System in package (SiP) is mainstream technology in the design of electronics system. Numerical simulation plays an important role in SiP applications. However, due to the specific complexity of SiP applications, existing algorithms for linear systems arising from time-harmonic Maxwell's equations are faced with great challenges, which become a bottleneck restricting efficiency of large-scale numerical simulations. In this paper, we review algorithms for time-harmonic Maxwell's equations in SiP applications. Based on capability assessment of existing algorithms for realistic SiP models, we propose a preconditioning strategy, and show its feasibility and efficiency. Furthermore, we analyze impact of such applications on performance behavior of current algorithms and the challenges we faced with.

We proposed a modified Green element method based on edge-based element concept. Unknowns and matrix dimension are reduced. Then, a local grid refinement method is proposed based on the improved Green element, which ensures precision in early flow regimes for fractured horizontal wells with complex fractured networks. Solution of a degradation model solved with the method is compared with solutions obtained with a semianalytical model and numerical simulation. It verifies accuracy and efficiency of the improved Green element method based on local grid refinement. Finally, effects of model parameters on transient behavior is analyzed. It shows that the Green element method is a high precision dynamic simulation method, which improves computational efficiency of dynamic simulation of fractured horizontal well by setting node on edge of grid. In addition, local grid refinement method applied to the modified Green element method is based on superposition principle, in which interpolation approximation is not needed. The method is of high accuracy. Under same grid systems, local grid refinement based on improved Green element method is better than using finite difference. On the other hand, conductivity of complex fracture, permeability and size of stimulated reservoir volume have great influence on transient behavior of fractured horizontal wells. These effects should be taken into consideration and interpreted in transient behavior analysis.

To avoid computation of multi-frequencies pressure in solution of frequency method, a dual reciprocity method (DRM) is applied to convert domain integral, which is related to derivatives of pressure over time, in boundary integral equation into boundary integral. Firstly, field points are collocated at all boundary nodes. With boundary condition, boundary quantities at boundary nodes can be represented linearly by quantities at domain nodes, which are applied for radial basis function interpolation and are arbitrarily distributed in DRM. Secondly, field points are further collocated at domain nodes. A system of ordinary differential equations (ODEs) of the second order is obtained. Thirdly, variation rate of pressure was introduced as an unknown quantity to reduce order of resulted ODEs. Finally, a precise integration method is adopted to solved the first ordered ODEs. Numerical examples demonstrated validity and stability of the method.

We propose a novel floating memristor chaotic circuit with serial connection between a charge-controlled memristor and an inductor.Basic dynamic properties of the system are investigated with conventional dynamic analysis method. It shows that the system produces a pair of "heart" type attractors about origin symmetry. Simulation results indicate that strange attractors like bow tie type are observed as voltage and electricity signal in observing chaotic attractors are generalized to power and energy signal. Hopf bifurcation behavior is analyzed and verified by numerical simulation. It shows that the system can produce two bifurcation behaviors by adjusting parameters. They are Hopf bifurcation and anti-period doubling bifurcation. Remarkable feature of the citcuit is that it adopts a floating memristor, and with different initial state it generates nonlinear phenomena including coexisting chaotic and coexisting chaotic-periodic attractors.

Time consuming analysis of grid search in JMCT, a continuous-energy Monte Carlo neutron transport code, is done. Region projecting acceleration is performed on grids of main energy, fission yield neutron and unresolved resonance data. Quality of project grid is shown with arithmetic progression and geometric progression. At last, optimized grid search is test on several k_∞ problems. The models are fuel lattice cell at begin of cycle and middle of cycle. It shows that optimized grid search improves speed of JMCT remarkably in middle of cycle problem.

Effects of gate underlap on electronic properties of conventional single-material-gate CNTFET (C-CNTFET) and heteromaterial-gate CNTFET (HMG-CNTFET) are investigated theoretically in a quantum kinetic model.The model is based on twodimensional non-equilibrium Green's functions (NEGF) solved self-consistently with Poisson's equations.It shows that intrinsic cutoff frequency of C-CNTFETs reaches a few THz.In addition,a comparison study was performed about C-and HMG-CNTFETs.Calculated results show that,C-CNTFETs with longer underlap have better switching speed but less on/off current ratios.For HMG-CNTFET,gate underlap improves sub-threshold performance and switching delay times,and decreases output conductance significantly.

