Analysis of an upscaling method based on conservation of dissipation

In this paper, a scaling method based on conservation of dissipation and use of periodic boundary conditions is presented. We prove that the method leads to a symmetric positive definite tensor. We also show that the method is identical to the method of Durlovsky and Chung, therefore an important property of the latter method is proved. Some other existing methods are also discussed in terms of their boundary conditions. For this purpose, the concepts of basis and class of boundary conditions for scaling methods are introduced.     

Non-periodic finite-element formulation of orbital-free density functional theory

We propose an approach to perform orbital-free density functional theory calculations in a non-periodic setting using the finite-element method. We consider this a step towards constructing a seamless multi-scale approach for studying defects like vacancies, dislocations and cracks that require quantum mechanical resolution at the core and are sensitive to long range continuum stresses. In this paper, we describe a local real-space variational formulation for orbital-free density functional theory, including the electrostatic terms and prove existence results. We prove the convergence of the finite-element approximation including numerical quadratures for our variational formulation. Finally, we demonstrate our method using examples.     

Spectral methods based on the least dissipative modes for wall bounded MHD flows

We present a new approach for the Spectral Direct Numerical Simulation (DNS) of Low-Rm wall-bounded magnetohydrodynamic (MHD) flows. The novelty is that instead of using bases similar to the usual Chebyshev polynomials, which are easy to implement but incur heavy computational costs to resolve the Hartmann boundary layers that arise along the walls, we use a basis made of elements that already incorporate flow structures such as anisotropic vortices and Hartmann layers. We show that such a basis can be obtained from the eigenvalue problem of the linear part of the governing equations with the problem’s boundary conditions. Since this basis is not always orthogonal, we develop a spectral method for non-orthogonal bases. We then demonstrate the efficiency of this method on the simple case of a laminar channel flow with periodic forcing. In particular, we show that this method eliminates the computational costs incurred this Hartmann layer, and this for arbitrary high magnetic fields B. We then discuss the application of our method to nonlinear, turbulent flows for which the number of modes required to resolve the flow completely decreases strongly when B increases, instead of increasing as in the case of currently employed Chebyshev-based methods.     

A formula of solution to the integral of rational functions

ＡＦＯＲＭＵＬＡＯＦＳＯＬＵＴＩＯＮＴＯＴＨＥＩＮＴＥＧＲＡＬＯＦＲＡＴＩＯＮＡＬＦＵＮＣＴＩＯＮＳＧｕｉＺｕ－ｈｕａ（桂祖华）（ＴｈｅＤｅｐａｒｔｍｅｎｔｏｆＡｐｐｌｉｅｄＭａｔｈｅｍａｔｉｃｓ,ＳｈａｎｇｈａｉＪｉａｏ－ｔｏｎｇＵｎｉｖｅｒｓｉｔ...     

Inverse electromagnetic scattering for a locally perturbed perfectly conducting plate

We consider a simple but fully three-dimensional inverse problem to determine the shape of a local perturbation of a perfectly conducting plate from far-field measurements of time harmonic electromagnetic fields. For this purpose we reformulate the model problem as an exterior Maxwell problem for a symmetric domain, and prove an equivalence between the model problem and its reformulation. Then, linear sampling method is applied to solve the reformulated problem. We illustrate the feasibility of this method by some numerical examples.     

Further numerical methods for the Falkner-Skan equations: Shooting and continuation techniques

We consider in this paper the numerical solution of the Falkner-Skan differential equation, modelling under some similarity assumptions the boundary layer equation. We look for the extremal solution of this third order differential equation. The methods we use are basically the Newton method with a shooting process, which is coupled with a continuation method: they allow us to follow the solution arcs which contain regular and turning point solutions.     

On one class of differential equations of fractional order

We consider the Darboux problem for a differential equation of fractional order that contains a regularized mixed derivative. Sufficient conditions for the existence and uniqueness of a solution of this problem are obtained in the class of continuous functions. We also propose a method for finding an approximate solution of this problem and prove the convergence of this method.     

Convergence analysis of a finite element method based on different moduli in tension and compression

When analyzing materials that exhibit different mechanical behaviors in tension and compression, an iterative approach is required due to material nonlinearities. Because of this iterative strategy, numerical instabilities may occur in the computational procedure. In this paper, we analyze the reason why iterative computation sometimes does not converge. We also present a method to accelerate convergence. This method is the introduction of a new pattern of shear modulus that was strictly derived according to the constitutive model based on the bimodular elasticity theory presented by Ambartsumyan. We test this procedure with a numerical example concerning a plane stress problem. Results obtained from this example show that the proposed method reduces the cost of computation and accelerates the convergence of the solution.     

