Hare Rama Hare rama Rama Rama Hare Hare Hare Krishna Hare Krishna Krishna Krishna Hare

Hare

PSEUDO-ELASTIC & PARTITION OF UNITY FINITE ELEMENT METHODS AND THEIR APPLICATION IN MATERIAL NON-LINEAR PROBLEMS
SEMINAR REPORT Submitted By LEKSHMI DINACHANDRAN S1 M TECH MACHINE DESIGN ROLL NO. ME 12 MD 12 To The University of Kerala in partial fulfillment of the requirements for the award of the degree of Master of Technology in Mechanical Engineering (Machine Design)

Department of Mechanical Engineering SCT College of Engineering, Thiruvananthapuram 18 JULY 2013

Hare Rama Hare rama Rama Rama Hare Hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

DEPARTMENT OF MECHANICAL ENGINEERING SCT COLLEGE OF ENGINEERING THIRUVANANTHAPURAM 18

CERTIFICATE
This is to certify that the report entitled Pseudo-Elastic & Partition Of Unity Finite Element Methods And Their Application In Material Non-Linear Problems, submitted by Lekshmi Dinachandran, Roll No. ME 12 MD 12, S1 M Tech to the University of Kerala in partial fulfillment of the requirements for the award of the Degree of Master of Technology in Mechanical Engineering (Machine Design) is a bona fide record of the seminar presented by her.

Prof. M.R Sarathchandra Das Professor, (Guide & Dean - Student Affairs)

Dr. Anilkumar S.H Professor, (PG Coordinator)

Prof. Ayyappan M HEAD OF THE DEPT. (Seal)

Hare Rama Hare rama Rama Rama Hare Hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

ACKNOWLEDGEMENTS
I would like to sincerely thank my guide Prof. M R Sarathchandradas who teaches and guides not just with his academic prowess but with his heart as well. I express my gratitude to Dr. Shaji Senadhipan, Principal, SCT College of Engineering who has been a constant source of inspiration right from my undergraduate years. I thank Prof. M. Ayyappan, HOD, Dept of Mechanical Engg and Dr. Anilkumar S H, PG Coordinator. I would like to take this opportunity to thank all my teachers especially Prof. Anoop Kumar S - for their guidance. I would also like to thank the authors whose works I have referred to and cited in this seminar. I specially thank Dr. N.R. Rajesh who so generously took the time to clear my numerous doubts. I gratefully acknowledge the generous help given by my classmates and friends. My husband Sri.Remesh S, my little son Ananthapadmanabhan, my parents and parents-in-law have been immensely supportive in all my endeavors. I believe God is with me each step of the way and I thank Him for all the grace that He has given me.

Hare Rama Hare rama Rama Rama Hare Hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

ABSTRACT
Finite Element Modeling of non-linear behavior presents a considerable challenge often due to the extensive demands on computational resources. A meshless methodology based on Partition of Unity Finite Element Method (PUFEM) and Pseudo Elastic Analysis for solving material non-linear fracture problems is presented. In Pseudo-Elastic Analysis, the stress field determination in elasto-plastic problems is done based on linear elastic finite element analysis. Henckys total deformation theory is used to define effective material parameters, which are treated as spatial field variables. Partition of unity method is used to define local enrichment base functions which describe the peculiar characteristics of fracture problems the crack tip stress field and the discontinuous displacement across the crack faces. The need for remeshing in order to model crack propagation is eliminated in PUFEM as the crack and the mesh are treated as independent entities. Results obtained from this methodology vis--vis conventional FEM and analytical solutions are also presented.

Hare Rama Hare rama Rama Rama Hare Hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

CONTENTS
Certificate ........................................................................................................................................ ii Acknowledgements ........................................................................................................................ iii Abstract .......................................................................................................................................... iv List Of Tables ................................................................................................................................. 7 List Of Figures ................................................................................................................................ 7 Nomenclature .................................................................................................................................. 9 1. Introduction ............................................................................................................................... 11 1.1 Motivation ........................................................................................................................... 11 1.2 Objective & Scope Of The Seminar .................................................................................... 11 1.2 Organization Of The Report ................................................................................................ 12 2. Literature Survey ...................................................................................................................... 13 3. Pseudo- Elastic Analysis .......................................................................................................... 16 3.1 Effective Material Parameters Determination ..................................................................... 16 3.2 Finite Element Implementation ........................................................................................... 19 4. Partition Of Unity Finite Element Method ............................................................................... 23 4.1 Application Of Pufem In Fracture Mechanics ..................................................................... 24 Modifications to numerical integration scheme .................................................................... 27 5. Evaluation Of Elastic-Plastic Crack Tip Parameters Using PUFEM And Pseudo Elastic Analysis......................................................................................................................................... 29 5.1 Finite Element Approximation & Modeling ....................................................................... 29 5.2 Numerical Implementation .................................................................................................. 33 5.3 Pseudo Elastic Implementation ........................................................................................... 36 5.4 Elastic-Plastic Singular Crack Tip Field And J-Integral ..................................................... 37 5.5 Numerical Examples............................................................................................................ 38 Boundary Layer Analysis ..................................................................................................... 38 Mode I Problem Center Crack Test Specimen .................................................................. 41

Hare Rama Hare rama Rama Rama Hare Hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare 6. Conclusion ................................................................................................................................ 47 References ..................................................................................................................................... 48

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

LIST OF TABLES
Table 1: Variation of J Intergral for different domains................................................................. 43

