0 Up votes0 Down votes

4 views42 pagesFEM Presentation

Dec 14, 2016

© © All Rights Reserved

PDF, TXT or read online from Scribd

FEM Presentation

© All Rights Reserved

4 views

FEM Presentation

© All Rights Reserved

- Finite Element Analysis for Piping
- ABAQUS tutorial
- Ever Stress Fe 10
- sdarticlesss
- FEM1 -Introduction.pdf
- Huber2009 Elementos Finitos
- l7 - Fdfd Method
- SAFIR Theory
- FEMM
- Introduction to FEM
- Www.bu.Edu Pasi Files 2011 01 LuisdelaCruz1!11!1700
- Creo Simulate Tutorial Releases 1 0 Amp 2 SDC Publications
- Dec 2010.pdf
- MIT2_092F09_exam2
- On FE and CV Upwinding Methods for High Pe Flows
- ANSYS-TUtorial-Day.pdf
- Sequences
- FEA on Furniture IKEA
- MECH413 - Course Syllabus - 2016-17 SPRING.pdf
- Intro to FEM Lab (Aditional Lab Sheet)-2

You are on page 1of 42

AND TECHNOLOGY, SURAT

Finite Element Method (FEM) And Finite Difference Method (FDM) Introduction And Difference

PREPARED BY:

Name

1. Ms. KRISHNA MODI

2. Mr. HITESH RANA

3. Mr. AVADHESH H. VYAS

Enrollment No.

(150090709009)

(150090709017)

(150090709018)

M.E., 2nd Semester, Machine Design,

GTU, Surat, Gujarat.

GUIDED BY:

Prof. SHRADHDHA R. MEHTA

CKPCET, SURAT.

AGENDA

Introduction

Case :1

Case :2

Introduction

Introduction

Methods of Solution

Results

Conclusion

INTRODUCTION

The FE method is a technique in which a given domain is represented as a collection of

simple domains, called finite elements, so that it is possible to construct the approximate

functions needed in a vibrational or weighted residual approximation of the solution of a

problem over each element.

Division of whole domain into sub-domains.

Derivation of approximate function.

Assembly of elements.

Introduction [1]

Steps of FEM

Step 1: Discretization of Domain

Step 2: Derivation of Element Equation

It involves three steps:

1. Construct the weighted residual or weak

form of G.D.E.

2. Assume the form of approximate solution

over a typical finite element

3. Derive the finite element equation using

approximate solution

Step 3: Assemble all the elements

5

Introduction [1]

What is FDM?

In FDM, the derivatives appearing in the differential equation and the boundary

conditions are replaced by their finite difference approximations.

Finite Difference Approximations of Various Derivatives

If y(x) and its derivatives are single valued continuous functions of x then by Taylors

expansion

6

Introduction [1]

Forward Difference

Backward Difference

Central Difference

Central Difference

Introduction [1]

For example

is the governing equation of the continuum shown below.

Introduction [1]

Diagonal 5-point formula

9

Introduction [1]

Jacobis Method

The accuracy of the calculations depends on the mesh-size i.e. smaller the mesh size, better

the accuracy. But if h is too small, it may increase the round off error.

10

Introduction [1]

Boundary Conditions

Fixed or Dirichlet Boundary Condition

Derivative or Neumann Boundary Condition

Mixed Boundary Conditions

11

Introduction [1]

FEM VS FDM

12

Introduction

FEM

FDM

13

Introduction

:1

14

Introduction

FRONTAL CRASH SIMULATION

COLUMNS USING FEM

Case :1

OF

VEHICLES

AGAINST

LIGHTING

There are a lot of utility poles, lighting columns and signposts in streets. These may cause

many severe and fatal crashes that result in vehicles colliding with lighting columns such

as street columns and street traffic lights. It is not matter that the accidents occur between

vehicle and vehicle or vehicles and any objects like lighting columns. It is found that the

existing lighting columns are fabricated from a material of higher yield strength steel. The

characteristics of this material have higher static strength and lower absorbing impact

energy.

This means that the vehicle must absorb the remaining kinetic energy from the crash.

The resulted high impact forces on such small area of the car to lead to high risk of injury

as well as endangering the lives of the vehicle occupants.

15

Case :1

In order that the occupants of the vehicle to survive a crash of this nature, it is necessary

that the vehicle must be able to absorb the energy created during the impact which already

achieved in modern design cars. Although modern cars are designed with composite

materials that absorb most of the developed energy resulted from the impact, yet the

risk of injury is still high.

A new design of lighting columns that would dramatically sustain the impact energy has

been proposed. The suggested design is aimed at reducing the internal energy that the

vehicle absorbs during impact is a demand.

