All issues

We consider an application of the Message Passing Interface (MPI) technology for parallelization of the program code which solves equation of the linear elasticity theory. The solution of this equation describes the propagation of elastic waves in demormable rigid bodies. The solution of such direct problem of seismic wave propagation is of interest in seismics and geophysics. Our implementation of solver uses grid-characteristic method to make simulations. We consider technique to reduce time of communication between MPI processes during the simulation. This is important when it is necessary to conduct modeling in complex problem formulations, and still maintain the high level of parallelism effectiveness, even when thousands of processes are used. A solution of the problem of effective communication is extremely important when several computational grids with arbirtrary geometry of contacts between them are used in the calculation. The complexity of this task increases if an independent distribution of the grid nodes between processes is allowed. In this paper, a generalized approach is developed for processing contact conditions in terms of nodes reinterpolation from a given section of one grid to a certain area of the second grid. An efficient way of parallelization and establishing effective interprocess communications is proposed. For provided example problems we provide wave fileds and seismograms for both 2D and 3D formulations. It is shown that the algorithm can be realized both on Cartesian and on structured (curvilinear) computational grids. The considered statements demonstrate the possibility of carrying out calculations taking into account the surface topographies and curvilinear geometry of curvilinear contacts between the geological layers. Application of curvilinear grids allows to obtain more accurate results than when calculating only using Cartesian grids. The resulting parallelization efficiency is almost 100% up to 4096 processes (we used 128 processes as a basis to find efficiency). With number of processes larger than 4096, an expected gradual decrease in efficiency is observed. The rate of decline is not great, so at 16384 processes the parallelization efficiency remains at 80%.

The paper considers the problem of stationary mixed convection and heat transfer of a viscous heatconducting fluid in a plane square lid-driven cavity. The hot top cover of the cavity has any temperature $T_\mathrm{H}$ and cold bottom wall has temperature $T_\mathrm{0} (T_\mathrm{H} > T_\mathrm{0})$, whereas in contrast the side walls are insulated. The fact that the fluid density can take arbitrary values depending on the amount of overheating of the cavity cover is a feature of the problem. The mathematical formulation includes the Navier–Stokes equations in the ’velocity–pressure’ variables and the heat balance equation which take into account the incompressibility of the fluid flow and the influence of volumetric buoyancy force. The difference approximation of the original differential equations has been performed by the control volume method. Numerical solutions of the problem have been obtained on the $501 \times 501$ grid for the following values of similarity parameters: Prandtl number Pr = 0.70; Reynolds number Re = 100 and 1000; Richardson number Ri = 0.1, 1, and 10; and the relative cover overheating $(T_\mathrm{H}-T_\mathrm{0})/T_\mathrm{0} = 0, 1, 2, 3$. Detailed flow patterns in the form of streamlines and isotherms of relative overheating of the fluid flow are given in the work. It is shown that the increase in the value of the Richardson number (the increase in the influence of buoyancy force) leads to a fundamental change in the structure of the liquid stream. It is also found out that taking into account the variability of the liquid density leads to weakening of the influence of Ri growth on the transformation of the flow structure. The change in density in a closed volume is the cause of this weakening, since it always leads to the existence of zones with negative buoyancy in the presence of a volumetric force. As a consequence, the competition of positive and negative volumetric forces leads in general to weakening of the buoyancy effect. The behaviors of heat exchange coefficient (Nusselt number) and coefficient of friction along the bottom wall of the cavity depending on the parameters of the problem are also analyzed. It is revealed that the greater the values of the Richardson number are, the greater, ceteris paribus, the influence of density variation on these coefficients is.

