流场有限元分析的并行计算

在流场有限元分离式求解法的基础上 ,导出了并行计算的算法流程 ,可利用并行计算机同时求解的相互耦合的流场描述非线性方程组 ,并在曙光 2 0 0 0并行计算机系统上成功实施 ,比之串行计算 ,较大幅度地提高了计算效率     

基于集群的并行有限元分析研究

利用面向对象的方法实现了基于集群的并行有限元分析.首先介绍构建高性能计算集群的系统分析新方法,并针对建成集群系统的特点,对并行消息传递库MPI进行了并行语义分析,进而对其主要的实现函数进行面向对象的重构而建立起OO消息传递库,以此为基础分别利用两种不同的策略实现了面向对象的并行有限元分析(OOParaFEA: Object-Oriented Parallel FEA),一种是通过对传统的基于域分解方法的并行PCG算法进行改造而在集群平台实现,另一种是对已有的基于子结构方法的串行有限元分析程序加以扩展,加入系统方程组并行求解器而达到有限元分析并行化的目的.多个分析算例表明,基于集群进行面向对象的并行有限元分析可以有效提高计算效率,为进一步的网络化CAD/CAE研究奠定良好基础.     

A finite element parallel algorithm of the parametric variational principle for elastic contact problems

Basetl on the finite element solution of the parametric varialional principle of elastic con/del problem, a corresponding parallel algorithm has been created bv utilizing the specialities of parallel computer and the architecture of concurrent processing in this paper. In this algorithm. the parallelisms have heen realized in the processes of creation and assembly of stiffness matrix, of the static condensation, of the solution of stresses and in many other aspects. The programme of this algorithm has been realized on ELXSI-6400 parallel computer of Xi'an Jiaotong University. The results of computation show that the computational time can be saved efficiently and it is an effective parallel algorithm for the analyses of contact problems.     

有限元子结构并行算法的效能分析

引入了有限元子结构的并行算法,该方法通过划分子结构,由多个CPU并行进行静凝聚,然后求解界面位移,最后返回各小结构并行求解内点位移和应力.该方法在ELXSI 6 400并行机上程序实现,表明能大幅度节省机时.本文同时也对该方法的效能进行了分析.     

面向对象的串并行有限元分析系统

研究了面向对象的串并行有限元分析系统的关键实现技术。首先介绍的有限元系统类库分析,重点阐明了面向对象方法应用于串并行有限元分析所具有的优势,基于此将集成系统归纳成多个组成类库:包括数学计算类库、数据处理类库、有限元基本类库、数值分析模型类库、图形类库以及应用框架工具类库等,进而介绍分析程序的面向对象实现方法及过程,最后形成一个集成的交互式并行有限元分析框架OOParaFEA(Object OrientedParallel FEA framework)。系统快速高效的实现,表明面向对象的方法的确从软件的可维护性、可扩充性以及减少复杂性等方面带来前所未有的优势,并且由于其固有的继承特性而使其特别适合团队协作开发大型有限元分析软件,从而给我国大规模科学计算软件的研制提供一个新思路。     

A fire field model implemented in a parallel computing environment

This paper describes the implementation of a fire field model in the parallel computing environment offered by multiple transputers. The fire model is built into the general purpose SIMPLE-based CFD code HARWELL-FLOW3D. The technique of domain decomposition has been applied tb convert the conventional serial version of FLOW3D into a code capable of efficiently utilizing an arbitrary number of transputers. Fire simulations consisting of up to 24 000 computational cells are performed on parallel systems with up to 15-processors. The run time for this simulation has been reduced from over 4 days on a single processor to just over 8 h on the 15-processor system. An interactive graphics system has also been developed which runs in parallel with the main computations.     

An iterative parallel algorithm of finite element method

In this paper, a parallel algorithm with iterative form for solving finite element equation is presented. Based on the iterative solution of linear algebra equations, the parallel computational steps are introduced in this method. Also by using the weighted residual method and choosing the appropriate weighting functions, the finite element basic form of parallel algorithm is deduced. The program of this algorithm has been realized on the ELXSI-6400 parallel computer of Xi'an Jiaotong University. The computational results show the operational speed will be raised and the CPU time will be cut down effectively. So this method is one kind of effective parallel algorithm for solving the finite element equations of large-scale structures.     