16

Introduction

Case :1

In this paper, Abaqus explicit code is used to numerically simulate frontal crash of

vehicle with lighting columns. The vehicle is assumed to move in right angle against the

lighting column.

The shape of the road, straight or curved road is not considered in this study. Main

parameters influencing maximal deceleration value are found to be column shell

thickness, outer diameter and material properties.

Therefore, lighting columns with different materials and shell thicknesses are

investigated. The resulted accelerations, deformed energies and predicted contact forces

after impact are calculated. The results of the analyses are graphically presented and the

predicted trend is found to be generally satisfied.

17

Introduction

Case :1

should be given appropriate attention. These problems are:

a) The selection of a mesh & type of element, which should be fine enough, especially at

frontal contact areas to get accurate results & to represent real simulation during impact.

b) Totally about 6000 elements have been generated in the entire model where the smallest

size of the element is about 0.01 m in the front contact area (bumberradiatorshood and

front doors). Coarse mesh may applied for areas located far from collision region to

reduce CPU time.

c) ABAQUS/Explicit Version 6.9-4 code is used to simulate the impact of vehicles with

lighting columns. The vehicle and lighting columns are modeled using four nodes, thin

shell double curved elements (S4R). The contact surface during collision, elements

based surfaces are applied to define contact region.

18

Case :1

The present column (rigid) has a material yield strength steel of 355 N/mm2 with

isotropic hardening to a strength of 490 N/mm2 at plastic strain of 0.025.

However, the vehicle is free to move in right angle with a translational velocity along the

X-axis equal to 14 m/s. The velocity is applied at all nodes of the vehicle. Point masses

are assumed to represent masses of mechanical and transmission components. The

vehicle gross weight is taken 980 kg.

19

FEM

Case :1

The present lighting column has a length of 8 m and its diameters at base and top are 200

mm and 10 mm respectively, while the shell thickness (ts) is 4 mm.

20

Case :1

Figure 2 : Deformation history of the lighting column with the new material

21

Case :1

ts= 4mm

t = 4mm

t = 3mm

t = 2mm

22

Case :1

23

Case :1

24

Case :1

The present rigid lighting column with high yield strength of material absorbs little

impact energy, in turn high injury risk is expected.

The lighting column fabricated from the new material decelerates the vehicle and absorbs

higher impact energy, in turn increasing the safety of vehicle occupant.

The peak of the acceleration value is happened at a time less than that expected. This

limitation is because the model lacks some of the details of the additional material parts

in the interior of the vehicle.

Lighting columns with internal stiffening systems have to be further investigated.

The validation of this study has to be checked in the future with real crash colliding test.

25

:2

26

Introduction

Case :2

DIFFERENCE METHOD FOR TRANSIENT SIMULATION OF A GAS PIPELINE

Transient flow of gas through a pipeline is given by a linear partial differential equation

of diffusion type. Finite difference and finite element methods are selected for solving this

equation.

For the finite element method linear and cubic interpolating polynomials are used.

Investigations are carried out for the single pipeline.

The calculations are made using a PRIME 9950 computer. The computation times and

accuracy are determined for every method.

Comparison between analyzed methods are carried out, taking as a criterion of comparison

the accuracy of the results and computation time.

27

Case :2

Assumptions:

the flow is one-dimensional

the cross-sectional area changes slowly along the path of the stream of gas

the radius of curvature of the pipe is large compared with the diameter

the shapes of velocity and temperature profiles are approximately constant along the

pipe

A p

M

=

c 2 t

x

p w 2 4 f

Fp sin = ( w) + ( w2 )

x

2 D

t

x

28

(1)

Case :2

pipe, Pa

M = M(x, t) = average mass flow (averaged over cross-sectional area)

in pipe, kg/s

A = cross-sectional area, m^2

c = speed of sound in gas, m/s

= (x, t) = average gas density (averaged over cross-sectional area) in pipe,

kg/m^3

w = w(x, t) = average gas velocity (averaged over cross-sectional

area) in

pipe, m/s

f = friction co-efficient

D = hydraulic mean diameter, m

F = net body force per unit mass

= angle between the horizontal and direction x

29

Math. Modeling

Case :2

when boundary conditions do not change rapidly or the capacity of the pipe is relatively

large, the transient flow through the horizontal pipe can be represented by equations of the

form

A p

M

=

c 2 t

x

(2)

M = wA = Q = s Qs & p = c 2

(3)

c Q

p

= 2 s s

t

A x

30

2 f s2 c 2Qs2

p

=

x

A2 Dp

(4)

Math. Modeling

Case :2

After performing transformations (see Ref. 4), we obtain the biquadratic model

(5)

31

(6)

Math. Modeling

Methods of Solution

Case :2

Finite difference method

Using finite difference method eq. 5 & 6

can be written as,

k+1

Pnk +1 Pnk

Pnk+11 2 Pnk +1 + Pnk++11

=

t

(x) 2

(7)

k

n-1

n+1

,

,

!"5$

!"6$

The finite difference (FD) schemes discussed above are based on the evaluation of pressures

at specified discrete mesh points in the pipe space-time domain. The pressure profile at points

not on the mesh is not considered.

