


Vol 55, No 7 (2019)
- Year: 2019
- Articles: 11
- URL: https://journal-vniispk.ru/0012-2661/issue/view/9375
Numerical Methods
Efficient Numerical Integration Methods for the Cauchy Problem for Stiff Systems of Ordinary Differential Equations
Abstract
The notion of stiffness of a system of ordinary differential equations is refined. The main difficulties encountered when solving the Cauchy problem for stiff systems are indicated. The advantages of switching to a new argument, the integral curve arc length, are demonstrated. Various mesh step selection criteria are discussed, and the integral curve curvature criterion is recommended. The most reliable implicit and explicit schemes suitable for solving stiff problems are presented. A strategy permitting an asymptotically accurate computation of the error of a numerical solution simultaneously with the solution itself is described. An analysis of chemical kinetics of hydrogen combustion in oxygen with allowance for 9 components and 50 reactions between them is provided as an illustration.



Boundary Value Problems for a Pseudoparabolic Equation with the Caputo Fractional Derivative
Abstract
We study boundary value problems for a third-order pseudoparabolic equation with variable coefficients and with the Caputo fractional derivative. A priori estimates are derived in the differential and difference settings. These estimates imply the uniqueness of the solution of these problems and its stability with respect to the initial data and right-hand side as well as the convergence of solutions of the associated difference problem to the solution of the differential problem at the rate O(h2 + τ), where h and τ are the steps in the space and time variables.



Numerical Model of Compression Plasma Flows in Channels under a Longitudinal Magnetic Field
Abstract
We consider a mathematical model of plasma flows in nozzle channels formed by two coaxial electrodes. The acceleration of plasma in an azimuthal magnetic field is accompanied by its compression and heating in the compression zone at the channel axis past the tip of the shorter central electrode. The mathematical apparatus of the model is based on numerically solving two-dimensional MHD problems using the Zalesak flux-corrected transport (Z-FCT) scheme. In the computations, we study the dependence of the compression phenomenon and its quantitative characteristics on the channel geometry, the problem parameters, and the additional longitudinal magnetic field present in the channel.



Three-Level Schemes for the Advection Equation
Abstract
The advection equation, which is central to mathematical models in continuum mechanics, can be written in the symmetric form in which the advection operator is the half-sum of advection operators in the conservative (divergence) and nonconservative (characteristic) forms. In this case, the advection operator is skew-symmetric for any velocity vector. This fundamental property is preserved when using standard finite element spatial approximations in space. Various versions of two-level schemes for the advection equation have been studied earlier. In the present paper, unconditionally stable implicit three-level schemes of the second order of accuracy are considered for the advection equation. We also construct a class of schemes of the fourth order of accuracy, which deserves special attention.



Development and Application of Numerical Methods for Equations of Mixed Type in an Unbounded Domain
Abstract
We propose a generalization of methods for solving problems for a linear elliptic operator with a known fundamental solution in an unbounded domain to the case of mixed-type problems. Two new methods are constructed, the method of setting integral boundary conditions and a three-stage iterative method. The scope of the methods constructed is limited to the cases where the operators have a known fundamental solution outside some finite domain. Numerical algorithms implemented as computer software are created to solve particular problems with the methods proposed. The applicability of the methods to differential and integro-differential equations in the two-dimensional case and the possible applicability to three-dimensional problems are demonstrated.



Stability Analysis of Solution Methods for a Phase Transition Problem
Abstract
We consider a model of crystallization process in a binary compound taking into account heat and mass transfer in the solid and liquid phases. Stability of various methods for the numerical implementation of nonlinear conditions on a moving internal boundary is analyzed. For a group of methods based on the successive solution of governing equations, the stability domains determined by the system thermodynamic parameters are indicated. A coupled algorithm for solving the problem is proposed. Nonlinear equations at the phase interface are solved by the Newton method. The coupled algorithm has a significant stability margin and provides reliable results in a broad range of parameters of practical interest.



Iterative Method for Solving an Inverse Problem for a Hyperbolic Equation with a Small Parameter Multiplying the Highest Derivative
Abstract
We consider the Cauchy problem for a hyperbolic equation with a small parameter multiplying the highest derivative. The inverse problem of finding an unknown function that is a coefficient of the equation and also occurs in the initial condition is posed. The values of the solution of the Cauchy problem and its derivative at x = 0 are given as additional information for solving the inverse problem. An iterative method for determining the unknown function is constructed, and its convergence is proved. Existence theorems are proved for the solution of the inverse problem.



Adaptive Numerical Methods for Solving the Problem about Scattering on a Force Center
Abstract
We construct families of adaptive symplectic conservative numerical methods for solving problems about scattering on a force center. The methods preserve the global properties of the exact solution of the problem and approximate the dependences of the phase variables on time with the second, fourth, or sixth approximation order. The variable time step is selected automatically in two different ways depending on the properties of the solution.



Boundary Value Problems of Shielding Magnetic Fields by Cylindrical Multilayered Film Screens with Nonlinear Properties of the Layers
Abstract
A mathematical model is proposed for shielding permanent magnetic fields by a cylindrical thin-walled multilayered film screen made of materials with permeability depending nonlinearly on the magnetic field intensity. Boundary value problems with Robin boundary conditions and integral boundary conditions on the screen surface are stated for the magnetostatic equations. Efficient numerical methods are developed for solving these problems with allowance for the matching of contact layers with different magnetic properties. A new form of the nonlinear permeability factor is derived from experimental data. The magnetic field potential and intensity in the multilayered screen, as well as the shielding efficiency factor, which characterizes the attenuation of an external magnetic field when penetrating into the cylindrical screen, are studied numerically.



Two-Layer Completely Conservative Difference Schemes for the Gasdynamic Equations in Eulerian Coordinates with Adaptive Solution Regularization
Abstract
A family of two-layer completely conservative difference schemes with space-profiled weight factors for approximations on a time grid is constructed for the gasdynamic equations in Eulerian coordinates. We propose a construction for regularized mass, momentum, and internal-energy fluxes that effectively removes nonphysical solution oscillations and does not violate the complete conservativeness of the difference schemes in this class.



Finite Element Method for 3D Deformation of Hyperelastic Materials
Abstract
We describe a finite element method for solving 3D problems of nonlinear elasticity theory in the framework of finite strains for a hyperelastic material. Constitutive equations written with the use of the polar and upper triangular (QR) decompositions of the deformation gradient are considered. Our method permits developing an efficient, easy-to-implement technique for the numerical analysis of the stress—strain state of any hyperelastic material.


