Localized Basis Function Method for Computing Limit Cycle Oscillations

An alternate approach to the standard harmonic balance method (based on Fourier transforms) is proposed. The proposed method begins with an idea similar to the harmonic balance method, i.e. to transform the initial set of differential equations of the dynamics to a set of discrete algebraic equations. However, as distinct from previous harmonic balance techniques, the proposed method uses a set of basis functions which are localized in time and are not necessarily sinusoidal. Also as distinct from previous harmonic balance methods, the algebraic equations obtained after the transformation of the differential equations of the dynamics are solved in the time domain rather than the frequency domain. Numerical examples are provided to demonstrate the performance of the method for autonomous and forced dynamics of a Van der Pol oscillator.     

Numerical modelling of thermal radiation absorption during the dispersion of dense gas clouds in the atmosphere

A numerical model is presented for the prediction of thermal radiation absorption in dense gas clouds during accidental release in the atmosphere. The model is based on the discrete transfer method (DT method) that was originally developed for modelling radiation in combustion chambers. The radiation model assumes a number of representative rays of predetermined orientation fired from each of the domain boundaries. Each ray is traced through the domain elements until reaching another boundary where it is terminated. Radiation sources are calculated for each element by performing an energy balance across each element for each ray passing through it. The energy sources recovered are fed into the finite element flow solver for every time step in the energy equation. The model proved accurate, and memory and computer time efficient. It showed that accounting for radiation effects lead to improved predictions. It also showed that in certain scenarios, radiation effects could be predominant. © 1998 John Wiley & Sons, Ltd.     

非定常流动模拟的时间离散方法

对于不同非定常流动问题, 采用合适的时间离散方法,可有效提高数值精度和计算效率. 本文在总结传统时间离散方法的基础上,对近些年发展的非线性频域法、谐波平衡法、经典时间谱方法、时间谱元法、时间有限差分法等进行了系统地总结.根据离散形式的不同,将上述方法分为时域推进法、频域谐波法、时域配点法和混合方法4大类.首先简要介绍了各类方法的数学思想以及研究进展,并重点比较了(准)周期性非定常流动计算中各方法的精度、效率以及适用范围.然后, 对各种时间离散格式的特点进行总结,并就不同的非定常流动问题如何选择合适的时间离散方法给予了建议.最后, 对这些新型时间离散格式在工程中的应用进行了简要介绍,并对其发展方向进行展望.      

Time discretization methods in the computation of unsteady flow

对于不同非定常流动问题, 采用合适的时间离散方法,可有效提高数值精度和计算效率. 本文在总结传统时间离散方法的基础上,对近些年发展的非线性频域法、谐波平衡法、经典时间谱方法、时间谱元法、时间有限差分法等进行了系统地总结.根据离散形式的不同,将上述方法分为时域推进法、频域谐波法、时域配点法和混合方法4大类.首先简要介绍了各类方法的数学思想以及研究进展,并重点比较了(准)周期性非定常流动计算中各方法的精度、效率以及适用范围.然后, 对各种时间离散格式的特点进行总结,并就不同的非定常流动问题如何选择合适的时间离散方法给予了建议.最后, 对这些新型时间离散格式在工程中的应用进行了简要介绍,并对其发展方向进行展望.     

Evolution of governing mass and momentum balances following an abrupt pressure impact in a porous medium

A mathematical model is developed of an abrupt pressure impact applied to a compressible fluid flowing through a porous medium domain. Nondimensional forms of the macroscopic fluid mass and momentum balance equations yield two new scalar numbers relating storage change to pressure rise. A sequence of four reduced forms of mass and momentum balance equations are shown to be associated with a sequence of four time periods following the onset of a pressure change. At the very first time period, pressure is proven to be distributed uniformly within the affected domain. During the second time interval, the momentum balance equation conforms to a wave form. The behavior during the third time period is governed by the averaged Navier-Stokes equation. After a long time, the fourth period is dominated by a momentum balance similar to Brinkman's equation which may convert to Darcy's equation when friction at the solid-fluid interface dominates.     

弹性动力学的多变量响应问题的解法

基于瞬时混合变分原理与乘积型二元三次 B样条函数 ,以板壳为例 ,建立样条动力方程。引入样条参数及其对时间的导数作为状态变量 ,导出状态方程。对空间域采用混合样条元法 ,对时间域采用现代控制论中的状态空间法。文末数值算例表明 ,计算精度与效率是令人满意的。本文方法对计算多输入与多输出 ,时不变与时变系统和线性与非线性系统等多变量动力响应问题 ,有广阔的应用与发展前景     

Simulation of two-dimensional elastodynamic problems using a new adaptive physics-based method

In this paper, an adaptive physics-based method is developed for solving wave motion problems in two dimensions (i.e., lamb waves). The solution of the problem has two main parts. In the first part, after discretization of the domain, a physics-based method is developed considering the conservation of mass and the balance of momentum. In the second part, adaptive points are determined using wavelet theory. This part is well done using Deslauries–Dubuc (D–D) like wavelets. After solving the problem in the first step, the domain of the problem is discretized by the same cells attending loading and characteristics of the structure. After the first trial solution, the D–D interpolation shows the lack and redundancy of points in the domain. These points will be added or eliminated for the next solution. This process may be repeated for achieving adaptive mesh for each step. Finally, the results of the proposed method are compared with the results available in the literature. This comparison shows excellent agreement between the obtained results and those already reported.     