LIST OF FIGURES
Figure 1: Projection method of eff. material property determination ........................................... 19 Figure 2 Arc length method of eff. material property determination ........................................... 20 Figure 3 Neuber method of eff. material property determination ................................................. 21 Figure 4: Classical approach: element faces aligned with crack .................................................. 24 Figure 5: Strain field with embedded discontinuities ................................................................... 24 Figure 6: classical fem with cohesive formulation ....................................................................... 24 Figure 7: Partition of unity approach ............................................................................................ 25 Figure 8: Elements & nodes associated with the crack................................................................. 25 Figure 9: Nodes which are enriched using different functions ..................................................... 26 Figure 10 : Nodal enrichment scheme and crack representation in a regular mesh ..................... 34 Figure 11: Partitioning of element intersected by crack element ................................................. 35 Figure 12: Projection method to evaluate effective material properties using uniaxial stress-strain curve .............................................................................................................................................. 36 Figure 13: Geometrical details - Physical domain & Finite Element Model ................................ 39 Figure 14: Variation of normalized stress ahead of crack tip along =0 degree (a) radial stress (b) tangential stress ............................................................................................................................. 40 Figure 15: Variation of normalized stress ahead of crack tip along =45 degree (a) radial stress (b) tangential stress ....................................................................................................................... 40 Figure 16: Variation of normalized stress ahead of crack tip along =90 degree (a) radial stress (b) tangential stress ....................................................................................................................... 41 Figure 17: Variation of normalized shear stress ahead of crack tip along (a) =45 degree & (b) =90 degree ................................................................................................................................... 41 Figure 18: Geometrical Configuration - CCT Specimen .............................................................. 42 Figure 19: (a) Domain Discretization (b) Domains Around Crack Tip For J Integral Evaluation 42 Figure 20: J-integral variation with the applied load for centre crack problem for various strain hardening exponents (a) Plane strain (b) Plane stress ................................................................... 44

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare Figure 21: J-integral variation with the applied load for centre crack problem, m=3 (a) Plane strain (b) Plane stress .................................................................................................................... 44 Figure 22: Stress variation ahead of the crack tip for centre crack problem underplane stress condition with m=3 and stress ratio=0.5 ....................................................................................... 45 Figure 23 : variation of normalized stress ahead of the crack tip for centre crack problem under plane stress condition with m=3 ................................................................................................... 46 Figure 24: Strain variation ahead of the crack tip for centre crack problem under plane stress condition ....................................................................................................................................... 46

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

NOMENCLATURE

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

1. INTRODUCTION
Non-linearity is inherent in nature. Most, if not all, phenomena in nature exhibit non-linear behavior to some extent. Linear models are oftentimes satisfactory or convenient approximations. This pervasiveness of non-linearity in nature reflects in engineering problems as well. One of the chief problems faced by engineers is that of accounting and adjusting for nonlinearity. Structural Mechanics problems such as those which involve plastic deformation of metals, yielding, creep and local buckling , heat transfer analyses where temperature-dependent material properties, heat absorption or liberation due to phase changes, radiation etc are examples of such problems Material non-linearity refers to cases wherein the material properties themselves are functions of the state of stress or strain. Due to the mathematical complexities involved in solving material non-linear problems, numerical methods are commonly employed. Analytical solutions are limited to simple geometry and boundary tractions. 1.1 MOTIVATION The modeling and simulation of moving discontinuities such as cracks are of great interest to a mechanical engineering. Even as most commercial software available today provides excellent point and click facilities, an understanding of both the physical phenomena and the solution strategy adopted is essential for meaningful analysis of the obtained results. This seminar aims to present one of the newer numerical techniques adopted for the modeling of cracks in materials. 1.2 OBJECTIVE & SCOPE OF THE SEMINAR The objectives of this seminar include

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare 1 Introduction to the principles of Pseudo-Elastic Analysis and Partition of Unity Finite Element Methods 2 Overview of the finite element implementation of the above 3 In-depth look at the application of these two methodologies for the determination of elastic-plastic crack tip parameters. 1.2 ORGANIZATION OF THE REPORT The report is organized into six chapters as follows Chapter 1 provides a brief introduction to the importance of considering nonlinearity in design decisions. The motivation, scope and objective of the seminar have been set forth. Chapter 2 gives a review of the literature in this field. Chapter 3 takes a look at the principles of Pseudo-Elastic Analysis and places it in context with respect to the modeling of material nonlinearity. The concept of effective material parameters, and the finite element implementation of this approach is also detailed. Chapter 4 explains Partition of Unity finite element method Chapter 5 details the application of Pseudo-Elastic analysis and Partition of Unity Finite Element Method in the determination of elastic-plastic crack tip parameters with the help of a numerical example. Chapter 6 points out the scope for further work in the field

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

2. LITERATURE SURVEY
Rice (1968) proposed a path independent line integral called J-integral encircling the crack tip to characterize the elastic-plastic fracture field. Using the line integral, Hutchinson (1968a, 1968b) and Rice and Rosengren (1968) studied stress/strain singularities dominant at the crack tip region of an elastic-plastic material. Their study showed the existence of stress/strain singularity, known as HRR singularity, well within the plastic zone surrounding the crack tip. Results also showed that the intensity coefficient of this singularity, J-integral, uniquely characterize the elasticplastic crack tip field Work of Shih and German (1981) using finite element analysis under small scale yielding condition for mode-I problems showed that though J-integral characterizes the elastic-plastic crack tip field, the region of influence of HRR singularity varies from geometry to geometry and also depend on the material hardening parameter. In recent years, considerable research is being carried out in meshless methods, a new class of numerical method, for solving partial differential equations associated with boundary value and initial value problems. These methods are highly suitable for crack propagation problems, large deformation studies and complex geometries where remeshing is time consuming and expensive. Various commonly used meshless methods are Smooth Particle Hydrodynamics (SPH), Diffuse Element Method (DEM), Element Free Galerkin method (EFGM), H-P cloud method, Reproducing Kernel Particle Method (RKPM), Meshless Local Petrov-Galerkin Method (MLPG), Partition of Unity Finite Element Method (PUFEM) [Melenk and Babuska (1996)], Generalized Finite Element Method (GFEM), Extended Finite Element Method (XFEM) etc. MLPGM originated by Atluri (2004) is considered as a true mesh free method since does not require any mesh either for the interpolation of the solution variables or for the integration of the

