Use of characteristics method with cubic interpolation for unsteady-flow computation

The specified-time-interval (STI) scheme has been used commonly in applying the method of characteristics (MOC) to unsteady open-channel flow problems. However, with the use of STI scheme, the numerical error for the simulation results can always be induced due to the interpolation used to approximate the characteristics trajectory. Hence, in order to remedy the numerical errors caused by the interpolation, one needs to seek some kind of interpolation technique with higher-order accuracy. Instead of the linear interpolation technique, which has been used very commonly and can induce serious numerical diffusion, the Holly--Preissmann two-point, method, which is a cubic interpolation technique with fourth-order of accuracy, is proposed here to integrate with the method of characteristics for the computation of one-dimensional unsteady flow in open channel. The concept of reachback and reachout in space and time directions for the characteristics is also introduced to assure the model stability. The computed results from this new model are compared with those computed by using the Preissmann four-point scheme and the multimode method of characteristics with linear interpolation.     

A cell‐based smoothed finite element method with semi‐implicit CBS procedures for incompressible laminar viscous flows

In this paper, the cell‐based smoothed finite element method (CS‐FEM) with the semi‐implicit characteristic‐based split (CBS) scheme (CBS/CS‐FEM) is proposed for computational fluid dynamics. The 3‐node triangular (T3) element and 4‐node quadrilateral (Q4) element are used for present CBS/CS‐FEM for two‐dimensional flows. The 8‐node hexahedral element (H8) is used for three‐dimensional flows. Two types of CS‐FEM are implemented in this paper. One is standard CS‐FEM with quadrilateral gradient smoothing cells for Q4 element and hexahedron cells for H8 element. Another is called as n‐sided CS‐FEM (nCS‐FEM) whose gradient smoothing cells are triangles for Q4 element and pyramids for H8 element. To verify the proposed methods, benchmarking problems are tested for two‐dimensional and three‐dimensional flows. The benchmarks show that CBS/CS‐FEM and CBS/nCS‐FEM are capable to solve incompressible laminar flow and can produce reliable results for both steady and unsteady flows. The proposed CBS/CS‐FEM method has merits on better robustness against distorted mesh with only slight more computation time and without losing accuracy, which is important for problems with heavy mesh distortion. The blood flow in carotid bifurcation is also simulated to show capabilities of proposed methods for realistic and complicated flow problems.     

A comparative study of lattice Boltzmann methods using bounce‐back schemes and immersed boundary ones for flow acoustic problems

In order to find applicable treatments of moving boundary conditions based on the lattice Boltzmann method in flow acoustic problems, three bounce‐back (BB) methods and four kinds of immersed boundary (IB) methods are compared. We focused on fluid–solid boundary conditions for flow acoustic problems especially the simulations of sound waves from moving boundaries. BB methods include link bounce‐back, interpolation bounce‐back and unified interpolation bounce‐back methods. Five IB methods are explicit and implicit direct‐forcing (Explicit‐IB and Implicit‐IB), two kinds of partially saturated computational methods and ghost fluid method. In order to reduce the spurious pressure generated by the fresh grid node changing from solid domain to fluid domain for BB methods and sharp IB methods, we proposed two new kinds of treatments and compared them with two existing ones. Simulations of the benchmark problems prove that the local evolutionary iteration (LI) is the best one in treatments of the fresh nodes. In addition, for standing boundary problems, although BB methods have a little higher accuracy, all the methods have similar accuracy. However, for moving boundary problems, IB methods are more appropriate than BB methods, because IB methods' smooth interpolation of pressure eld produces less disturbing spurious pressure waves. With improved treatments of fresh nodes, BB methods are also acceptable for moving boundary acoustic problems. In comparative tests in respective type, unified interpolation bounce‐back with LI, Implicit‐IB, and ghost fluid with LI are the best choices. Copyright © 2013 John Wiley & Sons, Ltd.     