32

Case :2

Finite element method

In the finite element (FE) technique, the pipe domain is viewed as an assembly (or sum)

of small pipe sections or elements joined by nodes at their common boundaries. Over

each element, the pressure profile is approximated as a linear combination of selected

trial functions and nodal pressures.

To derive the pipe (FE) equations, we write the pressure profile over each element as,

N

p ( x, t ) = wij ( x) pi (t )

i =1

where

&

and '& denote pressure and trial functions associated with node i. N is the

number of nodes along x. Trial functions can be chosen in the form of linear or higherorder polynomials. However, there is a tradeoff between computational efficiency and

order of the trial functions used for spatial discretization.

33

Methods

Case :2

Applying the Galerkin weighted residuals method to equation (6) on an element-byelement basis requires that

2 p ( x, t )

p ( x, t ) j

a

1

wi ( )d = 0

1 2

t

1

In a similar manner, the Galerkin method has been applied to equation (5) For

discretization purposes, the pipe is divided into as many elements as necessary to form an

FE mesh.

34

Methods

Case :2

Initial conditions are determined on the basis of the pipe state at the moment preceding the

start time of the simulation. Let us assume that the initial moment is t = 0. Thus Q = f(x),

the initial flow, and p(x, 0) = h(x), the initial pressure profile. It was assumed that Q =

constant (steady-state flow).

Boundary conditions are determined by the way in which the pipe is supplied and loaded.

For discussion purposes assume that the pipe section under investigation was supplied

from a compressor station at x = 0 and fed at x = L by a receiver having a load Q(t), a

discrete period function with a discrete step At = 2h, the time interval being t = [0, 24h]

It was assumed that by introducing appropriate changes of the capacity of the compressor

station, the value of the pressure at the beginning of the pipe may be kept invariant. Thus,

p(0, t) = constant

35

Q (L,t) = f(t)

Result

36

Case :2

Case :2

37

Result

Case :2

38

Result

Case :2

39

Result

Conclusion

Case :2

The results in Table I relate to equations (5) and (6). A comparison of computation times

between FE and FD difference shows a considerable advantage for FD method in each case.

Computation time increases for higher-order interpolation functions. In the event that

higher accuracy is of concern, the Galerkin cubic finite element may be used at the expense

of increased computational costs.

Analyzing the results (Figures 3-10), we see that the maximum difference methods do not

exceed 0.06 MPa. Taking into account errors of measurement devices and simplifications of

the mathematical model of transient flow, we can neglect these differences.

40

Reference

[1] Finite Element Method: A Mathematical Approach by Dr. Chaitanya K. Desai

[2] Yehia A. Abdel-Nasser, Frontal crash simulation of vehicles against lighting columns

using FEM, Alexandria Engineering Journal (2013) 52, 295299.

[3] Andrzej J. Osiadacz, Mohamed Yedroudj, A comparison of a finite element method

and a finite difference method for transient simulation of a gas pipeline, Appl. Math.

Modelling, 1989, Vol. 13.

