Archive

25 January 2019, Volume 36 Issue 1 Previous Issue    Next Issue

---

For Selected:

Download Citations EndNote Ris BibTeX

Toggle Thumbnails

Select

Parallel Numerical Simulations of Thermal Convection in the Earth's Outer Core Based on Domain Decomposition and Multigrid YIN Liang, YANG Chao, MA Shizhuang 2019, 36(1): 1-14.  DOI: 10.19596/j.cnki.1001-246x.7771 Abstract ( )   HTML ( )   PDF (14123KB) ( )   Numerical simulations of thermal convection in a rotating spherical shell play an important role in dynamo models. In this paper, we present a parallel numerical model with high performance for the Earth's outer core convection based on a homegrown supercomputer. An approximate factorization fractional method combined with Crank-Nicolson scheme and second-order Adams-Bashford formula is employed for temporal integration. Spatial terms are discretized by a second-order finite volume scheme based on cubed-sphere grid. Two resultant large sparse linear algebraic equations are solved by preconditioned Krylov subspace iterative method. To accelerate convergence rate and improve parallel performance, linear solver is preconditioned with multilevel restricted additive Schwarz preconditioner based on domain decomposition and multigrid. The preconditioner reduces compute time and improve parallel performance of the solvers, which scales well to over ten thousand processor cores. Numerical results are in good agreement with reference solutions of the benchmark Case 0.

Select

A Sound Speed-Based Rezoning Strategy for ALE Methods DAI Zihuan 2019, 36(1): 15-24.  DOI: 10.19596/j.cnki.1001-246x.7782 Abstract ( )   HTML ( )   PDF (10649KB) ( )   A sound speed distribution-based rezoning strategy is introduced for optimizing deformed Lagrangian grid in arbitrary Lagrangian Eulerian fluid computational methods which includes sound speed distribution-based grid relaxation functional, derivation, discretization and solution of modified gradient flow equations, conditions for starting/terminating grid optimizing procession based on subzonal time step, and algorithm of ALE method based on the new rezoning strategy. Advantage of the method is that it not only optimizes geometry form of grid but also improves time step of Lagrangian hydrodynamic computation. Numerical examples which include ALE calculations of Rayleigh-Taylor instability problem are given to show effectiveness of the strategy.

Select

Sufficient Conditions and Numerical Validations for Smooth Functions in Particle Approximation of SPH Scheme XU Chengjun, XU Shengli 2019, 36(1): 25-38.  DOI: 10.19596/j.cnki.1001-246x.7807 Abstract ( )   HTML ( )   PDF (2945KB) ( )   To obtain second order accuracy in SPH scheme, sufficient conditions for smooth functions are discussed and formula are presented. Uniqueness is analyzed for solution of the second order particle approximation. SPH approximate solutions are numerically validated to arrive at nearly second accuracy in the case of periodic boundary conditions. While particle number increases, L1-error decreases and L1-order approaches to the second order particle approximation of SPH scheme.

Select

Error Estimation of k-α Iteration Method for α Eigenvalue Problems WANG Yi, PAN Liujun, WANG Ruihong 2019, 36(1): 39-46.  DOI: 10.19596/j.cnki.1001-246x.7794 Abstract ( )   HTML ( )   PDF (3284KB) ( )   We analyze k-α iteration method for α eigenvalue problems, and point out that parameter in k-α iteration method affects error estimation of alpha eigenvalues, while error estimation has no relationship with the parameter. If the parameter isn't given properly, the estimated error deviates the real error estimation. Estimated error given by MCNP4C may deviates from real error estimation due to improper parameter. A parameter selecting method called average neutron fission lifetime method is proposed to make estimated error and real error estimation equal strictly in mathematics. It was tested with numerical experiments.

Select

Transient Pressure for a Fractured Well with Asymmetric Fracture JI Anzhao, WANG Yufeng, ZHANG Guangsheng 2019, 36(1): 47-52.  DOI: 10.19596/j.cnki.1001-246x.7781 Abstract ( )   HTML ( )   PDF (3782KB) ( )   Asymmetric vertical cracks in process of pressure was studied. Based on principle of seepage mechanics, well-test mathematical model, which is asymmetric finite-conductivity vertical fracture with three different outer boundary conditions, was established according to point source function and Green function. Typical curves were obtained with Laplace integral transformation and Stehfest numerical inversion. It shows that typical curve is divided into four flow stages, stage of wellbore storage,1/4-slope bilinear flow, 1/2-slope linear flow and 0.5-value horizontal line radial flow. Asymmetric factors affect mainly end time of bilinear flow stage. The greater the asymmetry factor is, the sooner the bilinear flow stage ends is. The greater the conductivity capability of fracture is, the more unobvious characteristic of pressure derivative curve of bilinear flow stage are. The bigger asymmetric factor is,the more uneven rate distribution of asymmetric fracture is. The model provides theoretical basis for well test analysis of asymmetric finite-conductivity vertical fracture.

