Since 1978
Published in Sarov (Arzamas-16), Nizhegorodskaya oblast
RUSSIAN FEDERAL
NUCLEAR CENTER -
ALL-RUSSIAN RESEARCH INSTITUTE
OF EXPERIMENTAL PHYSICS
 
 Русский |  English
ABOUT EDITORIAL BOARD PUBLICATION ETHICS RULES FOR AUTHORS AUTHORS ARCHIVE MOST RECENT ISSUE IN NEXT ISSUE PAPER OF THE YEAR


Issue No 1, 2015


A QUASI-1D MODEL OF THERMONUCLEAR IGNITION OF DENSE DT MIXTURE DRIVEN BY PROTON BEAMS

K. V. Khishchenko, A. A. Charakhch’yan
VANT. Ser.: Mat. Mod. Fiz. Proc. 2015. No 1. P. 3-15.

We consider a cylindrical DT target of thickness 2H and density ρ0 ≤ 100ρs, where ρs ≈ 0,22 g/cm3 is the solid DT density under atmospheric pressure and temperature of 4K. The target is supposed to be enclosed in a heavy magnetized shell preventing lateral expansion and heat transfer from the fuel to the shell. The target is ignited at both ends simultaneously by two equivalent monoenergenic proton beams with 1MeV kinetic energy. The beam pulse intensity and duration are 1019/cm2 and 50 ps or 1018/cm2 and 500 ps, respectively. We use a 1D one-velocity two-temperature hydrodynamic model with a wide-range equation of state of the fuel, electron and ion heat conductivities, DT reaction kinetics, self-radiation of plasma and its heating by α-particles. The latter is assumed to be the governing target ignition mechanism. The flight of α-particles beyond the fuel is incorporated within the track method. We propose a modification of this method, which approximates the known Cauchy problem for a homogeneous steady-state kinetic equation in the Fokker-Planck approximation. The paths of α-particles are constrained by a cylindrical surface of a given radius, which is the method’s parameter and which is identified with the target radius and the proton beam radius. This quasi-1D model allows us to estimate the ignition energy and the target mass. For the initial fuel density of ρ0 = 100ρs, the estimated ignition energy is approximately 10 times smaller than the corresponding estimate for the problem with a beam radius much smaller than the target size.


Key words: cylindrical inertial confinement fusion target, ignition energy, burnup factor, track method, Fokker-Planck approximation.
APPLICATION OF THE "PISO" ALGORITHM TO MOLECULAR-IMMISCIBLE FLUID DYNAMICS SIMULATIONS

S. V. Yatsevich, V. V. Kurulin, D. P. Rubtsova
VANT. Ser.: Mat. Mod. Fiz. Proc. 2015. No 1. P. 16-29.

The paper considers some modifications of the finite-volume implementation of the known PISO algorithm on Eulerian combined grids as applied to molecular-immiscible fluid dynamics simulations. Results of comparative benchmark simulations for the one-velocity model implemented in the LOGOS application package are reported. Based on the results obtained, the most acceptable algorithm modification for the given class of problems is identified.


Key words: immiscible fluids, one-velocity model, numerical simulations, interface, PISO, LOGOS.
VALUE-BASED MODELING TECHNIQUE FOR REACTOR PHYSICS SIMULATIONS OF DEEP PENETRATION AND DETECTION USING THE "PRIZMA" CODE

O. V. Zatsepin, Ya. Z. Kandiev
VANT. Ser.: Mat. Mod. Fiz. Proc. 2015. No 1. P. 30-36.

PRIZMA is a code that has been developed at RFNC-VNIITF for Monte-Carlo ionizing radiation transport simulations for more than thirty years. The code enables simulations of both separate and joint transport of neutrons, photons, electrons, positrons and ions in the 1D, 2D and 3D geometry. Functionals are calculated using visiting estimates, i.e. the estimates, the contributions to which are non-zero only for the particle paths intersecting the integration region/surface. To improve efficiency, the code uses value-based modeling for deeper penetration and detection simulations. Consideration of reactor physics problems belonging to these classes revealed special features that required special modeling schemes and revision of the existing value-based modeling technique.
       This paper provides an overview of the methods for deep penetration and detection simulations using the PRIZMA code and describes specific features of some reactor physics problems belonging to these classes and requiring the development of the value-based modeling technique. To prove the validity of the improved technique, the paper presents solutions of a number of test and applied problems. The solutions are shown to agree with asymptotic solutions, if available, or, in the other cases, with solutions obtained by the analog modeling method or using other codes. Applied simulations to estimate signals of out-of-core detectors from neutron sources located in the core of VVER-1 000 and in-core detectors of VVER-1 000 are considered.