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare weak form as compared to EFGM which requires background mesh for integration. The MLPG method has been further enhanced to solve 3D elasto-static problems non linear static and dynamic problems and 2D stationary and transient problems in piezoelectric and magnetoelectric-elastic material models with continuously varying material properties. Basic theory and applications of Partition of Unity Finite Element Method are explained in detail with numerical examples in the report of Melenk and Babuska (1996). PUFEM uses patch based interpolation approximations compared to element based interpolation approximations of standard FEM. Treating the entire domain as group of overlapping patches with local enrichment functions defined over each patch, PUFEM enforces continuity over the entire domain using a set of C0 functions called PU functions. Moes, John Dolbow and Belytschko (1999) proposed PUFEM based enrichment technique using Heaviside and trigonometric functions for modeling the cracks. The accuracy of the method was demonstrated by predicting crack growth with relatively coarse mesh. Strouboulis, Babuska and Copps (2000, 2001) discussed in detail the design and implementation of Generalized Finite Element Method as a direct extension of standard finite element method. Discontinuities and singularities were modeled using special enrichment functions under the framework of partition of unity. The versatility of PUFEM was demonstrated by Fan, Liu and Lee (2004) by directly extracting pure mode stress intensity factors in a mixed mode problem. They used enrichment functions only at the crack tip corresponding to truncated asymptotic series of linear elastic displacement field along with p-version FEM and used coincident nodes to simulate rest of the crack surface. Though substantial research has been carried out in elastic regime using partition of unity concept, study in elastic-plastic regime is still in the developing stage. Rao and Rahman (2004)

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare used elastic-plastic crack tip enrichment functions to capture the HRR singularity using element free Galerkin method. Elguedj et al. (2006) studied about various EPFM enrichment functions for XFEM modeling and used incremental non-linear solver for the elastic-plastic analysis. Elastic-plastic fracture solutions of various crack problems are compared with linear solutions in their work. Hagihara, et al. (2007) used element-free Galerkin method to calculateelastic-plastic fracture mechanics parameters, J-integral and T* -integral for both stationary crack problem and stable crack growth problem in 2D domain. Long, Liu and Li (2008) studied elastic-plastic fracture problems using meshless Petrov- Galerkin method and incremental plasticity analysis. Pseudo elastic method, initially developed by Desikan and Sethuraman (2000) is an iterative method based on the Henckys total deformation theory of plasticity, where material properties are treated as spatial field variables. Same methodology is used in elastic-plastic fracture study along with element free Galerkin method by Sethuraman and Reddy (2004, 2008) and in the inelastic analysis of 2D problems using Radial Point Interpolation Method by Dai, Liu, Han and Li (2006). In the work of Sethuraman and Rajesh (2009) pseudo elastic methodology is coupled along with partition of unity finite element method for elastic-plastic fracture characterization and simulation of HRR singular field at the crack tip. Effectiveness of the proposed method has been demonstrated by the elastic-plastic analysis of various fracture problems using different material hardening models.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

3. PSEUDO- ELASTIC ANALYSIS

In classical design of structural components, the norm defined on the stress tensor is bounded within the elastic limit; but in practice it would be advantageous to consider the elasto-plastic behavior of structural materials. This may be due to the unexpected presence of notches, voids, micro-cracks, overloads, etc. In pressure vessel design to enhance the pressure-carrying capacity, residual stresses are developed by introducing initial partial plastic yielding, i.e., autofrettageing. Thus, the behavior of the materials in the plastic range is essential in order to understand the structural behavior completely and to have reliable estimate of life. However, the nonlinear stress-strain relationship and the loading path dependency in the plastic range make the analysis tedious. Over the years, finite element method has been successfully employed in analyzing the material behaviors in elastic and elasto-plastic range. The method of Pseudo-Elastic Analysis was proposed by R. Sethuraman & V. Desikan (2000). It is based on Henckys Total Deformation Theory is applicable for proportional loading. In this method, a problem in the elastic-plastic or fully plastic regime is idealized as a series of linear elastic problems and solved iteratively. The field variables in this analysis method are the effective material parameters. Different algorithms such as projection method, arc-length method and Neuber rule can be used to determine the effective material parameters. 3.1 EFFECTIVE MATERIAL PARAMETERS DETERMINATION As stated earlier, in Pseudo Elastic Analysis, quantities known as effective material parameters are treated as field variables. These material parameters are functions of the final state of stress fields. Since the final state of stress at each point is unique, Eeff and eff the effective modulus of elasticity and effective Poissons ratio can be considered as field variables.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare The principle of virtual work for a static equilibrium body under elastic or elasto-plastic stress field, for every kinematically admissible virtual displacement field, is given as

The stress- strain relationship can be written as

where the total strain tensor is the sum of the elastic and plastic part given as

From Hookes Law for isotropic materials it can be seen that

Henckys total strain theory, on which this analysis method is based, postulates a one-on-one correspondence between the stress and strain tensors given by

Where S is the deviatoric strain tensor and

Equivalent plastic strain and equivalent stress are defined as

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare From equations (4) through (8) and equation (3), the strain tensor can be written as

Since all quantities within the brackets are associated with material properties, final equivalent plastic strain and equivalent stress, this can be written as

The equivalent material parameter values are obtained as follows: 1) For Isotropic Materials

2) For Elastic Perfectly Plastic Materials

3) For Ramberg Osgood Material Model which obeys the power law the effective material parameters are given by

Thus, for constant values of E

eff

and

eff

equation (10) describes the elastic behavior of the

body.It is known that every state of stress at point in a equivalent sense follows experimental uniaxial tensile curve. From that point of view the reciprocal of Eeff is nothing but secant modulus defined on the experimental uniaxial curve. Hence, the effective modulus can be obtained from the experimental material curve.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare 3.2 FINITE ELEMENT IMPLEMENTATION The discrete form of the equation is obtained by discreting the given domain into finite domains using constant strain triangular element -CST, which has the constant stress-strain field representation. Since the stress field is constant over the element domain, the effective material parameter is also constant over the element, thereby leading to the determination of finite number of material parameters over the entire domain. The element stiffens matrix for the CST element is given by

where [B] is the conventional strain-displacement matrix & [Deff] is obtained from constitutive equation (eq. 10). ALGORITHMS USED FOR DETERMINATION OF EFFECTIVE MATERIAL PROPERTIES Three algorithms have been proposed for the determination of effective material properties. They are 1) Projection Method