High‐order stable interpolations for immersed boundary methods

The analysis and improvement of an immersed boundary method (IBM) for simulating turbulent flows over complex geometries are presented. Direct forcing is employed. It consists in interpolating boundary conditions from the solid body to the Cartesian mesh on which the computation is performed. Lagrange and least squares high‐order interpolations are considered. The direct forcing IBM is implemented in an incompressible finite volume Navier–Stokes solver for direct numerical simulations (DNS) and large eddy simulations (LES) on staggered grids. An algorithm to identify the body and construct the interpolation schemes for arbitrarily complex geometries consisting of triangular elements is presented. A matrix stability analysis of both interpolation schemes demonstrates the superiority of least squares interpolation over Lagrange interpolation in terms of stability. Preservation of time and space accuracy of the original solver is proven with the laminar two‐dimensional Taylor–Couette flow. Finally, practicability of the method for simulating complex flows is demonstrated with the computation of the fully turbulent three‐dimensional flow in an air‐conditioning exhaust pipe. Copyright © 2006 John Wiley & Sons, Ltd.     

The improved surface gradient method for flows simulation in variable bed topography channel using TVD‐MacCormack scheme

This paper reports four different approaches to discretize the source terms for the simulation of one‐dimensional open‐channel flows with rapidly varied bottom topography using TVD‐MacCormack scheme. Compared with other high‐resolution shock‐capturing schemes, MacCormack‐type predictor–corrector method is easy to implement and does not present any additional difficulty in dealing with the source terms. To avoid the generation of artificial numerical waves, if the bottom topography shows strong variation, special treatment of the source terms is still required to eliminate or reduce the artificial numerical error caused by adding TVD corrections to the method. The computed results demonstrated that the improved surface gradient method is more suitable for simulating open‐channel flow with highly irregular bed topography by using the surface gradient instead of the depth gradient for TVD corrections and considering the balancing of the source terms and the flux gradients. Copyright © 2003 John Wiley & Sons, Ltd.     

On the use of the reach-back characteristics method for calculation of dispersion

The Holly-Preissmann two-point finite difference scheme (HP method) has been popularly used for solving the advection equation. The key idea of this scheme is to solve the dependent variable (i.e. the concentration for the pollutant transport problem) by the method of characteristics with the use of cubic interpolation on the spatial axis. The interpolating polynomials of higher order are constructed by use of the dependent variable and its derivatives at two adjacent grid points. In this paper a new interpolating technique is introduced for incorporation with the Holly-Preissmann two-point method. The new method is denoted herein as the Holly-Preissmann reach-back method (HPRB) and allows the characteristics to project back several time steps beyond the present time level. Through stability analyses it has been observed that the increase of the reach-back time step numbers for the characteristics indeed reduces the numerical damping and dispersive phenomena. A schematic model has been constructed to demonstrate the merits of this new technique for the calculation of the pure advection and dispersion equations. Numerical experiments and comparisons with analytical solutions which support and demonstrate this new technique are presented.     

An accurate computation for rapidly varied flow in an open channel

A new method combining the Preissmann four-point scheme and the Holly–Preissmann reach-back scheme is introduced to solve the rapidly varied flow problem in an open channel. The Preissmann four-point scheme is well known for the computation of one-dimensional unsteady flow. The Holly–Preissmann reach-back scheme integrates the Holly-Preissmann two-point scheme with the concept of reach-back characteristics, which allows the characteristics to project several time steps beyond the current time level. A spontaneous surge formation case is used to demonstrate and evaluate the applicability of the new method. It has been found that the results from this method are quite compatible with those of Preissmann four-point scheme. In addition, with the appropriate choice of the number of reach-back time steps, this new method can always avoid the numerical oscillation which usually exists when one uses the Preissmann four-point scheme for the condition of Courant number not close to unity.     

Investigation of use of reach-back characteristics method for 2D dispersion equation

The use of the Holly-Preissmann two-point scheme has been very popular for the calculation of the dispersion equation. The key to this scheme is to use the characteristics method incorporating the Hermite cubic interpolation technique to approximate the trajectory foot of the characteristics. This method can avoid the excessive numerical damping and oscillation associated with most finite difference schemes for advection computation. On the basis of the fundamental idea of the Holly-Preissmann two-point scheme, a new technique is introduced herein for the computation of the two-dimensional dispersion equation. This new scheme allows the characteristics projecting back several time steps to fall on the spatial or temporal axis, while the characteristics foot is still solved by the Holly-Preissmann two-point method. The diffusion portion of the dispersion equation is solved by the commonly used Crank-Nicholson method. The calculation for these two processes consisting of advection and diffusion is carried out separately but consecutively in one time step, a method known as the split operator algorithm. A hypothetical model was constructed to demonstrate the applicability of this new technique for the calculation of the pure advection and dispersion equation in two dimensions.     

