直立码头前船波浪力耦合计算模型

建立了外域用Boussinesq方程、内域用刚体运动方程的直立码头前二维船剖面波浪力的时 域计算耦合模型，内域与外域在交界面的匹配条件是流量连续和压力相等. 进行了相关模型 实验，并把计算结果与实验结果进行了对比. 推导了船体与水底和直立码头之间间隙内流体 运动的自振频率，研究了间隙内流体运动的共振现象.     

基于等几何分析的边界元法求解二维Laplace方程

针对二维Laplace问题,提出了基于非均匀有理B样条的等几何边界单元法(IGABEM),并利用径向积分法来处理奇异积分。该方法实现了几何与求解域的无缝融合,不仅实现了求解域与几何的完美匹配,而且节约了前处理时间。该方法可以很容易地实现模型的细分,并且在仅增加少量自由度的情况下获得更高的精度。数值算例表明,该方法能够有效地求解二维Laplace方程,且具有非常好的计算精度。     

比例边界坐标插值方法在谱元法中的应用——无穷域Euler方程的数值模拟

将比例边界坐标插值方法引入谱元法, 构成比例边界谱单元, 对无穷域Euler方程进行数值模拟.阐述了比例边界谱单元的基本使用方法以及基于比例边界谱元的Runge-Kutta间断Galerkin方法求解Euler方程的过程;计算了无穷域圆柱和NACA0012翼型绕流问题, 并与已有结果进行了比较, 显示了计算结果的正确性.用基于比例边界谱元的间断Galerkin方法求解无穷域Euler方程时, 最多只需将求解域划分为2个子域, 避免了一般谱方法将求解域划分为9个或者27个子域的麻烦. 比例边界谱单元为无穷域Euler方程的直接求解提供了一个可供参考的方法.     

三维势流场的比例边界有限元求解方法

比例边界有限元法(SBFEM)是线性偏微分方程的一种新的数值求解方法。该方法只对计算域边界利用Galerkin方法进行数值离散,相对于有限元方法(FEM)减少了一个空间坐标的维数,而在减少的空间坐标方向利用解析方法进行求解;相对于边界元法(BEM),比例边界有限元方法不需要基本解,避免了奇异积分的计算,所以它结合了有限元和边界元方法的优点。本文建立了利用比例边界有限元法求解三维Laplace方程的数值模型并用于计算三维物体周围的水流场,将计算结果与解析解和边界元方法进行了对比,结果表明此方法可以很好地模拟水流场,且具有较高的计算精度。     

基于MUSTA格式的全非线性Boussinesq波浪传播数值模型

建立了求解二维全非线性布氏（Boussinesq）水波方程的有限差分/有限体积混合数值格式. 针对守恒形式的控制方程,采用有限体积方法并结合 MUSTA格式计算数值通量, 剩余项则采用有限差分方法求解, 采用具有总变差减小（totalvariation diminishing, TVD）性质的三阶龙格-库塔法进行时间积分.该格式具备间断捕捉、程序实现简单、数值稳定性强、海岸动边界以及波浪破碎处理方便和可调参数少等优点.利用典型算例对数值模型进行了验证,计算结果与实验数据吻合较好.      

Helmholtz边界积分方程中奇异积分间接求解方法

提出了间接求解传统Helmholtz边界积分方程CBIE的强奇异积分和自由项系数,以及Burton-Miller边界积分方程BMBIE中的超强奇异积分的特解法。对于声场的内域问题,给出了满足Helmholtz控制方程的特解,间接求出了CBIE中的强奇异积分和自由项系数。对于声场外域对应的BMBIE中的超强奇异积分,按Guiggiani方法计算其柯西主值积分需要进行泰勒级数展开的高阶近似,公式繁复,实施困难。本文给出了满足Helmholtz控制方程和Sommerfeld散射条件的特解,提出了间接求出超强奇异积分的方法。推导了轴对称结构外场问题的强奇异积分中的柯西主值积分表达式,并通过轴对称问题算例证明了本文方法的高效性。数值结果表明,对于内域问题,采用本文特解法的计算结果优于直接求解强奇异积分和自由项系数的结果,且本文的特解法可避免针对具体几何信息计算自由项系数,因而具有更好的适用性。对于外域问题,两者精度相当,但本文的特解法可避免对核函数进行高阶泰勒级数展开,更易于数值实施。     

