FINITE ELEMENT Ni=aatleleie A Ma Vet)por diae4 FINITE ELEMENT METHODS For VI Semester - Mechanical Engineering As per New Syllabus of VTU $.B. Haleshw.rect, (eno, miste, Associate Professor Department of Mechanical Engineering, Sir M. Visvosvaraya institute of Technology Bengaluru -562 187 Sapna Book House 3d Main Road, Gandhinagar, Bengaluru -560 009 Pr :40114455 [REFERENCE 800k]Finite Element Methods - A Text Book For BE. - Mechanical Engineering Students, by S.B. Haleshand published by Sapna Book House (P) Lid., R.0. #11, 3" Main Roed, Gandhinagar, Bengaluru - 560009 Ph, : 080-40114455 |, (Engg -243) ISBN :978-81-260-2112-4 © Author WOTICE OF CAUTION "= Reproduction Rights Dubllealion which 1s materal protected by this copyright notice may be reproduced or transmitted or utlized or stored in any form or by any means now ‘known of hereinafter invented, electronic, digital or mechanical, including photocopying, scanning, ‘recording or by any information storage or retrieval syetem, without pricr written permission from the publisher" © Photocopying and Resale Prohibition : “Copying of the book and selling it after photocopying or reselling it 2s second hand book is legal and is nct alowed, under the copyright act This book s Sold subject to the condition that it shall not, by way of trade or otherwise, be lent, resold, hired out, or otherwise circulated without the publisher's prior writen consent. Pages : xiv + 473 First Ecition_: Jan., 2013 I Revisad & Enlarged Edition; Jan. 2015 Reprint: Feb, 2015 lil Revised & Enlarged Edition Jan., 206 Reprint; Jan., 2017 eS Liprany \% AIR 2EF [FOR COPIES CONTACT OUR FOLLOWING DISTRIRUTO . UaSPabishersDstibvters Li, Bonu, zeae zzones 2 les Back Nose Began zane esos | 4 aaa ag Bergt . Po atts 87088 4 Veyamara Boot Depot abt Praauan S Vnaata mets, Haba Panis ‘Ase avaleble at ou Stovrcons 4 SADASHVWAGAR, Near Bhastyan Clie, Bengaluru Ph: (090) 4123 0271 4 JAYANAGAR, 4 Blick, Bengalura Ph, (020) 4906 6700 4 INDIRANAGAR, Opp. Chomays Kushne Terie, Bengalis Ph (C80) 4065 $909 4 KORAMANGALA, 80M Road, 7Ih Block, Bengatar Ph. 4080) 40839399 4 RESIDENCY ROAD, No. 22, ACR Towers, Bengalur Ph, : 060) 49166905 +4 ROYAL MEENAKSH MALL, 2 Fler, Bamorhats ne Benga Ph, en) 47586299 + ELEMENTS MALL, Su Fly, Traresandte Nat Rcac Bengatu Ph, (08) 8729 4151 + MYSURU, Norayan Shasiy Rosd, Devoe halle, Ph. (0821) co0eeae 4 VANGALURU, Excel Mal, KS. Rae Road Ph, oe) «232800 4 HUBBALLI, Laxmi Nat, Coen Roa Ph. : 0886) 2eot06 4+ DHARWAD, EDM Callige of Maal Sconce end Hospital Ph: sneae 44002 4 KALABURKG), Cy Corior Nal, Maret Rood, Near Jagath Cree rs, : (472) 2756081218811 + COIMBATORE, No 1, Ett Perysowerny Road, RS Puram a, azz) 4620608 Tpesetting and Cover Design by. CCOMATAUNK. Bengsiury-20. email: [email protected] Printed af : Print Rast, Bengalurs,SYLLABUS PART - A. UNIT - 1 INTRODUCTION Equilibrium equations in elasticity subjected to boxly force, traction forces, and stress-strain relations for plane stress and plane strains, General description of Finite Element Method, Application and limitations. Types of elements based on geometry. Node numbering, Half band width, 07 Hours UNIT - 2 Basie Procedure Euler ~ Lagrange equation for bar, beam (cantilever / simply supported fixed) Principle of virtual work, principle of minimum potential energy, Raleigh's Ritz method, Direct approach for stiffness matrix formulation of bar element, Galerkin's method, 07 Hours UNIT <3 Interpolation Models Tmterpolation polynomials- Linear, quadratic and cubic, Simplex complex. and multiplex elements, 2D PASCAL's triangle, CST elements-Shape functions and Nodal load vector, Strain displacement matrix and Jacobian for tiangular and rectangular element. 07 Hours UNIT +4 Solution of 1-D Bars Solutions of bars and stepped bars for displacements, reactions and stresses by using penalty approach and elimination approach, Gauss~ climination technique. 06 HoursPART - B Higher Order Elements Lageange’s interpolation, Higher order one dimensional elements- Quadtatic and cubic element and their shape fianetions. Shape furecion of 2-D quadrilateral element-linews, quadric element lsoparametric, Sub parametric and Super parametric elements. Numerical integration : 1, 2 and 3 gauge point for 1D and 2D eases 06 Hours UNIT -6 Trusses Stiffness matrix of Truss element, Numerical problems, 06 Hours UNIT -7 Beams Hermite shape functions for beam clement, Derivation of stitness matrix Numerical problems of beams carrying concentrated, UDL and linearly varying loads. 06 Hours Heat Transfer Steady state heat ransfer, ID heat conduction governing equations, Functional approach for heat conduction. Galerkin’s approach for heat conduction, 1D heat transfer in thin fins. 07 HoursCONTENTS UNIT 01 - INTRODUCTION 1-31 1.1 Overview 1 1.2 Basie Equations of Elasticity 2 1.2.1. Force Distribution 2 1.3. Stress At A Point 3 1.4 Differential Equations of Equilibrium 4 1.4.1. For A Two Dimensional Body 4 1.4.2 For A Three Dimensiona Body 6 1.5 Strain-Displacement Relations 8 1.6 Strass-Strain Relations 9 1.7 Plane-Stress 1 1.8 Plane - Strain 12 1.9 General Description of The Finite Element Method 14 1.10 Basic Steps In Finite Element Method 15 1.11. Background of Finite Element Method 16 1,12 Numerical Methods 17 1.13 Advantages of Finite Element Method 17 1.14. Disadvantages of Finite Element Method 18 1.15. Limitations of Finite Element Method 18 1.16 Applications of Finite Element Method 19 4.17. Discretisation Process 20 1.17.1 Type of The Element 20 1.17.2 Size of The Elements 23 1.17.3 Number of Elements 23 1.17.4. Location of Nodes 24 1.17.5 Node Numbering Scheme 25 1.17.6 Banded Matrix 26 1.18 Phases of FEA 27 1.18.1 Pre-Processing 27 1.18.2 Solution 27 1.18.3 Post Processing 28 1.18.4 Software Packages 29 Review Questions 30UNIT 02 - BASIC PROCEDURE 32-97 2.1 Introduction 22 2.2. Variational Approach 32 2.3 Calculus of Variations 2 2.4 Euler-Lagrange’s Equation 33 2.4.1 Boundary Conditions 8 2.8 Principle of Virtual Work 6 2.6 Total Potential Energy (MN) 8 2.6.1 Potential Energy Functional For A Three Dimensional Body 38 2.6.2 Principle of Minimum Potential Energy 40, 2.7 Rayleigh - Ritz Method aa 2.8 Galeskin’s Method - Method of Weighted Residucls 3 2.9. Stiffness Matrix of A Bar Element 88 2.9.1. Direct Method 89 2.9.2. Using Principle of Minimum Potential Energy £0 2.10 Global Stiffness Matrix 92 2.11 Properties of Stiffness Matrix 93 Review Questions 95 UNIT 03 - INTERPOLATION MODELS 98 - 160 3.1 Introduction 98 3.2. Types of Interpolation Models 98 3.3 Polynomial Form of Interpolation Model 9g 3.4 Simplex, Complex And Multiplex Elements 101 3.4.1. Simplex Elements 101 3.4.2 Complex Elements 101 3.4.3 Multiplex Elements 102 3.5. Order of The Polynomial Interpolation Model 102 3.6 Geometric Invariance - Pascal Triangle 103 3.7 Convergence Requirements toa 3.8 Compatibility Condition 105 3.9 Coordinate Systems 105 3.10 Relation Between Cartesian Coordinate And Natural Coordinate 106 3.11 Relation Between The Generalized Coordinates And Nodal Displacements 107 3.12 Shape Functions 1083.13 Finite Element Formulation of 1-D Bar Element 110 3.13.1 Derivation of Shape Function In Global Co-Ordinates/Cartesian Co-Ordinates 110 3.13.2. Derivation of Shape Function In Local Co-Ordinates 112 3.13.3 Derivation of Shape Funetion In Natural Co-Ordinates 113 3.13.4 Derivation of Strain And [B] - Strain Displacement Matrix 115 3.13.5. Derivation of Stiffness Matrix (Using Strain Energy and Strain Displacement Matrix) 17 3.13.6 Derivation of Load Vector 18 3.14 Finite Element Formulation of 2-D Constant Strain Trienguler Element (CST Element) 121 Review Questions 168 UNIT 04 - SOLUTION OF 1-D BARS 161 - 230 4.1 Methods of Handling Boundary Conditions 161 4,2 Types of Boundary Conditions 161 4.3 Elimination Method of Handling Boundary Condition 162 4.4 Penalty Method of Handling Boundary Conditions 164 4.5 Value of C 165 4.6 Effect of Temperature 211 4.7. Gauss Elimination Method 215 4.7.1. Generalizetion of The Method : General Alg 216 Review Questions 226 UNIT 05 - HIGHER ORDER ELEMENTS 231-270 5.1 Introduction 231 5.2. Lagrange Interpolation Function 231 5.3 Lagrange Interpolation Function For Linear Element 232 5.3.1 Shape Function For Two Node Bar Element 232 5.3.2. Shape Function For 2-D Elements - Quadrilateral / Rectengular Element 234 5.4 Lagrange Interpolation Function For Higher order Element 236 5.4.1. Shape Function For Three Node Bar Element (Quadratic Bar Element) : 236 5.4.2. Shape Function For Four Node Bar Element (Cubic Bar Element) 2395.4.3. Shape Function For Nine Node Rectangular