Simple treatment of non‐aligned boundaries in a Cartesian grid shallow flow model

A simple method is proposed for treating curved or irregular boundaries in Cartesian grid shallow flow models. It directly evaluates fictional values in ‘ghost’ cells adjacent to boundary cells and requires no interpolation or generation of cut cells. The boundary treatment is implemented in a dynamically adaptive quadtree grid‐based solver of the hyperbolic shallow water equations and validated against several test cases with analytical or alternative numerical solutions. The method is easy to code, accurate, and demonstrably effective in dealing with irregular computational domains in shallow flow simulations. Results are presented for still water in a basin of complicated geometry, steady hydraulic jump in an open channel with a converging sidewall, wind‐induced circulation in a circular shallow lake, and shock wave diffraction in a channel containing a contraction and expansion. Copyright © 2007 John Wiley & Sons, Ltd.     

A novel finite point method for flow simulation

A novel finite point method is developed to simulate flow problems. The mashes in the traditional numerical methods are supplanted by the distribution of points in the calculation domain. A local interpolation based on the properties of Taylor series expansion is used to construct an approximation for unknown functions and their derivatives. An upwind‐dominated scheme is proposed to efficiently handle the non‐linear convection. Comparison with the finite difference solutions for the two‐dimensional driven cavity flow and the experimental results for flow around a cylinder shows that the present method is capable of satisfactorily predicting the flow separation characteristic. The present algorithm is simple and flexible for complex geometric boundary. The influence of the point distribution on computation time and accuracy of results is included. Copyright © 2002 John Wiley & Sons, Ltd.     

Comparisons of CIP,compact and CIP‐CSL2 schemes for reproducing internal solitary waves

Six different models were evaluated for reproducing internal solitary waves which occur and propagate in a stratified flow field with a sharp interface. Three stages were used to compute internal solitary waves in a stratified field: (1) first‐phase computation of momentum equations, (2) second‐phase computation of momentum equations, which corresponds to computing the Poisson's equation, and (3) density computation. The six models discussed in this paper consisted of combinations of four different schemes, a three‐point combined compact difference scheme (CCD), a normal central difference scheme (CDS), a cubic‐polynomial interpolation (CIP), and an exactly conservative semi‐Lagrangian scheme (CIP‐CSL2). The residual cutting method was used to solve the Poisson's equation. Three tests were used to confirm the validity of the computations using KdV theory; i.e. the incremental wave speed and amplitude of internal solitary waves, the maximum horizontal velocity and amplitude, and the wave form. In terms of the shape of an internal solitary wave, using CIP for momentum equations was found to provide better performance than CCD. These results suggest one of the most appropriate scheme for reproducing internal solitary waves may be one in which CIP is used for momentum equations and CCD to solve the Poisson's equation. Copyright © 2005 John Wiley & Sons, Ltd.     

车桥耦合振动分析的两种常用插值方法比较

利用模态综合法分析车辆与桥梁之间的相互作用时,合理地构造桥梁的插值振型函数可以大幅提高计算精度.其中,分段三次Hermite插值函数和三次样条插值函数较为常用.为研究二者的异同,以简支梁桥为例分别采用这两种插值函数构造结构梁单元模型的一维插值振型函数和板单元模型的二维插值振型函数.基于以上两类插值振型函数,分析单自由度簧上质量匀速过桥时,桥梁的跨中位移、跨中梁底正应力和轮-桥接触力时程响应.结果表明:无论是一维问题还是二维问题,由三次样条插值法构造的插值振型函数与结构的实际振型较为吻合,计算结果具有较高的收敛性和精度.而要达到相同的精度,分段三次Hermite插值法则须加密单元网格,但其误差仅存在于独立网格内,不会累积放大.     

