Partition Of Unity Finite Element Method Research Articles

Dynamic analysis of a viscoelastic orthotropic cracked body using the extended finite element method

The extended finite element method (XFEM) is found promising in approximating solutions to locally non-smooth features such as jumps, kinks, high gradients, inclusions, voids, shocks, boundary layers or cracks in solid or fluid mechanics problems. The XFEM uses the properties of the partition of unity finite element method (PUFEM) to represent the discontinuities without the corresponding finite element mesh requirements. In the present study numerical simulations of a dynamically loaded orthotropic viscoelastic cracked body are performed using XFEM and the J-integral and stress intensity factors (SIF’s) are calculated. This is achieved by fully (reproducing elements) or partially (blending elements) enriching the elements in the vicinity of the crack tip or body. The enrichment type is restricted to extrinsic mesh-based topological local enrichment in the current work. Thus two types of enrichment functions are adopted viz. the Heaviside step function replicating a jump across the crack and the asymptotic crack tip function particular to the element containing the crack tip or its immediately adjacent ones. A constitutive model for strain-rate dependent moduli and Poisson ratios (viscoelasticity) is formulated. A symmetric double cantilever beam (DCB) of a generic orthotropic material (mixed mode fracture) is studied using the developed XFEM code. The same problem is studied using the viscoelastic constitutive material model implemented in ABAQUS through an implicit user defined material subroutine (UMAT). The results from XFEM correlate well with those of the finite element method (FEM). Three cases viz. static, dynamic and viscoelastic dynamic are studied. It is shown that there is an increase in the value of maximum J-integral when the material exhibits strain rate sensitivity.

Numerical simulation of acoustic waves in air and poroelastic media using the partition of unity finite element method

Foams and fibrous materials are used in a large range of applications such as automotive or building acoustics. Their properties can be described with either a poroelastic model or an equivalent fluid model. These models, also used in geophysics, are now widely spread and are also available in some commercial finite element software applications. However, the discretization level required to achieve reasonable accuracy is not always acceptable in the mid-frequency range. In such case, the Partition of Unity Finite Element Method (PUFEM) using plane wave functions seems appropriate to avoid this limitation. The PUFEM has been recently applied to rigid frame materials with and without coupling with an acoustic cavity. It has also been applied efficiently to poroelastic materials. The present work focuses on the coupling between a poroelastic material, i.e., described with Biot's equations, and an air cavity. Some practical examples are tested to demonstrate the efficiency of the PUFEM for solving noise control problems at medium frequency, but also to underline the precautions that must be taken when dealing with an air-porous interface.

Performances of the Partition of Unity Finite Element Method for the analysis of two-dimensional interior sound fields with absorbing materials

The paper deals with the numerical simulation of the acoustic field in two-dimensional cavities in which absorbing materials are present. Though Finite Element Method (FEM) could be employed for this purpose, the discretization level required for achieving reasonable accuracy renders the method impractical in the mid-frequency range. To alleviate this limitation, the Partition of Unity Finite Element Method (PUFEM) using plane wave functions has been shown to be very effective for solving short wave Helmholtz problems. In the present work, the method is extended to the computation of the pressure wave field within the absorbing media which is modeled as a bulk-reacting material characterized by a complex-valued and frequency dependent mean density and dynamic compressibility. Lagrange multipliers are used to enforce the transmission conditions at the air–material interface. Performances of the PUFEM are compared with a standard FEM in several examples of practical interests. It is shown that the technique is a good candidate for solving noise control problems at high frequency.

EXTENSION OF THE PUFEM TO ELASTIC WAVE PROPAGATION IN LAYERED MEDIA

This work deals with the extension of the partition of unity finite element method (PUFEM) "(Comput. Meth. Appl. Mech. Eng.139 (1996) pp. 289–314; Int. J. Numer. Math. Eng.40 (1997) 727–758)" to solve wave problems involving propagation, transmission and reflection in layered elastic media. The proposed method consists of applying the plane wave basis decomposition to the elastic wave equation in each layer of the elastic medium and then enforce necessary continuity conditions at the interfaces through the use of Lagrange multipliers. The accuracy and effectiveness of the proposed technique is determined by comparing results for selected problems with known analytical solutions. Complementary results dealing with the modeling of pure Rayleigh waves are also presented where the PUFEM model incorporates information about the pressure and shear waves rather than the Rayleigh wave itself.