大型结构有限元并行分析程序——PARNFAP

本文介绍了作者在西安交通大学ELXSI-6400并行机调试的大型结构有限元分析程序—PARNFAP。该程序利用子结构方法实现有限元计算的并行化,计算结果表明其具有良好的并行性能。     

Analysis of a 6-DOF redundantly actuated 4-legged parallel mechanism

In this paper, we propose a new 4-legged 6-DOF redundantly actuated UPS parallel mechanism. Kinematics, singularity and dynamics analyses of the proposed mechanism are addressed. Inverse kinematics has a closed-form solution, whereas the forward kinematics has a semi-closed form solution. Also, the Jacobian matrix has been determined using the concept of reciprocal screws. Using the principle of virtual work and minimum norm method, a general formulation for the inverse dynamics of redundantly actuated parallel manipulator is presented. Moreover, the proposed redundant mechanism has been compared with a similar but nonredundant mechanism in three aspects: reachable points, and actuation forces under static and dynamic conditions. We show that the redundancy decreases singular points, and dramatically reduces the actuators’ forces and torques.     

Transient finite element analysis of generalized Newtonian coextrusion flows in complex geometries

A two-dimensional transient finite element model capable of simulating problems related to two-layer polymer flows has been developed. This technique represents an effective tool which can be used to study the possibility of the onset of interfacial instability in coextrusion flows, considering melt rheology as well as the fluid–geometry interaction. A code has been developed to solve the transient problem of the flow of bi-component systems of Newtonian and generalized Newtonian fluids through parallel plates and complex geometries, such as: 2:1 abrupt expansion, 2:1 (30°) expansion, 4:1 abrupt contraction and 4:1 tapered (30°) contraction. Solutions are compared with experimental data from the literature and results provided by linear stability analysis (LSA) for the case of parallel plate flows. Numerical results are in agreement with LSA results for the parallel plate geometry cases studied. The expansion geometries tend to stabilize flows in the parallel plate section downstream of the expansion. Contractions may give rise to break-up of the interface depending on the flow conditions. © 1998 John Wiley & Sons, Ltd.     

结构有限元分析并行处理的研究进展

在过去20年, 由于并行计算机的出现, 对结构工程应用的有限元算法研究已取得很大的进展. 为了将串行机上运行的算法改成并行和研制适应于并行计算机的新方法, 一直在进行大量的研究. 本文论述了该领域一些已有的有限元并行算法, 对适应结构工程的一些有限元并行算法进行概括. 对该领域的未来发展趋势作了展望.      

Investigation of Entry Region Turbulent Flows of Non-Gray Gases

The entrance region turbulent flow between parallel planes has been investigated numerically. Variation of fluid properties with temperature has been considered. Monte Carlo simulation has been used with a narrow band non-gray model to investigate radiative transfer. The effects of radiative transfer and variable properties have been separately and systematically considered. Representative results are included for carbon monoxide and carbon dioxide. For the conditions considered, the results show that the variable properties exert a stronger influence compared to radiative transfer alone.     

Thermal effect on vibration of non-homogenous visco-elastic rectangular plate of linearly varying thickness

An analysis for vibration of non-homogenous visco-elastic rectangular plate of linearly varying thickness subjected to thermal gradient has been discussed in the present investigation. For visco-elastic, the basic elastic and viscous elements are combined. We have taken Kelvin model for visco-elasticity that is the combination of the elastic and viscous elements in parallel. Here the elastic element means the spring and the viscous element means the dashpot. The governing differential equation of motion has been solved by Galerkin’s technique. Deflection, time period and logarithmic decrement at different points for the first two modes of vibration are calculated for various values of thermal gradients, non homogeneity constant, taper constant and aspect ratio for non-homogenous visco-elastic rectangular plate which is clamped on two parallel edges and simply supported on remaining two edges. Comparison studies have been carried out with homogeneous visco-elastic rectangular plate to establish the accuracy and versatility.     