In this work we have developed a new efficient program for the numerical simulation of 3D global chemical transport on an adaptive finite-difference grid which allows us to concentrate grid points in the regions where flow variables sharply change and coarsen the grid in the regions of their smooth behavior, which significantly minimizes the grid size. We represent the adaptive grid with a combination of several dynamic (tree, linked list) and static (array) data structures. The dynamic data structures are used for a grid reconstruction, and the calculations of the flow variables are based on the static data structures. The introduction of the static data structures allows us to speed up the program by a factor of 2 in comparison with the conventional approach to the grid representation with only dynamic data structures.

We wrote and tested our program on a computer with 6 CPU cores. Using the computer microarchitecture simulator gem5, we estimated the scalability property of the program on a significantly greater number of cores (up to 32), using several models of a computer system with the design “computational cores – cache – main memory”. It has been shown that the microarchitecture of a computer system has a significant impact on the scalability property, i.e. the same program demonstrates different efficiency on different computer microarchitectures. For example, we have a speedup of 14.2 on a processor with 32 cores and 2 cache levels, but we have a speedup of 22.2 on a processor with 32 cores and 3 cache levels. The execution time of a program on a computer model in gem5 is 104–105 times greater than the execution time of the same program on a real computer and equals 1.5 hours for the most complex model.

Also in this work we describe how to configure gem5 and how to perform simulations with gem5 in the most optimal way.

The paper is devoted to the study of wave and relaxation effects during the pulsed outflow of a gas mixture with a high content of solid particles from a cylindrical channel during its initial partial filling. The problem is formulated in a two-speed two-temperature formulation and was solved numerically by the hybrid large-particle method of the second order of approximation. The numerical algorithm is implemented in the form of parallel computing using basic Free Pascal language tools. The applicability and accuracy of the method for wave flows of concentrated gas-particles mixtures is confirmed by comparison with test asymptotically accurate solutions. The calculation error on a grid of low detail in the characteristic flow zones of a two-phase medium was 10^-6 . . . 10^-5.

Based on the wave diagram, the analysis of the physical pattern of the outflow of a gas suspension partially filling a cylindrical channel is performed. It is established that, depending on the degree of initial filling of the channel, various outflow modes are formed. The first mode is implemented with a small degree of loading of the high-pressure chamber, at which the left boundary of the gas-particles mixture crosses the outlet section before the arrival of the rarefaction wave reflected from the bottom of the channel. At the same time, the maximum value of the mass flow rate of the mixture is achieved. Other modes are formed in cases of a larger initial filling of the channel, when the rarefaction waves reflected from the bottom of the channel interact with the gas suspension layer and reduce the intensity of its outflow.

The influence of relaxation properties with changing particle size on the dynamics of a limited layer of a gas-dispersed medium is studied. Comparison of the outflow of a limited gas suspension layer with different particle sizes shows that for small particles (the Stokes number is less than 0.001), an anomalous phenomenon of the simultaneous existence of shock wave structures in the supersonic and subsonic flow of gas and suspension is observed. With an increase in the size of dispersed inclusions, the compaction jumps in the region of the two-phase mixture are smoothed out, and for particles (the Stokes number is greater than 0.1), they practically disappear. At the same time, the shock-wave configuration of the supersonic gas flow at the outlet of the channel is preserved, and the positions and boundaries of the energy-carrying volumes of the gas suspension are close when the particle sizes change.

In real problems of exploration seismology we deal with a heterogeneity of the nature of elastic waves interaction with the surface of a fracture by the propagation through it. The fracture is a complex heterogeneous structure. In some locations the surfaces of fractures are placed some distance apart and are separated by filling fluid or emptiness, in some places we can observe the gluing of surfaces, when under the action of pressure forces the fracture surfaces are closely adjoined to each other. In addition, fractures can be classified by the nature of saturation: fluid or gas. Obviously, for such a large variety in the structure of fractures, one cannot use only one model that satisfies all cases.