Wavelet based ILU preconditioners for the numerical solution by PUFEM of high frequency elastic wave scattering

Daubechies family of wavelets combined to the Incomplete Lower–Upper (ILU) factorization are considered as preconditioners for a block sparse linear system arising from the approximation of the time harmonic elastic wave equations by the Partition of Unity Finite Element Method (PUFEM). After applying the discrete wavelet transform (DWT) to each dense block in the final matrix and the known right-hand side, due to the local enrichment by pressure ( P) and shear ( S) plane waves, the resulting linear system is solved by the restarted Generalized Minimum RESidual method (GMRES) with ILU preconditioners allowing fill-in elements in the L and U matrix factors. A reordering algorithm of the vertices based on Gibbs method is also introduced. It leads to a significant reduction of the bandwidth of the wave based Finite Element (FE) matrix and enables the ILU preconditioners to be more effective. To study the performance of the proposed preconditioners, a problem of a horizontal S plane wave scattered by a rigid circular body in an infinite elastic medium is considered. The numerical tests show the good performance of the DWT based ILU preconditioners in improving the rate of convergence of GMRES, for high numbers of approximating P and S plane waves, on coarse mesh grids containing multi-wavelength sized elements. Moreover, the Haar DWT enhances the scalability with respect to the problem size, when the number of the nodal points increases, of the ILU preconditioner which uses the threshold strategy in the control of the fill-in elements. Despite the high level of the conditioning, a good accuracy may be achieved for a discretization level of around 1.9 degrees of freedom per S wavelength, which is far below the rule of thumb of 10 nodal points per wavelength, adopted for the conventional FE.

A q-adaptive partition of unity finite element method for the solution of the 2-D Helmholtz equation

The Partition of Unity Finite Element Method (PUFEM) is used to solve a wave scattering problem governed by the Helmholtz equation. The number q of the enriching plane waves is usually considered to be constant all over the computational domain which is a reasonable choice when a uniform mesh grid is considered. However, for nonuniform mesh grids where the element size may well vary throughout the computational domain this choice can considerably increase the condition number of the resulting linear system. It is shown that varying the number q as the element size varies may significantly reduce the required total number of degrees of freedom and improve the conditioning, when a fixed accuracy is considered.

Improvement of PUFEM for the numerical solution of high‐frequency elastic wave scattering on unstructured triangular mesh grids

AbstractThe purpose of this paper, which builds on previous work (Int. J. Numer. Meth. Engng 2009; 77:1646–1669), is to improve a numerical scheme based on the partition of unity finite element method (PUFEM) for the solution of the time harmonic elastic wave equations. The approach consists to approximate the displacement field by the standard finite element shape functions, enriched locally by superimposing pressure (P) and shear (S) plane waves. The aim is to accurately model two‐dimensional elastic wave problems on relatively coarse mesh grids, capable of containing many wavelengths per nodal spacing, for wide ranges of frequencies. This allows us to relax the traditional requirement of about 10 nodal points per S wavelength. In this work, an exact integration scheme for the linear triangular finite element is developed to evaluate the oscillatory integrals arising from the use of the PUFEM. The main contribution here consists in developing an explicit closed‐form solution for two‐dimensional wave‐based integrals, when the phase variation is linear in the local coordinate element system. The evaluation of the element mass matrix is performed from appropriate edge integrals. All other element matrices, obtained by adequate splitting of the element stress tensor matrix, are simply deduced from the element mass matrix entries.The results show clearly that the proposed integration scheme evaluates accurately the entries of the global matrix with drastic reduction of the computational time. Numerical tests dealing with the scattering of S elastic plane waves by a circular rigid body show that, for the same discretization level, it is possible to improve the accuracy by using large elements associated with high numbers of approximating plane waves rather than using small elements with less plane waves. However, this increases the conditioning and the fill‐in of the global matrix. At high frequency, it is even possible to push the number of degrees of freedom per S wavelength under 2 and still achieve good accuracy. Finally, some remarks on the choice of the numbers of P and S plane waves leading to better accuracy and conditioning are discussed. Copyright © 2010 John Wiley & Sons, Ltd.

A semianalytic meshless approach to the transient Fokker–Planck equation

In this paper, a semianalytic partition of unity finite element method (PUFEM) is presented to solve the transient Fokker–Planck equation (FPE) for high-dimensional nonlinear dynamical systems. Meshless spatial discretization of PUFEM with local p -refinement (discussed in a previous paper) is employed to develop linear ordinary differential equations for the time varying coefficients of local shape functions. A similarity transformation to modal coordinates is shown to reveal numerous spurious modes in the eigenspace of the discretized FPE operator. Identification and elimination of these modes leads to an analytical solution of ODEs obtained from spatial discretization in terms of the remaining admissible modes, and a significant reduction in the size of the discretized transient problem. Initial equation error resulting from the set of admissible modes is shown to be an upper bound for all time, meaning that the reduced admissible set is sufficient for the FPE approximation for all time.