Element (Quadratic Rectangular Element) zai 5.5 Serendipity Elements 243 5.8.1 Shape Function For Eight Node Rectangular Element (Quadratic Rectangular Element) 245 5.6 Properties of The Shape Function 246 5.7 Isoparametric Formulation 247 5.7.1 Necessity of isoparametric Formulation 247 5.7.2 Coordinate Transformation 250 5.7.3 Iso-, Sub- And Super Parametric Element 250 5.8 Numerical Integration 252 5.9 Gaussian Quadrature For 1-D Integral 252 5.9.1 One - Point Formula 253 5.9.2. Two - Point Formula 254 5.9.3. Three - Point Formula 255 5.9.4 Change of Intervals 257 5.10 Gaussian Quadrature For 2:D Integral 258 Review Questions 270 UNIT 06 - TRUSSES 271- 310 6.1 Introduction 2m 6.2 Element Stiffness Matrix [K] 272 6.3 Element Stress Matrix [SI 275 6.4 Finite Element Equation For A Truss Element 276 6.5 Reaction Forces {R} 276 Review Questions 308 UNIT 07 - BEAMS 311-364 7.1 Introduetien ant 7.2 Hermite Shape Function 3 7.2.1 Derivation of Hermite Shape Function of Beam Element 312 7.3 Displacement Equation 317 7.4 Stain, Stress And Bending Moment In A Beam Element 318 7.5 Potential Energy Functional For Beam Element 319 7.6 Element Stiffness Matrix 320 7.7 Derivation of Shape Function 3237.8 Derivation of Element Stiffness Matrix 327 7.8.1 Method - 1; In Global Co-Ordinate/Cartesian Co-Ordinate 327 7.8.2. Method - 2 : In Global Co-Ordinate/Cartesian Co-Or 329 7.9 Derivation Load Vector Due To Uniformly Distributed Load 332 Worked Example 335 UNIT 08 - HEAT TRANSFER 365 -420 8.1 Introduction 365 8.2 Basic Equations 365 8.2.1 Energy Balance Equation 308 8.2.2 Rate Equation (Laws of Heat Transfer) 306 8.3 One Dimensional Heat Conduction 307 8.4 Differential Equation For An 1-0 Heat Conduction 307 8.5 Differential Equation For A Three-Dimensionel Body 368 8.5.1 Boundary And Initial Conditions (Heat Boundary Conditions) 370 8.6 Finite Element Governing Differential Equation 372 8.7 Finite Element Formulation - 3d 372 t 8.7.1. Variational Approach 372 t 8.7.2. Galerkin Approach (Wr) 375 | B.8 Finite Element Formulation of An 1-D Heat Conduction 377 ; 8.8.1 Derivation of Shape Function 377 ; 8.8.2 Derivation of Stiftness Matrix or Conductivity Matrix 378 | 8.8.3 1-D Heat Conduction With Free End Convection 380 | 8.8.4 1-D Heat Conduction With Convection And Internal Heat Heneration 381 8.9 Boundary Conditions 385 | 8.9.2 Specified Heat Flux Boundary Condition 385 | 8.10 Galerkin Approach For An 1-D Heat Conduction 386 8.11 Heat Transfer In Thin Fins 387 812 Heat Transfer Through Tapered Fin 391 8.13 Heat Transfer Through Composite Wall 393 Review Questions 419 APPENDIX 421-434 ee PREVIOUS YEAR QUESTION PAPERS © 435-471 eee REFERENCES 472-473UNIT nk lc INTRODUCTION 1.1 OVERVIEW ‘Various phenomena treated in science and engineering are often described in terms of differential ‘equations formulated by using their continuum mechanics models. Solving differential equations under various conditions such as boundary or initial conditions leads to the understanding of the phenomena and can predict the future of the phenomena. Exact solutions for differential equations, however, are generally difficult to obtain, Numerical methods are adopted to obtain approximate solutions for differential equations. The finite elemen: method is peshaps the most popular numerical technique for solving enginecring problems. The method is general enough to handle any complex shape or geometry, any material properties, any boundary conditions and any loading conditions, ‘The generality of the finite element method fits the analysis requirements of today’s complex ‘engineering system and design, where closed form solution of governing equilibrium equations are usually not available. In addition, it is an efficient design (ool by which designers can perform parametric design studies by considering various design cases (¢ifferent shapes, materials, loads ic), analyses them and can choose the optimum design. ‘The finite element method is a nomerical technique for obtaining approximate solution 10 a wide variety of engineering problem. The method originated in the aerospace industry as a tool to study siresées in complex air-frame structure, The method has gained increased popularity among both researcher and practitioners, Since its birth, the finite element method has gone through various developmental stages. Started ‘with formulation of finite element theory, varistional and Galerkin’s approaches for two dimensional and three dimensional problems were formulated. Isoparametric formulations for both linear and nonlinear, static and dynamic analysis were developed, Also computational methods required by the finite element method were established. This method includes numerical integration of element cquations, solution of the global system of algebraic equations, study of numerical stability and error analysis solution algorithms. .2__—Finite Element Methods 4.2 BASIC EQUATIONS OF ELASTICITY ‘The theory of elasticity is the branch of solid mechanics which deals with stresses and deformations in elastic bodies produced by external forces or change in temperature. Generally speaking, elasticity problems are reduced to solving the partial differential equations known as the equilibrium equations together with the stress-strain relations or the constitutive equations, the strain-cisplacement relations, ‘and the compatibility equation under given boundary conditions, The exact solutions can be obtained in quite limited cases only and in general cannot be solved in closed forms, In order io overcome these difficulties, the Finite Element Method (FEM) has been developed as one of the powerful ‘numerical methods to obiain approximate solutions for various kinds of elasticity problems. The study of and application of FEM requires a sound knowledge of the basic equations of elasticity. 4.2.4 Fonce Distmipurion ‘The forces acting on a body can be classified as follows: Internal forces - The intemal forces are the reactive forces which are setup due to external forces, And within elastic limits, the internal forces are numerically equal to the external forces. External forces - The state of stress and strain in a body arises due to external influences. The external forces acting an a body can be divided into three types as shown in figure(1.1). ‘T-Surface force/Traction Fig. 1.1: Forces acting on a body 1. Body force () It isa distributed force acting on every elemental volume of the body. Unit is Force / Unit volume (Nimm’). Ex: Self weight due to gravity, electromagnetic forces, centrifugal Forces, inertia forces etc, Body force vector fis given by fe #¥ (1) Where f,, {and f, are the body force components in x,y, and z direction respectively, ‘The work potential is the work done by the external force. Therefore work potential (WP) due to body force is given by WP due to body force (1.2) je “falIntroduction __3 Where U'={u vw} and .vand ware the displacement components in x, and z directions respectively. 2. Traction (7) I isa distributed force acting on the surface of the body. Unitis Force / Unit area (Minn). But for one dimensional problem, unit is Force / Unit length, Ex: Frietional resistance, viscous drag and Surface shear. Force vector T due to traction is given by Where 7, T,, and 7_ate the traction components in 2, y, and z directions respectively. Therefore work potential (WP) due to traction/surface force is given by (4) WP cue 10 Traction T= Ju'rds Where UT={u vw} and u ,vand w are the displacement components in x, y and z directions respectively. 3. Point lod (P) It isa force acting at a particular point which causes displacement, Point load or concentrated load P is given by Pa as) RO Jat Bak). The work potential due to concentrated load ‘P" is given by WP due to Point load P= Durr (16) +. The total work potential due to all these force is given by we = -| {ura formu Sura) tL 4) a Note: Work. Potential is negative since energy is obtained from the external sources 1,3STRESS AT A POINT - Stress in its most general form can be defined as the force per unit area. Stress at a point can be defined in terms of six components of stress and is resolved into normal and shear components: ‘The normal stress 0, acts on the plane perpendicular to the x-direction (This applies to 6, and o4__—Finite Element Methods also) and shear components acts at arbitrary angle to the coordinate axes, The shear stress has ‘ovo components and need two subscript; the first subscript represents the plane in which the stress acts and second subscript represents the direction in which siress acts. For example*t, is the shear stress in the plane perpendicular io x-axis in the y-direction. In order to establish the state stress a point, nine quantities musi defined; 6, 0, , 0, (three normal siresses) and €,, 5%, fst, (O18 shear stresses), If stresses are slowly varying cross the infinitesimal cube, moment equilibrium, about the centroid of the cube requires that t,,= 7, .