This article is concerned with description of developed mathematical fracture models which can be used for numerical solution of exploration seismology problems using the grid-characteristic method on unstructured triangular (in 2D-case) and tetrahedral (in 3D-case) meshes. The basis of the developed models is the concept of an infinitely thin fracture, whose aperture does not influence the wave processes in the fracture area. These fractures are represented by bound areas and contact boundaries with different conditions on contact and boundary surfaces. Such an approach significantly reduces the consumption of computer resources since there is no need to define the mesh inside the fracture. On the other side, it allows the fractures to be given discretely in the integration domain, therefore, one can observe qualitatively new effects, such as formation of diffractive waves and multiphase wave front due to multiple reflections between the surfaces of neighbor fractures, which cannot be observed by using effective fracture models actively used in computational seismology.

The computational modeling of seismic waves propagation through layers of mesofractures was produced using developed fracture models. The results were compared with the results of physical modeling in problems in the same statements.

Considered the application of physics-informed neural networks using multi layer perceptrons to solve Cauchy initial value problems in which the right-hand sides of the equation are continuous monotonically increasing, decreasing or oscillating functions. With the use of the computational experiments the influence of the construction of the approximate neural network solution, neural network structure, optimization algorithm and software implementation means on the learning process and the accuracy of the obtained solution is studied. The analysis of the efficiency of the most frequently used machine learning frameworks in software development with the programming languages Python and C# is carried out. It is shown that the use of C# language allows to reduce the time of neural networks training by 20–40%. The choice of different activation functions affects the learning process and the accuracy of the approximate solution. The most effective functions in the considered problems are sigmoid and hyperbolic tangent. The minimum of the loss function is achieved at the certain number of neurons of the hidden layer of a single-layer neural network for a fixed training time of the neural network model. It’s also mentioned that the complication of the network structure increasing the number of neurons does not improve the training results. At the same time, the size of the grid step between the points of the training sample, providing a minimum of the loss function, is almost the same for the considered Cauchy problems. Training single-layer neural networks, the Adam method and its modifications are the most effective to solve the optimization problems. Additionally, the application of twoand three-layer neural networks is considered. It is shown that in these cases it is reasonable to use the LBFGS algorithm, which, in comparison with the Adam method, in some cases requires much shorter training time achieving the same solution accuracy. The specificity of neural network training for Cauchy problems in which the solution is an oscillating function with monotonically decreasing amplitude is also investigated. For these problems, it is necessary to construct a neural network solution with variable weight coefficient rather than with constant one, which improves the solution in the grid cells located near by the end point of the solution interval.

The paper contains numerical simulation of nonisothermal nonlinear flow in a porous medium. Twodimensional unsteady problem of heavy oil, water and steam flow is considered. Oil phase consists of two pseudocomponents: light and heavy fractions, which like the water component, can vaporize. Oil exhibits viscoplastic rheology, its filtration does not obey Darcy's classical linear law. Simulation considers not only the dependence of fluids density and viscosity on temperature, but also improvement of oil rheological properties with temperature increasing.

To solve this problem numerically we use streamline method with splitting by physical processes, which consists in separating the convective heat transfer directed along filtration from thermal conductivity and gravitation. The article proposes a new approach to streamline methods application, which allows correctly simulate nonlinear flow problems with temperature-dependent rheology. The core of this algorithm is to consider the integration process as a set of quasi-equilibrium states that are results of solving system on a global grid. Between these states system solved on a streamline grid. Usage of the streamline method allows not only to accelerate calculations, but also to obtain a physically reliable solution, since integration takes place on a grid that coincides with the fluid flow direction.

In addition to the streamline method, the paper presents an algorithm for nonsmooth coefficients accounting, which arise during simulation of viscoplastic oil flow. Applying this algorithm allows keeping sufficiently large time steps and does not change the physical structure of the solution.

Obtained results are compared with known analytical solutions, as well as with the results of commercial package simulation. The analysis of convergence tests on the number of streamlines, as well as on different streamlines grids, justifies the applicability of the proposed algorithm. In addition, the reduction of calculation time in comparison with traditional methods demonstrates practical significance of the approach.