Characteristic fast marching method on triangular grids for the generalized eikonal equation in moving media

The governing equation of the first arrival time of a monotonically propagating front (wavefront or shock front) in an inhomogeneous moving medium is an anisotropic eikonal equation, called the generalized eikonal equation in moving media. When the ambient medium is at rest, this equation reduces to the well-known (isotropic) eikonal equation in which the characteristic direction coincides with the normal direction of the propagating front. The fast marching method is an efficient method for computing the first arrival time of a propagating front as the approximate solution of the isotropic eikonal equation. The fast marching method inherits the property that the characteristic direction coincides with the normal direction at every point on the propagating wavefront and therefore is well suited for the eikonal equation. Due to anisotropic nature, this property does not hold in the case of front propagation in a moving medium. Thus, the fast marching method cannot be directly used for the generalized eikonal equation and needs some suitable modifications. We recently proposed a characteristic fast marching method on a rectangular grid for the generalized eikonal equation (Dahiya et al., 2013) and shown numerically that this method is stable, accurate, and easy to update to second order approximations. In the present work, we generalize the method on structured triangular grids. We compare the numerical solution obtained using our method with the ray theory solution to show that the method captures accurately the viscosity solution of the generalized eikonal equation. We use the method to study some interesting geometrical features of an initially planar wavefront propagating in a medium with Taylor–Green type vortices.     

离散元与壳体有限元结合的多尺度方法及其应用

在深入研究复杂结构和非均质材料冲击响应和破坏机理的过程中,往往遇到多尺度计算问题。本文尝试建立三维离散元与壳体有限元结合的多尺度方法用于处理圆柱壳问题,该方法采用三维离散元对感兴趣的局域进行局部模拟,利用平板壳体有限元进行整体模拟,采用一种特殊的过渡层使离散元区和有限元区能很好的衔接。我们将这一方法应用于激光辐照下充压柱壳的热/力耦合冲击破坏响应,得到的模拟结果与文献报道有较好的吻合。     

Analysis of Newton’s Method to Compute Travelling Waves in Discrete Media

We present a variant of Newton’s method for computing travelling wave solutions to scalar bistable lattice differential equations. We prove that the method converges to a solution, obtain existence and uniqueness of solutions to such equations with a small second order term and study the limiting behaviour of such solutions as this second order term tends to zero. The robustness of the algorithm will be discussed using numerical examples. These results will also be used to illustrate phenomena like propagation failure, which are encountered when studying lattice differential equations. We finish by discussing the broad application range of the method and illustrate that higher dimensional systems exhibit richer behaviour than their scalar counterparts.     

On higher order multigrid methods with application to a geothermal reservoir model

This paper considers the multigrid iterative method applied to the solution of finite difference approximations to a linear second-order self-adjoint elliptic equation. It represents an extension of work by Dinar and Brandt. We compare two methods to obtain fourth-order convergence. The first is local error extrapolation developed by Brandt, the second is iterative improvement developed by Lindberg. This work considers non-separable problems, but only on a rectangular domain with Dirichlet boundary conditions. We consider test cases with non-smooth (i.e. discontinuous second derivatives) as well as smooth solutions. We also apply the multigrid method to an elliptic equation with non-separable coefficients which occurs in a geothermal model. In this case an analysis of the error fails to show any advantage in a fourth-order difference scheme over a second-order scheme. However, we do demonstrate that the multigrid iteration performs well on this problem. Also, this example shows that the multigrid iteration can be combined with iterative improvement to create an efficient fourth-order method for a non-separable elliptic equation which is coupled with a marching equation. Other work has found an advantage in this fourth-order scheme for a similar geothermal model.     

A field theoretical approach to the quasi-continuum method

The quasi-continuum method has provided many insights into the behavior of lattice defects in the past decade. However, recent numerical analysis suggests that the approximations introduced in various formulations of the quasi-continuum method lead to inconsistencies—namely, appearance of ghost forces or residual forces, non-conservative nature of approximate forces, etc.—which affect the numerical accuracy and stability of the method. In this work, we identify the source of these errors to be the incompatibility of using quadrature rules, which is a local notion, on a non-local representation of energy. We eliminate these errors by first reformulating the extended interatomic interactions into a local variational problem that describes the energy of a system via potential fields. We subsequently introduce the quasi-continuum reduction of these potential fields using an adaptive finite-element discretization of the formulation. We demonstrate that the present formulation resolves the inconsistencies present in previous formulations of the quasi-continuum method, and show using numerical examples the remarkable improvement in the accuracy of solutions. Further, this field theoretic formulation of quasi-continuum method makes mathematical analysis of the method more amenable using functional analysis and homogenization theories.     