虚拟区域法在流固耦合问题中的应用

将多相流领域内的虚拟区域法引入到流固耦合问题的分析中,将固体视为应变率为零的虚拟流体,对流体和虚拟流体均以速度和压强作为基本变量,采用Navier-Stokes方程作为控制方程,同时求解流体域和虚拟流体域, 得到整个计算域的流场分布,应用分布式拉格朗日乘子法在虚拟流体域上施加刚体约束, 以保持虚拟流体的刚体外形和运动形式,最终建立一种流固耦合模型及其数值求解方法. 通过对粒子流问题和流固耦合问题进行数值模拟,验证了此模型的正确性和求解大变形/运动流固耦合问题的有效性.      

曲线坐标下连通域河道平面二维水流的数值模拟

在边界拟合曲线坐标系下,运用B型交错网格模式和动边界扫描技术建立了基于连通域的二维水流数学模型,并提出了模型中有关参数的处理方法.采用贴体坐标变换将复杂的物理域变换成规则的计算域,在计算域上采用控制容积法离散方程,应用SIMPLEC算法计算速度-压力耦合.研究结果表明:采用控制容积法和SIMPLEC算法离散求解方程,具有良好的守恒性和稳定性; 该模型能够较准确地模拟连通域河段的流场变化、水位变化等过程,可供实际工程应用.     

基于Boussinesq方程的波浪模型

从欧拉方程出发，提供了另一种推导完全非线性Boussinesq方程的方法，并对方程的 线性色散关系和线性变浅率进行了改进. 改进后方程的线性色散关系达到了一阶Stokes波 色散关系的Pad\'{e}[4,4]近似，在相对水深达1.0的强色散波浪时仍保持较高的准确性，并且方程的非线性和线性 变浅率都得到了不同程度的改善. 方程的水平一维形式用预估-校正的有限差分格式求解， 建立了一个适合较强非线性波浪的Boussinesq波浪数值模型. 作为验证，模拟了波浪在潜 堤上的传播变形，计算结果和实验数据的比较发现两者符合良好.     

不同发射深度下导弹水下点火气水流体动力计算

从流体动力角度研究了不同发射深度下，导弹水下点火这一非定常非线性过程。整个系统分为外部水流场、喷管流场和燃气泡流场三个区域加以考虑。水流场采用不可压势流模型，用边界元方法求解；喷管内流场采用非定常一元流动模型，用特征线差分法求解，并设置了激波检测功能；燃气泡采用基于质量和能量守恒的零维计算模型。在时间域中用步进方法实现了三个流场的耦合求解。给出了四种发射深度下的数值计算结果，展示了导弹水下点火的一     

An experimental and numerical study of heave added mass and damping of horizontally submerged and perforated rectangular plates

Forced harmonic heave motions of horizontally submerged and perforated rectangular plates are studied experimentally and numerically at both a deep and shallow submergence. The steady-state vertical forces are expressed in terms of added mass and damping coefficients. The numerical results are partly obtained by combining potential flow with linear free-surface conditions and a nonlinear viscous pressure loss condition at the mean oscillatory plate position. A domain decomposition technique is applied with a boundary element method in the inner domain and an analytical representation of the velocity potential in the outer domain. A drag term accounts for the vortex shedding at the outer plate edges. The numerically predicted Keulegan–Carpenter number dependent heave added mass and damping coefficients agree reasonably with experimental values, in particular for the deeper submergence.     

Hybrid-Trefftz displacement element for spectral analysis of bounded and unbounded media

The hybrid-Trefftz displacement element is applied to the elastodynamic analysis of bounded and unbounded media in the frequency domain. The displacements are approximated in the domain of the element using local solutions of the wave equation, the Neumann conditions are enforced directly and the surface forces are approximated on the Dirichlet and inter-element boundaries of the finite element mesh. Two alternative elements are developed to model unbounded media, namely a finite element with absorbing boundaries and an unbounded element that satisfies explicitly the Sommerfeld condition. The finite element equations are derived from the fundamental relations of elastodynamics written in the frequency domain. The numerical implementation of these equations is discussed and numerical tests are presented to assess the performance of the formulation.     