= t,, and, = t,, The nine components now reduces to six independent quantities 6, .6,, 0, and t,, and which can be represented as Oy, gee Gy=:[ty 0, ity (18) Te te 1.4 DIFFERENTIAL EQUATIONS OF EQUILIBRIUM A body is said (0 be in equilibrium if under the static loads, the reactive forces and moments developed balances the externally applied force and moments. In other words, overall body must satisfy the force and moment equilibrium equations. 1.4.1 For & Two Dinensionat. Boor Consider a two dimensional element of unit thickness as shown in figure (1.2). It is subjected to normal stresses. 6, and , shear stress t,, and body forces f, and f.. The stresses acting on the clement are assumed to be constant as they act on the width of each face, but vary form one face as. to the opposite. For example, 6, is acting on the left vertical face, where 6, + dv is acting on the right vertical face, Fig, £2 : Two Dimensional stress elementIntroduction — 5 For equilibrium of the system, EF, = 0 and ZF, =0, Thus, Considering, 2F, = 0 .e., adding all the forces acting on the element in x-direction, we have 6, + aby 6,dr4| 1, + ayer tyr funy = 0 as) a te ty - 6 ,dy By ly ya fet 2 20, Ie : 8+ Seid — a y+ tye 5 dye tye faded =" Fe aay + dydx+ faardy = 0 Since, dedy #0 OG, | My 0, , My Oe + oF af (1.10) ‘Similarly, S Considering, EF, = 0 ic., adding all the forces acting on the element in y-direction, we have a, ary 0,15 ls odes] ty +e birt Fey = 0 =.) 6 ts es 6 + ty + dey et Fey daty + 5 dys fe Wr etre Sidi + Gedy + feds = 0 Since, dxdy 40 a0, , ey erie ps ea thts, 20 (12) Considering, 2M =, i.c.,Taking moments about the origin ‘O" , we have Se ea [>. is Be acla( y+) (tody)r+ [+e a }+an) fonts er' if, 22 aye) feat -[ ag aude) fy dxa{y09 +4) ee vard( re) og qu)6_—Finite Element Methods ‘Neglecting the higher order terms, we get aa, , a, ao, (c,.-t, )axay {® +oE4y, Je te a Fes pled = f, Jeary = 0 ay ae But, from Eg.0 and veo * tnt, =0 ‘Thos, 1,21, (114) ~ 15} a tf =0 == (116) dy 1.4.2 For a THREE DimeNsiona Booy Consider a three dimensional element of size dx, dy and dz as shown in figure(1.3). tis subjected to normal stresses 6,, 0,and 6, shear stresses t,t, and 1,, and body forces ff, and f, The stresses acting on the element are assumed to be constant as they act on the width of each face But vary form one face to the opposite. For example, 6, i8 acting on the left vertical face, where ot ae is acting on the right vertical face. For equilibrium of the system, EF =0 and ZM = 0 in all the direction. Thus, Considering, 2F, = Di.e., adding all the forces acting on the element in x-direction, we have a [o. eas =6,dyde {e Se =t dade fn + te de dy aay + Sudrdyde =) (1.17) On isc ody + © pide apriere gies” Seated tye eu didy + daddy — egal + fede =0 +a 1, ser =0 azIntroduoton —_7 * ay| ff Z Fig. 1.3 : Three dimensional stress element Since, drdydee0 dO, , My, te * ae ay + ry + feed a> (1.18) Similarly, considering, EF, = 0. and EF, = 0 i.e., adding all the forces acting on the element in yedirection and z-direction, we get 20, Hey, Be 4p a Wy oe tae TH =o Seely) or, 2s Me ayy og (1.20) oz ax fy q.(1.18), Fq(1.19) and £q.(1.20) are known as the differential equation of equilibrium for a three dimensional body.8_—Finite Element Methods 1.5 STRAIN-DISPLACEMENT RELATIONS Consider a two dimensional element abed which lies in the x-y plane as shown in figure(|.4), ‘When a force acts on the element, it undergoes deformation and it becomes abic'd’, Displacement in the x-direction is wand y-direct wee eformation of a element abed In general, normal strain is defined as the ratio of change in length to original length of the body. Considering elementab in.x-ditection, a'b-ab Strain, 20 From figure (1.4) a feat cet = [ue Bal oa] 0am fae) (avy Using the binomial theorem and neglecting the higher order terms, ie.{ 5) and [5° We have, au ait’ = dev Shas (1.23) Substituting Eq (1.22) and Eq1.23) in Eq.(1 21), we get ax. x de de aya {ie 34-1) Sig SeIntroduction 9 Strain inx-direction, == (1.24) Similarly, considering the element ad in y-direction, av Strain in y-direction, By (1.25) ‘The shear strain represent the amount of distortion abcd. The angle dab is a right angle in the undeformed state but has been distorted to angled'ab' by shearing, The change in angle is denoted composed of two parts denoted by o and B, given by slopes of ab! and ad', respectively as dv/dx and du/ay. Thus, : = 4H 1.26 Shear strain, Yeah (1.26) (1.24), Bo.(1.25) and Eq.(1.26) are the strain-displacement relations for a two dimensional element For a three dimensional element, considering w as the displacement in z-direction strain- displacement relations are obtained by extending two dimensional analysis. Strain in z-direction, ae ™ (127) % aw ae Shear strain, ye eS 1.28) ear strain, Y ye (1.28) Shea Su, Oe 129) year strain, Eta (1.29) 1.6 STRESS-STRAIN RELATIONS. ‘The equations between stress and strain applicable (o a particular material are known as the constitutive equations for that material. In the most general type of material possible, itis shown thatthe constitutive equations can contain up to 81 independent material constants, However, for a homogeneous, isotropic, linearly elastic material, it is shown that only two independent material constants are required to completely specify the relations, These two constants are modulus of ‘elasticity (Young’s modulus) and Poisson’s ratio, Again referring to the simple uniaxial tension test, the modulus of elasticity is defined as the slope of the stress-strain curve in the elastic region or Si Bee Where, itis assumed that the axis of loading corresponds to the x axis, As strain is dimensionless, the modulus of elasticity has the units of stress usually expressed in mega pascal (MPa). This relation between the stress and strain known as defined by Hooke’s Law, which is independent of time and loading history. 0.30)10_—Finite Element Methods Poisson's ratio is defined as unit lateral contraction unit axial elongation “Thus, in the tension test, if €, represents the strain resulting from applied Joad, the induced strain ‘components are given by &, = €, =-ve, “The general stress-strain relations for a homogeneous, isotopic, linearly elastic material subjected toa general three-limensional deformation are as follows: Normal strains are given by (131) ~~ (1.32) (133) Shear strains are given by 34) (135) (1.36) Similarly, we have Normal siresses are given by 6, = awit +(e, +e.) (137) : 8, = Tacall ve tee te] 038 3, = aay. +e. +e,)] “139 Shear stresses are given by = Oy, 1.40)Introduction — 11 E Te = Whew reer tA) E » = ey" (1.42) Where, G is the shear modulus or modulus of rigidity, defined by E C= ey) te) ‘Thestrest-strain relations can easily be expressed ift matrix form by defining the material property ‘matrit (D) as (ey) ov v 0 0 0 vo (ey oy 0 0 0 vow (Hy 0 ° 0 E D-tagem 0 oO | «4s 0 0 0 0 {o} = {ox 95 a (145) fo} = [2] fe} (146) 1.7PLANE-STRESS. Plane stress case-is represented by a thin plate in:x~ plane; plane subjected to in-plane loads along x- and/or y- direction, and no load along the normal tothe plane (iz direction) is as shown infigure 1.5. Definition: Plane stress is defined to be a state of stressin which the normal stress (6) and shear stress (x) directed perpendicular to the plane are.assumedito be zero. ‘Example: Thin plate with hole, Thin cantilever plate: For plane stress, we have Ope t=t=0) and G8} st, FG) a (14) ‘A plate-loaded im its mid-plane:is said to be:ina state-of plane stress ora membrane state, if tie following assumptions hold:12 _—Finilo Element Methods Assumeri0ns 1. All loads applied to the plate actin the mid-piane direction, and are symmetric with respect tothe ‘micl-plane. 2. All support conditions are symmetric about the mid-plane. 3, In-plane displacements, strains and stresses can be taken to be uniform through the thickness. 