A new adaptive finite-element numerical method has been developed for the unsteady Navier-Stokes equations of incompressible flow in two dimensions. The momentum equations combined with a pressure correction equation are solved employing a non-staggered grid. The solution is advanced in time with an explicit/implicit marching scheme. An adaptive algorithm has been implemented, which refines the grid locally in order to resolve detected flow features. A combination of quadrilateral, as well as triangular cells provides a stable and accurate numerical treatment of grid interfaces that are located within regions of high gradients. Applications of the developed adaptive algorithm include both steady and unsteady flows, with low and high Reynolds numbers. Comparisons with analytical, as well as experimental data evaluate accuracy and robustness of the method.

The proposed algorithm is based on the fourth-order compact discretization schemes for the Navier-Stokes equations in streamfunction-vorticity-pressure formulation. The equations are expressed in terms of a general orthogonal curvilinear coordinate system which allows for modelling non-standard geometries. Two distinct parallelization strategies are considered. The first one relies on the domain decomposition approach, in which each subdomain is served by a different processor. In the second strategy, suitable for massively parallel computers, each processor serves a single grid point. The comparison of the performance of various computing platforms is presented, including a 2048-processor MasPar computer.

This paper presents a local mesh refinement (LMR) technique and its application to incompressible fluid flows with or without free surface boundaries. In the LMR method patches of fine grid are embedded in arbitrary regions of interest. Hence, more accurate solutions can be obtained with a lower number of computational cells. LMR is very suitable for the simulation of free surface movements because free surface flow problems generally require a finer computational grid to obtain adequate results. By means of this technique one can place finer grids only near surfaces and thus greatly reduce the total number of cells and computational costs. This paper introduces two LMR codes. Numerical examples calculated with the codes demonstrate well the advantages of the LMR method.

New methods of the neutron and photon transport Monte Carlo simulation suitable for vector computers have been investigated and general purpose multigroup and continuous energy codes have been developed. On vector supercomputers, the codes achieved high speed-up gains of an order of ten or more compared with the conventional scalar codes. To achieve more speed-up, the Monte Carlo codes are applied to three types of parallel processing environments; (1) a massively parallel computer, (2) a vector-parallel type supercomputer, and (3) a cluster of workstations connected to a network. On the massively parallel computer and the vector-parallel supercomputer, speed-ups almost proportional to the number of processors are achieved by simply assigning particles uniformly to each processor, and the speed-up with the vector-parallel supercomputer is enhanced by vector processing. On the other hand, in the workstation cluster, the computational power of each workstation may differ and the simple particle assignment may not be successful. By modifying particle assignment methods, effective parallel processings are made possible in such an environment.

The JAERI Monte Carlo Machine has been developed mainly to enhance the computational performance of numerical simulations with particle models such as Monte Carlo methods. The features of the JAERI Monte Carlo machine are i) vector processing capability for arithmetic operations, ii) special pipelines for fast vector processing in categorizations of particles, iii) enhanced load/store pipelines for indirectly addressed vector elements, iv) parallel processing capability for spatially and phenomenologically independent particles. This paper describes the design philosophy and architecture of the JAERI Monte Carlo machine and its effective performance through practical applications of the multi-group criticality safety code KENO-IV, the continuous-energy neutron/photon transport code MCNP and other codes for particle simulation.

Passive designs of several proposed light water reactors rely on containment cooling by condensation heat transfer in a high-concentration, noncondensable gas environment. To evaluate the safety of these plants, methods for analyzing the performance of the cooling systems must be developed. The current study discusses the physics of condensation in the presence of noncondensable gases and a method for predicting the accompanying heat transfer rates based on experimental data. The resulting experimental correlation for heat transfer coefficients has been implemented into the TRACG code, and a specific application has been made to the Simplified Boiling Water Reactor for conditions following a Design Basis Accident LOCA (Loss-of-Coolant Accident).

The paper presents selected results of the analysis of thermal and mass flow transient processes within the containments of the WWER-440 and the WWER-1000 nuclear reactors during Loss-of-Coolant Accidents based on the mathematical model and computer code for LOCA simulation. General assumptions of the mathematical model (with lumped parameters) are briefly presented. Changes of thermal variables (temperature, pressure etc.) are governed by the fundamental thermodynamic equations. All these equations have the nonlinear, integral form. The whole area of the containment is divided into several control volumes. Control volumes are joined in a given mode (orifices, valves, siphon closures etc.). The liquid phase (water) and the gaseous phase (air, steam and hydrogen) can appear in a control volume. Thermal equilibrium within an individual phase and a non-equilibrium state between phases is assumed. Heat accumulation in the walls and internal structures of the containment is taken into account and heat transfer between liquid and gaseous phases is also considered. The working mathematical model can be used for the analysis of different scenarios of LOCA within the containment of the PWR and BWR reactors. Later on the sample results of calculations of changes of pressure and temperature within the containment of the WWER-440 nuclear reactor and within the full containment of the WWER-1000 reactor are presented.

In the engineering processes it is important to simulate thermo-hydraulic phenomena numerically in the limited time before design of equipment. In thermo-hydraulic problem, as it may be time-consuming to solve elliptic equation numerically, a heavy burden is imposed on the computer. A SOR method is one of the effective methods to solve elliptic equation. As it is difficult to find the optimum relaxation factor, the value of this factor for the practical problems used to be estimated by the expertise. In this paper, the implications about the relaxation factor are translated into fuzzy control rules on the basis of the expertise of numerical analysts, and then the fuzzy controller is incorporated into the numerical algorithm. A Dirichlet problem of the Poisson's equation and the cavity flow problem are chosen to verify the feasibility of fuzzy controller for relaxation. Numerical experiments with the fuzzy controller resulted in generating a good performance.