Ballistic performance of single-material-gate and triple-material-gate graphene nanoribbon field effect transistors (GNRFETs) with various doping profiles is studied in a quantum kinetic model based on non-equilibrium Green's functions (NEGF) solved self-consistently with Poisson's equations.It shows that triple-material-gate GNRFET with linear doping (TL-GNRFET) exhibits significant advantage in reducing SCEs and DIBL effects,as well as achieving better subthreshold slope and better on/off current ratio.In addition,asymmetric gate underlap is also discussed.It is revealed that as top and bottom gates are both shifted towards source on/off state current performance is improved.

To overcome disadvantage of Levenberg-Marquardt(LM) algorithm,which can not keep geometric correlation of defect parameters,a modified method with fast convergence is presented.In a heat transfer model for a two-dimensional test piece with a subsurface rectangular defect,numerical experiments are conducted to verify effectiveness of the modified LM algorithm.Factors of initial guess and infrared temperature measurement error are considered.Conclusions are drown as follows: The modified LM algorithm keeps geometric correlation of defect parameters well and influence of initial guess is negligible;Identification accuracy of different defect parameters differs due to infrared temperature measurement errors.

Band structures,density of states and optical properties of pure,N-doped,S-doped and N-S co-doped rutile TiO₂ are studied with plane-wave ultrasoft pseudopotential method based on first-principles density functional theory.It shows that the gap of N-doped TiO₂ reduces to 1.43 eV,and there is an impurity band on top of the valence band.S-doped TiO₂ results in Fermi level moves to the conduction band,and the gap decreases to 0.32 eV.N-S co-doped TiO₂ results in two impurity bands located in the band gap.One impurity band is apart from the bottom of conduction band by 0.35 eV,the other is apart from the top of valence band by 0.85 eV,which is mainly formed by N-atom 2p orbital and S-atom 3p orbital.It causes a red-shift in absorption wavelength,and gives rise to extremely great absorption coefficient in visible-light region,which shows a strong photocatalytic activity.

Possible geometrical structures and relative stability of alkali metal nitride compound (K₃N)_n(n=1,…,5) clusters are studied with a hybrid density functional theory(B3LYP)with 6-311*-basis sets.Bond property,charge distribution,vibrational property and stabilities of the stablest isomers of (K₃N)_n(n=1,…,5)clusters are analyzed.It shows that the stablest isomers of (K₃N)_n(n=1,…,5) clusters are changing gradually from planar structure to three-dimensional structure.(K₃N)₄ and (K₃N)₅ are layered structure similar to crystal.Coordination rlnmhers of N atoms are usually 5 or 6.Natural charges of N atom and K atom are about-1.608e and+0.550e.respectively.A strong ionic bond exists between K atom and N atom.Dynamic stability of (K₃N)₄ cluster is higher than that of other clusters.

Based on CFD and heat transfer theory,a model of the nozzle in a plasma sprayer is developed.The flow and heat transfer of the cooling water are simulated with coupling.It is found that the wall near heat source and the gasket in the back of the nozzle are the most deficient cooling parts.Then different parameters of cooling water are analyzed,and the optimum velocity and direction of cooling water are found.

A hybrid method which collects the advantages of both iterative method and optimization method is introduced.This method is adopted in designing the diffractive optical device used in target illumination of ICF system,and better design results are obtained.The diffractive efficiency is up to 97.5%,the mean square error is 0.026 and the focal plane profile has an perfectly flat top.

By direct numerical simulation of Navier-Stokes equations the reconnection of two anti-parallel vortex tubes has been studied in terms of the pseudo-spectral method.The influence of different disturbances for initial vortex tubes on reconnection is observed. "Curved vortex belts" are found,and the physical mechanism for the amplification of background vorticity is also investigated.