4, The normal and shear stress components in the z direction are zero or negligible. ¥y y tiles M “ | |» Fa E jen tht tt WH t Fig. 1S : Plane stress : Thin plate with in-plane loading, ‘The governing equations are 4) Equilibrium Equations a (1.48) = (1.49) ~~ (1.59) 1.8 PLANE - STRAIN Plane strain case is represented by 2 thin plate in. y plane, which is constrained along the normal to the plane is z-direction (ic., no strain along the normal). Its subjected in - plane loads along x and/or y direction and no load along the normal to the plane {and hence no stress) as shown in. figure 1.6 Definition: Plone strain is defined to-be a state of strain in which normal sirain and shear strain normal to the xy piane are assumed to he veraIntroduction — 19 For plane strain, we have €, = 4-120 and u(x, V=V Gr yu, =0 ~ (51) Assumerions 1. The length of the structure is very large in comparison with the other two dimensions. ‘2. The loads are applied only normally to the longitudinal axis (the loads are perpendicular to the z-axis), The loads are uniformly distributed along the length of the body. 3, The support conditions are the same along the z-axis Examples~ A long cylindrical tube under internal pressure, a dam restrained at both ends, tunnels, long stip footings Fig. 1.6 A long cylinder «plane strain ‘The governing equations are }) Equilibrium Equations as, , 2, ay a, Rt +f =0 and at +f,=0 w= (1.52) fi) Strain « displacement relation ee Oe Dr te Fea IM UE ae t By (1.53) Stress - Strain Relation14 _—Finte Element Metiods. Differentiate between plane stress and plain strain problems with examples. Write the stress strain relations for both. Dec’ 2013-2014(8M) Plane stress I Plane strain ~ Applicable to bodies whose dimension is very small in one of the coordinate dimensions. Thin plates loaded in [Applicable to bodies that are very long and whose geometry and joading do not vary significantly in the longitudinal direction, the plane of the plate. and t. y = = = = Be p> Le Plate with hole Dam subjected to Horizontal loading i) be #00, = Fu -9y) 0, 40 5, =VG qu — i p. Example: in-plane loading of thin Example: Analysis of dams, cylinders, | late, ring press fitted on a shaft retaining walls | | On, Ty © |fex| | fo, in v | p. {o,, VL 0 Nest ent = Gaia) I-v oo 4, 0 ty Zo} | tte 1.9 GENERAL DESCRIPTION OF THE FINITE ELEMENT METHOD In the field of Engineering Design we come across many complex problems, the: Mathematical Formulation of which is tedious and usually notpossible by analytical methods. At such instants we resort 10 thewuse of Numerical techniques.’ Here lies the importance of FEM, which is a very powerful tool for getting the Numerical solution of a widerange of engineering problems. Thebasic concept is that a body or structure may be divided into smaller elements of finite dimensions called as “Finite Elements”. The origins! body or structure is then considered as an assemblage of these elements connected ata finite number of joints called as “Nodes” or “Nodal Points”, The properties f the elements are formulated and combined to obtain the propectics of the entire body. ‘The equations of equilibrium for the entire structure or body are then obtained by combining the j ‘equilibrium equation of each element such that the continuity is ensured at each node. The necessary boundary conditions are then imposed and the equations of equilibrium are the solved to obsain the required variables such as Stress, Strain, Temperature Distribution or Velocity Flow depending on. the application.Introduction 18 ‘Thus, instead of solving the problem for the entire structure or body in onc operation, in the method aitention is mainly devoted to the formulation of properties of the constituent elements. Acommon, ‘procedure is adopted for combining the elements, solution of equations and evaluation of the required variables in all fields, Thus the modular structure of the method is well exploited in variousdisciplines of Engineering, Definition: The Finite Element Method (REM) is a numerical analysis technique used by engineers, scientists, and mathematicians t0 obtain solutions (approximate solutions) 10 the differential equations or partial differential equations thai describe, or approximately describe a wide variety of physical and non-physical problems. Physical problems tange in diversity {from solid, fluid and soil mechanics, to electromagnetism or dynamics ' 110 BASIC STEPS IN FINITE ELEMENT METHOD ‘The solution of a general continuum, problem by the finite element method always follows an orderly step-by-step process. The step-by-step procedure with reference the static structural ‘problems can be stated as follows: 1. Diseretisation of the continuum: The first step in the finite element method is to divide the given continuum into smaller regions of finite dimensions called as “Finite elements" The original Continuum (body or structure) is then considered as an assemblage of these elements connected at finite number of joints called as “Nodes” or “Nodal Points”. Ateach node, unknown displacements are 10 be prescribed. The type, size, number, and arrangement of the elements depend on the accuracy of the solution required. 2, Selection of approximating functions: ‘Approximating functions are also known as the displacement function or interpolation model. Displacement function is the starting point of the ‘mathematical analysis. This:represents the variation of the displacement within the element. The displacement function may be approximated in the form a linear function or @ higher-order function. ‘A-convenient way toexpress tis by polynomialexpressions. The shape or geometry ofthe element may also'be approximated. 3.Formation of the element stiffness matrix: After continuum is discretised with desired element Shapes, the individual element stiffness matrix is formulated: Basically it isa minimization procedure ‘whatever may be the approach adopted. The geometry of the element is defined in reference to the global frame, Coordinate transformation must be done for elements where it if necessary. 4: Formation of overall stiffness matrix: After the element stiffness matrices in global coordinates are formed, they are assembled to form the overall stiffness matrix. The assembly is done through the nodes which are common to.adiacent elements. The overall stiffness matrit is symmetric and banded Overall stiffness matrix is also known as the global stiffness matrix. ‘8. Formation of the element losiding matrix: The loading forms an essential parameter in any structural engineering problem. The loading inside:an element is transferred atthe nodal points and. consistent element matrix is formed, 6:Formation of the overall loading matrix: Like the overall stiffness matrix, the element loading ‘matrices are essembied to form the overall loading matrix. This matrix has one column per loading case and it is either a.column vector or ¢ rectangular matrix depending on the number of loading ses.16 _—Finito Elomont Methods equilibrium equation: Overall equilibrium equation is the systematic arrangement of the overall stiffness matrix, overall load vector and overall displacement vector to get set of simultaneous equations. Overall equilibrium equation can be expressed as shown below: le} = 4} OSD Where, [dis a overall or global stiffness matrix (Square matrix) {Q}isa overall or global displacement vector (Column matrix) {F} is a overall or global force vector (Column matrix) 8, Incorporation of boundary conditions: The boundary restraint conditions are to be imposed in the stiffness matrix to avoid the condition of singularity. The solution cannot be obtained unless support conditions are included in the stiffness matrix. This is because. if all the nodes of the structure are included in displacement vector, the stiffness matrix becomes singular and cannot be solved ifthe structure is not supported amply, and it cannot resist the applied loads. 