FIGURE 1: PROJECTION METHOD OF EFF. MATERIAL PROPERTY DETERMINATION

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare Suppose, for a particular element, the equivalent stress is calculated from linear finite element analysis and point A, as shown in Fig. 1, is obtained. This point has crossed the yield stress. Keeping the strain value the same, i.e., strain controlled, and projecting point A on the experimental uniaxial curve, the effective value of Youngs modulus for the next iteration is obtained. Substituting this effective value in Eq. 12, the effective Poisson ratio is obtained. These effective values are obtained for all the elements which have yielded. With this new set of effective material parameters, the next linear elastic finite element analysis is performed. This iterative procedure is repeated and elastic analysis with currently evaluated Eeff and eff is performed until all the effective material parameters converge and equivalent stress falls on the experimental uni-axial stress-strain curve. 2) Arc Length Method

FIGURE 2 ARC LENGTH METHOD OF EFF. MATERIAL PROPERTY DETERMINATION

In arc-length method, the yielded point A, as shown in Fig. 2, is projected onto the experimental uniaxial curve, considering OA as the arc radius and O as the center. The intersecting point B on the experimental uniaxial curve is used to obtain the new effective modulus for the next iteration. Substituting this effective modulus in Eq. 12, the new effective Poisson ratio is obtained and is used for the next iteration.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare 3) Neuber Method

FIGURE 3 NEUBER METHOD OF EFF. MATERIAL PROPERTY DETERMINATION

In Neuber method, the yielded point A, as shown in Fig. 3, is projected onto the uniaxial curve at point B. While projecting the total strain, energy is assumed to be conserved; i.e., the area ODAEO is the same as area OCBFO. Thus, a corresponding point on the uniaxial material curve which has the same strain energy is found. With this point, the effective Young modulus is obtained and is substituted in Eq. 12 to get the effective Poisson ratio for the next iteration. The method can be easily adapted to the existing linear elastic finite element code with suitable updating of elastic material properties with effective material parameters which are readily obtained from one-dimensional uniaxial tension curve of a given material. Stress fields obtained are found to be in good agreement with the available nonlinear finite element results. The rate of convergence for the three algorithms were studied, and it was found that the Neuber rule has better convergence rate.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

4. PARTITION OF UNITY FINITE ELEMENT METHOD

Finite Element Methods are widely used in the analysis of many different kinds of problems. However, modeling of moving discontinuities such as cracks using conventional FEM often requires extensive mesh modification between iterations. This in turn leads to diminished computational performance. Better computational efficiency is offered by methods which minimize or avoid the need for remeshing. Partition of Unity Finite Element Method (PUFEM) is one such technique. PUFEM is a class of meshless method in that a mesh in the classical sense is not created. There is no division of the domain into elements. There is only a background mesh for integration. This method offers an easy way to incorporate a priori knowledge and analytical info about the problem into the solution process. Local enrichment functions are used to modify a standard finite element approximation to achieve this end. Enrichment involves the provision of additional degrees of freedom using certain typical functions in the region of interest. The choice and usage of these functions dictate the applicability of this method to various types of problems. PUFEM uses patch based interpolation approximations compared to element based interpolation approximations of standard FEM. Treating the entire domain as group of overlapping patches with local enrichment functions defined over each patch, PUFEM enforces continuity over the entire domain using a set of C0 functions called PU functions. Moes, John Dolbow and Belytschko (1999) proposed PUFEM based enrichment technique using Heaviside and trigonometric functions for modeling the cracks

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare 4.1 APPLICATION O F PUFEM IN FRACTURE MECHANICS Various strategies are used for the finite element modeling of cracks. The classical approach is to align the element faces along the crack surface. As the crack progresses, the nodes split and duplicate as required. A modification of this method uses special kind of elements known as cohesive elements along the crack. These are
FIGURE 4: CLASSICAL APPROACH: ELEMENT FACES ALIGNED WITH CRACK

interface elements which are capable of modeling the atypical properties associated with crack faces. Cracks or other kinds of

displacement discontinuities can also be modeled as a strain field with embedded discontinuities. are defined The as

discontinuities
FIGURE 6: CLASSICAL FEM WITH COHESIVE FORMULATION

localized modes in the general or standard strain field. i.e. local variations in the strain are used to define the discontinuous displacements associated with the problem.
FIGURE 5: STRAIN FIELD WITH EMBEDDED DISCONTINUITIES

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare In Partition of Unity Method the mesh and

discontinuity are separate entities. The mesh is simply for integration. The last two methods are discontinuous field
FIGURE 7: PARTITION OF UNITY APPROACH

approaches as opposed to conventional methods illustrated in the first two figures. In fracture mechanics, modeling of cracks are of primary importance. Along a crack, there will be a localized jump or discontinuity in the displacement. This can be represented effectively using PUFEM.

FIGURE 8: ELEMENTS & NODES ASSOCIATED WITH THE CRACK

As mentioned earlier, PUFEM utilizes enrichment functions to provide additional degrees of freedom for salient nodes. In figure 8, elements through which the crack passes are indicated in yellow while the element which contains the crack tip is shown in green. The nodes associated with these elements are the ones which are selected for enrichment. A discontinuous function is used to represent the jump in displacement across the crack line. In the near-tip region, asymptotic displacement fields exist. These are modeled using the appropriate enrichment functions.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare An important question in PUFEM is to determine which nodes to enrich. Enrichment involves more computations, and if more than the optimum number of nodes is chosen, it will be counterproductive in terms of computational efficiency. The enrichment is done as follows: 1. All nodes whose supports are cut by the crack are enriched using jump functions. 2. Nodes surrounding the crack tip are enriched using asymptotic enrichment functions.