Energy cost that bounce shock wave on its propagation through outer iron core must pay for dissociation of heavy nuclei is considered as an adjusted parameter. Based on this, the method of artificial prompt explosion model is proposed to research effects of iron core mass of SN Ⅱ on prompt explosion. Range of iron core mass of presuperonva models which can produce prompt explosion is estimatd. The preliminary investsgation shows that iron core mass of presupermova models of WW(1993) is too large to produce prompt explosion.

In terms of the three basic forms and the eight-various cases of artifical viscosity for the same progenitor madel-"WW88", the numerical simulation has been performed for the whole processes of collapse and explosion of supernovae, and the artifical viscosity influences on compulation of supernovae explosion are investigated through analysis of numerical experimental results.

FIDAP software Package is the first genaral software for simulation of incompressible viscous flow by using the finite element method.There are 120000 Fortran senterces in it,A brief introduction is given for its theorical models and algorithm functions,the treatment tricks for implanting FIDAP into a PC computer are given also.

As development of equation of state supernova core collapse will reach high density above nuclear density, the general relativitic effects are most important. The density of core collapse is higher in the general relativitic (GR) core than in the Newtonian (NR) case. The higher the density reached the deeper the shock wave digs into the gravitational well and hance the shock wave will be launched with a larger energy. The formation mass point for shock wave is further from core boundary in GR case than in NR case. The disintegration of iron on the way of shock wave to propagate out will be increased. Once the core centre density reaches it maximum, the imner parts of the cross mass point for two velocity distribution in GR case and in NR case (which slight greater than the sonic mass point in NR case). The velocity value in GR case is greater than one in NR case, but the outer parts of the velocity cross point the velocity value is smaller in GR case than one in NR case, when the shock wave reaches the outer parts where the density and velocity is lower in GR case than in NR, which makes it easier for the shock wave to propagate.

There is a restriction to the solution of the epuation of θ (the ratio between the density of the nuclear matter and the density of the nuclear matter at saturation) in the J. Cooperstein's equ ation of state. Starting from the expression of nucleons pressure near the saturation density at zero temperature, a expression of nucleon internal energy is obtained by us. the re striction mentioned above in J. Cooperstein's equation of state has been removed by substituting the expression of nucleon internal energy obtained by us for the expression used by J.Cooperstein. Me present results of hydrodynamic cal culation of core collapse and subsequent shock wave formation and propagation using the equation of state of us and using the 15M_⊙ initial model of WZW.The relesse of gravitational binding energy of core can be increased by proper softening the equation of state above the sqturation density of nucl ear matter.

A numerical solution of the initial value problem of regularized long-Wave equation (RLW Equation) was made by a Petrov-Galerkin finite elemnte method. The numerical resuts are consistent with the exact solution of propagation of single solitary wave, due to second-order and fourth-order accuracy for the time and spacing variable respectively. For the collision of two solitons, there is a slight oscilatoiy wave trail after their collision, so it doesn't have the quality of soliton strictly.

The analysis contrasting the feature of BBAL's equation of state (eos) with the feature of equation of state EOS(1) is presental. At range of low density of matter(ρ<10¹¹g cm^-3), the chemical potential of neutron and proton of BBAL's(eos) differs notablly from EOS(1).This results in that two equations of state mentioned above differ greatly in abundance of free neutron and free proton at low density.The difference between the balance expression for finite temperature correct (thermal effect) and for "zero temperature appoximation" has been investegated.The effect for finite temperature correct on the chemical potential of neutron and proton is small, but the effect on the mass number and charge number of the average heavy nucleus is marked.Having taken 15M_① model of weaver et al.as initial model,the collapse process of stellar core between the onset of collapse and core bounce has been calculated by using both the (eos) of BBAL and EOS(1).The results show that the difference betwean the calculational results of stellar collapse (at ρ_c~10¹⁴g cm^-3) with EOS (1) and with BBAL's (eos)is not so large, in spite of that BBAL's (eos) differ greatly from EOS(l) in abundance of free nucleon at low density. The difference between calculational result (at ρ_c~10¹⁴g cm^-3) of stellar collapse resulting from finite temperature correct and "zero temperature approximation" for both BBAL's (eos) and EOS(l) is more notable.