Finite element solution to passive scalar transport behind line sources under neutral and unstable stratification

This study employed a direct numerical simulation (DNS) technique to contrast the plume behaviours and mixing of passive scalar emitted from line sources (aligned with the spanwise direction) in neutrally and unstably stratified open‐channel flows. The DNS model was developed using the Galerkin finite element method (FEM) employing trilinear brick elements with equal‐order interpolating polynomials that solved the momentum and continuity equations, together with conservation of energy and mass equations in incompressible flow. The second‐order accurate fractional‐step method was used to handle the implicit velocity–pressure coupling in incompressible flow. It also segregated the solution to the advection and diffusion terms, which were then integrated in time, respectively, by the explicit third‐order accurate Runge–Kutta method and the implicit second‐order accurate Crank–Nicolson method. The buoyancy term under unstable stratification was integrated in time explicitly by the first‐order accurate Euler method. The DNS FEM model calculated the scalar‐plume development and the mean plume path. In particular, it calculated the plume meandering in the wall‐normal direction under unstable stratification that agreed well with the laboratory and field measurements, as well as previous modelling results available in literature. Copyright © 2005 John Wiley & Sons, Ltd.     

A collocated grid,projection method for time‐accurate calculation of low‐Mach number variable density flows in general curvilinear coordinates

A time‐accurate algorithm is proposed for low‐Mach number, variable density flows on curvilinear grids. Spatial discretization is performed on collocated grid that offers computational simplicity in curvilinear coordinates. The flux interpolation technique is used to avoid the pressure odd–even decoupling of the collocated grid arrangement. To increase the stability of the method, a two‐step predictor–corrector time integration scheme is employed. At each step, the projection method is used to calculate the hydrodynamic pressure and to satisfy the continuity equation. The robustness and accuracy of the method is illustrated with a series of numerical experiments including thermally driven cavity, polar cavity, three‐dimensional cavity, and direct numerical simulation of non‐isothermal turbulent channel flow. Copyright © 2012 John Wiley & Sons, Ltd.     

基于RBF插值技术的CFD/CSD非线性耦合分析方法研究

基于非结构混合网格的N-S方程求解器和结构柔度影响系数法,发展了一种考虑气动、结构非线性的基于RBF插值技术CFD/CSD耦合分析方法,适用于解决现代大展弦比飞机的非线性静气动弹性问题。该方法采用时间相关法（即求解非定常方程组,用长时间的渐近解趋于定常状态）求解静气弹分析时的定常流动。考虑大展弦比飞机结构变形问题为大变形小应力问题,在利用柔度系数法求解结构方程时,假设每次求解结构方程时应力与应变为线性关系,整体静气弹分析过程为非线性关系,因此每次求解结构方程时要更新柔度影响系数矩阵。在非定常N-S方程每求解一个时间步耦合一次结构有限元分析,由于结构有限元分析的时间相对于气动分析时间是很短的,所以这种方法实际上近似使用了一次求解非定常气动力的时间完成了整个静气动弹性分析的过程。对于气动网格与结构有限元网格不一致性,本文采用径向基函数（RBF）插值方法中的TPS方法进行结构弹性变形和气动载荷插值,采用虚功原理完成气动载荷数据交换。为了节省气弹分析时间,采用动网格方法对气动网格进行更新,本文基于RBF插值方法发展一种适用于混合网格（四面体、三棱柱、金字塔和六面体）变形的动网格方法,可以保证附面层网格的质量与分布从而准确模拟其流动。利用该方法对M6机翼、DLR-F6翼身组合体和某大型客机机翼进行了静气动弹性特性分析,结果验证了本文开发的非线性CFD/CSD耦合分析方法的可行性、精确性和高效性。     

On the use of characteristic‐based split meshfree method for solving flow problems

This study presents characteristic‐based split (CBS) algorithm in the meshfree context. This algorithm is the extension of general CBS method which was initially introduced in finite element framework. In this work, the general equations of flow have been represented in the meshfree context. A new finite element and MFree code is developed for solving flow problems. This computational code is capable of solving both time‐dependent and steady‐state flow problems. Numerical simulation of some known benchmark flow problems has been studied. Computational results of MFree method have been compared to those of finite element method. The results obtained have been verified by known numerical, analytical and experimental data in the literature. A number of shape functions are used for field variable interpolation. The performance of each interpolation method is discussed. It is concluded that the MFree method is more accurate than FEM if the same numbers of nodes are used for each solver. Meshfree CBS algorithm is completely stable even at high Reynolds numbers. Copyright © 2007 John Wiley & Sons, Ltd.     