Water surface wave radiation generated by multiple cylinders oscillating with identical frequency

ＩｎｔｒｏｄｕｃｔｉｏｎＡｆｕｎｄａｍｅｎｔａｌｓｕｂｊｅｃｔｉｎｔｈｅｓｔｕｄｙｏｆｉｎｔｅｒａｃｔｉｏｎｂｅｔｗｅｅｎｗａｖｅｓａｎｄｂｏｄｙｉｓｔｏｓｏｌｖｅｔｈｅｔｗｏ＿ｄｉｍｅｎｓｉｏｎａｌｒａｄｉａｔｉｏｎｐｒｏｂｌｅｍｇｅｎｅｒａｔｅｄｂｙｃｙｌｉｎｄｅｒｓｏｓｃｉｌｌａｔｉｎｇｉｎｏｒｂｅｌｏｗｔｈｅｆｒｅｅｗａｔｅｒｓｕｒｆａｃｅ,ｗｈｉｃｈｈａｓａｔｔｒａｃｔｅｄｗｉｄｅａｔｔｅｎｔｉｏｎ．ＥｖｅｒｓｉｎｃｅＵｒｓｅｌｌｐｕｂｌｉｓｈｅｄｔｈｅｐａｐｅｒ［１］ｉｎｖｏｌｖｉｎ…     

Finite element computations for unsteady fluid and elastic membrane interaction problems

The interaction between the hydrodynamic forces of a flow field and the elastic forces of adjacent deformable boundaries is described by elastohydrodynamics, a coupled fluid–elastic membrane problem. Direct numerical solution of the unsteady, highly non-linear equations requires that the dynamic evolution of both the flow field and the domain shape be determined as part of the solution, since neither is known a priori. This paper describes a numerical algorithm based on the deformable spatial domain space–time (DSD/ST) finite element method for the unsteady motion of an incompressible, viscous fluid with elastic membrane interaction. The unsteady Navier–Stoke and elastic membrane equations are solved separately using an iterative procedure by the GMRES technique with an incomplete lower-upper (ILU) decomposition at every time instant. One-dimensional, two-dimensional and deformable domain model problems are used to demonstrate the capabilities and accuracy of the present algorithm. Both steady state and transient problems are studied. © 1997 John Wiley & Sons, Ltd.     

A numerical modelling of stator–rotor interaction in a turbine stage with oscillating blades

In real flows unsteady phenomena connected with the circumferential non-uniformity of the main flow and those caused by oscillations of blades are observed only jointly. An understanding of the physics of the mutual interaction between gas flow and oscillating blades and the development of predictive capabilities are essential for improved overall efficiency, durability and reliability. In the study presented, the algorithm proposed involves the coupled solution of 3D unsteady flow through a turbine stage and the dynamics problem for rotor-blade motion by the action of aerodynamic forces, without separating the outer and inner flow fluctuations. The partially integrated method involves the solution of the fluid and structural equations separately, but information is exchanged at each time step, so that solution from one domain is used as a boundary condition for the other domain. 3-D transonic gas flow through the stator and rotor blades in relative motion with periodicity on the whole annulus is described by the unsteady Euler conservation equations, which are integrated using the explicit monotonous finite volume difference scheme of Godunov–Kolgan. The structural analysis uses the modal approach and a 3-D finite element model of a blade. The blade motion is assumed to be constituted as a linear combination of the first natural modes of blade oscillations, with the modal coefficients depending on time. A calculation has been done for the last stage of the steam turbine, under design and off-design regimes. The numerical results for unsteady aerodynamic forces due to stator–rotor interaction are compared with results obtained while taking into account blade oscillations. The mutual influence of both outer flow non-uniformity and blade oscillations has been investigated. It is shown that the amplitude-frequency spectrum of blade oscillations contains the high-frequency harmonics, corresponding to the rotor moving past one stator blade pitch, and low-frequency harmonics caused by blade oscillations and flow non-uniformity downstream from the blade row; moreover, the spectrum involves the harmonics which are not multiples of the rotation frequency.     