6. Calculation of unknown nodal displacements: After incorporation of boundary conditions, Anode lonode 20 nods o- Prism Elements Tnode Inde Wade Hexahedral Elements ep rae a Z S node 20 node 32node Fig. 1.10 : Three dimensional 4, Axisymmetric elements: ‘The axisymmetric elements are developed by rotation of a planar (triangle or quadrilateral) about an axis located in a plane of the element through 360°. These elements are used when the geometry and loading of the problems are axisymmetric such as analysis of shaft, cylindrical storage tanks ete. Figure 1.11 shows axisymmetric elements, N Ny Ol Fig. 1.1: Axis symmetric elementsIntroduction 28 1.17.2 Sze OF THE ELEMENTS The size of the elements chosen in the finite element model plays-an vital role inthe finite element analysis because, the size of the elements influences the convergence of the solution, ‘Therefore it has to be chosen with lot of care, Ifthe size of the elements is fine Le. small the final solution is expected (0 be more accurate but ar the same time, we have to remember that use of fine mesh (small clement size) will also mean more computational time, As or as possible we have touse the same element size throughout the body but some times, we may have to use different sizes in the same body. For example, in the case of stress analysis of a plate with a hole elements of different sizes have to be used as shown in figure (1.12). The size of the elements has to be very small near the hole (where the stress concentration is expected because of the change in cross section) compared to the far away places. In general, whenever there is a change in cross section or the ‘where the stress concentration in expecied, the size of the elements should be small in those s. Another characteristic related to the size of the elements which affects the finite element solution is the aspect ratio of the elements. The aspect ratio describes the shape of ihe element in the finite element model. For a two-dimensional elements, the aspect ratio is taken as the ratio of Jongest dimension of the element to the smallest dimension of the same element. Elements with an aspect ratio of nearly unity generally yields best results. Plate with a hole (Geometric model) FE model with different elem Fig. 112 4.17.3 Numozn of Etemenrs Selection of the number of elements in the finite element model is related to the size of element, number of degrees of freedom involved and more importantly accuracy required, Although an inerease in the number of clement (use of fine mesh size) generally yields more accurate results for a given problem, but thece will be a certain number of elements beyond which the accuracy cannot be improved by any significant amount. This behaviour in shown graphically in figure (1.13) Importantly, we have to remember that use large number of elements will also mean large number of degrees of freedom, more computational time and more the computer memory requited to store resulting data,24__—Finte Element Methods Accuracy Number of Elements No significant improvement beyond No Fig, Lt. 1.17.4 Locanion oF Noves Node is a point where the properties are defined and determined for an element. Therefore location ‘of node is very important, Ifthe body has no discontinuity, then the body can be divided into equal number of elements and hence the spacing of the nodes ean be uniform. On the other hand, if the body has discontinuity such as discontinuity in loading, discontinuity in geometry, or discontinuity in ‘material properties, then the nodes hive to be intracluced or placed at those discontinuity as shown + Kffect of varying the number of elemenis, x e ; ; tote j 4 P (aipdieted tabs le (h oicanucy aces [Node {7} fae eee : Yo Kiet Nodes to Cae jnoonees (snes a posta nea discontinuity in geometryer {ttt Node ee (e) Plate with a crack - discontinuity in materiat Fig, LI 1.17.5 Nove Numerine ScHeme ‘The size of the global stiffness matrix resulting from the finite element analysis depends upon the nade numbering scheme and the number of degrees of freedom considered per node. If we can minimize the size, both solution time and storage tequirements can be minimized. To achieve this minimization, we have to perform a systematic subdivision and adopt an appropriate numbering system for nodes. The band widih (Strictly speaking half band width) of the global stiffness matrix can be defined as Half band width B = (D4 DF (1.58) Where, D is the maximum of largest difference between node numbers in an clement and all the clements are considered for its determination and f'is the number of degrees of freedom at each node. Bq.(1.58) indicates that D has to be minimised in order to minimize the ban width. As an example consider various numbering scheme for two-dimensional problem. Assume that there are 2 degrees of freedom at each node, Let the node numbering of this structure in different ways one along the shorter dimersion and other along the longer dimension as shown in figure 1.15(a), (b), () and @. DIneNsIONING ALONG THE LonGeA EDGE (SIDE) B=(I7+1)x2=36 B= (0+ 1)x2=20 Pig. (2) Fig.)26 __—Firite Element Mothods Dimenstontnc ALONG THE SHORTER EDGE (IDE) : : e 7 + a cr ey : es a Perret Fig. 1.15 : Node numbering scheme Thus a shorter band width can be obtained simply by numbering the nodes across the shortest dimension of the body. 1.17.6 Banoeo Marax Ina banded matrix non-zero elements are contained within the band and outside the band elements are zero, Inthe matrix, all non-zero elements (coefficients) are on either side of the main diagonal joining the elements k,, and k,,, and these are within the band whose width is given by “abl called the hall ‘band wideh, The half band widthor semi-band width is defined as the greatest number of coefficients ‘of matrix in any row of the matrix from and including the leading diagonal to the right hand-side non-zero coefficients, Its value can be calculated by the Fiq.(1.58) Uses ‘The main use of banded matrix lies in reducing the space required to store the given matrix ancl minimize the solution time. Thus banded matrix is used to store only non zero elements, These be stored compactly in nx Temuative Solation,u @ i j Fig, 21 + Region (x, x, ] From figure 2.1), let the values of wr) be preseribed on the boundaries as n(r,) =, and u(s,) = u,, These are the called boundary conditions of the problem or trial solutions i7(x) are tried for a given problem and the functional is expressed! in terms of each of the tentative solutions, From all such possible solutions satisfying the boundary conditions, the solution which satisfies the variational principle governing the behaviour will be the one which makes the functional asstume a stationary or extreme In variational procedure, a series of tent ‘The mathematical procedure used to select correct solution from a number of tentative solution is called the calculus of variations. 2.4 EULER-LAGRANGE’S EQUATION - (2.2) Ae A= J? Posies e)te be the functional Where, variable and its first and second derivatives with respect tothe variablex, 1’ and 1” are functions of x. and [x, x, isthe region in which A” is defined. ‘The condition for the functional A to be mi zero. uum is the variation is functional must be equal to ic, 84 =0 ‘Thus, from maximization or minimization of simple function in calculus (OF OF. OF 8A = iE (Favs Moved Bu ‘\is<0 (2.3) oe 84 = ["8rds=0 - (24)84 _—Finte Element Methods Integration by parts for the respecti f° auas (25) |, Ou 2 OF 5 OF Seed = [%s 26) 2 OF syne [oul [Af 2E fa)” and [° Ze8urae = (em: (ae + en ‘Substituting into Eq, (2.3) and simplifying, we get [OF d (dF OF f (2-22 -S) me (ese) Since Su is arbitrary, each term must vanish individually so that ar_d (ar), d? (ar [= -4)-5(]] ny en) The Eq.(2.9) is know as the Euler equation or the Euler - Lagrange’s equation ~ [&seyl- (S=] =0 —[email protected]) Eq, (2.10) and Eq.(2.11) gives the boundary conditions. ‘The conditions on (2.12), and = (2.13) are called the natural boundary conditions.Basic Procedure —_35 fied, they are called free boundary conditions. If hey are not satisfied, we should, Ifihey are sat have bu (x) Bux, =0 . : 2.14 bu OK) = 0. Bea =0 a in order to satisfy Eq.(2.10) and Eq (2.11), These are called geometric or forced or essenti boundary conditions. Thus, when the Finite element equations are derived on the basis of the variational principles, the natural boundary conditions are automatically incorporated in the formulation; lence, only the geometric boundary conditions are-to be enforced on the solution. 