基于季节气候可预报分量的相似误差订正方法和数值实验

基于由历史相似信息对模式预报误差进行预报的思路,发展了一种针对季节气候可预报分量的相似误差订正新方法（FACEPC）,目的是识别模式预报结果中对初值相对不够敏感的可预报分量,将其作为对象进行历史相似选取和误差订正.该方法被应用到国家气候中心业务季节预报模式实验中,对不同区域给出有针对性的相似选取指标和预报方案.25年的交叉检验结果表明,夏季降水和环流的预报技巧评分相对于系统误差订正有明显提高,在发生中等及以上强度ENSO事件年和可预报分量贡献较大地区的预报技巧提高更为显著.特别是中国区域降水和关键区环流的 关键词： 短期气候预测 相似误差订正 可预报分量     

基于多个参考态更新的动力相似预报方法及应用

为了更有效地利用历史资料中的相似性信息提高数值模式预报水平,提出了一种新的动力相似预报方法——多参考态更新(MRSU) 法.该方法基于“更新”观点,通过引入相似更新周期(PAU)的概念,将整个预报时段按PAU分成若干小的子时段,在预报进行到PAU时重新选取多个参考态,并采用超平面近似法将相似-动力模式产生的多个预报估计成最佳预报向量,这样就形成了“选取-估计”的循环,不断重复这一过程直到完成整个时段的预报.进一步将简化的MRSU方法应用于T63全球谱模式.月预报试验结果表明,与控制试验相比,MRSU法对逐 关键词： 相似-动力模式 参考态 相似更新周期 月预报     

An approach to estimating and extrapolating model error based on inverse problem methods:towards accurate numerical weather prediction

Model error is one of the key factors restricting the accuracy of numerical weather prediction(NWP). Considering the continuous evolution of the atmosphere, the observed data(ignoring the measurement error) can be viewed as a series of solutions of an accurate model governing the actual atmosphere. Model error is represented as an unknown term in the accurate model, thus NWP can be considered as an inverse problem to uncover the unknown error term. The inverse problem models can absorb long periods of observed data to generate model error correction procedures. They thus resolve the deficiency and faultiness of the NWP schemes employing only the initial-time data. In this study we construct two inverse problem models to estimate and extrapolate the time-varying and spatial-varying model errors in both the historical and forecast periods by using recent observations and analogue phenomena of the atmosphere. Numerical experiment on Burgers’ equation has illustrated the substantial forecast improvement using inverse problem algorithms. The proposed inverse problem methods of suppressing NWP errors will be useful in future high accuracy applications of NWP.     

一种基于误差补偿的多元混沌时间序列混合预测模型

针对多元混沌时间序列的预测问题, 考虑到单纯改进储备池算法无法明显地提高预测精度, 提出一种基于误差补偿的时间序列混合预测模型. 实际观测的数据既包含线性特征又包含非线性特征. 首先利用自回归移动平均模型预测线性特征, 使得残差数据仅含非线性特征; 然后, 建立正则化回声状态网络模型预测; 最后, 将非线性部分的预测值与线性部分的预测值相加, 以实现高精度的多元混沌时间序列预测. 基于Lorenz和太阳黑子-黄河径流量时间序列的仿真实验验证了本文所提模型的有效性. 关键词： 回声状态网络 混沌 多元时间序列预测 误差补偿     

动态最优多因子组合的华北汛期降水模式误差估计及预报

利用国家气候中心季节预报1983—2009年27年模式预报结果,结合74项环流指数及美国国家海洋局和大气管理局提供的40个气候指数和美国气候预报中心实际降水分析资料,采用资料诊断分析和数值模拟实验相结合的方法,通过多因子的历史相似信息提取预报相似年,获得预报场的误差订正项.在这一订正思路的基础上,考虑前期关键影响因子的选取、多因子组合的优化配置,构建适用于不同预报年的区域动力-统计模式预报误差订正方案.以华北为例,探索多因子最优组合的多元客观相似判据,发展基于多因子动力-统计模式预报误差的动态订正新技术,改善华北夏季降水预报效果,提高预报技巧.通过2005—2009年独立样本回报结果表明,动态最优多因子组合相似订正方法距平相关系数评分相对于系统订正方法有着显著的提高,该订正方案对华北地区的夏季降水预测有着很好的业务前景,具有重要的应用价值,即将投入业务运行. 关键词： 关键因子集 模式误差估计 汛期降水 预测     

Application of retrospective time integration scheme to the prediction of torrential rain

The retrospective time integration scheme presented on the principle of the self-memory of the atmosphere is applied to the mesoscale grid model MMS,constructing a mesoscale self-memorial model SMMS, and then the shortrange prediction experiments of torrential rain are performed in this paper. Results show that in comparison with MM5 the prediction accuracy of SMM5 is obviously improved due to its utilization of multiple time level past observations,and the precipitation area and intensity predicted by SMM5 are closer to observational fields than those by MM5.     