连体结构基于时域显式法的瞬时最优控制算法

本文针对受到非平稳地震激励作用下的连体结构提出一种基于时域显式法的瞬时最优控制算法。该算法可以利用有限元模型施加脉冲荷载快速建立受控模型,可以避免系统状态部分可观测的问题,而且可以根据需要单独调控子结构受控效果的优点。一连体高层建筑结构算例用来说明该算法,并将该算法与传统瞬时最优控制算法比较,结果表明本文算法具有良好的控制效果。     

Numerical solution of solidification in a square prism using an algebraic grid generation technique

The solidification of an infinitely long square prism was analyzed numerically. A front fixing technique along with an algebraic grid generation scheme was used, where the finite difference form of the energy equation is solved for the temperature distribution in the solid phase and the solid–liquid interface energy balance is integrated for the new position of the moving solidification front. Results are given for the moving solidification boundary with a circular phase change interface. An algebraic grid generation scheme was developed for two-dimensional domains, which generates grid points separated by equal distances in the physical domain. The current scheme also allows the implementation of a finer grid structure at desired locations in the domain. The method is based on fitting a constant arc length mesh in the two computational directions in the physical domain. The resulting simultaneous, nonlinear algebraic equations for the grid locations are solved using the Newton-Raphson method for a system of equations. The approach is used in a two-dimensional solidification problem, in which the liquid phase is initially at the melting temperature, solved by using a front-fixing approach. The difference of the current study lies in the fact that front fixing is applied to problems, where the solid–liquid interface is curved such that the position of the interface, when expressed in terms of one of the coordinates is a double valued function. This requires a coordinate transformation in both coordinate directions to transform the complex physical solidification domain to a Cartesian, square computational domain. Due to the motion of the solid–liquid interface in time, the computational grid structure is regenerated at every time step.     

Non-linear free periodic vibrations of variable stiffness composite laminated plates

Steady-state free vibrations, with large amplitude displacements, of variable stiffness composite laminated plates (VSCL) are analysed. The intentions of this research are: (1)?to find out how the natural frequencies and (mode) shapes evolve with the displacement amplitude in this new type of laminated composite material; (2)?to describe modal interactions in VSCL due to energy interchanges under the coupling induced by non-linearity; (3)?to compare the VSCL with traditional, constant stiffness, laminated plates. The VSCL of interest here have curvilinear fibres and the numerical analysis carried out is based on a recently developed p-version finite element with hierarchic basis functions. The element follows first-order shear deformation theory and considers Von Kármán??s non-linear terms. The time domain equations of motion are first reduced using the linear modes of vibration and then transformed to the frequency domain via the harmonic balance method. These frequency domain equations are solved by an arc-length continuation method.     

考虑膨胀力的非饱和介质热-水-应力耦合二维有限元分析

从建立应力平衡方程、水连续性方程、能量守恒方程和弹塑性矩阵入手,使用Galerk in方法,将各控制方程分别在空间域和时间域进行离散,开发出了一个可考虑膨胀力的用于分析非饱和介质中热-水-应力耦合弹塑性问题的二维有限元程序.通过对一个假定的核废料地下处置库的热-水-应力耦合问题的数值计算,比较了无、有膨胀力时的情况,在定性上验证了该程序的正确性.     

A time domain collocation method for obtaining the third superharmonic solutions to the Duffing oscillator

In this study, a simple time domain collocation method (TDC) is applied to investigate the third superharmonic solutions of the Duffing oscillator. Upon using the proposed scheme, the multivaluedness, jump phenomenon, and transitional region of the third superharmonic response are explored. The amplitude frequency response curves for various values of damping, nonlinearity, and external force are obtained and compared. In addition, instead of collocating at N points so that the resulting nonlinear algebraic system is well determined, we extend the time domain collocation method to a new version by collocating at M>N points. The resulting over determined system is solved by the least square method. The extended time domain collocation method can significantly relieve the nonphysical solution phenomenon, which may be severe in the time domain collocation method, and its equivalent high dimensional harmonic balance method. Finally, numerical examples confirm the simplicity, efficiency, and accuracy of the proposed scheme.     

