Professional Documents
Culture Documents
STUDY MATERIAL
FINITE ELEMENT ANALYSIS
DEPARTMENT OF MECH
JUNE – 2010
R S
Vel Tech
Vel Tech Multi Tech Dr.Rangarajan Dr.Sakunthala Engineering College
Vel Tech High Tech Dr. Rangarajan Dr.Sakunthala Engineering College
SEM - VII
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
INDEX
I. Introduction 06
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
2
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Student Strength of Vel Tech increased from 413 to 10579, between 1997 and 2010.
Our heartfelt gratitude to AICTE for sanctioning highest number of seats and highest number of courses for the
academic year 2009 – 2010 in Tamil Nadu, India.
Consistent success on academic performance by achieving 97% - 100% in University examination results during
the past 4 academic years.
Tie-up with Oracle Corporation for conducting training programmes & qualifying our students for International
Certifications.
Permission obtained to start Cisco Networking Academy Programmes in our College campus.
Satyam Ventures R&D Centre located in Vel Tech Engineering College premises.
Signed MOU with FL Smidth for placements, Project and Training.
Signed MOU with British Council for Promotion of High Proficiency in Business English, of the University of
Cambridge, UK (BEC).
Signed MOU with NASSCOM.
MOU’s currently in process is with Vijay Electrical and One London University.
Signed MOU with INVICTUS TECHNOLOGY for projects & Placements.
Signed MOU with SUTHERLAND GLOBAL SERVICES for Training & Placements.
Signed MOU with Tmi First for Training & Placements.
Companies Such as TCS, INFOSYS TECHNOLOGIES, IBM, WIPRO TECHNOLOGIES, KEANE SOFTWARE & T
INFOTECH, ACCENTURE, HCL TECHNOLOGIES, TCE Consulting Engineers, SIEMENS, BIRLASOFT,
MPHASIS(EDS), APOLLO HOSPITALS, CLAYTON, ASHOK LEYLAND, IDEA AE & E, SATYAM VENTURES,
UNITED ENGINEERS, ETA-ASCON, CARBORANDUM UNIVERSAL, CIPLA, FUTURE GROUP, DELPHI-TVS
DIESEL SYSTEMS, ICICI PRULIFE, ICICI LOMBARD, HWASHIN, HYUNDAI, TATA CHEMICAL LTD, RECKITT
BENKIZER, MURUGAPPA GROUP, POLARIS, FOXCONN, LIONBRIDGE, USHA FIRE SAFETY, MALCO,
YOUTELECOM, HONEYWELL, MANDOBRAKES, DEXTERITY, HEXAWARE, TEMENOS, RBS, NAVIA MARKETS,
EUREKHA FORBES, RELIANCE INFOCOMM, NUMERIC POWER SYSTEMS, ORCHID CHEMICALS, JEEVAN
DIESEL, AMALGAMATION CLUTCH VALEO, SAINT GOBAIN, SONA GROUP, NOKIA, NICHOLAS PHARIMAL,
SKH METALS, ASIA MOTOR WORKS, PEROT, BRITANNIA, YOKAGAWA FED BY, JEEVAN DIESEL visit our
campus annually to recruit our final year Engineering, Diploma, Medical and Management Students.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
3
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
This edition is a sincere and co-ordinated effort which we hope has made a
great difference in the quality of the material. ‚Giving the best to the students,
making optimum use of available technical facilities & intellectual strength‛ has
always been the motto of our institutions. In this edition the best staff across the
group of colleges has been chosen to develop specific units. Hence the material, as a
whole is the merge of the intellectual capacities of our faculties across the group of
Institutions. 45 to 60, two mark questions and 15 to 20, sixteen mark questions for
Asst. Professor.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1. INTRODUCTION 9
Historical background – Matrix approach – Application to the continuum – Discretisation – Matrix algebra – Gaussian
elimination – Governing equations for continuum – Classical Techniques in FEM – Weighted residual method – Ritz
method
Finite element modeling – Coordinates and shape functions- Potential energy approach – Galarkin approach –
Assembly of stiffness matrix and load vector – Finite element equations – Quadratic shape functions – Applications to
plane trusses
Introduction – Finite element modelling – Scalar valued problem – Poisson equation –Laplace equation – Triangular
elements – Element stiffness matrix – Force vector – Galarkin approach - Stress calculation – Temperature effects
4. AXISYMMETRIC CONTINUUM 9
Axisymmetric formulation – Element stiffness matrix and force vector – Galarkin approach – Body forces and
temperature effects – Stress calculations – Boundary conditions – Applications to cylinders under internal or external
pressures – Rotating discs
The four node quadrilateral – Shape functions – Element stiffness matrix and force vector – Numerical integration -
Stiffness integration – Stress calculations – Four node quadrilateral for axisymmetric problems.
TEXT BOOKS
1. Chandrupatla T.R., and Belegundu A.D., ‚Introduction to Finite Elements in Engineering‛, Pearson Education
2002, 3rd Edition.
