THE PROPERTIES OF GRAVITATIONAL TURBULENT MIXING MODELS FOR NON-SELF-SIMILAR PROBLEMS

V.E.Neuvazhaev
VANT. Ser.: Mat. Mod. Fiz. Proc. 1994. No 3. P. 3-10.

      New results obtained by the author when studying the properties of k- and kε-models commonly used for applications problems related to IF, are given. Four types of problems are considered. These problems are relatively complicated and are not only interesting by themselves, but may serve the tests for comparison with other models. All results are obtained in analytical form which allows to formulate general rules.

V.V. Evdokimov, R.M. Shagaliev
VANT. Ser.: Mat. Mod. Fiz. Proc. 1994. No 3. P. 11-17.

      A consistent flow convergence acceleration method (FCA) is proposed for simple iterations (collision integral iterations) for numerical solution of 2-D transport equations with DSn -method schemes. The method is developed for general case, when space meshes are nonorthogonal and weight factors of difference scheme are space point functions. For some test problems, proving the efficiency of the proposed acceleration method comparative calculational results are included.

M.Yu. Kozmanov
VANT. Ser.: Mat. Mod. Fiz. Proc. 1994. No 3. P. 18-20.

      In this article a maximum principle for the frequency dependent radiative transfer equations is formulated and proved. Medium is considered fixed, but electronic and ionic heat-conductions are taken into account.

N.V. Ivanov, S.N. Nikonov, V.N. Piskunov
VANT. Ser.: Mat. Mod. Fiz. Proc. 1994. No 3. P. 21-25.

      A prediction technique based on Monte Carlo method is presented for environment contamination due to industrial and abnormal release in air. An air parameter model used for urgent prediction is given. Also discussed are the resources of definition for emission solid source and aerosol dispersed composition in it. The characteristics of the applications program for calculations on PC/AT are briefly described.

E.V. Diyankova, O.S. Shirokovskaya
VANT. Ser.: Mat. Mod. Fiz. Proc. 1994. No 3. P. 26-32.

      A modification of a well-known Boris scheme is proposed for Neuton-Lorents equation solution when simulating self-consistent motion of charged particles. In general case a modificated scheme has the 2^nd order accuracy and in case of nonrelativistic motion and constant fields the accuracy reaches the 4^th order. In nonlinear case the scheme is completed with iteration algorithm. The examples of numerical calculations presented here show that the modification proposed and iteration algorithm allow to obtain more precise characteristics of a moving particle.

I.A. Gasilova
VANT. Ser.: Mat. Mod. Fiz. Proc. 1994. No 3. P. 33-40.

      An algorithm for automated generation of the initial approximation of a curvilinear coordinate grid for single-connection two-dimensional star-shaped domains with a boundary consisting of a finite number of line segments is proposed. Whether the domain belongs to the star-shaped domain class is determined by means of R-functions and simple geometric constructions. Examples of constructed grids for various star-shaped domains are given.

Yu.A. Bondarenko
VANT. Ser.: Mat. Mod. Fiz. Proc. 1994. No 3. P. 41-46.

      General form of 3-D flow of nonviscous incompressible fluid with constant pressure is found. The fact, that there are 3 types of such solutions, is proved. The first type - is the constant velocity. In the second type two velocity components are the functions of the third component (such solutions are already known). In the third type of solutions one velocity component is the function of two other components. General solution of the third type is implicitly defined from a system of 3 equations with 3 arbitrary functions, each of two arguments.

O.V. Ushakova
VANT. Ser.: Mat. Mod. Fiz. Proc. 1994. No 3. P. 47-56.

      Algorithm and program for generation of curvilinear optimal adaptive grids in two-dimensional single- connection regions are described. The algorithm represents a special procedure of minimizing the grid quality functional which is constructed by means of weight function equidistribution principle and geometrical approach. The results of computations are given.

L.F. Gudarenko, V.A. Zherebtsov
VANT. Ser.: Mat. Mod. Fiz. Proc. 1994. No 3. P. 57-64.

      The paper seeks to construct a tabulated-analytical equation of state on the basis of the tables with thermodynamic data for water and water vapour. Table data approximation is performed with polynomial spline functions. The equation of state allows to compute thermodynamic functions together with their derivatives with respect to any of the following three pairs of variables: density - temperature, density - energy, density - pressure. The domain described by the equation of state is the following: density — 0 < ρ < 1 g/cm³; pressure - 0 < P < 0,1 GPa; specific energy - 0 < E <3,7kJ/g; temperature - 0 < T < 800ºC. The equation is intended to compute a compressible medium flow when choosing structural elements of power systems and to substantiate their carrying capability.

Yu.A. Bondarenko, V.I. Selin
VANT. Ser.: Mat. Mod. Fiz. Proc. 1994. No 3. P. 65-77.

      The generation of multiparameter finite-difference CREST schemes using sequential variational method for isotropic elasticity at rectangular meshes is described. Stability and conservation of momentum and total energy are proved. Nonstability is discovered in linear analog of a well-known Wilkins scheme.

S.M. Bakhrakh, S.P. Egorshin
VANT. Ser.: Mat. Mod. Fiz. Proc. 1994. No 3. P. 78-80.

      The implementation experience is described for algorithms of generation and calculational support for regular meshes, used in EGAK complex on PS-2100 multiprocessor.