘˜π‘™ , πœ‰π‘  ∈ π‘‰Λœπœ‰π›Ό , 𝛼 = 1...𝑁, (10) (𝛼) where πΆπ‘–π‘—π‘˜π‘™ are the components of the tensor of elasticity moduli of the com- posite components (they are constants). According to the formula (10) if mi- crodestruction occurs at the point πœ‰π‘  , elasticity modulus is equal to zero at this point. To calculate the strength of the composite as a whole, we need to calculate the limit values of average stresses πœŽΒ―π‘˜π‘™ . At this stresses an initial microdestruction occurs at least in one of its components (fillers or matrix) in a point πœ‰π‘ * ∈ π‘‰Λœπœ‰ at a time 𝑑* , and then complete destruction occurs. For the calculation of limit values of stresses in experimental research usually implement a process of linear load, in which the average stresses are proportional the time: 𝜎 Β―π‘˜π‘™ (𝑑) = 𝜎 Λœπ‘˜π‘™ 𝑑, where 𝜎 Λœπ‘˜π‘™ are the components of the stress gradient tensor. Substituting (5) in the strength criterion of the matrix or fillers (7) we obtain the initial failure condition of composite (𝛼) max {𝑧(π΅π‘™π‘›π‘˜π‘š (πœ‰π‘  )Β―πœŽπ‘˜π‘š (𝑑* ))} = 1, (11) πœ‰π‘  βˆˆπ‘‰Λœπœ‰ where πœ‰π‘  = πœ‰π‘ * are coordinates of the point in the PC2, 𝑑* is the time point at Β―π‘˜π‘š (𝑑* ) are limit stresses. which the condition (11) is executed first, 𝜎 282 Mathematical and Information Technologies, MIT-2016 β€” Mathematical modeling After appearance of the initial failure elastic moduli are changed in the de- stroyed areas of the matrix and/or fillers in accordance with the model described above. With further increase in average stress values 𝜎 Β―π‘˜π‘š (𝑑) failure condition (11) is satisfied in a large number of points of PC2, that is, there is the process of propagation of microdestruction. Some area π‘‰πœ‰ * (𝑑) of the partial destruction of the composite is formed in the periodicity cell 2. For modeling of effective elastic and strength properties of PC3 and PC1 is used a similar method. The stresses occur due to the thermal strain πœ€0π‘˜π‘™ = π›Όπ‘˜π‘™ (πœƒ βˆ’ πœƒ* ) of the ceramic composite during cooling after the laser sintering. 6 Details of numerical simulation The local tasks (1), (2) are solved by a finite element method which is described in [16, 17, 20]. We use 4-node tetrahedral finite elements, generated by open-source grid generators. The meshes contain different numbers of nodes (from 104 to 106 ). Meshes with a large number of finite elements are used in the calcula- tion of effective elastic moduli, when micro destruction is not happened. After microdestruction is beginning, the local tasks become nonlinear, because the elastic modulus of the matrix or fillers is changed, so we use iterative method to solve it. The number of iterations to achieve complete destruction is about 103 , so for these tasks we use meshes with a smaller number of elements to reduce the time of the numerical experiments. A numerical solution of large systems of linear algebraic equations, preprocessing and postprocessing, including 3D visu- alization and animation, was implemented in the software package, developed by the scientific and educational center ”Supercomputer Engineering Simulation and Development of Software Packages” of the Bauman Moscow State Technical University. 