A two-dimensional mathematical model was developed for estimating the stresses in welded joints formed during multipass welding of multilayer steels. The basis of the model is the system of equations that includes the Lagrange variational equation of incremental plasticity theory and the variational equation of heat conduction, which expresses the principle of M. Biot. Variational-difference method was used to solve the problems of heat conductivity and calculation of the transient temperature field, and then at each time step – for the quasi-static problem of thermoplasticity. The numerical scheme is based on triangular meshes, which gives a more accuracy in describing the boundaries of structural elements as compared to rectangular grids.

In this paper, the system of equations of magnetic hydrodynamics (MHD) is considered. The exact solutions found describe fluid flows in a porous medium and are related to the development of a core simulator and are aimed at creating a domestic technology «digital deposit» and the tasks of controlling the parameters of incompressible fluid. The central problem associated with the use of computer technology is large-dimensional grid approximations and high-performance supercomputers with a large number of parallel microprocessors. Kinetic methods for solving differential equations and methods for «gluing» exact solutions on coarse grids are being developed as possible alternatives to large-dimensional grid approximations. A comparative analysis of the efficiency of computing systems allows us to conclude that it is necessary to develop the organization of calculations based on integer arithmetic in combination with universal approximate methods. A class of exact solutions of the Navier – Stokes system is proposed, describing three-dimensional flows for an incompressible fluid, as well as exact solutions of nonstationary three-dimensional magnetic hydrodynamics. These solutions are important for practical problems of controlled dynamics of mineralized fluids, as well as for creating test libraries for verification of approximate methods. A number of phenomena associated with the formation of macroscopic structures due to the high intensity of interaction of elements of spatially homogeneous systems, as well as their occurrence due to linear spatial transfer in spatially inhomogeneous systems, are highlighted. It is fundamental that the emergence of structures is a consequence of the discontinuity of operators in the norms of conservation laws. The most developed and universal is the theory of computational methods for linear problems. Therefore, from this point of view, the procedures of «immersion» of nonlinear problems into general linear classes by changing the initial dimension of the description and expanding the functional spaces are important. Identification of functional solutions with functions makes it possible to calculate integral averages of an unknown, but at the same time its nonlinear superpositions, generally speaking, are not weak limits of nonlinear superpositions of approximations of the method, i.e. there are functional solutions that are not generalized in the sense of S. L. Sobolev.

Numerical simulation of hull flow, marine propellers and other basic problems of ship hydrodynamics using Cartesian adaptive locally-refined grids is advantageous with respect to numerical setup and makes an express analysis very convenient. However, when more accurate viscous phenomena are needed, they condition some problems including a sharp increase of cell number due to high levels of main grid adaptation needed to resolve boundary layers and time step decrease in simulations with a free surface due to decrease of transit time in adapted cells. To avoid those disadvantages, additional boundary layer grids are suggested for resolution of boundary layers. The boundary layer grids are one-dimensional adaptations of main grid layers nearest to a wall, which are built along a normal direction. The boundary layer grids are additional (or chimerical), their volumes are not subtracted from main grid volumes. Governing equations of flow are integrated in both grids simultaneously, and the solutions are merged according to a special algorithm. In simulations of ship hull flow boundary layer grids are able to provide sufficient conditions for low-Reynolds turbulence models and significantly improve flow structure in continues boundary layers along smooth surfaces. When there are flow separations or other complex phenomena on a hull surface, it can be subdivided into regions, and the boundary layer grids should be applied to the regions with simple flow only. This still provides a drastic decrease of computational efforts. In simulations of marine propellers, the boundary layer grids are able to provide refuse of wall functions on blade surfaces, what leads to significantly more accurate hydrodynamic forces. Altering number and configuration of boundary grid layers, it is possible to vary a boundary layer resolution without change of a main grid. This makes the boundary layer grids a suitable tool to investigate scale effects in both problems considered.