Improved linear interpolation practice for finite‐volume schemes on complex grids

Methods for the computation of flow problems based on finite‐volume discretizations and pressure‐correction methods frequently require the interpolation of control volume face values from nodal values. The simple, often employed central differencing scheme (CDS) leads to a significant loss in accuracy when the numerical grid is non‐regular as it is usual when modelling complex geometries. An alternative technique based on a multi‐dimensional Taylor series expansion (TSE) is proposed, which preserves the CDS‐like sparsity pattern of the discrete system. While the TSE scheme computationally is only slightly more expensive than the CDS approach, it results in a significantly higher accuracy, where the difference increases with the grid irregularity. The method is investigated and compared to the CDS approach for some representative test cases. Copyright © 2002 John Wiley & Sons, Ltd.     

Modal decomposition‐based global stability analysis for reduced order modeling of 2D and 3D wake flows

The method for computation of stability modes for two‐ and three‐dimensional flows is presented. The method is based on the dynamic mode decomposition of the data resulting from DNS of the flow in the regime close to stable flow (fixed‐point dynamics, small perturbations about steady flow). The proposed approach is demonstrated on the wake flows past a 2D, circular cylinder, and a sphere. The resulting modes resemble the eigenmodes computed conventionally from global stability analysis and are used in model order reduction of the flow. The designed low‐dimensional Galerkin model uses continuous mode interpolation between dynamic mode decomposition mode bases and reproduces the dynamics of Navier–Stokes equations. Copyright © 2015 John Wiley & Sons, Ltd.     

Large eddy simulation and a simple wall model for turbulent flow calculation by a particle method

A turbulent channel flow and the flow around a cubic obstacle are calculated by the moving particle semi‐implicit method with the subparticle‐scale turbulent model and a wall model, which is based on the zero equation RANS (Reynolds Averaged Navier‐Stokes). The wall model is useful in practical problems that often involve high Reynolds numbers and wall turbulence, because it is difficult to keep high resolution in the near‐wall region in particle simulation. A turbulent channel flow is calculated by the present method to validate our wall model. The mean velocity distribution agrees with the log‐law velocity profile near the wall. Statistical values are also the same order and tendency as experimental results with emulating viscous layer by the wall model. We also investigated the influence of numerical oscillations on turbulence analysis in using the moving particle semi‐implicit method. Finally, the turbulent flow around a cubic obstacle is calculated by the present method to demonstrate capability of calculating practical turbulent flows. Three characteristic eddies appear in front of, over, and in the back of the cube both in our calculation and the experimental result that was obtained by Martinuzzi and Tropea. Mean velocity and turbulent intensity profiles are predicted in the same order and have similar tendency as the experimental result. Copyright © 2012 John Wiley & Sons, Ltd.     

One‐dimensional simulation of supercritical flow at a confluence by means of a nonlinear junction model applied with the RKDG2 method

We investigate the one‐dimensional computation of supercritical open‐channel flows at a combining junction. In such situations, the network system is composed of channel segments arranged in a branching configuration, with individual channel segments connected at a junction. Therefore, two important issues have to be addressed: (a) the numerical solution in branches, and (b) the internal boundary conditions treatment at the junction. Going from the advantageous literature supports of RKDG methods to a particular investigation for a supercritical benchmark, the second‐order Runge–Kutta discontinuous Galerkin (RKDG2) scheme is selected to compute the water flow in branches. For the internal boundary handling, we propose a new approach by incorporating the nonlinear model derived from the conservation of the momentum through the junction. The nonlinear junction model was evaluated against available experiments and then applied to compute the junction internal boundary treatment for steady and unsteady flow applications. Finally, a combining flow problem is defined and simulated by the proposed framework and results are illustrated for many choices of junction angles. Copyright © 2007 John Wiley & Sons, Ltd.