FIGURE 9: NODES WHICH ARE ENRICHED USING DIFFERENT FUNCTIONS

In figure 9, the circled nodes are enriched with the discontinuous function and the squared nodes are enriched with the tip displacement functions. Enrichment with only the discontinuous function shortens the crack to point P. Mathematically, the finite element approximation in PUFEM is a superimposition of the enrichment functions on a standard finite element approximation.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare The first term on the right hand side represents the standard finite element approximation. The second term is the discontinuous enrichment and the third term denotes the crack-tip asymptotic enrichment. The enrichment functions H and F are defined as

That is, the finite element space defined above can be described as the sum of one which does not model the crack, a discontinuous enrichment and an asymptotic enrichment function. In the case of higher order elements i.e. elements with interior nodes, each DOF is enriched as opposed to each corner node in the case of first order elements. The criteria for selection of DOFs to be enriched are given as a degree of freedom is enriched if its support is cut by the crack. In the case of an edge degree of freedom, support refers to the elements connected to it. For an element degree of freedom, the support is the element itself. MODIFICATIONS TO NUMERICAL INTEGRATION SCHEME

In partition of unity finite element method, the meshing is done only for integration purposes. As with normal FEM, the domain itself is divided into a number of elements. i.e. the domain can be defined as the summation of a number of elements e

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare If conventional Gauss quadrature is used for integration, it may generate spurious nodes and stress singularities. In order to avoid this, a modified scheme is often chosen so that each element e which is cut by the crack is further divided into sub-polygons aligned to the crack path for integration purposes alone.

It is worth reiterating that this kind of subdivision has nothing to do with the actual computation of the field variable but only with the integration of these computed values to obtain the final result.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

5. EVALUATION OF ELASTIC-PLASTIC CRACK TIP PARAMETERS USING PUFEM AND PSEUDO ELASTIC ANALYSIS

Most of the modern engineering structural materials undergo large scale inelastic deformation at the crack tip and conventional SIF based fracture design is no longer valid. Instead, Elasticplastic fracture characterization is required for life prediction and damage tolerant design of engineering structures. Rice (1968) proposed a path independent line integral called J-integral encircling the crack tip to characterize the elastic-plastic fracture field. Using the line integral, Hutchinson (1968a, 1968b) and Rice and Rosengren (1968) studied stress/strain singularities dominant at the crack tip region of an elastic-plastic material. Their study showed the existence of stress/strain singularity, known as HRR singularity, well within the plastic zone surrounding the crack tip. Results also showed that the intensity coefficient of this singularity, J-integral, uniquely characterize the elastic-plastic crack tip field. Sethuraman & Rajesh (2009) proposed the combination of Pseudo- elastic analysis and Partition of Unity FEM for the evaluation of elastic-plastic crack tip parameters. Pseudo elastic methodology is coupled with Partition Of Unity Finite Element Method for elastic-plastic fracture characterization and simulation of HRR singular field at crack tip. Iterative linear elastic analysis using PUFEM is carried out for determination of elastic-plastic stress fields with effective material parameters as the field variables. 5.1 FINITE ELEMENT APPROXIMATION & MODELING For modeling of a two dimensional crack, the finite element approximation equation similar to the one in 4.1 is given by

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

In the case of Elastic-Plastic Fracture Mechanics problems, the asymptotic enrichment functions are given by

The discontinuous function is the Heaviside or Haar function

That is, on obtaining the values of the DOFs of an enriched crack-tip node, we can see from the sign of the DOF corresponding to the second term of the approximation whether the corresponding displacement is above ( y>0) or below (y<0) the crack. Also, the next four DOFs defined by the trigonometric asymptotic enrichment functions listed above provide the resultant displacement of those nodes as a function of the crack tip orientation in local polar co-ordinates and the strain hardening exponent m. As per the normal course of Finite Element procedure, the finite element equations are formed as

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

where [K] is the stiffness matrix, [B] is the strain displacement matrix, [D] is the appropriate constitutive matrix, [N] is the shape factor matrix and [F] is the load vector. The subscript en refers to enriched matrices, u refers to the un-enriched matrices and el refers to the elemental matrices. Shape functions of the vertex node, corresponding to its support elements in the patch, are used to construct the partition of unity function for that patch. These shape functions have value unity at the vertex and zero on the boundary of the patch. Together these shape functions form C0 partition of unity over the entire domain ensuring inter element continuity. Therefore trial function at any interior point depends on the non zero shape functions of four surrounding nodes and associated degrees of freedom. Thus construction of stiffness matrix and force vector can be done element wise and assembled in the usual finite element procedure. The element stiffness matrix is defined as

where,

Size of element stiffness matrix varies with its associated nodal enriched degrees of freedom. For an element with no enriched nodes, size of stiffness matrix will be same as that of conventional

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare element i.e. 8x8. If all four nodes of an element are enriched with Heaviside function, size of stiffness matrix will be 16x16 and for elements with all four nodes asymptotically enriched, size of stiffness matrix willbe 40x40. Elemental load vector is given by

with components

Assuming plane stress condition, the constitutive matrix

For elastic-plastic problems, in evaluating the stiffness matrix K, the effective Poissons ratio and effective Youngs modulus are used appropriately. In stiffness matrix and subsequent stress evaluation, the constitutive matrix is changed from D to the effective constitutive matrix [Deff] using the material parameters Eeff and eff . The element stiffness matrix for elastic-plastic analysis can be written as