41

- Finite Element Analysis for PipingUploaded byMahesh Rathore
- ABAQUS tutorialUploaded bypetrus_lioe84
- Ever Stress Fe 10Uploaded byKharisma Rahman Fadillah
- sdarticlesssUploaded byJingjin Shen
- FEM1 -Introduction.pdfUploaded byvhn1988
- Huber2009 Elementos FinitosUploaded byFelipe Lucena
- l7 - Fdfd MethodUploaded byMohamad Ali Ansari
- SAFIR TheoryUploaded byDaniel Pedrosa
- FEMMUploaded byShah Khan
- Introduction to FEMUploaded byminhnguyenvonhat
- Www.bu.Edu Pasi Files 2011 01 LuisdelaCruz1!11!1700Uploaded bytsar_philip2010
- Creo Simulate Tutorial Releases 1 0 Amp 2 SDC PublicationsUploaded byVictor Possamai
- Dec 2010.pdfUploaded byskptcae
- MIT2_092F09_exam2Uploaded byDaniel
- On FE and CV Upwinding Methods for High Pe FlowsUploaded byHernan Tesorieri
- ANSYS-TUtorial-Day.pdfUploaded byRuchi Dhule
- SequencesUploaded byMaya Rahayu
- FEA on Furniture IKEAUploaded byOak Thanawat
- MECH413 - Course Syllabus - 2016-17 SPRING.pdfUploaded byAlaa Shammaa
- Intro to FEM Lab (Aditional Lab Sheet)-2Uploaded bysaifulsabdin
- faurecia cae.pdfUploaded byVimal Menon
- Lecture Finite DifferenceUploaded bykanuni41
- lect_11Uploaded byLuis Inga Marquez
- IEEE PCS Series Book Proposal Template 4 17updateUploaded byडॉ. कनिष्क शर्मा
- Different Models of Soil-structure Interaction AndUploaded byMuhammad Arfan Jadoon
- Modelling and Analysis Laboratory Manual VTU.pdfUploaded byHareesha N G
- syllabus.pdfUploaded byShivan Dosky
- femUploaded byManoj SP
- What is FEM (Finite Element MethodUploaded bycoolbks
- Lesson - 12 - Some Miscellaneous Topics in FEMUploaded byAnimesh

- Plane TrussesUploaded byHitesh Rana
- Tribology Presentation(OEP) BearingUploaded byHitesh Rana
- Tribology PresentationUploaded byHitesh Rana
- EO PresentationUploaded byHitesh Rana
- Excel to LaTexUploaded byHitesh Rana
- syllabus PV.pdfUploaded byHitesh Rana
- Effect of Fibre OrientationUploaded byHitesh Rana
- FMLUploaded byHitesh Rana
- Simulation-of-mode-I-crack-growth-in-polymers-by-crazing_2000_International-Journal-of-Solids-and-Structures.pdfUploaded byHitesh Rana
- FDMUploaded byHitesh Rana
- Material Requirement PlanningUploaded byHitesh Rana

- vol5no6_1Uploaded byBryanMonangWiener
- notes-t2 (2)Uploaded byAkshay Singh
- 1617 Level H Mathematics Exam Related Materials T2 Wk5.pdfUploaded byAndrew
- 14th-Dec-2014.pdfUploaded bysreenu
- 3-D inversion of gravity dataUploaded byrivaldovieira
- mathlinks9 ch 2 textbookUploaded byapi-171445363
- OscillatorUploaded byludihemicar
- 04Chapter 1WISE EditedUploaded bydvarsastry
- Requirement Enginering Software Requirement Tutorial 3Uploaded bySoleh My
- Graphical Method of Linear ProgrammingUploaded byMohitSharma
- MMW-REVIEWER.pdfUploaded byClarence Odsinada
- DuoMod 5045 Product CharacterizationUploaded byskinnyy
- ArticlesUploaded bySparsh Shrivastava
- DailyDoublePlan Math AUSL Week7 PatternsUploaded byLNMadorsky
- Geometrical Concepts in Teaching Electromagnetics_Russer_2005Uploaded byschlemihl69
- 2005 Effects of Petrophysical Rock Properties on Tortuosity Factor - Attia M. Attia.pdfUploaded byHafizhan Abidin Setyowiyoto
- question_paper_1336 (2).docUploaded byShashi Shekhar
- parent guide math enUploaded byapi-370473727
- Video Inpainting SynopsisUploaded bySumit Gharat
- EN(434)Uploaded byreacharunk
- Section 1Uploaded byAnonymous 5JTn4vCF
- GUIDELINES FOR DRAWING A FLOW CHARTUploaded byMaedenne Jamne
- Quantitative Process AnalysisUploaded bydjchar55
- CH 29_Case Studies - Java Machine Learning Software Available on the Web.pdfUploaded byVALDEMIR PRAXEDES TRINDADE
- lab 4mathematicsofmusic physicUploaded byapi-263500375
- 11. LeibnizUploaded byPtyx
- Solutions to Concepts of Physics by HC Verma Chapter 5Uploaded byJaspal Singh
- Taking Input from the User - Java Tutorial - Java With Us.pdfUploaded byMilds Modesto
- Transient_Analysis_Of_A_PV_Power_Generator_Charging_A_Capacitor_For_Measurement_Of_The_I–V_CharacteristicsUploaded byMatthias Belloeil
- Colpitts Oscillator Design and Performance Optimization 2332 0796.1000132Uploaded byAshish Sharma

## Much more than documents.

Discover everything Scribd has to offer, including books and audiobooks from major publishers.

Cancel anytime.