Based on maximum principle analysis together with definitions of restriction position and restriction stencil, a unification of multidimensional limiter construction process for unstructured mesh was presented. To enhance flow resolution, a new type of multidimensional limiter was developed. It was compared with several conventional limiters. It shows that due to less restriction in gradient reconstruction process, the new limiter is less dissipative and has better resolution for complex flow with shock and contact discontinuity interaction.

With principle of conformal transformation and mirror reflection, a mathematical model of flow field in a mixed injection production well pattern was established. Five point horizontal well and vertical well pattern are used for an example. Mathematical expressions of potential function and stream function are derived. Distribution maps of potential line and flow line in a mixed injection production well network with five point method are obtained. Influence of key parameters on internal flow field distribution of mixed injection production well network is analyzed. It shows that at both ends of horizontal well flow lines are relatively dense, and liquid producing capacity at both ends is stronger. The longer the horizontal section, the greater the discharge area of flow line control. As horizontal well is 45 degree angle with vertical wells connection line, displacement effect of horizontal well is relatively good. As horizontal well is perpendicular to direction of main permeability, streamlines are concentrated to production well. Linear displacement is formed and the displacement effect is good. As horizontal well is off center position greater, streamline distribution density differentiation is more obvious.

A silicon micro actuator model with large-displacement and low-voltage is presented based on principle of verticalhorizontal bending. Dynamic equation of micro actuator with axial and transverse loading is built based on Lagrange-Maxwell electromechanical analysis dynamics. Influence of temperature stress,axial electrostatic force and squeezing force are analyzed. Transverse distributed load and axial load are transformed equivalently into transverse centralized load with Runge-Kutta algorithm and finite difference method respectively. Relationships of deformation with driving voltage,regulation voltage,and axial compression quantity and temperature difference are simulated. It shows that displacement is as large as 10. 861 μm as driving voltage is only 16 V. Displacement is much greater than that of existing micro actuators. Simulation results are verified by experiments.

To study nonlinear aerodynamics of merging vortex pair, a simplified model is employed to obtain merging at different separating distances by solving N-S equations in an LES turbulence model. Interactions of vortex merging, effects of separation distance on merging and control of vortex position by merging of vortex pair with different vortex strength ratios are analyzed with topological structure of streamlines. It is shown that merging distance increases nonlinearly with separating distance. Increasing separation distance decreases ability to control vortex position.

The spin reorientation of magnetic thin film on a simple cubic lattice is simulated by employing Monte Carlo method.The influence of the magnetic anisotropy and the long-range dipolar interaction on the spin configurations of the system is also studied in detail.Meanwhile,the phase diagrams and the distributions of spin configuration,magnetization,specific heat of the systems are obtained.The results show that in a range of parameters,a transition from out-of-plane to in-plane ordering with the increase of the temperature is possible.

Some natures have been studied for the propagation process of very intense and very short laser pulse in underdense plasma.The wave equation is derived and solved using the method of 2-D Fast Fourier Transformation.The basic equations for relativistic self-focusing include the nonlinear source term and the effect of diffraction.As a result of ponderomotive forces and relativism,the plasma frequency is reduced and so the refractive index is changed.It affects the propagation of laser in plasma.When the incident laser power (p) exceeds the critical power (p_c),self-focusing appears during the propagation of laser.On the contrary,laser damps gradually and self-focusing doesn't appear.

Abstract] 2-D and 3-D numerical simulation methods of filamentation instability in laser-target interaction are presented.The Landau-fluid models are adopted with a simplified 3-D numerical method,it simulates filamentation process for laser beams in underdense plasma.Plasma background resulted by 1D simultaneously calculation of the laser-target coupling is used to investigate the filamentation process and several reasonable results have been obtained.Nonlinear and relativistic modification of transport coefficients,and nonlinear hydrodynamics computation prove to be needed for the case of intense laser beams.

The diffusion coefficients and interaction viscosities for Ar in N₂ are calculated based on physical dynamics theory, by employing quantum mechanical IOSA approximation and Ar-N₂ effective interaction potential. The calculation results are in good agreement with the experimental values, and better than those of Gianturco and others. T he calculation interaction viscosity values are very close to experimental ones in the caluculation range.

In the framework of finite element approach, a new calculation scheme is proposed through combining the conventional finite element approach and the basis set expansion method in quantum chemistry. Under the new scheme, total energies in the ground states for diatom molecules Li₂, LiH, and BH are calculated on parallel computer DAWN1000. In comparison with the traditional finite element approach,it is found that the more accurate results can been obtained through the new scheme with same computational efforts.

The canonic equations of an intense field model are constructed by substituting symmetric difference quotient for partial derivative in Hamiltonian and the numerical solution is found by using square-preserving and symplectic scheme.The result is consistent with theoretical analyses,and shows that the method is successful.It probably overcomes the difficulties that often encountered to intense field model by using conventional difference methods.

To compute the multi-particle productions in the Meson-Meson.Meson-Hadron, Hadron--Hadron and electron-positron reactions, a unified Monte Carlo simulation was made. The routine has been examined practically in the reactions mentioned above with DPM model in the intervals of 20-200 Gev. It is easy to reform it for another model.

A REDUCE program based on the two step approach is developed to solve the Schrodinger equation with polynomial potential. The program is very simple and suitable to be carried out on microcomputer, The transformation for Hamiltonian, the obtaining of matrix elements, the derivation of the equation which must be satisfied by the parameter in variation method and even the Fortran program for numerical calculation are all completed or generated by computer automatically, As an example the pure quartic oscillator is calcuated and the results are compared with those available in literatures, Agreements are excellent.

On the basis of a large amount of experimental data and results from numerical Calculation, an exponetial function for the relation between the ray tube cross-section and Shock wave mach number is presented in this paper. With Sucn a faction, the Calculation of the ground Surface Peak Overpressure distribution in the mach reflection region is greatly Simplfied, and the results of the Calculation agree fairly Well With experimental data. This method will be rery useful in the design of hardened protection Structuses.

An iterative procedure for getting the numerical solution of Schrod-inger equation on stationary bond states is introduced. The theoretical foundation, the practical steps and the details of the method are presented. An example calculated is added at end. Comparing with other methods, this one requires less storage, less running time but posesses higher, accuracy. It can be run on personal computer or microcomputer with 256K internal memory and 16 bit word lenght such as IBM/PC, MC68000/83/20, PDP11/23 etc.