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

The effective constitutive matrix is obtained iteratively using Pseudo-Elastic Methodology and Projection Method as outlined in previous chapters. 5.2 NUMERICAL IMPLEMENTATION PUFEM implementation presented in the current study is based on the work of Sukumar and Prvost (2003). Domain is discretized with mesh without considering crack geometry. Crack geometry is considered as a virtual segment across interiors of a group of elements and is assumed to be terminating in the interior of an element. The crack face discontinuity and crack tip singular field are characterized using Heaviside and asymptotic enrichment functions. A sample mesh is shown in Fig. 10 with an edge crack terminating at interior of an element. 4 nodes around the crack tip usually form the node set Nc for asymptotic enrichment. Alternatively a fixed circular domain around the crack tip can also be used to identify nodes for asymptotic enrichments. Nodes that belong to elements that are split by the crack segment, but not selected for asymptotic enrichment, are included in the set Nf for Heaviside enrichment. Nodes that are enriched with asymptotic functions have 10 degrees of freedom per node (two conventional degrees of freedom and eight degrees of freedom associated with four enrichment base functions). Nodes with Heaviside enrichment have 4 degrees of freedom per node (two conventional degrees of freedom and two degrees of freedom associated with enrichment basefunction). Nodes 1, 2, 5, 6 of the above mesh belong to node set Nc and nodes 3, 4, 7, 8 belong to node set Nf . Signed area of the triangle formed with material point (either gauss point /node) and end

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare points of the crack segment as vertices is used to assign the value of Heaviside function. Construction of element stiffness matrices and element force vectors for partitioned elements include integration of discontinuous functions. The partitioned elements are ones whose all nodes are enriched, thus encompassing the discontinuity (crack). The elements A, B, C are the partitioned elements for the mesh shown in Fig. 10.

FIGURE 10 : NODAL ENRICHMENT SCHEME AND CRACK REPRESENTATION IN A REGULAR MESH

Integration of stiffness matrix in case of elements with crack segment in the interior requires special treatment because of the discontinuous enrichment functions involved. Element is partitioned into smaller subdomains conforming to the crack segment. Stiffness evaluation requires looping over smaller partitioned subdomains as shown in Fig. 11. Integration over domain abcd requires evaluation of shape functions, which are defined only for the parent element, at local gauss points. This is done by mapping physical co-ordinates of gauss points in the subdomain abcd to mapped quadrilateral of the parent element i.e.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

FIGURE 11: PARTITIONING OF ELEMENT INTERSECTED BY CRACK ELEMENT

Contribution of subdomain abcd to elemental stiffness matrix is given by

where ( 1; 2) is the parent element co-ordinate system and (1; 2 ) is the subdomain coordinate system. Global co-ordinate system is aligned with crack tip co-ordinate system for calculating of asymptotic enrichment functions and post processing of results. If any one node of an element is enriched, its contribution to element stiffness matrix is to be added in addition to contribution associated with conventional degrees of freedom. Strain-displacement matrices of partitioned elements A, B, C are

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare Global stiffness matrix and force vector are constructed similar to standard finite element method. Global nodal degrees of freedom array is created first, using all nodal degrees of freedom including both conventional and enriched degrees of freedom. Element stiffness matrices are assembled in the global stiffness matrix according to the global degrees of freedom of its member nodes. 5.3 PSEUDO ELASTIC IMPLEMENTATION Pseudo elastic method is an iterative procedure based on the elastic-plastic stress strain relation given as

. All material points are assigned uniaxial material properties, E and n in the initial iteration. After the first iteration, effective modulus, Ee projection method as illustrated in Fig. 12.
f f

, at all material points are calculated using

FIGURE 12: PROJECTION METHOD TO EVALUATE EFFECTIVE MATERIAL PROPERTIES USING UNIAXIAL STRESS-STRAIN CURVE

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare Point A corresponds to equivalent stress at a material point after the first iteration and point B is the corresponding point on experimental uniaxial tensile curve. The effective modulus, Eeff (secant modulus of experimental uniaxial curve) is the slope of the line OB. Effective Poisson ratio,
eff

is calculated. Values of Ee f f and e f f are used as modified

material parameters in the next analysis. Iterations are repeated until the state of stress and strain of all material points follow the uniaxial material curve. Points D, E, F represent corresponding points on the material curve in subsequent iterations. 5.4 ELASTIC-PLASTIC S INGULAR CRACK TIP FIELD AND J-INTEGRAL Hutchinson (1968a, 1968b), Rice and Rosengren (1968) showed that in elasticplastic materials within the plastic zone, elastic strains are very small and stress strain behavior reduces to pure power law even for general Ramberg-Osgood material model. Based on deformation theory of plasticity, they proposed path independent J-integral which characterizes the stress/strain field (HRR singularity) in non-linear materials as given below where r, are polar coordinates centered at the crack tip, In is a dimensionless integration constant which is a function of strain hardening exponent, m i j, i j and ui are dimensional functions which depend on and m and J is the amplitude of the crack tip singular field The above relations are based on the power law hardening material response given by - material constant 0 - the reference yield stress m hardening exponent

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare For linear elastic materials, m=1 and for perfectly plastic response, m = . The J-integral defines the severity of the crack tip stress/strain field both in the elastic and elastic-plastic regime. For a general cracked body subjected to remote loading, J-integral in the absence of body forces is given by

where is any arbitrary closed contour around the crack tip starting from one crack face to the other. W is the strain energy density defined by

while t is the traction vector and u is the displacement vector on path . 5.5 NUMERICAL E XAMPLES Initially, the effectiveness of the present method is demonstrated using a boundary layer analysis for an edge crack problem with its outer boundary subjected to mode-I elastic displacement field. In the original work, a number of Mode I and Mode II problems are further considered for the analysis. J-integral and stress/strain fields are evaluated using the present pseudo-elastic method and compared with the results available in literature and also with the solution obtained from non-linear finite element analysis. However, this seminar dealt with the analysis of only a single Mode I Center Crack Test specimen. BOUNDARY LAYER ANALYSIS Boundary layer approach is adopted here to predict the J-integral dominant HRR field and K dominant elastic field around the crack tip under plane stress conditions. In this method, displacement field corresponding to mode-I elastic solution is imposed on the outer boundary of

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare the domain [Rao and Rahman (2004)]. Physical domain of the problem consists of a square plate, size 100 mm, with an edge crack, length 50 mm, terminating at centre of the plate. Physical domain and corresponding mesh used are shown in Fig 13. Quadrilateral mesh of 405 elements and 416 nodes are used to discretize the domain. Crack tip co-ordinate system is considered to be coincident with global coordinate system. Finer mesh is used around the crack tip region. EPFM asymptotic enrichment functions are used for modeling crack in the domain.