Dirac方程的数值解法及其在原子总能量计算中的应用

中心势近似下径向Dirac方程的求解是相对论性原子(离子)结构计算的基础.本文通过相对论性方程中径向波函数大分量与非相对论方程径向波函数的类比,提出了径向Dirac方程的一种数值解法.为了验证数值解法的精度和可靠性,首先将数值结果与类氢势作用下的解析解进行比较.然后,将这种算法扩展到基于解析势的相对论性原子结构计算中,并将计算出的总能量与实验结果和其他方法得到的结果进行对比.     

张量重正化群方法及其应用

张量重正化群方法是近年来发展起来的一种新的数值计算方法,它将经典配分函数和量子波函数的张量网络表示与重正化群方法相结合,在强关联系统的数值研究中,发挥着越来越重要的作用。文章以经典统计模型和量子格点模型为例,简要介绍了张量重正化群的一些基础知识和研究给定物理模型的一般性思路,并对张量重正化群未来可能的发展方向和亟待解决的问题进行了讨论。     

Reformulation of the path probability method and its application to crystal growth models

The microscopic master equation of a system is derived within the framework of the path probability method (PPM). Then, by extending Morita's method in equilibrium statistical mechanics, the path probability function constructed microscopically can be systematically decomposed to result in the conventional path probability function of cluster approximation when correlations larger than the chosen basic cluster are neglected. In order to critically compare the master equation method with the PPM, the triangle approximation is treated by both methods for crystal growth models. It is found that the PPM gives physically satisfactory kinetic equations, while the master equation (supplemented with a cluster probability in the superposition approximation) does not. The triangle PPM calculation considerably improves the result of the pair approximation for crystal growth velocity in the solid-on-solid model, and compares well with Monte Carlo results.     

一种二维自适应网格构造方法及其实现

讨论了一种二维自适应网格的构造方法,在二维插值法的基础上做出了自适应网格。最后利用自适应网格技术对一个二维流体力学问题的实例进行了计算。     

Level Set interface treatment and its application in Euler method

Level Set interface treatment method is introduced into Euler method,which is employed for interface treatment method for multi-materials. Combined with the ghost fluid method,the moving interface is tracked. Fifth-order WENO spatial discretization and third-order TVD Runge-Kutta time discretization methods are used. Shock-wave action on bubble,implosion and velocity field Shock effect bubbles; implosion and velocity field are simulated by means of LS-MMIC3D programmed by C++. Nu-merical results show that t...     

Artifacts introduced by zero order phase correction in proton NMR spectroscopy and a method of elimination by phase filtering

In in-vivo applications of proton NMR spectroscopic imaging, an oscillatory "ringing" artifact has been observed in some of the spectra. The source of this artifact was found to be the presence of a harmonic "beating" effect in the amplitude of the water reference free induction decay (FID) which was used for zero order phase correction for B0 inhomogeneity and eddy current compensation. The source of the beats was found to be the presence of distinct populations of spins resonating at slightly different frequencies. When the common method of zero order phase correction was implemented using such an FID, the resulting phase-corrected, water-suppressed spectra displayed ringing. Examination of the unwrapped phase correction angle revealed unexpected jumps in phase at points in time corresponding to nodes in the amplitude of the FID. Low-pass filtering of the phase correction angle of the reference FID was found to smooth out these unanticipated phase jumps. When used as a reference for phase correction, the filtered phase information gave a phase-corrected, water-suppressed spectrum free from ringing.     

具有选择与遗忘机制的极端学习机在时间序列预测中的应用

针对训练样本贯序输入时的极端学习机 (ELM)训练问题,提出一种具有选择与遗忘机制的极端学习机 (SF-ELM),并研究了其在混沌时间序列预测中的应用. SF-ELM以逐次增加新训练样本的方式实现在线训练,通过引入遗忘因子以减弱旧训练样本的影响,同时以泛化能力为判断依据,对其输出权值进行选择性递推更新. 混沌时间序列在线预测实例表明,SF-ELM是一种有效的ELM在线训练模式. 相比于在线贯序极端学习机,SF-ELM具有更快的在线训练速度和更高的在线预测精度,因此更适于混沌时间序列在线预测. 关键词： 混沌时间序列 时间序列预测 神经网络 极端学习机     

Nonlinear scaling and its application to ferromagnetic systems

Making use of renormalization-group ideas, a scaling equation of state applicable to ferromagnetic systems and involving the nonlinear scaling variables =ε/t and =h/t, instead of the usual linear scaling variables ε=(T-T_c)/Tc = t-1 and h=H (ordering field), has been derived. The magnetic equation of state so obtained is then generalized to take into account the effect of nonlinear relevant and irrelevant scaling fields. To facilitate a comparison with experiments, the analytic (non-analytic) corrections to the dominant singular behaviour of spontaneous magnetization (order parameter) M(T, 0), ‘zero-field’ susceptibility χ(T, 0), and specific heat in zero field that the nonlinear relevant (irrelevant) scaling fields give rise to are explicitly calculated up to third order in Due consideration is also given to the modifications in the Arrott-Noakes form of the scaling equation of state and the Kouvel-Fisher definition of the effective susceptibility exponent brought about by these scaling fields. A detailed analysis of the M(T, 0) and χ(T, 0) data for crystalline and amorphous ferromagnets in terms of the theoretical expressions derived in this work reveals that i) in conformity with the theoretical predictions, the “;non-analytic”; corrections to the singular behaviour dominate over the “;analytic”; ones for temperatures in the immediate vicinity of the critical point Tc, whereas reverse is the case for temperatures far away from Tc and (ii) the expression for χ(T, 0), based on the nonlinear scaling arguments, which includes the leading “;analytic”; correction, reproduces closely the observed variation of χ with T over a wide range of temperatures T_c ≤ T ≤ 1.5T_c (in some cases, up to 3T_c) for both ordered as well as quench-disordered ferromagnets.     