2.4.1 BounoaRy CoNomions ‘A problem in structural mechanics. not completely specified unless boundary conditions are specified {In fact without the imposition of boundary conditions, the element and giobal stiffness matrix are singular that is, their determinants vanish and their inverses do not exist. The physical significance of this is that a loaded body or siructure is free to experience unlimited rigid body motion unless some support or kinematic constraints are imposed that will ensure the equilibrium of loads. These constraints are the boundary conditions. ‘There are wo types of boundary conditions i) Kinematic or Displacement Boundary Condition: In this type of problems only displacement and slope i.e, derivative of displacement may be specified. i Static or Force Boundary Condition : In this case, where forces and moments may be specified. Exampce oF Bounpary Conommons. Figure (2.2) shows a cantilever beam AB subjected to a uniformly distributed load, Ifthe vertical displacement v at any point is taken as afield variable, it must satisfy the differential equation ~~ (2.15) Which is an equilibrium condition. ‘The solution to the above equation must also satisfy the boundary condition at A and B as follows Kinewanc Bounoary Conomons ar A: a Displacemen: v= O.and slope or its derivative F° = 2.16)36_—Finite Element Methods rane pouNoARY coNoMmON Ar Bs Deflection equation of the beam is given by dh ete — (2,17) Differentiating the beam equation with respect tox , we get dy d(M. EF = =F = (2.18) de de viet a() : ee Where, “is the rate of change of bending moment = Shear force = F ay Le. Shear fo i a, one (2,19) ie, Shear force at 19) and dy Bending moment eI 0 se (2.20) Inma.. Conorions In these problems where time is involved init ly may be specified 2.5 PRINCIPLE OF VIRTUAL WORK. ‘The word ‘virtual’ means ‘having the effect of, but not the actual form of, what ix specified’. Thus ‘we can imagine ways in which to impose virtual work, without worrying about how it might be achieved in the physical world, “The principle of virtual work isa very useful approach for solving varieties of structural mechanics problem. When the force and displacement are unrelated to the cause and effect relation, the work is called virtual work, Therefore, the vinual work may be caused by true force moving through imaginary displacements or vice versa. Thus, there are two ways to define virtual work, as follows:eC Basic Procedure — 37 |. Prineiple of Virtual Displacements: Virtual work is the work done by the actual forces acting | onthe body moving through a virtual displacement. This means we solve an equilibrium problem through geometry, 2. Prineiple of Virtual Forces: Virtual work is the work done by the virtual force acting on the body moving through the actual displacements. This mcans we solve a geometry problem through equilibrium. ‘The principle of virtual forces establishes the compatibility conditions. The principle of virtual éisplacements establishes the conditions of equilibrium and is used in the displacement model of the finite element technique. Avawvsis oF THE STATEMENT ‘Consider a three dimensional body subjected to surface force and body forces, The external virtual ‘work is the work done by real load moving through imaginary displacements in a structiare. These Toads include both the load distributed over the entire surface and volume, Thus, the virtual work done by the external force is: qT, iF, aw, = [Bu 8 Bw} iT, pds + [fu By Bw))F, pay : Tho == (221) Where, 8u, 8v and 8 are the components of the viral displacements in x, y and = direction respectively. 7,, 7, and 7, are the surface forces and F,, F, and F_ are the body forces in.x, y and direction respectively, In the Eq, (2.21), the integration is carried out over the entire surface in the first term and over the entire volume in the second term, The Eq. (2.21) can be rewritten as aw, = [6OF {rlas+ JatoF {Fw P 2.22) Where, {OF = fou ov aw}, het, 7% 7Paddr}= i, ww For the three dimensional stress-strain condition, there are six components of stresses (6, 0, 0, t,, ‘_£,) and six components of strains in virtual displacement fields (8e, Be, 5e, 87, Sy, 8y,). The virtual internal work can be expressed as follows: Bu = J Be, Be, 8, Hy Ye Byah pV ..223)38 _—Finite Element Methods ou (2.24) J 8eF fhav According to principle of virtual work, the work done by external forces due tothe viral displacement ofa structure in equilibrium is equal to the work done by the internal forces for the virtual internal lisplacement. . BW, = 8U ~~ (2.25) Thus, JOY fas [SloY rw = foley fohav 2.29) 2.6 TOTAL POTENTIAL ENERGY (11) ‘The total potential enerey of an elastic body is defined as the sum of the strain energy due to internal stresses produced and the work potential due to the external force. ‘Total potential energy(PE) = Strain Energy (SE) + Work Potential (WP) e, PE functional, [1 = SE +WP 27) 2.6.1 Porenriat Eweray Fuxcriowat. For a THREE Dinensionat Booy Consider a three-dimensional elastic body of volume v, subjected to body force, surface force and a point loads, Let «, and w be the displacement components in.x, », zdirection respectively. From Eq.(2.27), we have, Potential Energy Functional = SE + WP ie 1 = sE+wP + (2.28) ‘The strain energy of the body is given by area under the curve (Figure 2.3) (For linear elastic materials) Stress Strain = —> Fig.23: Stress-Strain diagramBasic Procedure 38 ie, Strain Energy, SE = 5 6%e x volume L229) Fora small element of volume dv within a body SE= ore dy 1 2 For entiebody, Toul SE = 5 Jot ea = (230) Where, a"=[0,, 9), O. Gy Ox. Fu] ‘Work Porenriat, Work Potential due to extemal forces, i., Body force (), Surface force/Traction (7), and Point Joad*P* is givenby §) Work Potential due to Body force Forsmall element WP = (faut fy fow)dv For entire Body, we = f(t Fv4 fu)av we = [uray 231) Where, U7 =(uvwland f= Uf, £1" il) Work Potential due to Traction/Surface force Forsmallelement, we = (Tu+T,v+T.w)ds For entire Body, we = -[(Tus Tue Twas we = -fuTras w= (2.32) Where, T= 17, 1, TI" fil) Work Potential due to Point loads WP = {Pu P.v+ Pw) =-U'P40_—Finte Element Methods If there / number of point loads, then we = -Yure, «Total Work Potential due to external forces is aes {jer fave furresSure] gas Substituting Eq (2.30) and Eq,(2.34) into Eq.(2.28), we got PE functional for a three-dimensional body is, : x pore aor raefurrae Sur] e a “Thus. Eq 2.35) gives expression for the potential energy. functional fora three-imei 2.6.2 PrINciPLE OF MiNMum POTENTIAL ENERGY Ik states that “Of all the displacement configuration a body can assume which satisfy compatibility conditions and boundary conditions, the configuration which satisfying equilibrium condition is one which will have the minimum potential energy” ‘Thus, fromE9(2.27), Potential Energy functional, 1 =SE+ WP For Pé functional to be minimum, 8G) = BSE) + HWP) =0 ~~ 2.36) For example: Consider two bodies X and ¥ having the same weight falling through the different heights and h, as shown in figure (2.4), satisfies boundary conditions 2 and compatibility conditions C. Out of these two bodies, the body Which has minimum potential energy is body Y because it attain the equilibrium state earlier than the body X, when allowed to fall freely. Thus any body which has the minimum potential energy satisfies stability condition orany body which isin stable state will have minimum potential energy. Stable Bquilibrium position Pig. 2.4 : Equilibrium stateBasic Procedure — 41 WORKED EXAMPLES Example 1 For the spring system shown in figure E2.1 using the Principle of minimum potential energy, determine the nodal displacements. Take: F,= 75V and F, = 100N. E =50Nimm | 3 E k,=70N/mm K,-60Nimim 8 oy Fig. 62.1 Solution: We kwow Thar Potential Energy Functional = SE+ WP ie, Tl =SE+WP But (SE), dag? rg Hue ee Need 2d , (SE) yuan ™ 3 M81 + 9 KBD +9 89 From figure(1.10) 8 = us B= i eyes ad : SE = 7 hul + 7 hut + 5 by (- F and Fu, — Fie 5 [ei hat? + halts? +14? — 2a) Fy — By Potential Energy functional, ts? (ky + Ry + hy) yey — 2atuegy |= Flag, — Fhtty —@ Forpotential energy functional to be minimumVe 42_—Finte Element Methods al a By, = Om 5, differentiating Eq.(a) with respect u,, we get ; Path +h, +h,)~2te by] - 1 hytkybiegkyteky 1 3 [2usk; -2u,k,]-% =O = mt, ky + a ky FA+H 75+100 Kk 30460 uw, = 159mm, ‘From Eq. (c) uw, = 3.019 mm ‘Thus, nodal displacements au, = 159mm u, = 3.019 mm. 0) © Example 2 For the spring shown in figure E2.2, determine the nodal dis of minimum potential energy. Jacements using principleBasic Procedure — 43 Solution: We know tar Potential Energy Functional E+ WP ie. T= SE+WP 1 Ske But SByeig =F 7 1 I 1 Lpgry liga diee pes 2 ABE + 7 BP + 5 8 + > LB Lot a, 4, and 1, be the nodal displacement, From figure (1.1) 5, Hy By = yt By = y= my B= 1 1 1 1 = FAC WPF yg? + 7 y= uy 9 a! = Fu, Fis, (Note: F,=0) (SE), c and we), a peste tag wim thease thyme d tag 1a : 1 Le ease 1s AU 2+ t= 2a tH) + 3 Ay ty? # hy 2+ Wem 2 H+ > hy = Fay = Fy, “ For potential energy functional to be minim Differentiating Eq.(A) with respect w,. we get 1 eH -2) ou, Fy ky kay = () mit and mite [ku(Quy ~200,)+2ks ty hy (2 ~ 2t4)]=0 2 0 =~ (ey a4) + sty Ryltty = a) 3) om L Also, Sig [Aa (2e5 — 2g) + 2k, w)-F=0 i + + A ay ~@44_—Firite Element Mothods Rewriting equation (1), 2) and (3), we have Kyte, — Aytty = (4) ~ Kit, + Ut hat k= y= F, = kil Ct hy) My (6) Substituting values 40 Nimm } &, = 40 N/mm 80 Nimm ; 0 v,— 40.0, = 60 = 40 4, + 160 1-804, = 0 = 804, +170 uy = 50 On simplification, we get Nodal displacements 4, = 2514mm 4, = 1.014 mm 4, = 0771 mm Explain clearly as to why the potential energy function is not expressed in terms of "mgh" Solution : Potential energy is a functional as it is. function of a another function. i.