基于改进量子粒子群的分布式并行计算框架设计

为了实现用户任务在大规模计算机集群上进行高效地处理,并克服现有并行计算框架通用性不强的缺点,提出了一种基于改进量子群算法和Map-Reduce模型的通用并行计算框架;首先,对经典的Map-Reduce分布式并行计算框架以及并行计算流程进行了具体描述;然后,基于改进的量子粒子群算法设计了改进的Map-Reduce模型,在Map阶段通过多种群并行搜索并计算所有粒子适应度,在Shuffle和Sort 阶段实现粒子的排序和种群的重新划分,然后在Reduce阶段更新控制系数和粒子位置,当最优解不变时,通过混沌扰动对其进行扰动;仿真实验表明同,文中设计的基于改进量子粒子群算法和Map-Reduce模型能高效地执行任务,较传统的Map-Reduce模型具有较少的执行时间,具有很强的可行性,是一种有效的通用并行计算模型。     

光子两自由度超并行量子计算与超纠缠态操控

光子系统在量子信息处理和传输过程中有非常重要的应用. 譬如, 利用光子与原子(或人工原子)之间的相互作用, 可以完成信息的安全传输、存储和快速的并行计算处理等任务. 光子系统具有多个自由度, 如极化、空间模式、轨道角动量、时间-能量、频率等自由度. 光子系统的多个自由度可以同时应用于量子信息处理过程. 超并行量子计算利用光子系统多个自由度的光量子态同时进行量子并行计算, 使量子计算具有更强的并行性, 且需要的量子资源少, 更能抵抗光子数损耗等噪声的影响. 多个自由度同时存在纠缠的光子系统量子态称为超纠缠态, 它能够提高量子通信的容量与安全性, 辅助完成一些重要的量子通信任务. 在本综述中, 我们简要介绍了光子系统两自由度量子态在量子信息中的一些新应用, 包括超并行量子计算、超纠缠态分析、超纠缠浓缩和纯化三个部分.     

声学Burton-Miller方程边界元法GPU并行计算

基于GPU,对声学Burton-Miller积分方程的边界元解法进行并行计算.提出并行计算格式和程序实现方法,以及Burton-Miller方程中各类奇异(包括强奇异、超奇异)积分的GPU计算和局部修正方法.典型算例结果表明,在特征频率处可获得正确的解,具有较高精度,可在普通个人计算机上快速完成自由度超过2×10⁵的声学边界元分析.为计算声学及相关工程领域的中、大规模声场分析问题提供一种快速、高效、简便的数值计算工具.     

U-RANS/PDF方法的并行计算研究

本文对非结构网格中U-RANS/PDF混合方法的并行计算进行了研究,通过对方柱绕流的数值模拟,讨论了并行程序的有效性和计算效率.非结构网格上U-RANS/PDF混合方法的并行计算,提高了PDF方法的计算效率和处理复杂几何结构的能力,使得PDF方法能够方便地应用于实际燃烧问题.     

高分辨率空间相机的工程分析

以大型结构设计软件SolidEdge建立了高分辨率空间相机的精确的几何模型, 在此基础上用大型有限元分析软件MSC.Patran建立了相机的有限元模型, 并利用MSC/Nastran对该相机整机进行了工程分析. 其计算结果与试验结果相差10%以内.     

行波管多信号非线性注波互作用的网络并行计算

 把网络并行计算技术应用于微波管的研究领域，提出了用于计算2.5维行波管多信号非线性注波互作用的基于TCP/IP协议的网络并行计算模型，编写了模拟行波管多信号非线性注波互作用过程的网络并行计算程序，其可在工作站和微机组成的计算机网络环境中实现并行计算。计算结果表明，该网络并行算法能够有效地减少行波管多信号非线性注波互作用的计算时间，提高工作效率。     

大规模声学边界元法的GPU并行计算

提出一种大规模声学边界元法的高效率、高精度GPU并行计算方法.基于Burton-Miller边界积分方程,推导适于GPU的并行计算格式并实现了传统边界元法的GPU加速算法.为提高原型算法的效率,研究GPU数据缓存优化方法.由于GPU的双精度浮点运算能力较低,为了降低数值误差,研究基于单精度浮点运算实现的doublesingle精度算法.数值算例表明,改进的算法实现了最高89.8%的GPU使用效率,且数值精度与直接使用双精度数相当,而计算时间仅为其1/28,显存消耗也仅为其一半.该方法可在普通PC机(8GB内存,NVIDIA Ge Force 660 Ti显卡)上快速完成自由度超过300万的大规模声学边界元分析,计算速度和内存消耗均优于快速边界元法.     