利用线化和校正法求非线性单摆运动的周期

应用线化和校正方法,研究了单摆的非线性振动,作出了周期比和相对误差随摆角的变化曲线.将所得近似解与精确解比较可知,该方法具有简单实用,精度高,相对误差低等优点,对于求解非线性振动问题具有一定的实用价值.     

Quantitative measurement of displacement and strain by the numerical moiré method

The numerical moié method with sensitivity as high as 0.03 nm has been presented.A quantitative displacement and strai,analysis program has been proposed by using this method.It is applied to an edge dislocation and a stacking fault in aluminum.The measured strain of edge dislocation is compared with theoretical prediction given by Peierls-Nabarro dislocation model.The displacement of stacking fault is also obtained.     

非协调数值流形方法的稳定性和收敛性分析

在流形元的基础上,提出了非协调数值流形方法,非协调数值流形方法的优点是在不增加广义节点自由度的前提下,大大提高数值流形方法的计算精度和计算效率.利用内部自由度静力凝聚处理,推导了消除内参后的单元应变矩阵和单元刚度矩阵.在Hilbert空间内,从最小势能原理出发对非协调数值流形方法的稳定性和收敛性进行了分析和讨论,得到了保证非协调流形元解唯一存在和收敛的基本条件,完善了非协调数值流形方法的理论基础.数值试验表明,新单元构造过程简单,有较高的精度,从而证明了本方法的可行性. 关键词： 数值流形方法 非协调元 稳定性分析 收敛性分析     

探测系统能量分辨在中子能谱数值模拟中的应用

针对中子能谱实验的数值模拟研究中涉及到的探测系统能量分辨问题,基于系统探测伽玛和中子的对应关系以及对伽玛射线的能量分辨率,介绍了一种简便实用的数值模拟方案,利用探测系统对中子、伽玛的光响应对应关系以及其对伽玛的能量分辨率,得到系统对中子的能量分辨率,再进一步把它应用到中子能谱的数值模拟中,并对数值计算结果与相关实验数据进行比较分析。结果表明,该方法可以很好地应用于中子能谱的数值模拟研究。     

半解析对偶棱边元及其在波导不连续性问题中的应用

给出电磁波导的对偶变量变分原理,并采用对偶棱边元对波导的横截面进行半解析离散. 将波导中沿纵向均匀的区段视为子结构,运用基于Riccati方程的精细积分算法求出其出口刚度阵,然后与不均匀区段的常规有限元网格拼装即可对波导不连续性问题进行求解. 半解析对偶棱边元的采用可以在最大程度上对有限元网格进行缩减,并且能够在不增加计算量的前提下任意增加子结构的长度,从而可以将截断求解区域的人工边界设置在距离不均匀区段充分远的地方,极大地减少了近似边界条件所带来的误差. 数值算例证明这种方法具有很高的精度与效率. 关键词： 波导的不连续性 半解析辛分析 对偶棱边元 精细积分     

Multiple temperature model for the information preservation method and its application to nonequilibrium gas flows

The information preservation (IP) method has been successfully applied to various nonequilibrium gas flows. Comparing with the direct simulation Monte Carlo (DSMC) method, the IP method dramatically reduces the statistical scatter by preserving collective information of simulation molecules. In this paper, a multiple temperature model is proposed to extend the IP method to strongly translational nonequilibrium gas flows. The governing equations for the IP quantities have been derived from the Boltzmann equation based on an assumption that each simulation molecule represents a Gaussian distribution function with a second-order temperature tensor. According to the governing equations, the implementation of IP method is divided into three steps: molecular movement, molecular collision, and update step. With a reasonable multiple temperature collision model and the flux splitting method in the update step, the transport of IP quantities can be accurately modeled. We apply the IP method with the multiple temperature model to shear-driven Couette flow, external force-driven Poiseuille flow and thermal creep flow, respectively. In the former two cases, the separation of different temperature components is clearly observed in the transition regime, and the velocity, temperature and pressure distributions are also well captured. The thermal creep flow, resulting from the presence of temperature gradients along boundary walls, is properly simulated. All of the IP results compare well with the corresponding DSMC results, whereas the IP method uses much smaller sampling sizes than the DSMC method. This paper shows that the IP method with the multiple temperature model is an accurate and efficient tool to simulate strongly translational nonequilibrium gas flows.