Select

Numerical Study of Response Performance of Vortex-Induced Vibration on a Rigid Cylinder Based on Finite Difference Method GAO Yun, ZOU Li, ZONG Zhi 2019, 36(1): 53-59.  DOI: 10.19596/j.cnki.1001-246x.7815 Abstract ( )   HTML ( )   PDF (6047KB) ( )   Numerical study based on a wake oscillator model is conducted to determine response performance of vortex-induced vibration (VIV) on a rigid cylinder. A coupled model of structural oscillator of a rigid cylinder and wake oscillator is established, and then the model is discretized and solved based on a standard central finite difference method of the second order. VIV response characteristics including dimensionless displacement, dimensionless lift, and frequency ratio and lock-in region varied with mass ratios and damping ratios are compared. It can be found that the numerical method simulates response performances of vortex-induced vibration on a rigid cylinder very well. As dimensionless mass ratio increases, onset of lock-in region occurs later, lock-out point occurs earlier, and so the lock-in region becomes smaller.

Select

LBM Simulation of Heat and Mass Double Diffusion, Fluid-Solid Conjugate Heat Transfer and Adsorption LUO Zhuqing, LOU Qin, XU Hongtao, YANG Mo 2019, 36(1): 60-68.  DOI: 10.19596/j.cnki.1001-246x.7790 Abstract ( )   HTML ( )   PDF (32120KB) ( )   We adopted lattice Boltzmann method to simulate double diffusion mixed convection, fluid-solid conjugate heat transfer and adsorption process in a lid-driven composite enclosure filled with homogeneous medium at pore scale. Influences of buoyancy ratio Br (-100 ≤ Br ≤ 100) and adsorption rate constant k1 (0.001 ≤ k1 ≤ 0.005) on characteristics of heat and mass transfer were compared at porosity ε=0.79, Prandtl number Pr=0.7, Grashof number Gr=104 and Lewis number Le=1.0. Streamlines, isotherms, isoconcentrations, adsorption capacity (CS), average Nusselt number Nuav and Sherwood number Shav of the left heated wall under different parameters were discussed in detail. It shows that Br affects adsorption by changing concentration distribution around medium in flow field; Increasing k1 improves significantly adsorption efficiency and adsorption capacity.

Select

Simultaneous Synthesis of Heat Exchanger Network by Random Walk Algorithm with Compulsive Evolution Based on Trilevel Protection Strategy LI Jian, CUI Guomin, CHEN Jiaxing, XIAO Yuan 2019, 36(1): 69-79.  DOI: 10.19596/j.cnki.1001-246x.7814 Abstract ( )   HTML ( )   PDF (5292KB) ( )   To avoid the problem of being disturbed by stochastic acceptance of imperfect solution for evolution process of individual optimal solution existing in optimization of heat exchanger network by random walk algorithm with compulsive evolution, an improved RWCE based on trilevel protection strategy is proposed. Individuals in population are divided into three levels. The lower-level is optimized by basic RWCE to protect global search ability of individuals. The middle-level reads historical optimal solution of the lower-level's individuals, and optimized by RWCE with fine tuning to protect evolution process of each individual's optimal solution from disruption. All individuals in the upper-level are initialized by solution of the best individuals in the middle-level, and optimized by RWCE with automatic fine search to ensure that the best individuals are fully searched. Finally, result of the upper-level is passed to corresponding individual at the lower-level. Two cases are optimized by using the algorithm,and results are better than those in literature. Evolution process of individual optimal solution is protected while accepting imperfect solution, therefore, both global search ability and local search ability are realized.