¢., Potential energy functional is a function of strain energy andl work potential where es both strain energy and work potential are functions of displacements experienced by the structure. Potential can be expressed interms of "f" and 74" only if potential energy is considered as SE+WP. Ithelps in deriving finite clement equations 2.7 RAYLEIGH - RITZ METHOD Rayleigh-Ritz method isa most widely used variational method 6 obtain the approximate solution for the given problems. In this method, an approximate solution of the following type is assumed for the field variable u(x): we) = Yass) 3) Where, f(x) are known Linear independent functions are called trial function defined over a entire domain and boundary and a, are unknown parameters to be determined, ‘This approximate solution (x) is to be substituted into the functional, then functional becomes. Tunction of unknown parameter @, Then the potential energy is minimized by using necessary conditions for the functional to be stationary. Let potential energy functional IT, thenBasic Procedure — 45 ant (2.38) Which yields conditions, the displacements NOTE: GENERAL STEPS INVOLVED 1, Formulate the PE nal for the given problem 2, Assume a trial displacement function for the given problem, which should satisfy the boundary conditions, 3, Substitute the displacement function into PE funetional 4, Minimize the PE funetional so asto obtain the equilibrium conditions. 5, Determine the unknown displacements hence the strains and stresses ‘n° equations in 1 unknown ¢, with these values of a, ane f, (x) satisfying boundary claimed there by sirains and stresses. Exaniple 3 By R-R method, for a bar of cross sectional area A clastic modulus £, subjected to uniaxial loading P, show that at a distance x from fixed end is 1 = ( Z s and hence determine the end deflection and the stress to which the bar is subjected to. Soluion: nee bee Fig. E23: Axial loaded bar element Lat whe the axial displacement at any point x from the fixed end. When the bar subjected to uniaxial loading P at point 1. and let w, be the displacement at load point 1. #) Formulate the PE functional PE functional is given by T1=SE4 WP Where, Strain energy stored in bar is given by Work potential is,46_—Finito Element Methods AE p* (du) ‘i = S(t} e-em, Assume a polynomial displacement function We aqtag (2) Where, a, and a, are the generalized coordinates to be determined from the boundary condition, ie, at 1H=0; =a, Eg.(2) becomes atx=L, ++ Bq(3) becomes 4 sab Differentiating Eq(3) with respect to x, we get ou a4 lil) Substitute the displacement function into PE functional Substituting Eg.(4) and Eq.(5) into (1), we get AE ft na fate Puy AE a T= > axl = Pay 7 Wilehy = Pave AE = ah- Pol iv) Minimize PE functional risks eae asia oath an Condition forminimizationis, 5,~ = 0 a, From Eq.(6), we have (©Basic Procedure —_47 ¥) Determination of displacement, strain and stress To find Displacement ‘Thus, substituting value of a, into Fq.(3), we get Ge) w= [|x : AE Hence the Proof. To find end deflection and Stress We know that, = Lat the end a ae Ans. ‘Therefore, from Eq.(7), we have Ths, | Hooke’s law, we have Stee ie 8s, 6 =F units Ans. Example 4 A bar of length L, cross sectional aren A and modulus of elasticity £, is subjected to distributed axial load q = cr, where c is a constant as shown is figure E24, Determine the displacement of the bar at the end using Rayleigh - Ritz method, aoe Pl arer i bes Fig, E24 : Axial bar43_— Finite Element Methods Solution: Let u - be the axial displacement at any point x 1) Formulate the PE functional PE functional is given by T=se+WP ‘Where, Strain energy stored in bar is given by ia SE “ ly (*) a ax, “ 1 Work potential, we = -[‘quae=-f (cx)udr AE (du)? t = AE (SY as [teu —) il) Assume a polynomial displacement function wear ax Q) Where, anda, are the generalized coordinates are to be determined from the boundary condition. ie, atx=0;u=0; =4,= 2. From Eq.(2) wax —@) Differentiating Eq.(3) with respect to.x, we get au a ert 4) ili) Substitute the displacement funetion info PE functional ‘Substituting Eg.(3) and Eq.(4) into (1), we get AB ploy) oft n->he dx- [oxaads —-() iv) Minimize PE functional ‘The condition forthe minimization is —=0 a,Basic Procedure —_49 From Eq(5), we have s cane ©) '¥) Determination of displacement Thus, substituting value of a, into Eq,(3), we get Eis tinecassem vy = Lae The disp iu Example 5 ‘Use the Rayleigh - Ritz method to find the displacement at the mid point of the rod shown in figure E2.5. Jan/Feb. 2006 , Dec.06/Jan.07 oe yl it Yt |p rome et pee) UTA SR a Ht it Wd I | ‘ Solution: Letu- be the axial displacement at any point x i) Formulate the PE functional PE functional is given by Tl =SE+ WP Where, Strain energy stored in bar is given by Ie)50_—Finte Element Mothods ‘Work potential due to body force is, WP= -[fudve —[fudde beget 4 AE ¢ (Ou n= = (Ja -Jfuads —()) Since A=E=J= (given); and substituting the limits e., 0 and 2; we have ! Lhyau?. n=3 1) ae ~ fuas = (2) ii) Assume a polynomial displacement function weatax Where,q, anda are the generalized coordinates are to be determined from the boundary condition. ie. 4 u = Oatx = a,=0 From Bq.) wax a= (4) Differentiating Eq.(4) with respect to a, we get au unt 9 iif) Substitute the displacement function into PE functional Substituting Eq (4) and Eq.(5) into (2). we get —O iv) Minimize PE functional Mn Condition for minimization is S-- = 0 From Eq.(6), we have Nl =4)-2a, = By, 7241-220 = a,=1Basic Procedure —_§1 ¥) Determination of displacement, at the mid point Thus, substituting value of a, into Eq.(4), we get ‘The displacement u at any point is given by wear ‘and displacement, at mid point ie., at w= units Ans. Note: If A= E=f #1, Then consider a Eq. (1) ie. AE ph 9 adie ne> [ard F, Jaw de ‘Minimizing the functional ie., oe 0 = AELa = ae 2E ‘Therefore, displacement wat any point x is fl sfx ‘ = usoe 5 L Displacement at mid point ie. atx= > and Displacement at end,52_—Finita Element Methods Example 6 Use Rayleigh-Ritz method to find stress and displacement at the mid point of a bar shown in figure E2.6, Take £ = 70 GPa; A = 100mm?, Assume the displacement model to be a 2"! order polynomial. displacement function (2nd order ++ given) we atarta,e ) a, and a are the generalized coordinates are to be determined from the boundary i) Agsume a polyne i i 20mm 20mm ; rns Data : E = 70 GPa =70 x 10° MPa, A= 100 mm? ; P= 8 x 10° N ; 2nd order polynomial | Solution: Let u be the axial displacement at any point x from the fixed end When the bar subjected to | ‘uniaxial loading P at point 1, u, be the displacement at loading point | j i) Formulate the PE functional | PE functional is given by Tl =SE4 wp ‘Where, Strain energy stored in bar is given by AE (4 (uy se = AEF (3) ds ‘Work potential, WP = Pu, | . net (Ja Pi 2 ty Lax. ited ar (any | 0 90870 1G S| de —-8 x10", ax j 6 (Qu) , M1 = 35x10 1) dx=Bx10%14 aoe | () w=Oatx=0, + a,=0(b) w= 0 at v= 40, = = ~ 3) = Se = Ha, + aye sue (4) © wan, a x= 20 +. Bq.(3) becomes Oa, x 20 + a, (207 ie, uw, =~ 400a, 6) lil) Substitute the displacement function into PE funetional ‘Substituting Eq.(3); Eq.(4) and Bq,(5) into Eq.(1), we get 3.5 x 106, 404, +248) de =8 10" + (—4004,) 1 eS er > . 45K ionf (16004? +4a;"x" ~160.,?x)dx-+32 x10‘, ls b Th = 7.465 x 10! ¢,? +32 x 108 ay on 6) |v) Minimize PE funetionat Condition for minimization is a ant From Fa.(6), we have SE = 7.465% 10% 2a, +32 x 10*= 0 aa, = a, = ~2.143 x 10 ¥) Determination of displacement and stress Substituting value of «into Eq.(5), we get ty = ~400 a, =~ 400 (2.143 x 10-5) = 8.572 x 107 :. Displacement at loading point a, = 8.572 1O°mm To find stress From Hooke’s Law 254_—Finite Element Methods Qu oneee(2)e = (A0a, +2a,x)E (-40{-2.143 107) + 2{~-2.143 x10" }20}70%10" At mid point = 20; o= o=0 ° Ans Atr=0, o = ~ 40a, E = ~40 (- 2.143 x 10%) 70 10" 6 = +60 Nim? At 0 = (-40(-2.143 10") +2(-2:143.x10")40}70%10" 6 =-60 Nim Example 7 Using Rayleigh - Ritz, find the displacement at the mid point of a beam as shown in figure E2.7(a) carries a central point load “P", having Young's modulus Z and cross sectional area “we f 2 2 Fig, E2.7(a) Solution: Fig. £2.70) i) Formulate the PE functional PE functional is given by T= SE+ WP Where, strain energy stored in the beam is given byBasic Procedure —_§5 Brgy \t 4 Tay J (3) Pru —W) ii) Assume a polynomial displacement function vrataxtatta sta —-Q 14, cand a,ate the generalized coordinates are to be determined from the boundary = a,<0 Also, Fe rt ae + Bay at + day --@) and a i) y=Oand 5 = Ojat x =L From Eq.