FIGURE 13: GEOMETRICAL DETAILS - PHYSICAL DOMAIN & FINITE ELEMENT MODEL

Ramberg-Osgood material model is considered for the analysis. The values of material parameters used are Youngs modulus (E) =200 GPa, Poissons ratio () =0.3, yield stress (0) =700 MPa, yield offset =0.1 and hardening exponent m =3. Each element cut by the crack is partitioned into 16 smaller quadrilateral sub domains, conforming to the crack edges and a higher order 6x6 Gauss quadrature is used for the numerical integration. Penalty method is used to enforce the displacement field along the outer periphery of the domain. Results obtained from the present study are compared with both HRR singularity

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare solution and linear elastic crack tip field. Computed elastic-plastic stress field from the present study under plane stress condition along the radial lines at = 00, = 450 and = 900 ahead of the crack tip are presented below. The variation of normalized radial stress ( rr=0) and tangential stress ( =0) with respect to a normalized radial distance (r0/J) ahead of the crack tip for = 00, = 450 and = 900 under plane stress condition are shown in Figs. 14, 15 and 16 respectively. The distribution of normalized shear stress r=0 along radial lines = 450 and = 900 are also given in Fig. 17

FIGURE 14: VARIATION OF NORMALIZED STRESS AHEAD OF CRACK TIP ALONG =0 DEGREE (A) RADIAL STRESS (B) TANGENTIAL STRESS

FIGURE 15: VARIATION OF NORMALIZED STRESS AHEAD OF CRACK TIP ALONG =45 DEGREE (A) RADIAL STRESS (B) TANGENTIAL STRESS

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

FIGURE 16: VARIATION OF NORMALIZED STRESS AHEAD OF CRACK TIP ALONG =90 DEGREE (A) RADIAL STRESS (B) TANGENTIAL STRESS

FIGURE 17: VARIATION OF NORMALIZED SHEAR STRESS AHEAD OF CRACK TIP ALONG (A) =45 DEGREE & (B) =90 DEGREE

MODE I PROBLEM CENTER CRACK TEST SPECIMEN

A Center Crack Test (CCT) specimen subjected to remote tension in both plane stress and plane strain conditions are considered for the pseudo-elastic analysis and fracture characterization. Crack length (a) =50 mm, width (W) =100mmand L/W=10 are used as geometrical dimensions

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare for modeling the specimens. Ramberg-Osgood material model with Youngs modulus (E) = 200 GPa, Poissons ratio () = 0.3, yield stress (0) = 200 MPa, yield offset = 3/7 and hardening exponent m = 3, 5 and 10 is considered for the present study. Half model of the physical domain is used for analysis. Remote load

corresponding to 80% of the yield

FIGURE 18: GEOMETRICAL CONFIGURATION - CCT SPECIMEN

stress is applied. Domain is discretized

with 345 elements and 372 nodes. In present study, all nodes in the rectangular region of size 100x100 mm around the crack tip are enriched with asymptotic enrichment functions. Penalty method is adopted for enforcing symmetric boundary and also for enforcing constraint equations connecting enriched degrees of freedom of crack face nodes lying on the plane of symmetry. Various domains, as shown in Fig. 19 are used for J-integral evaluation to check the domain independence. J-integral

values, calculated for the normalized far field stress =0=0.5 and

FIGURE 19: (A) DOMAIN DISCRETIZATION (B) DOMAINS AROUND CRACK TIP FOR J INTEGRAL EVALUATION

hardening

exponent

m=3

are

compared with results evaluated from

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare EPRI estimation scheme [Anderson (2000)] in Tab. 1 for plane strain condition. Variation of Jintegral values for different domains is found to be less than 1%. The specimen is analyzed for various normalized boundary stresses/tractions (=0) and for different strain hardening exponents. Normalized stresses are varied from 0.1 to 0.8 for all cases with hardening exponent, m = 3, 5 and10. Domain No. 1 2 3 4 5 J kJ/m2 JEPRI = 12.877 12.8654 12.9181 12.9080 12.8800 12.8365 % deviation from EPRI scheme 0.09 0.32 0.24 0.02 0.31

TABLE 1: VARIATION OF J INTERGRAL FOR DIFFERENT DOMAINS

J-integral values for different normalized far field stresses with m=3 for centre crack problem under plane strain and plane stress conditions are presented in Fig 20. J-integral values from the present analysis are found to be in good agreement with results of EPRI estimation scheme over the considered range of loading. J-integral variation with different strain hardening exponents for same crack configuration and loading, under plane strain and plane stress loading are presented in Figs. 21(a) and 21(b) respectively. The effect of m on J is pronounced when =0 is more than 0.6.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

FIGURE 21: J-INTEGRAL VARIATION WITH THE APPLIED LOAD FOR CENTRE CRACK PROBLEM, M=3 (A) PLANE STRAIN (B) PLANE STRESS

FIGURE 20: J-INTEGRAL VARIATION WITH THE APPLIED LOAD FOR CENTRE CRACK PROBLEM FOR VARIOUS STRAIN HARDENING EXPONENTS (A) PLANE STRAIN (B) PLANE STRESS

Stress and strain field evaluated ahead of the crack tip using the present method are compared with HRR solution and non-linear finite element solution in Figs. 22 and 23. It is observed that the normalized stress fields 22=0 and 11=0 obtained from the present method matches well with the HRR solution closer to crack tip (r0/J <5) but slightly under predict when compared with the HRR solution at larger distances from the crack tip (r0/J >5). But better agreement