二维双曲型方程初边值问题的块三对角并行求解算法

讨论用块三对角线性方程组的可扩展并行算法,求解带Dirichlet边界条件的一阶二维双曲型方程初边值问题。用本文方法在上海大学超级计算机"自强3000"上进行数值实验,实验结果与理论分析一致。在保证精度的前提下,得到线性加速比,并行效率达到90%以上。     

The structure of parallel layers in steady two-dimensional magnetohydrodynamic flows in sudden duct expansions and contractions

The structure of non-linear, steady, two-dimensional parallel layers at high values of the Hartmann number, Ha, the Reynolds number, Re, and the interaction parameter, N, for duct expansions and contractions has been investigated. The magnetic field is transverse to the flow. For the flow regime, in which the electromagnetic force balances the inertial force in the layer, a viscous sublayer at the solid wall parallel to the magnetic field has been obtained, in which the flow is driven by the pressure gradient induced in the outer, inviscid layer.     

Transient boiling flow instabilities in a multi-channel upflow system

In this study, the transient boiling flow instabilities in a four parallel channel upflow system with and without cross-connections have been experimentally investigated using Freon-11 as the test fluid. Several series of transient experiments have been run using a step increase in power input with a constant flow control valve setting and a step decrease in flow rate with a constant heat input. Two major modes of oscillations, namely density-wave type and pressure-drop type oscillations, have been observed. Total system pressure drop has been plotted as a function of mass flow rate for various heat inputs and flow control valve settings, and the stable and unstable regions for transient density-wave type and pressure-drop type oscillations have been determined and the boundaries of sustained instabilities have also been indicated. Experiments have been conducted with equal and unequal heat inputs and the results have been compared with each other. A comparison of the transient instabilities observed in four parallel channel system with and without cross connections has also been made.     

Property of intrinsic drift coefficients in globally-evolving-based generalized density evolution equation for the first-passage reliability assessment

Acta Mechanica Sinica - The turbulent Poiseuille flow between two parallel plates is one of the simplest possible physical situations, and it has been studied intensively. In this paper, we propose...     

A methodology to achieve the set of operation modes of reconfigurable parallel manipulators

The demand for increasingly more versatile machinery has boosted the development of the so-called reconfigurable mechanisms. In this paper, the authors present a general methodology to assess the multioperational capacity of a 6-DOF parallel manipulator basing on the possible motion patterns having Lie group structure that the manipulator owns. This ability of having different operation modes enables the manipulator to adapt to diverse tasks. To show the potential of the methodology, this approach has been applied to the 6-DOF 3-CPCR which is capable of generating multiple motion patterns. In addition to carrying out the complete theoretical study in which all the different operation modes are obtained, and validating the procedure with GIM software designed for kinematic analysis and design of mechanisms, a demonstration prototype of the 3-CPCR parallel manipulator has been also built.

     

Three-dimensional analysis of fluid flow and heat transfer in single- and two-layered micro-channel heat sinks

A three-dimensional numerical analysis of laminar fluid flow and conjugate heat transfer has been conducted for single- and two-layered micro-channel heat sinks. The validity of the numerical model has been confirmed by comparison with available experimental data. Results for the overall thermal resistance, pumping power, the maximum temperature difference on the heat-sink surface where the heat flux is applied, and an overall performance parameter were obtained for single- and two-layered sinks. The effects of Reynolds number, inlet velocity profile, and flow arrangement in the channels (parallel and counter) on these results are presented and discussed. A special emphasis was placed on the comparison between the thermal performances of the parallel and counter flow arrangements and further results were obtained in order to quantify and explain the relative performance under these flow arrangements.     

Elastodynamics of a spatial redundantly actuated parallel mechanism constrained by two higher kinematic pairs

Meccanica - A spatial redundantly actuated parallel mechanism (RAPM) constrained by two point-contact higher kinematic pairs (HKPs) has been designed for food texture evaluation during the...