7 Results 7.1 Numerical simulation of microdestruction of ceramics for periodicity cell of the third level Consider the numerical simulation of microdestruction of ceramics for periodicity cell of the third level with the following properties of SiC matrix π‘š3 : elastic modulus πΈπ‘š = 320 GPa, Poisson’s ratio πœˆπ‘š = 0.35, ultimate strength πœŽπ‘‡0 π‘š = 0 0 0, 07 GPa; πœŽπΆπ‘š = 4 GPa; πœŽπ‘†π‘š = 0, 06 GPa. We suppose that the pores have a spherical shape. Fig. 3 shows some of the results of microstresses calculations (𝛼) in the PC3. Fig. 3a) shows the distribution of component 𝐡1111 of the stress concentration tensor in the PC3, where concentration of pore before the start failure is equal to 20%. Fig. 3b) shows the distribution of parameter of damage 𝑧 in the PC3 under tension in the direction of 𝑂π‘₯1 . Fig. 4 shows the process of microdestruction in the PC3 (matrix with defect) under compression. The failure of the periodicity cell starts on the surface of the pore (defect) and at first is spread in a direction perpendicular to load direction, 283 Mathematical and Information Technologies, MIT-2016 β€” Mathematical modeling a) b) (𝛼) Fig. 3. a) Distribution of the components 𝐡1111 of the stress concentration tensor in the PC3 (before the start of failure); b) distribution of the parameter of damage 𝑧 in the PC3 under tension in the direction of 𝑂π‘₯1 . and then the failure zone is turned round and spread in the direction of load action to complete destruction of PC3. a) b) c) Fig. 4. The process of microdestruction of PC3 (matrix with defect) depending on compressive stress. 7.2 Numerical simulation of microdestruction of ceramics for periodicity cell of the second level Fig. 5 shows the results of numerical solution of the process of microdestruction in the periodicity cell of the second level under compression. These results are 284 Mathematical and Information Technologies, MIT-2016 β€” Mathematical modeling calculated taking into account initial technological stresses, which occurs from the application of laser sintering. a) b) c) Fig. 5. The process of microdestruction of PC2 under compression taking into account initial technological stresses. Figure 6 shows the tensile strength of ceramic material depending on con- centration of coarse-grained fractions of SiC particle taking into account initial technological stresses. Fig. 6. The tensile strength of ceramic material depending on concentration coarse- grained fractions of SiC particle taking into account initial technological stresses. 8 Conclusions A mathematical model of microdestruction of reaction bonded silicon carbide has been developed. This model is based on the homogenization method and the finite element method for solution of local problems on periodicity cells. The new strength criterion of ceramic materials has been applied. The comparison with 285 Mathematical and Information Technologies, MIT-2016 β€” Mathematical modeling experimental data has shown that this criterion is applicable to solve the problem of microdestruction of the reaction-bonded silicon carbide. It is demonstrated that the developed model allows to simulate the processes of microdestruction of the ceramic composite and can be used as a tool for research and design of new materials with specified properties. References 1. Balokhonov, R.R., Romanova, V.A., Schmauder, S.: Computational analysis of de- formation and fracture in a composite material on the mesoscale level. Computa- tional Materials Science, vol. 37, pp. 110–118 (2006) 2. Bakhvalov, N. S. and Panasenko, G. P.: Averaging Processes in Periodic Media. Nauka Press, Moscow (1984) 3. Bardzokas, D. I., Zobnin, A. I.: Mathematical Modeling of Processes in Composite Materials with a Periodic Structure. URSS Press, Moscow (2003) 4. Clayton, J.D., Kraft, R.H., Leavy, R.B.: Mesoscale modeling of nonlinear elasticity and fracture in ceramic polycrystal under dynamic shear and compression. Int. J. of Solids and Structures, vol. 49, pp. 2686 –2702 (2012) 5. Dimitrienko, Yu. I.: Mechanics of Composite Materials at High Temperatures. Mashinostroenie Press, Moscow (1997) 6. Dimitrienko, Yu. I., Kashkarov, A. I.: Calculation of effective characteristics of com- posites with a periodic structure by the finite element method. Vestn. MGTU N. E. Baumana, Ser. Estestv. Nauki, no. 2, pp. 95–108 (2002) 