(2) forx=L3v=0 O=atabraLta lal Buta, =Oanda,=0 => a= -(mb+al’) av and for. x=, 5~ = 0 Eq(3) becomes 0 =a, +20, b+ 3ah +44, 0 ‘Substituting the values of a, and a,, we have 0 = 04 2L (-ayL-ayl?)+3a,l? +44, E 0 = 22 a, - 21 4, +30, 2+ dab? = 6, = = 2a, Eq, (4) becomes a,54,L ‘Thus substituting the values of «4, 4. a, and a, into Eq.(2), we havewo 86 —Finite Element Methods vsalx-a eta = v= a (4-200 +22") ov eee Also. eo (42° -6Lx? +217x) SY oo, (i208 126428) ee i and vis maximum at = att = ae You Rint Ail) Substitute the displacement function into PE functional ‘Substituting q.(6) and Eq.(7) into Eq.(1), we get n-= [eae -raussaeyf a-o(%2) 16 (is) pi ri = ay) Sf tae ea! 4144? x 2a Le a8 x4 a8! <2) ft) a Nea? 2 [isZearie Maar Same awe Lane] . 1 El o{ Seen tte tt (=) yw “Lie ) 48EIL | > no yt 1y) Minimizing the PE functional an Condition far minimization is 37 an a, 4 —@) © --@ (8)Basic Procedure = _ 57 pale O64) BL ‘Thus, substituting value of a, into E9.(5) and Eq.7). we get ‘ 3) Pts pudyh 3 Deflection at any point is v ( ) di #Exb=2Lx) a P 5) Pe oy sini Cozaar Note: Pe he exact value is (i The exact value «(] ce at both ends, the displacement function is given by 1, (x6 + Px? -2tx) Gi) For the beam, ve Example & . Using Rayleigh-Ritz method, determine the maximum deflection of the beam fixed at both ends subjected to a uniformly distributed load P, over entire length as shown is figure E2.8. Fig. E28 ‘Take £ - Young's modulus, /~ moment of inertia and A — Area of cross section Solution: i) Formulate the PE functional PE functional is given by Tl =SE+ WP Where, Strain energy stored in the beam is given by Er (& Sea ot 2 2 elu as Work potential due to uniformly distributed Ps58_—Finile Element Methods reo ii) Assume a polynomial displacement function ‘We know that, for a beam fixed at both ends, displacement function is given by = a, (x¢+Px"=2ix') [refer example 7 and Eq. (5)] ili) Substitute the displacement function into PE functional av Sp 7 (4x4 Gt? +272) a, (12x? = 124+ 28) pale! + Px! = 210 ae Pl “(30 )" ie, iv) Minimizing the PE functional ‘Condition for minimization is < Displacement at any point, Pandonoe 5) at Note ; Exact solution a( a SHON 384) “EyBasic Procedure — 59 Example 9 Using Rayleigh Ritz method, derive an expression for the displacement of a cantilever beam subjected to uniformly distributed load P, over entire length as shown is figure E29 Hex op P.Nham ABL ba Fig. E2.9 : Cantilever boa (OF CROSS SFETION Were, £- Younc’s MopuLus, J - MOMENT OF INERTIA AND A + Aw Solution: ') Formulate the PE functional PE functional is given by Tl =SE4 WP ‘Where, strain energy stored in the beam is given by ze (B)« and. Work potential due to uniformly distributed P, is (SE : ae we = -['R vas EL fay 3 2 (S)« 1 fm vac fa) ii) Assume a displacement function Ve gta etek tae 4 ae Where, dy,2, 4,54, anda are the generalized coordinates arc to be determined from the boundary condition. av ay ay (@)at =O, v=05 (yarx=0, 5 = 0; @ata= SF Oil arxnh Sy =0 ie, Ve qtort aes ae eax 0 a wy Fo ney + day + Bayt + to, 2 -@ a ge bog ae ays? —-®60_—Finite Exement Methods a ae 24m, Applying Boundary condition, Using BC (a) ic., at From Eq.(1) we get, Using BC (b) ic., at From Bq.(2), we get, 2 Using BC (c) ie. at xan Zico From Eq, (3), we get 0 = 2a, + 6a,! +12 4,8 Using BC (A) ic, at From Bq.(4), we get = Substituting value of a, into Eq(S), we get . 2ast 61 (- Al a,) + 12a, P =0 > a, = 6P a, ‘Thus, a, = 0:4, =6P.a,)a,=~-4 a, Substituting these values into Eq.(1) and Eq.) we get vsO40+GP ape t Ala) x + a,4 v =a, (f= 481 + Ge P) ‘This is the polynomial displacement function for a cantilever beam subjecied 10 UDL iii) Substitute the displacement function into PE functional ay oF Substituting Eq.(6) and Eq.(7) into Potential Energy functional, we get ei n-Z lag (1202 —24ad + 120) dr — B f'ag(24 40°14 652 ae = a ag: feat 4+576x°F? + 14at* =576x"1—S76x1+288x°F" Jae = a, (1202-24 at + 12 F) y [aa (a4 42°14 672) ds a 5) 6 —(DBasic Procedure — Integrating between the limits D and J, we get 3 ‘ 2 EL a | a4 +5767 + 144i -5761—76P 4298" ars 3 4 2 3 sa sy Sha tea) eee ® ra ® Ha +6 Onsimplification, EL T= J ag Px O88) -1.2 P, al? iv) Minimizing the PE functional Condition for minimization is, a =0 OT 44 080 20,-12 Py =0 aa, = 1A = TRREIP fe 005 MET ‘Thus, displacement at any point, Fea V = pgp (t' -4x'+6r'?) Example 10 . Using Rayleigh-Ritz method, derive an expression for maximum deflection of the simply supported beam subjected load P as shown in figure E2.10(a). Use trignometric function. May/Tune. 08, Dec. 07/Jan.08, May/June 2010 P A, EL Fig. £2.10) ‘Take E - Young's modulus, I - moment of inertia and A - Area of cross section of the beam62_—Finito Element Methods Solution: Fig. £2.10(b) i) Formulate the PE functional PE functional is given by Tl = SE+WP =) ‘Where, Strain energy stored in the beam is given by ELH avy aime a (Prem = 3 (3) ‘Work potential due to point load Pis WP = -Py. | ea EL 4(@v a ==> fla] a el ] s net (= PY @ ii) Assume a trignometric displacement function le, ¥ = a, sin ™ + 0, sin —@) 1 ‘Which should satisfy the following boundary condition ay G)atx=0, y= 0: ()atx=0, 57 =O; Cars Lv=0; (ax 1 | j “Thus, veg sin +o, sin | av | os ~@) ay oad —O ‘ame ‘sin sin3 ( 7 Jee) heathBasic Procedure — 63 iii) Substituting displacement function into PE functional i et (avy ie, n Bt Byor, Considering, tee. [fet 0 Considering, each terms sin? Mar = eT64 __—Finte Element Methods er{ 2m Bint ‘Thus, n= Bay Feat P(a,~a,) iv) Minimizing PE functional an 2a, ant ie, 3a 2a 60] pa -o)=0 > om ef Bin‘ 3a, [ote PO-1) =0 peer * = Git ET ‘Thus, displacement at any point is given by spi 2Pt \ein wx En’) 2p) \sin 3x Sin ‘EL t 2PP \ Fig RU sin 3RX wife ‘sin _ 1 sin, eleait HL aI | i maximum displacement is obtained from Eq.(6) ie, Vou =, é 2PP | 2p me TET” Bin ET ote Pe om 4818EF © —oec Basic Proceaure—_ 65 Example 11 Using Rayleigh-Ritz method, derive an equation for maximum deflection at the center using trignometric function of a simply supported beam is subjected to a uniformly distributed load P, N/m as shown in figure E2.11(a), EIA -Py Nico i Fig. E2.11(a) Solution: x Fig, £2.1106) i) Formulate the PE functional PE functional is given by Tl =SE+ WP ‘Where, strain energy stored in the beam is given by tony EL ¢( av (SByoan = ae) dx and Work potential due to uniformly distributed load P, is we = -[ A vde EVt(as) af ney (& de [vax i) li) Assume a trignometric displacement function @ 7 ie, v= a)sin Tt +0, sin SR Which should satisfy by the following boundary conditions66 _—Finite Element Methoos 2 = syatxno, & = O;o)atead, 57 2050) atz=L, 57 =O am 3x Thus, v= a,sin= 4a, sin a(n) we ») 3me Eran F ea (Ze ili) Substituting displacement function into PE functional Substituting Eq.(2) and Eq,(3) into Eq.(1)we get 2 fl) on) oe oma) tad -P,f, (sint+epsin 2 —@ ‘Thus, Considering each term, we have 2nx sin 2% fiw = Bax = [4 4 (1-<0 ee =| 5 Es sin * 2 3E, PE de Or et ry od sin? 2B aso [5 (1- con farm | x | 73 tokBasic Provedure— 87 Substituting above values into Eq.(4), we get nt oft) afro aL) of nae] oS) tcd 6) iv) Minimizing the PE functional Condition for minimisation, io am (et) (# aoe da, bay [o+162a]-P, ie APP DRE Substituting the values of a, and a, into Bq.(2), we get Displacement at any point in given by i val Maximum deflection willbe at x =4h [! we ] fos = eer L248. a ae neglecting the term [is 24 ] Example 12 Determine the displacement at mid point and stress in linear one-dimensional rod as _ shown in Fig. E2.12. Use second degree polynomial approximation, for the displacement, ‘December 2010 J Betas! Fig, £2.12 Data; E=1,A =1,P=2x 10° Nj; 2° order polynomial Solution: Let w be the axial displacement at any point x from the fixed end. When the bar subjected to ‘uniaxial loading P at point I, u, be the displacement at loading point 1. i) Formulate the PE functional PE functional is given by 1 =SE+ WP ‘Where, Strain energy stored in bar is given byli) Assume a polynomial Ws atortae vise) Where, a, . a, al a, are the generalized coordinates are to be determined from the boundary condition. i.¢., (a) «=O at r=0, > (b) =O at x= 40, = fad 1 4 au y ae and u =u, at x= 1 (at midpoint and loading point) +. Bq.) becomes 4, =- 2a, +a, ie, 4, fii) Substitute the displacement function into PE functional ‘Substituting Eo.(3); Eq (4) and Eq(5) into Eq(1), we get Ty Lfaay(e—tfas-2x10°-) 3 = 2a} f(z? -2x41)ar42x 10a, re $-a42]s2-107, 1 = $a} +210", oe © iy) Minimize PE functional Condition for minimization is a =0 From Bq.(6), we have = = ba +2xi0 = a, = ~750