A hybrid parallel technology on message-passing interface (MPI) platform in axisymmetric cylindrical coordinates, combining parallel finite-difference time-domain(FDTD) method and near- to near-field transformation with Kirchhoff surface integral representation(KSIR), is presented for fast computing radiation fields of large vertically polarized EMP radiating-wave simulator. Transient response of electric field at testing points as the simulator excited by double exponential function is given. Calculation results are consistent with those by software and parallel FDTD method, while the computing time is saved about 80%. Speedup factor and efficiency of hybrid parallel technology are measured. The hybrid method has advantages of saving computing time and CPU storage especially for larger simulators. Radiation far-fields besides near-fields in time-domain of simulator can be obtained as well.

Parallel FDTD method is used in time-domain simulation of large horizontally polarized electromagnetic pulse(EMP) bounded wave simulator with discrete resistors, in which the earth is set as lossy media truncated by uniaxial perfectly matched layer (UPML) absorbing boundary. Simulator's radiation fields in time-domain with different earth conductivity, earth relative dielectric constant and cone's radius are given. And electric field with a 10 m-length cylinder effector in simulator is also presented. The number of total cells in parallelized FDTD computing is about 1.8 billion. It shows that reflection from the earth becomes great as the earth's dielectric constant or conductivity increasing; The cone's radius has great effect on peak value of E-field in simulator, and fields at same horizontal plane becomes uneven as the radius becomes large; Reflection and loss from the earth have great effect on peak value and pulse width of E-fields in a horizontal plan above the ground 1 m, while they have little effect on E-fields in a horizontal plan above the ground 5 m. Energy entering into effector through a square hole is great as the hole is at the same side of simulator fields' leak.

γ-Re_θt crossflow extension model was applied to conduct numerical simulation of crossflow transition on NLF(2)-0415 infinite swept wing. Effects of Reynolds number, surface roughness, sweep angle on crossflow instability of 3-D boundary layer were studied mainly. It found that γ-Re_θt crossflow extension model predicts 3-D boundary layer transition correctly and calculation results agree well with experiment data. It shows that with increase of Reynolds number or surface roughness, location of crossflow transition onset moves forward. However, with increase of sweep angle, location of crossflow transition onset moves forward firstly and then moves back.

An entropy stable scheme based on moving meshes is proposed for hyperbolic conservation laws. The method employs equidistribution principle to redistribute mesh points. Numerical solutions on new meshes are updated by using a conservative-interpolation formula. Entropy stable fluxes and third order strong stability-preserving Runge-Kutta time evolution method are employed to obtain numerical solutions at next time level. Several test problems are presented to demonstrate that the method not only improves resolution in discontinuous areas, but also reduces possible spurious oscillations.

A high resolution scheme is presented for shallow water equations. The scheme is based on entropy stable numerical flux with high order weighted essentially non-oscillatory (WENO) reconstruction at cell interfaces. A strong stability-preserving Runge-Kutta method is employed to advance in time. Several benchmark numerical examples demonstrate that the scheme is accurate and non-oscillatory.

Newton-Raphson iteration is used in semi-classical model of carbon nanotube field effect transistors(CNTFET) based on ballistic transport for self-consistent potential.In each iteration re-integration of state density is requred.Support vector regression (SVR) is applied for the relationship of self-consistent potential and carrier density.Numerical integrations are avoided.And gradient descent algorithm (GDA) is used to get self-consistent potential.It shows that compared with Newton-Raphson iteration,the proposed method reduces computation effectively while maintains high accuracy.It lays theoretical foundation for designing and applying CNTFET devices in large scale integrated circuits.

Based on the cluster multiple labeling technique,a novel cluster detection algorithm is presented as an analysis subroutine in two-and three-dimensional molecular dynamic simulations of ejecta that take place as a planar shock wave encounters a free metal surface.The algorithm is described,tested,and used to detect cluster distribution of ejecta from copper and aluminum under a shock loading.The information obtained about the size,distribution,evolution of the cluster is helpful in the understanding of ejection.

Second-order and fourth-order methods for approximate solutions of viscous Hamilton-Jacobi equations are developed on the basis of the weighted essentially non-oscillator (WENO) scheme.By modifying the numerical flux,constructing the second-order approximation based on nonlinear limiter and fourth-order approximation based on Taylor expansion for viscosity term, the one- and two-dimensional viscous Hamilton-Jacobi equations are solved successfully. Numerical tests demonstrate the desired high-resolution,robustness and non-oscillatory behaviors of the schemes.