Fictitious domain methods for two‐phase flow energy balance computations in nuclear components

This paper is dedicated to the numerical simulation of nuclear components (cores and steam generators) by fictitious domain methods. The fictitious domain approach consists in immersing the physical domain under study in a Cartesian domain, called the fictitious domain, and in performing the numerical resolution on this fictitious domain. The calculation times are then efficiently reduced by the use of fast solvers. In counterpart, one has to handle with an immersed boundary, generally non‐aligned with the Cartesian mesh, which can be non‐trivial. The two fictitious domain methods compared here on industrial simulations and developed by Ramière et al. deal with an approximate immersed interface directly derived from the uniform Cartesian mesh. All the usual immersed boundary conditions (Dirichlet, Robin, Neumann), possibly mixed, are handled through a unique formulation of the fictitious problem. This kind of approximation leads to first‐order methods in space that exhibit a good ratio of the precision of the approximate solution over the CPU time, which is very important for industrial simulations. After a brief recall of the fictitious domain method with spread interface (Ramière et al., CMAME 2007) and the fictitious domain method with immersed jumps (Ramière et al., JCP 2008), we will focus on the numerical results provided by these methods applied to the energy balance equation in a steam generator. The advantages and drawbacks of each method will be pointed out. Generally speaking, the two methods confirm their very good efficiency in terms of precision, convergence, and calculation time in an industrial context. Copyright © 2011 John Wiley & Sons, Ltd.     

Stability analysis of duffing oscillator with time delayed and/or fractional derivatives

The periodic motions of the fractional order and/or delayed nonlinear systems are investigated in the frequency domain using a harmonic balance method with the analytical gradients of the nonlinear quality constraints and the sensitivity information of the Fourier coefficients can also obtained. The properties of fractional order derivatives and trigonometric functions are utilized to construct the fractional order derivatives, delayed and product operational matrices. The operational matrices are used to derive the analytical formulae of nonlinear systems of algebraic equations. The stability of periodic solutions for the delayed nonlinear systems is identified by an eigenvalue analysis of quasi-polynomials characteristic equations. Sensitivity analysis is performed to study the influence of the structural parameters on the system responses. Finally, three numerical examples are presented to illustrate the validity and feasibility of the developed method. It is concluded that the proposed methodology has the potential to facilitate highly efficient optimization, as well as sensitivity and uncertainty analysis of nonlinear systems with fractional derivatives and/or time delayed.     

An electromechanical atomic-scale finite element method for simulating evolutions of ferroelectric nanodomains

In this paper, a novel atomic-level computational method of perovskite ferroelectrics is established by combining the shell model and atomic-scale finite element method (AFEM). Its applicability is carefully testified for both bulk and nanoscale ferroelectrics, by comparing the calculated structural parameters and polarizations with the molecular dynamics (MD) simulations, first-principles calculations and experiment results. A comparison of the CPU time demonstrates that the developed method has a computational speed about 10 times over that of shell model MD method and its advantage becomes more evident as the computational scale becomes larger. Moreover, two effective calculation skills of long-range Coulomb force are introduced which can further enhance the computational efficiency by about 10 times. Using the developed atomic-level method, we investigate the various patterns of nanoscale domain structures in BaTiO₃ and their evolutions under electrical loadings. A domain structure with coexistence of vortex and streamline polarization patterns is revealed. Furthermore, the simulations of domain evolutions not only reproduce well the two-step 90° domain switching process observed in experiments on single domain under an anti-parallel electric field, but also provide a full evolution diagram among different domain patterns under various electric fields. A quantitative analysis indicates that the direction-dependent coercive field of multi-domain structure can be well described by that of single domain based on a simple analytical model. This study on domain patterns and evolutions may help us understand the behaviors of ferroelectrics from the atomic level.     

Approximate Group Analysis and Multiple Time Scales Method for the Approximate Boussinesq Equation

This paper is devoted to investigation of the approximate Boussinesq equation by methods of the approximate symmetry analysis of partial differential equations with a small parameter developed by Baikov, Gazizov and Ibragimov. We combine these methods with the method of multiple time scales to extend the domain of definition of approximate group invariant solutions of the approximate Boussinesq equation.     

An explicit time domain solution for ground stratum response to harmonic moving load

Based on the thin layer method originally proposed in frequency domain, an explicit time domain semi–analytical solution has been developed for simulating three-dimensional layered ground responses to harmonic moving loads. The Fourier–Laplace transforms were applied to derive the transformed solution that satisfied the boundary conditions of horizontal infinities. The eigenvalue decomposition was performed with respect to Laplace parameter to express the ground motion corresponding to the eigenmodes. The formulation for each eigenmode incorporating the moving load expression was transformed back into time domain analytically, and the global system responses were given by means of the general mode superposition method. The proposed explicit time domain solution is suitable for studying various types of moving load acting on or inside the ground. In this paper a moving harmonic load with rectangular distribution was adopted to demonstrate the ground response simulation. Two illustrative examples for moving load with speeds below or above the ground Rayleigh wave velocity were presented to test the computational accuracy and efficiency of the proposed approach. A parametric study was also performed to investigate the influences of soil properties on the ground responses.The project is partially supported by the National Natural Science Foundation of China (50538010). The English text was polished by Yunming Chen.     