7. Dimitrienko, Yu. I., Sokolov, A. P.: Development of the numerical method for calcu- lations of ef- fective elasticity characteristics of composite materials. Vestn. MGTU N. E. Baumana, Ser. Estestv. Nauki, no. 2, pp. 56 – 67 (2008a) 8. Dimitrienko, Yu. I., Sokolov, A. P.: The system for automated prediction of prop- erties of composite materials. Informats. Tekhnol., 8, pp. 31 –38 (2008b) 9. Dimitrienko, Yu. I., Sokolov A. P.: Development of Automated Technology of Calcu- lation of Effective Elastic Characteristics of Composites by Method of Asymptotic Averaging. Herald of the Bauman Moscow State Technical University. Series: Nat- ural sciences, no 2, pp. 56 – 67 (2008) 10. Dimitrienko, Yu. I.: Nonlinear Mechanics of Continuum. Fizmatlit Press, Moscow (2009) 11. Dimitrienko, Yu. I., Sokolov, A. P.: On elastic properties of composite materials. Mat. Model., vol.21, no. 4, pp. 96–110 (2009) 12. Dimitrienko, Yu. I., Sokolov, A. P.: Elastic properties of composite materials. Math. Models Computer Simul., vol. 2, no. 1, pp. 116–130 (2010) 13. Dimitrienko, Yu. I., Sokolov, A. P.: Modern numerical analysis of mechanical prop- erties of composite materials. Izv. RAN, Fiz. Ser., vol. 75, no. 11, pp. 1551–1556 (2011) 14. Dimitrienko, Yu. I., Sokolov, A. P.: Numerical modeling of composites with multi- scale microstructure. Bulletin of the Russian Academy of Sciences: Physics, V. 75, no 11, pp. 1457–1461 (2011) 15. Dimitrienko, Yu. I., Sokolov, A. P.: Multiscale modeling of elastic composite ma- terials. Mat. Model., vol. 24, no. 5, pp. 3 –20 (2012) 16. Dimitrienko, Yu. I., Sokolov, A. P., Shpakova, Yu.V., Yurin, Yu.V.: Modeling of surfaces of composite strength on the basis of the microstructural finite element analysis. Science and Education., no 11 (2012) 286 Mathematical and Information Technologies, MIT-2016 β€” Mathematical modeling 17. Dimitrienko, Yu. I., Sborshchikov, S.V., Sokolov, A.P.: Computational Modeling of Microdestruction and Strength of Multidimensional Reinforced Composites. Journal on Composite Mechanics and Design, V.19., no3, pp. 365–383 (2013) 18. Dimitrienko, Y. I., Sborshchikov, S.V., Sokolov, A.P., Shpakova,Y.V.: Computa- tional modeling of failure of textile composites. Computational continuum mechan- ics, Vol.6, no 4, pp. 389–402 (2013) 19. Dimitrienko, Yu. I., Sborshchikov, S.V., Belenovskaya, Yu.V. , Aniskovich, V.A., Perevislov, S.N.:Modeling microstructural destruction and strength of ceramic com- posites based on the reaction-bonded SiC. Science and Education, no11, pp. 475–496 (2013) 20. Dimitrienko, Yuriy I., Sokolov, Alexandr P., Shpakova, Yulia V.: Computer-aided Analysis of Micromechanics and Damage of Composite Materials Based on Mul- tiscale Homogenization Method. MRS Proceedings, Volume 1535, mmm12–a–0232 (2013) 21. Khokar, N.: 3D weaving: theory and practice. J. Textile Inst., vol. 92, no. 2, 193–207 (2001) 22. Neshpor, V.S., Zaitsev, G.P., Maystrenko, A.L. et al: The armor ceramics ballistic efficiency evaluation. 8th CIMTEC-World Ceramics Congress and Forum on New Materials, June 28July 4, . 102, Florencia, Italy (1994) 23. Pobedrya, B. E.: Mechanics of Composite Materials. MGU Press, Moscow (1984) 24. Rozenberg, O.A., Maystrenko, A.L., Kulich, V.G., Shestakov, S.I. Peculiarities of manufacturing and properties of high-density silicon carbide based armored ceramics. Proc. 5th Int. Armament Conf., p. 11. Waplevo, Poland (2004) 25. Sanchez-Palencia, E.: Nonhomogeneous Media and Vibration Theory. Mir Press, Moscow (1984) 26. Skripnyak, E. G., Skripnyak, N. V., Skripnyak, V. A., Skripnyak, V. V., Vaganova, I. K.: Multiscale computational model for simulation of mechanical behavior of heterogeneous brittle ceramics under dynamic loading. Zbornik radova konferencije MIT-2013. pp. 661–670. Beograd (2013) 287