基于时域多分辨率的非球形气溶胶散射并行计算模型设计及验证EI北大核心CSCD

为解决串行时域多分辨率(MRTD)散射模型运行时间长和内存消耗大的问题,基于消息传递接口(MPI)技术设计了一种非球形气溶胶散射并行计算模型。介绍了MRTD散射模型的基本框架和2种并行数据通信方案,并基于MPI重复非阻塞通信技术实现了MRTD散射模型的并行化设计;搭建了网络并行计算平台,实现了模型的并行化计算。将MRTD散射模型与Mie散射模型、T矩阵法进行了对比,验证了并行MRTD散射模型的计算准确性。结果表明,MRTD模型可较准确地模拟非球形粒子散射特性,并行计算技术可显著提高计算效率;电磁场分量同时交换的并行设计方案的计算效率略高于仅交换磁场分量的方案;通过增加中央处理器核数,程序的并行加速比随之增大,但单核运行效率却略有降低。随着粒子尺度参数的增大,单核计算效率随之增加,复折射率的改变并不会显著影响并行计算效率。     

基于时域多分辨率的非球形气溶胶散射并行计算模型设计及验证

为解决串行时域多分辨率(MRTD)散射模型运行时间长和内存消耗大的问题,基于消息传递接口(MPI)技术设计了一种非球形气溶胶散射并行计算模型。介绍了MRTD散射模型的基本框架和2种并行数据通信方案,并基于MPI重复非阻塞通信技术实现了MRTD散射模型的并行化设计;搭建了网络并行计算平台,实现了模型的并行化计算。将MRTD散射模型与Mie散射模型、T矩阵法进行了对比,验证了并行MRTD散射模型的计算准确性。结果表明,MRTD模型可较准确地模拟非球形粒子散射特性,并行计算技术可显著提高计算效率;电磁场分量同时交换的并行设计方案的计算效率略高于仅交换磁场分量的方案;通过增加中央处理器核数,程序的并行加速比随之增大,但单核运行效率却略有降低。随着粒子尺度参数的增大,单核计算效率随之增加,复折射率的改变并不会显著影响并行计算效率。     

基于JASMIN的地下水流大规模并行数值模拟

针对具有精细网格剖分、长时间跨度特征的地下水流模拟中计算时间长、存储开销大等瓶颈问题,基于MODFLOW三维非稳定流计算方法,提出基于网格片的核心算法以及基于影像区的通信机制,并在JASMIN框架上研制了大规模地下水流并行数值模拟程序JOGFLOW.通过河南郑州市中牟县雁鸣湖水源地地下水流的模拟,对程序正确性和性能进行了验证;通过建立一个具有精细网格剖分的假想地下水概念模型对可扩展性进行测试.相对于32核的并行程序,在512以及1 024个处理机上的并行效率分别可达77.2%和67.5%.数值模拟结果表明,JOGFLOW具有较好的计算性能与可扩展性,能够有效使用数百上千计算核心,支持千万量级以上网格剖分的地下水流模型的大规模并行计算.     

Development of a 3D finite element acoustic model to predict the sound reduction index of stud based double-leaf walls

Building standards incorporating quantitative acoustical criteria to ensure adequate sound insulation are now being implemented. Engineers are making great efforts to design acoustically efficient double-wall structures. Accordingly, efficient simulation models to predict the acoustic insulation of double-leaf wall structures are needed. This paper presents the development of a numerical tool that can predict the frequency dependent sound reduction index R of stud based double-leaf walls at one-third-octave band frequency range. A fully vibro-acoustic 3D model consisting of two rooms partitioned using a double-leaf wall, considering the structure and acoustic fluid coupling incorporating the existing fluid and structural solvers are presented. The validity of the finite element (FE) model is assessed by comparison with experimental test results carried out in a certified laboratory. Accurate representation of the structural damping matrix to effectively predict the R values are studied. The possibilities of minimising the simulation time using a frequency dependent mesh model was also investigated. The FEA model presented in this work is capable of predicting the weighted sound reduction index R_w along with A-weighted pink noise C and A-weighted urban noise C_tr within an error of 1 dB. The model developed can also be used to analyse the acoustically induced frequency dependent geometrical behaviour of the double-leaf wall components to optimise them for best acoustic performance. The FE modelling procedure reported in this paper can be extended to other building components undergoing fluid–structure interaction (FSI) to evaluate their acoustic insulation.     

A Coupling Framework for Multi-Domain Modelling and Multi-Physics Simulations

This paper describes a coupling framework for parallel execution of different solvers for multi-physics and multi-domain simulations with an arbitrary number of adjacent zones connected by different physical or overlapping interfaces. The coupling architecture is based on the execution of several instances of the same coupling code and relies on the use of smart edges (i.e., separate processes) dedicated to managing the exchange of information between two adjacent regions. The collection of solvers and coupling sessions forms a flexible and modular system, where the data exchange is handled by independent servers that are dedicated to a single interface connecting two solvers’ sessions. Accuracy and performance of the strategy is considered for turbomachinery applications involving Conjugate Heat Transfer (CHT) analysis and Sliding Plane (SP) interfaces.     