弹性力学中的一种非协调数值流形方法

通过引入数学和物理双重网格，将插值域与 积分域分别定义在不同的覆盖上，即在数学网格上进行插值函数的构造，物理网格上完成 系统能量泛函积分运算，最后通过覆盖权函数将二者联结在一起. 它的优点是单元网格划 分随意，不受复杂边界形状和二相材料界面的限制，单元可以是任意形状，是较之于有限 元方法更一般的数值模拟方法. 在4节点四边形数值流形方法中，由于单元总体位移函数 包含的完全多项式不完全，使得计算精度不够精确，为此，在单元总体位移函数上附 加非协调位移基本项，使之趋于完全，提出了弹性力学问题的一种改进的数值流形 方法------非协调数值流形方法. 通过内部自由度静力凝聚处理，导出了消除内参后的单元应变矩阵 和单元刚度矩阵，使得在不增加广义节点自由度的前提下，大大提高了数值流形方法的计 算精度和计算效率. 同时对非协调项进行了显式处理，可以对工程实践起到更切实的帮助. 数值试验表明，它们能够保证收敛，有较高的精度，对畸变不敏感，从而证明了该方法的 可行性.     

A Hybrid Domain Decomposition and Multigrid Method for the Acceleration of Compressible Viscous Flow Calculations on Unstructured Triangular Meshes

This paper is concerned with the formulation and the evaluation of a hybrid solution method that makes use of domain decomposition and multigrid principles for the calculation of two-dimensional compressible viscous flows on unstructured triangular meshes. More precisely, a non-overlapping additive domain decomposition method is used to coordinate concurrent subdomain solutions with a multigrid method. This hybrid method is developed in the context of a flow solver for the Navier-Stokes equations which is based on a combined finite element/finite volume formulation on unstructured triangular meshes. Time integration of the resulting semi-discrete equations is performed using a linearized backward Euler implicit scheme. As a result, each pseudo time step requires the solution of a sparse linear system. In this study, a non-overlapping domain decomposition algorithm is used for advancing the solution at each implicit time step. Algebraically, the Schwarz algorithm is equivalent to a Jacobi iteration on a linear system whose matrix has a block structure. A substructuring technique can be applied to this matrix in order to obtain a fully implicit scheme in terms of interface unknowns. In the present approach, the interface unknowns are numerical fluxes. The interface system is solved by means of a full GMRES method. Here, the local system solves that are induced by matrix-vector products with the interface operator, are performed using a multigrid by volume agglomeration method. The resulting hybrid domain decomposition and multigrid solver is applied to the computation of several steady flows around a geometry of NACA0012 airfoil.     

Nonlinear dynamics of mistuned bladed disks with ring dampers

In turbomachinery applications, rotating bladed disks (blisks) are often subject to high levels of dynamic loading, such as traveling wave excitations, which result in large response amplitudes at resonance. To prevent premature high cycle fatigue, various dry friction dampers are designed for blisk systems to reduce the forced responses. Ring dampers are located in the disk, underneath the blades, and are held in contact with the blisk by centrifugal loading. Energy is dissipated by nonlinear friction forces when relative motions between the ring damper and the blisk take place. To investigate the dynamic responses of blisk–damper systems in the presence of the nonlinear frictional contacts, conventional methods based on numerical time integration are not suitable since they are computationally expensive. This paper presents a reduced-order modeling technique to efficiently capture the nonlinear dynamic responses of the blisk–damper systems. Craig–Bampton component mode synthesis (CB-CMS) serves as the first model reduction step. A novel mode basis that mimics the contact behavior under sliding and sticking conditions is developed to further reduce the CB-CMS model while maintaining its accuracy. The resulting reduced nonlinear equations of motion are solved by a hybrid frequency/time domain (HFT) method. In the HFT method, the contact status and friction forces are determined in the time domain by a three-dimensional contact model at each contact point, whereas the reduced equations of motion are solved in the frequency domain according to a harmonic balance formulation. Moreover, to investigate the effects of blade mistuning, which can lead to drastic increase of forced responses, an extension of the reduced-order models (ROMs) is developed based on component mode mistuning. Forced responses computed by the proposed ROMs are validated for both tuned and mistuned systems. A statistical analysis is performed to study the effectiveness of ring dampers under random blade mistuning patterns.