Development of a numerical wave tank for analysis of nonlinear and irregular wave field

A numerical wave-absorption filter has been developed for an open boundary condition in the analysis of nonlinear and irregular wave evolution. The filter is composed of a simulated sponge layer and Sommerfeld's radiation condition at the outer edge of the layer. The wave-absorption characteristics of the filter have been investigated by applying the linear potential theory and a two-dimensional nonlinear boundary element model. In both cases, the filter is found to he applicable for a wide range of wave parameters. In order to realize an idealized “numerical wave tank”, the present model also incorporates a nonreflective wave generator in the computational domain composed of a series of vertically aligned point sources. Numerous numerical experiments demonstrate that the present approach is effective in generating an arbitrary wave profile without reflection not only at the open boundaries but also at the wave generator.     

BOUNDARY CONDITIONS IN SHALLOW WATER MODELS—AN ALTERNATIVE IMPLEMENTATION FOR FINITE ELEMENT CODES

Finite element solution of the shallow water wave equations has found increasing use by researchers and practitioners in the modelling of oceans and coastal areas. Wave equation models, most of which use equal-orderC⁰ interpolants for both the velocity and the surface elevation, do not introduce spurious oscillation modes, hence avoiding the need for artificial or numerical damping. An important question for both primitive equation and wave equation models is the interpretation of boundary conditions. Analysis of the characteristics of the governing equations shows that for most geophysical flows a single condition at each boundary is sufficient, yet there is not a consensus in the literature as to what that boundary condition must be or how it should be implemented in a finite element code. Traditionally (partly because of limited data), surface elevation is specified at open ocean boundaries while the normal flux is specified as zero at land boundaries. In most finite element wave equation models both of these boundary conditions are implemented as essential conditions. Our recent work focuses on alternative ways to numerically implement normal flow boundary conditions with an eye towards improving the mass-conserving properties of wave equation models. A unique finite element formulation using generalized functions demonstrates that boundary conditions should be implemented by treating normal fluxes as natural conditions with the flux interpreted as external to the computational domain. Results from extensive numerical experiments show that the scheme does conserve mass for all parameter values. Furthermore, convergence studies demonstrate that the algorithm is consistent, as residual errors at the boundary diminish as the grid is refined.     

Three-dimensional coupled numerical model of creeping flow of viscous fluid

A three-dimensional coupled numerical model is developed to describe creeping flow in a computational domain that consists of a thick viscous layer overlaid with a thin multilayered viscous sheet. The density of the sheet is assumed to be lower than that of the layer. The model couples the Stokes equations describing the flow in the layer and the Reynolds equations describing the flow in the sheet. We investigate the long-time behavior of the flow in the sheet by using an asymptotic method and derive an ordinary differential equation for the sheet boundary displacements and the velocities at the interface between the sheet and the layer. The Stokes and Reynolds equations are coupled by applying the resulting equation as an internal boundary condition. Numerical implementation is based on a modified finite element method combined with the projection gradient method. The computational domain is discretized into rectangular hexahedra. Piecewise square basis functions are used. The model proposed enables different-type hydrodynamic equations to be coupled without any iterative improvements. As a result, the computational costs are reduced significantly in comparison with available coupled models. Numerical experiments confirm that the three-dimensional coupled model developed is of good accuracy.     

NUMERICAL STUDY OF THE BLOCKAGE EFFECTS ON VISCOUS FLOW PAST A CIRCULAR CYLINDER

In various numerical solutions of flow around bluff bodies the unbounded physical domain is replaced by a restricted computational one whose extent depends on the size of the computational grid network. The truncation of the solution domain in the cross-flow direction reduces the computer time required for the solution, but introduces numerical blockage effects which influence considerably the values of the various flow parameters. In the present paper the finite element solution of steady and unsteady flow around a circular cylinder at Re=106 is presented for blockage ratios of 0·05, 0·15 and 0·25. A boundary condition was tested for which the streamfunction values at the outer boundaries were those of the irrotational solution around a circular cylinder. The size of the standing vortices decreases with the blockage ratio when the flow is steady, while the spacing of the vortices decreases in both directions with increasing blockage ratio when the wake becomes unsteady. The hydrodynamic forces on the cylinder and the Strouhal number are magnified as the blockage ratio increases. The application of the streamfunction values derived from the irrotational solution at the outer boundaries reduced blockage effects only at high blockage ratio.