Approximation of the inductionless MHD problem using a stabilized finite element method

In this work, we present a stabilized formulation to solve the inductionless magnetohydrodynamic (MHD) problem using the finite element (FE) method. The MHD problem couples the Navier–Stokes equations and a Darcy-type system for the electric potential via Lorentz’s force in the momentum equation of the Navier–Stokes equations and the currents generated by the moving fluid in Ohm’s law. The key feature of the FE formulation resides in the design of the stabilization terms, which serve several purposes. First, the formulation is suitable for convection dominated flows. Second, there is no need to use interpolation spaces constrained to a compatibility condition in both sub-problems and therefore, equal-order interpolation spaces can be used for all the unknowns. Finally, this formulation leads to a coupled linear system; this monolithic approach is effective, since the coupling can be dealt by effective preconditioning and iterative solvers that allows to deal with high Hartmann numbers.     

Flexoelectric effect modelling

A statistical theory of dipole flexoelectric (FE) polarization in liquid crystals is used to calculate temperature dependence of order parameters, elastic constants and FE coefficients. Two systems with polar wedge-shaped and banana-shaped molecules are investigated. In both cases, the FE coefficients are proportional to the dipole moment component parallel to the molecule symmetry axis. It results from the symmetries of interactions and of the Mayer function. The origin of the FE effect and microscopic pictures of the distorted phases are discussed.     

地球外核热对流运动的区域分解多重网格并行数值模拟

旋转球层中热对流运动的数值模拟是地球发电机模型的重要组成部分,对研究地球发电机作用机理具有重要意义.本文设计一个基于国产超级计算平台并行性能良好的地球外核热对流运动并行数值模型.时间积分方案采用与Crank-Nicolson格式和二阶Adams-Bashford公式相结合的近似分解分步法,空间离散基于立方球网格的二阶精度有限体积格式.所得到的两个大规模稀疏线性代数方程组采用带预处理的Krylov子空间迭代法进行求解.为加速迭代求解过程及提高并行性能,迭代过程采用区域分解多重网格的多层限制型加法Schwarz预处理子,减少了求解程序的计算时间,提高了数值模型的并行性能,模型被很好地扩展到上万处理器核数.数值模拟结果与基准模型算例0的参考值吻合得很好.     

Finite element method for solving Kohn-Sham equations based on self-adaptive tetrahedral mesh

A finite element (FE) method with self-adaptive mesh-refinement technique is developed for solving the density functional Kohn-Sham equations. The FE method adopts local piecewise polynomials basis functions, which produces sparsely structured matrices of Hamiltonian. The method is well suitable for parallel implementation without using Fourier transform. In addition, the self-adaptive mesh-refinement technique can control the computational accuracy and efficiency with optimal mesh density in different regions.     

基于PANDA平台的多点基础激励谐响应的并行计算

基于PANDA自主并行计算平台,采用模态叠加法对多点基础激励作用下的谐响应分析开展算法设计和并行实现研究,构建相应的并行求解模块.结合商业有限元软件对研发模块的正确性进行验证;将该模块应用于某光机装置,实现11.88亿自由度超大规模的高效并行求解.结果表明：研发的多点基础激励谐响应分析模块可用于复杂装备的高效精细数值模拟,具备强大的并行计算能力,最大可扩展CPU核数可达数千个,远超现有通用商业有限元软件.     

Vibration analysis and measurement of a gradient coil insert in a 4 T MRI

High speed switching of current in gradient coils within high magnetic field strength Magnetic Resonance Imaging (MRI) scanners may result in high acoustic sound pressure levels in and around these machines. Many studies have already been conducted to characterize the sound field in and around MRIs and various methods have been investigated to attenuate the noise generated. To characterize the vibration properties of the gradient coil, a modified Finite Element (FE) model was developed according to the dimensional design of an available gradient coil insert and the concentration of the copper windings in the coil. The finite element analysis results were verified through experimental modal testing of the same gradient coil in a free-free state (no boundary constraints). Comparisons show that the FE model predicts the vibration properties extremely accurately. Based on the verified FE model, boundary conditions (supports) were added to the model to simulate the operating condition when the gradient coil insert is in place in an MRI machine. Vibration analysis results from the FE model were again verified through experimental vibration testing with the gradient coil insert installed in a 4 T MRI and excited using swept sinusoidal time waveforms. Through a comparison of the vibration signals generated it was found that the vibration resonances, both from the FE model and the experimental vibration testing, shift to higher frequencies after the boundary constraints were applied, as was expected. The predicted vibration response was very close to that measured from the gradient coil insert in operation. The FE modeling procedure that has been developed could easily be used to accurately predict the vibration properties of other gradient coil designs. Furthermore, the vibration analysis results from the FE model could be used in acoustic noise analysis to predict the sound pressure level produced by different types of input current pulse sequences.     

Research on the field emission mechanism of nano-structured carbon film

The field emission(FE) characteristics of nano-structured carbon films(NSCFs) are investigated.The saturation behaviour of the field emission current density found at high electric field E cannot be reasonably explained by the traditional Fowler-Nordheim(F-N) theory.A three-region E model and the curve-fitting method are utilized for discussing the FE characteristics of NSCFs.In the low,high,and middle E regions,the FE mechanism is reasonably explained by a modified F-N model,a corrected space-charge-limited-current(SCLC) model and the joint model of F-N and SCLC mechanism,respectively.Moreover,the measured FE data accord well with the results from our corrected theoretical model.