Numerical analysis of two plane wave finite element schemes based on the partition of unity method for elastic wave scattering

This work deals with a numerical model based on the partition of unity finite element method (PUFEM), for the two-dimensional time harmonic elastic wave equations. This approach consists in enriching the polynomial finite element spaces locally by superimposed shear (S) and pressure (P) plane waves. Two plane wave Finite elements schemes are particularly investigated here and compared from the point of view of levels of accuracy and conditioning. The problem considered is a horizontal S plane wave scattered by a rigid circular body, in an infinite elastic medium. A finite square domain in the vicinity of the scatterer is considered with the analytical solution of the problem imposed on its boundary through a Robin type boundary condition. An error analysis with respect to the mesh size and the plane wave enrichment is carried out and some numerical aspects, related to the conditioning and its behaviour as a function of the frequency and the number of approximating plane waves, are outlined. The aim is to gain a better practical understanding of the conditioning behaviour of the PUFEM in order to improve its reliability. The results show, for the same number of degrees of freedom per wavelength, higher order elements lead to better accuracy than low order elements. However, this is achieved in detriment of the conditioning especially when the number of degrees of freedom per wavelength is too large.

A comparison of NRBCs for PUFEM in 2D Helmholtz problems at high wave numbers

In this work, exact and approximate non-reflecting boundary conditions (NRBCs) are implemented with the Partition of Unity Finite Element Method (PUFEM) to solve short wave scattering problems governed by the Helmholtz equation in two dimensions. By short wave problems, we mean situations in which the wavelength is a small fraction of the characteristic dimension of the scatterer. Various NRBCs are implemented and a comparison of their performance is carried out based on the accuracy of the results, ease of implementation and computational cost. The aim is to accurately model such problems in a reduced computational domain around the scatterer with fewer elements and without refining the mesh at each wave number.

Reformulation of XFEM based on PUFEM for solving problem caused by blending elements

A lack of numerical accuracy in the standard extended finite element method (XFEM) is caused by ‘blending elements’, whose nodes are partially enriched. ‘The corrected XFEM’ proposed by Fries showed the effective improvement of this problem with a lot of numerical results. The theoretical approach of this proposal was however not sufficiently described. In the present paper, an approximation of the XFEM is reformulated based on the concept of the partition of unity finite element method (PUFEM) approximation, which assures the numerical accuracy in the entire domain, for solving the problem of blending elements. The form of the reformulated XFEM results in the coincidence with that of the corrected XFEM. It is therefore found out that the theoretical validation of the corrected XFEM is based on the PUFEM approximation. It is also found out that the problem of the blending elements in the application to two dimensional linear fracture mechanics has been sufficiently solved for actual use by the XFEM based on the PUFEM.

The partition of unity finite element approach with hp-refinement for the stationary Fokker–Planck equation

In this paper, the stationary Fokker–Planck equation (FPE) is solved for nonlinear dynamical systems using a local numerical technique based on the meshless partition of unity finite element method (PUFEM). The method is applied to stationary FPE for 2-, 3- and 4-D systems and is argued to be an excellent candidate for higher dimensional problems and the transient problem. Local refinement is applied by introducing higher order polynomials in selected subdomains (local p -refinement) to keep the problem size small while ensuring high approximation accuracy. Various local approximations are blended using novel pasting functions that provide any desired order of continuity. Results are compared with existing global and local techniques. Local p -refinement is touted as an important step towards breaking the curse of dimensionality in numerical solution of FPE.

Computational Nonlinear Stochastic Control

Computational Nonlinear Stochastic Control

Structural dynamics of viscoelastic sandwich plates by the partition of unity finite element method

The scope of this research concerns the passive damping of structural vibrations by the use of viscoelastic layers. It is motivated by the need for efficient numerical tools to deal with the medium frequency behaviour of industrial viscoelastic sandwich products. The sandwich modelling technique is based on the use of an interface element: the two deformable plates are modelled by special plate elements while the intermediate dissipative layer is modelled with interface elements. This interface element is based on the first-order shear deformation theory and assume constant peel and shear stresses in the polymer thickness. This element couples the lower and upper layers without additional degrees of freedom. The partition of unity finite element method (PUFEM) is applied to the development of enriched Mindlin plate elements. The element shape functions are obtained as the product of partition of unity functions with arbitrary chosen enrichment functions. Polynomial enrichment leads to the generation of high-order polynomial shape functions and is therefore similar to a p-FEM technique. Numerical examples illustrate the use of both PUFEM Mindlin plate elements and interface elements for the simulation of viscoelastic sandwich structures.