Select

Molecular Dynamics Study on Growth Mechanism of Pure Metals Solid-Liquid Interface During Solidification ZHANG Haiyan, YIN Xinchun 2019, 36(1): 80-88.  DOI: 10.19596/j.cnki.1001-246x.7779 Abstract ( )   HTML ( )   PDF (4352KB) ( )   Solidification of solid-liquid interface of simple metals (Ni and Al) is investigated with molecular dynamics simulation. Two metals show a common feature that growth rate approach to a maximum at a critical temperature (T*). Above T*, growth behaviors are similar as reflected by monotonically increased growth rate with increasing undercooling temperature. However, below T*, grow rate is nearly constant for Ni, while it rapidly decreases to zero for Al with increasing undercooling temperature. High-temperature Broughton-Gilmer-Jackson(BGJ) model and low-temperature Wilson-Frenkel (W-F) model are used to describe growth mechanism. At small undercooling and deep undercooling temperature zone, solid-liquid interface is formed by gradual change between collision and diffusion mechanism, and the dominated one varies in different temperature ranges.

Select

Analysis of Environmental Influencing on Acoustic Relaxation Frequency in Multi-component Excitable Gases ZHANG Kesheng, ZHANG Xiangqun, SHAO Fang, TANG Wenyong 2019, 36(1): 89-98.  DOI: 10.19596/j.cnki.1001-246x.7791 Abstract ( )   HTML ( )   PDF (4851KB) ( )   Analytic relations between acoustic relaxation frequency and external temperature and pressure in multi-component excitable gases are deduced. Theoretical and calculational results show that relaxation frequency is inversely proportional to relaxation time of primary relaxation process, proportional to vibration coupling heat capacity of primary relaxation process, and inversely proportional to external heat capacity. Increase of temperature is related to increase of heat transfer capacity and energy transfer rate of internal and external degrees of freedom, which leads to decrease of relaxation time. It results that relaxation frequency is proportional to ambient temperature. Increase of pressure increases molecular collision rate and causes relaxation time to decrease, which brings about relaxation frequency being linearly proportional to ambient pressure.

Select

Electronic Properties of Graphene Nanoribbons Doped with Rhombus Boron Nitride Segment MA Rui, ZHANG Hualin 2019, 36(1): 99-105.  DOI: 10.19596/j.cnki.1001-246x.7803 Abstract ( )   HTML ( )   PDF (10443KB) ( )   Electronic properties of graphene nanoribbons (GNRs) doped with rhombus boron nitride segments are investigated by using first-principles method based on density functional theory. It is shown that band gaps of AGNRs increase owing to doping, and band gaps of AGNRs vary slightly as doped at different positions. In nonmagnetic states, ZGNRs are metal whether or not it is doped. In ferromagnetic states, ZGNRs change from metal to semiconductor due to doping. In antiferromagnetic states, ZGNRs are semiconductor whether or not it is doped, but band gaps are changed owing to doping. Structures of doped AGNRs and ZGNRs are stable, and ground state of doped ZGNRs is antiferromagnetic. It suggests that rhombus boron nitride segments doping controls effectively properties of GNRs, which is positive on future graphene nanoelectronic devices.

Select

Spectral Constants and ab Initio Potential Energy Curves of POX(X=1,2) HAN Xiaoqin, XIAO Xiajie 2019, 36(1): 106-112.  DOI: 10.19596/j.cnki.1001-246x.7797 Abstract ( )   HTML ( )   PDF (4398KB) ( )   Many methods are used to optimize possible ground state structures of PO、PO2 with multi-basis set. It was found that DFT method is the most suitable for calculation of PO,PO2. For PO free radical, M-S potential energy functional constants are calculated with 6-311G(3df). For PO2 free radical, equilibrium structure, harmonic frequency, dissociation energy and force constant are calculated with 6-311+G(3df). Potential energy functions of PO2 are derived from many-body expansion theory. It was pointed out that there is a saddle point in reaction kinetics O+PO→OPO, and a stable PO2 molecule could be formed as O atom with energy surpassing 0.55 eV. A stable PO2 molecule could be formed through two equivalent channels over barrier.

Select

A Modified Surface Hopping Method for Schrödinger Equation with Conical Crossings LI Xinchun 2019, 36(1): 113-126.  DOI: 10.19596/j.cnki.1001-246x.7773 Abstract ( )   HTML ( )   PDF (3562KB) ( )   We present a modified surface hopping method for Schrödinger equation with conical crossings. At the crossing manifold, Landau-Zener formula gives probability that electrons hop to another energy level. Kammerer and Mehats raised a scheme with jump operators ensuring good energy conservation for transport of electrons in a graphene layer which is similar to the physics above. We employ these jump operators to improve a surface hopping method and obtain accurate numerical results.