Key words: particle transport, Monte Carlo method, value-based modeling, deep penetration, detection, PRIZMA code, VVER-1 000.
PARALLELING OF THE "D" CODE FOR 2D GAS DYNAMICS SIMULATIONS WITH DYNAMIC BALANCING OF ARITHMETIC PROCESSOR LOAD

I. M. Epishkov, P. V. Egorov
VANT. Ser.: Mat. Mod. Fiz. Proc. 2015. No 1. P. 37-46.

The paper describes multi-fragment regular-block decomposition algorithms and the basic principles of dynamic balancing of processor arithmetic load in multiprocessor simulations using the Lagrangian D code.
       The multi-fragment regular-block decomposition is done by splitting the domain into blocks in two directions: main (the number of nodes in this direction is larger) and auxiliary. The first splitting is done for the main direction as applied to the entire domain producing decomposition layers. The second splitting is done for the auxiliary direction and only within one decomposition layer. Each processor can process several task fragments belonging to different decomposition layers and different domains.
       Dynamic balancing algorithms are executed with a view to ensure uniform arithmetic load balancing between processors. The paper describes the procedures of identifying this load and performing a new decomposition and the structure of processor communications.
       The paper also presents results of test simulations demonstrating the applicability of the algorithms in the code.


Key words: D code, multi-fragment regular-block decomposition, dynamic balancing of processor arithmetic load.
UNSTRUCTURED PRISMATIC DISCRETIZATION OF COMPLEX GEOLOGICAL STRUCTURES IN THE PARALLEL MODE

M. L. Sidorov, V. A. Pronin
VANT. Ser.: Mat. Mod. Fiz. Proc. 2015. No 1. P. 47-55.

The paper presents a parallel grid generation method for prismatic unstructured grids used for the discretization of complex geological structures in oil and hydroecology simulations. The method enables grid matching to various types of objects (complex-trajectory wells, geological fractures, reservoirs etc.) and offers high grid generation speed. The method is implemented in the NIMFA code.


Key words: thermohydrodynamic simulator, unstructured grid, parallel grid generator, matching, frontal method of spheres, generalized angular point method, geological structures, NIMFA code, MPI paralleling.
SOME GRID GENERATION ALGORITHMS FOR UNSTRUCTURED POLYGONAL GRIDS IN THE "TIM-2D" CODE

A. I. Panov, A. V. Shurygin
VANT. Ser.: Mat. Mod. Fiz. Proc. 2015. No 1. P. 56-65.

Some grid generation algorithms for unstructured polygonal preprocessing grids for the TIM-2D code are presented. The format of grid topology data storage is identified. The method of cell cutoff by domain boundaries and the grid cross-linking method to integrate different types of grids within the same domain are described. Practical recommendations are given for the use of some or other types of grids depending on problem features.


Key words: TIM-2D code, unstructured grid, preprocessing, cross-linking of different types of grids.
2D SIMULATION DATA INTERPOLATION TECHNOLOGY IN THE "BAZIS" PROGRAM PACKAGE

S. V. Gagarin, N. V. Galitskaya, O. V. Belomestnykh, S. I. Kuzmina, D. V. Mogilenskikh, E. A. Prib, A. A. Ushkova
VANT. Ser.: Mat. Mod. Fiz. Proc. 2015. No 1. P. 66-77.

The difference grid used for Lagrangian calculations in mathematical continuum mechanics modeling can get severely distorted and make further calculations impossible. In this case, the user must create a new grid generally preserving the geometry of the model’s domains and run a program to interpolate (remap) physical quantities from the old difference grid to the new one.
       RFNC-VNIITF has developed a program package called BAZIS, which includes the following subsystems:
  1. descriptive (tabulated) representation of simulation data;
  2. graphic representation of simulation data;
  3. comparison of simulation graphs;
  4. preprocessing (GeomGrid2);
  5. interpolation job setup;
  6. interpolation initiation;
  7. calculations of mean (balance) values of physical quantities.

       The paper describes the procedures of interactive setup of interpolation jobs, their initiation and analysis of their results.
       The first three subsystems enable 2D simulation data analysis in the descriptive and tabulated form. GeomGrid2 generates a new difference grid. The interpolation job setup subsystem enables interactive definition of remap jobs. The interpolation initiation subsystem allows the user to launch the remap program both on the local computer and on the remote computational server. The mean value subsystem enables the comparison of balance values of physical quantities (mass, volume, density etc.) in simulations before and after interpolation.
       Introduction of BAZIS reduced the time spent on interpolation and the probability of remapping errors.


Key words: continuum mechanics, view, STDV (standard two-dimensional view), interpolation, initial geometry, difference grid, physical quantities.
DIGITAL SIGNAL RECOGNITION IN LOW SIGNAL/NOISE RATIO ENVIRONMENTS

O. N. Naryshkina, A. V. Svetikov, V. V. Shubin
VANT. Ser.: Mat. Mod. Fiz. Proc. 2015. No 1. P. 78-87.

The paper discusses digital signal recognition capabilities for low signal/noise ratio environments and results of developing a corresponding signal structuring algorithm. The algorithm is implemented in the C++ language in a signal processing program used for Fast Ethernet and FDDI signals. Experimental results of program tests for Fast Ethernet and FDDI signals are presented (125Mb/s rate). The error rate (BER) was decreased compared to the theoretical error probability for direct detection.


Key words: fiber optic link, signal/noise ratio, error probability, error rate, optical signal structuring.
[ Back ]
 
 
© FSUE "RFNC-VNIIEF", 2000-2024