Validation of 3D crack propagation in plain concrete -Part II: Computational modeling and predictions of the PCT3D test

The discrete crack-concept is applied to study the 3D propagation of tensile-dominated failure in plain concrete. To this end the Partition of Unity Finite Element Method (PUFEM) is utilized and the strong discontinuity approach is followed. A consistent linearized implementation of the PUFEM is combined with a predictor-corrector algorithm to track the crack path, which leads to a robust numerical description of concrete cracking. The proposed concept is applied to study concrete failure during the PCT3D test and the predicted numerical results are compared to experimental data. The proposed numerical concept provides a clear interface for constitutive models and allows an investigation of their impact on concrete cracking under 3D conditions, which is of significant scientific interests to interpret results from 3D experiments.

On solving singular interface problems using the enriched partition‐of‐unity finite element methods

It has been well recognized that interface problems often contain strong singularities which make conventional numerical approaches such as uniform h‐ or p‐version of finite element methods (FEMs) inefficient. In this paper, the partition‐of‐unity finite element method (PUFEM) is applied to obtain solution for interface problems with severe singularities. In the present approach, asymptotical expansions of the analytical solutions near the interface singularities are employed to enhance the accuracy of the solution. Three different enrichment schemes for interface problems are presented, and their performances are studied. Compared to other numerical approaches such as h‐p version of FEM, the main advantages of the present method include: easy and simple formulation; highly flexible enrichment configurations; no special treatment needed for numerical integration and boundary conditions; and highly effective in terms of computational efficiency. Numerical examples are included to illustrate the robustness and performance of the three schemes in conjunction with uniform h‐ or p‐refinements. It shows that the present PUFEM formulations can significantly improve the accuracy of solution. Very often, improved convergence rate is obtained through enrichment in conjunction with p‐refinement.

A modified element‐free Galerkin method with essential boundary conditions enforced by an extended partition of unity finite element weight function

AbstractIn this work we propose a method which combines the element‐free Galerkin (EFG) with an extended partition of unity finite element method (PUFEM), that is able to enforce, in some limiting sense, the essential boundary conditions as done in the finite element method (FEM). The proposed extended PUFEM is based on the moving least square approximation (MLSA) and is capable of overcoming singularity problems, in the global shape functions, resulting from the consideration of linear and higher order base functions. With the objective of avoiding the presence of singular points, the extended PUFEM considers an extension of the support of the classical PUFE weight function. Since the extended PUFEM is closely related to the EFG method there is no need for special approximation functions with complex implementation procedures, and no use of the penalty and/or multiplier method is required in order to approximately impose the essential boundary condition. Thus, a relatively simple procedure is needed to combine both methods. In order to attest the performance of the method we consider the solution of an analytical elastic problem and also some coupled elastoplastic‐damage problems. Copyright © 2003 John Wiley & Sons, Ltd.

An improved partition of unity finite element model for diffraction problems

AbstractThe partition of unity finite element method (PUFEM) is explored and improved to deal with practical diffraction problems efficiently. The use of plane waves as an external function space allows an efficient implementation of an approximate exterior non‐reflective boundary condition, improving the original proposed by Higdon for general diffraction problems. A ‘virtually’ analytical integration procedure is introduced for multi‐dimensional high‐frequency problems which exhibits a dramatic decrease in the number of operations for a given error compared with standard integration methods. Suitable conjugate gradient type solvers for the whole range of wavenumbers are used, including such cases in which PUFEM can produce nearly singular matrices caused by ‘round‐off’ limits. Copyright © 2001 John Wiley & Sons, Ltd.

The partition of unity finite element method: Basic theory and applications

The paper presents the basic ideas and the mathematical foundation of the partition of unity finite element method (PUFEM). We will show how the PUFEM can be used to employ the structure of the differential equation under consideration to construct effective and robust methods. Although the method and its theory are valid in n dimensions, a detailed and illustrative analysis will be given for a one-dimensional model problem. We identify some classes of non-standard problems which can profit highly from the advantages of the PUFEM and conclude this paper with some open questions concerning implementational aspects of the PUFEM.