定向爆破治理滑坡

本文介绍了用定向爆破治理滑坡新技术。研究了在不稳定土体上进行爆破的许多特殊问题,探讨了如何防止爆破瞬间可能引起的诱发滑坡问题,并将此技术用于工程实践中,获得成功。为治理滑坡提供了一种经济、快速、有效的新方法。     

A swept intersection‐based remapping method for axisymmetric ReALE computation

We use here a reconnection ALE (ReALE) strategy to solve hydrodynamic compressible flows in cylindrical geometries. The main difference between the classical ALE and the ReALE method is the rezoning step where we allow change in the topology. This leads for ReALE to a polygonal mesh, which follows more efficiently the flow. We present here a new displacement of generators in order to keep the Lagrangian features, which are usually lost using ALE with fixed topology. The reconnection capability allows to deal with complex geometries and high‐vorticity problems contrary to ALE method. The main difficulty of ReALE is the remapping step where we have to remap physical variables on a mesh with a different topology. For this step, a new remapping method based on a swept intersection algorithm has been developed in the case of planar geometries. We present here the extension of the swept intersection‐based remapping method to cylindrical geometries. We demonstrate that our method can be applied to several numerical examples up to problem representative of hydrodynamic experiments. Copyright © 2015 John Wiley & Sons, Ltd.     

Brinkmann Model and Double Penalization Method for the Flow Around a Porous Thin Layer

In this paper we study a penalization method used to compute the flow of a viscous fluid around a thin layer of porous material. Using a BKW method, we perform an asymptotic expansion of the solution when a little parameter, measuring the thickness of the thin layer and the inverse of the penalization coefficient, tends to zero. We compare then this numerical method with a Brinkman model for the flow around a porous thin layer.      

A 3D multi-field element for simulating the electromechanical coupling behavior of dielectric elastomers

We propose a multi-field implicit finite element method for analyzing the electromechanical behavior of dielectric elastomers. This method is based on a four-field variational principle, which includes displacement and electric potential for the electromechanical coupling analysis, and additional independent fields to address the incompressible constraint of the hyperelastic material. Linearization of the variational form and finite element discretization are adopted for the numerical implementation. A general FEM program framework is developed using C ++ based on the open-source finite element library deal.II to implement this proposed algorithm. Numerical examples demonstrate the accuracy, convergence properties, mesh-independence properties, and scalability of this method. We also use the method for eigenvalue analysis of a dielectric elastomer actuator subject to electromechanical loadings. Our finite element implementation is available as an online supplementary material.     

Mixed boundary value problem of elasticity for a quarter space

We consider the static elasticity problem for a quarter space with zero displacements on one of its surfaces and with given stresses on the other. The method for solving this problem is based on the use of newunknown functions in the formof a linear combination of the desired displacements, which reduces the system of three Lamé equations to two equations to be solved simultaneously and one equation to be solved separately. The exact solution of this problem was obtained earlier by the same method [1]. But it was shown in [2] that such a solution is exact only under certain restrictions on the given functions. In the present paper, the solution of this problem is constructed without restrictions on the given functions, which necessitates solving a one-dimensional integro-differential equation; this can be done approximately by the orthogonal polynomial method. We present numerical results obtained on the basis of our solution.     

Momentum‐based approximation of incompressible multiphase fluid flows

We introduce a time stepping technique using the momentum as dependent variable to solve incompressible multiphase problems. The main advantage of this approach is that the mass matrix is time‐independent making this technique suitable for spectral methods. A level set method is applied to reconstruct the fluid properties such as density. We also introduce a stabilization method using an entropy‐viscosity technique and a compression technique to limit the flattening of the level set function. We extend our algorithm to immiscible conducting fluids by coupling the incompressible Navier‐Stokes and the Maxwell equations. We validate the proposed algorithm against analytical and manufactured solutions. Results on test cases such as Newton's bucket problem and a variation thereof are provided. Surface tension effects are tested on benchmark problems involving bubbles. A numerical simulation of a phenomenon related to the industrial production of aluminium is presented at the end of the paper.     

Differential-algebraic equations of programmed motions of Lagrangian dynamical systems

We suggest a method for constructing the dynamic equations of manipulator systems in canonical variables. The system of differential dynamic equations has an integral manifold corresponding to the holonomic and nonholonomic constraint equations. The controls are determined so as to ensure the stability of this manifold. We state conditions for the exponential stability of the manifold and for constraint stabilization when solving the dynamic equations numerically by a simplest difference method. We also present the solution of the problem of control of a plane two-link manipulator.