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare between results of present study and non-linear finite element solution is observed at larger distances compared to HRR results at the crack tip region. Prediction of normal strains (11 and 22) based on the present method matches well with the HRR and FEA solutions. The present study is extended to study the effect of normalized far field stresses on crack tip field. The normal stress distribution 22=0, ahead of the crack tip for three different loadings with m=3 for centre crack problem are plotted in Fig.24. Invariance of the stress field with normalized far field stresses for a particular hardening exponent and r0/J support the fact that the single parameter J can characterize the entire crack tip field in non-linear materials. It can also be observed that stress field from the present study is in close agreement with HRR solution near the crack tip region (r0/J <5) and tend to deviate more from HRR solution at higher values of r0/J.

FIGURE 22: STRESS VARIATION AHEAD OF THE CRACK TIP FOR CENTRE CRACK PROBLEM UNDERPLANE STRESS CONDITION WITH M=3 AND STRESS RATIO=0.5

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

FIGURE 24: STRAIN VARIATION AHEAD OF THE CRACK TIP FOR CENTRE CRACK PROBLEM UNDER PLANE STRESS CONDITION

FIGURE 23 : VARIATION OF NORMALIZED STRESS AHEAD OF THE CRACK TIP FOR CENTRE CRACK PROBLEM UNDER PLANE STRESS CONDITION WITH M=3

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

6. CONCLUSION
A partition of unity finite element method coupled with pseudo elastic analysis is presented for the elastic-plastic fracture characterization of material non-linear problems. In the partition of unity finite element method enrichment functions are used to model the crack face discontinuity and crack tip asymptotic field. The applicability of the pseudo elastic method has been exploited within the partition of unity finite element method framework for solving material non-linear problems in a linear fashion. Effectiveness of the method is illustrated considering various case studies. The resulting stress field of the boundary layer approach, along various radial lines starting from the crack tip, matched well with J dominated near tip field and K dominated far field. Fracture parameter evaluated for case of CCT specimen showed close agreement with results of EPRI estimation scheme. Stress and strain field obtained for from the present study matched well with HRR singular field for normalized distance, (r0/J) less than 5 and tend to deviate more at larger normalized distances, but matches well with the non-linear finite element solution. J-integral values, estimated using the decomposition method, for edge crack problem subjected to remote shear traction are found to be in good agreement with the available literature results. The asymptotic crack tip stress fields obtained by the present method for the shear are compared with both HRR singular stress fields and also with non-linear finite element solution and are found to be in good agreement.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare

REFERENCES
1. Belytschko T, Black T (1999). Elastic crack growth in finite elements with minimal remeshing. International Journal for Numerical Methods in Engineering ; 45(5):601- 620. 2. Chakrabarty J (2006), Theory of Plasticity. Elsevier Butterworth-Heinemann 3. Fan, S. C.; Liu, X.; Lee, C. K. (2004): Enriched partition of unity finite element method for stress intensity factors at crack tips. Comput. Struct., vol. 82, pp. 445-461 4. Kumar P (2009), Elements of Fracture Mechanics. McGraw-Hill India 5. Malvern, L. E.(1969): Introduction to the Mechanics of a Continuous Medium, PrenticeHall, Engelwood Cliffs, NJ. 6. Melenk, J. M.; Babuska, I. (1996): The partition of unity finite element method: basic theory and applications. Comput. Methods Appl. Mech. Engrg., vol. 139, pp.289-314. 7. Moes, N.; John Dolbow; Ted Belytschko (1999): A finite element method for crack growth without remeshing. Int. J. Numer. Methods Eng., vol. 46, pp. 131-150. 8. Rao, B. N.; Rahman, S. (2004): An enriched mesh less method for non-linear fracture mechanics. Int. J. Numer. Methods Eng., vol. 59, pp. 197-223. 9. Raju Sethuraman; Cherku Sridhar Reddy (2008): Pseudo elastic analysis of elastic-plastic crack tip fields using element free Galerkin method. Int. J. Comput. Methods, vol. 5, pp. 91-117. 10. Raju Sethuraman; N.R.Rajesh (2009) :Evaluation of Elastic-Plastic Crack Tip Parameters using Partition of Unity Finite Element Method and Pseudo Elastic Analysis. Computer Modeling in Engineering & Sciences, Vol.39, No.1 11. Raju Sethuraman; V. Desikan (2000) : Analysis of Material Nonlinear Problems Using Pseudo-Elastic Finite Element Method. J. Pressure Vessel Technology, Vol. 122 12. Raju, I. S.; Shiva Kumar, K. N. (1990): An equivalent domain integral method in the two dimensional analysis of mixed mode crack problems. Eng. Fract. Mech., vol. 37(4), pp. 707-725. 13. Rice, J. R.; Rosengren, G. F. (1968): Plane strain deformation near a crack tip in a power law hardening material. J. Mech. Phys. Solids, vol. 16, pp. 1-12.
14. Shih, C. F.; German, M. D. (1981): Requirements for one parameter characterization of crack tip fields by the HRR singularity. Int. J. Fract., vol. 17, pp. 27-43.

Hare Rama Hare Rama Rama Rama Hare hare Hare Krishna Hare Krishna Krishna Krishna Hare Hare
15. Strouboulis, T.; Babuska, I.; Copps, K. (2000): The design and analysis of the generalized finite element method. Comput. Methods Appl. Mech. Engrg., vol. 181, pp. 43-69.

16. Strouboulis, T.; Copps, K.; Babuska, I. (2001): The generalized finite element method. Comput. Methods Appl. Mech. Engrg., vol. 190, pp. 4081-4193.

17. Zeinkiewicz, O. C., and Taylor, R. L. (1989) : The Finite Element Method, Vol. 1, McGraw-Hill International Editions