2. David V Hutton ‚Fundamentals of Finite Element Analysis‛2004. McGraw-Hill Int. Ed.
REFERENCES
1. Rao S.S., ‚The Finite Element Method in Engineering‛, Pergammon Press, 1989
2. Logan D.L., ‚A First course in the Finite Element Method‛, Third Edition, Thomson Learning, 2002.
3. Robert D.Cook., David.S, Malkucs Michael E Plesha, ‚Concepts and Applications of Finite Element
Analysis‛ 4 Ed. Wiley, 2003.
4. Reddy J.N., ‚An Introduction to Finite Element Method‛, McGraw-Hill International Student Edition, 1985
5. O.C.Zienkiewicz and R.L.Taylor, ‚The Finite Element Methods, Vol.1‛, ‚The basic formulation and linear
problems, Vol.1‛, Butterworth Heineman, 5th Edition, 2000.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
5
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
UNIT – I
PART – A
A complex region defining a continuum is discredited into simple geometric shapes called finite
elements.
FEA is numerical method, which can be used to find location and magnitude of critical stress
and reflection in a structure. FEA method can be applied to structure that have no theoretical
solution available, and without FEA we will have to use experimental techniques, which can be
consuming and expensive.
F f
4. Define nodes:
The finite element procedure reduces such unknown to a finite number by dividing the
solution region into small parts called elements and by expressing the unknown field variables
interms of assumed approximating functions within each element. The approximating functions
are defined interms of field variables of specified points called nodes or nodal points.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
6
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
6. Define Matrix:
A rectangular array of numbers with a definite number of rows and columns is a matrix.
In FEA, an engineering structure is divided into smaller regions, which have simpler geometry
and theoretical solution. Collectively the regions represent the entire structure, and the individual
element contributes to the solution of the structure. Challenge lies in representing the exact
geometry of the structure. Especially, the sharp curves. Generally, a multi-degree polynomial is
approximated by a high Number of straight edges.
If A = [aij], then the transpose of A, denoted as AT, is given by AT = [aji]. Thus the rows of a are the
columns of AT.
1 5
0 6
e.g : A then A T 1 0 2 4
2 3 5 6 3 2
4 2
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
7
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
{ F} = [K] {u}
In computers, a line, connecting two nodes at its ends as shown, represents a line element. The
cross, sectional area is assumed constant throughout the elements.
Line elements
ii) 2-D solid elements: Elements that have geometry similar to a flat plate.
2-D solid elements are plane elements, with constant thickness, and have either a triangular or
quadrilateral shape, with 3 nodes or 4 nodes. e.g: plane stress, plain strain, plates shells and
axisymmetric elements.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
8
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The basic 3-D solid elements have either a tetrahedral (4 focus) or hexahedral (6 faces) shape.
13. Give the relation ship between matrix & Algebra, algebraic equation:
Matrix form:
The Rayleigh – Ri+z method of expressing field variables by approximate method clubbed with
minimization of potential energy has made a big break through in finite element analysis.
The Rayleigh – Ri+z method involves the construction of an assumed displacement field,
u ai i (x,y,z) i 1 tol
v a j j (x,y,z) j l 1 to m.
w ak k (x,y,z) k m 1 to n
n m l.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
9
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1
v dv v u f d v s u Tds ui pi
T T T T
2 i
Where r = number of independent unknowns. Now, the extremum with respect to ai, (I = 1
to r) yields the set of r equations.
0 i 1,2, ...,r.
ai
from the solutions of r equation, we get these values of all ‘a’. With these values of ai and I
satisfying boundary conditions, the displacements are obtained.
A body is in equilibrium if the internal virtual work equals the external virtual work for every
kinematically admissible displacement field (,()).
For conservative systems, of all the kinematically admissible displacements fields, these
corresponding to equilibrium extremize the total potential energy. If the extremum condition is a
minimum, the equilibrium state is stable.
i) In classifial methods exact equations are formed and exact solutions are obtained where
as in finite element analysis exact equations are formed but approximate solutions are
obtained .
ii) Solutions have been obtained for few standard cases by classical methods, where as
solution can be obtained for all problems by finite element analysis.
iii) When material property is not isotropic, solutions for the problems become very
difficult in classical method. Only few simple cases have been tried successfully by
researchers. FEM can handle structures with anisotropic properties also without any
difficulty.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
10
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
iv) If structure consists of more than one material, element can be used without any
difficulty.
v) Problems with material and geometric non-linearities cannot be handled by classical
methods but there is no difficulty in FEM.
18. Distinguish between finite Element method (FEM) vs Finite Difference method (FEM):
i) FDM makes point wise approximation to the governing equations i.e it ensures
continuity only at the node points. Continuity along the sides of grid lines are not
ensured.
FEM makes piecewise approximation i.e it ensures the continuity at node points as well as
along the sides of the element.
ii) FDM needs larger number of nodes to get good results while FEM needs fewer nodes.
iii) With FDM fairly complicated problems can be handled where as FEM can handle all
complicated problems.
19. Write a short note on plane stress models and plane strain models:
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
11
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Aspect ratio is defined as the ratio of largest to smallest size in an element. Aspect ratio should
be as close to unity as possible.
For a two dimensional rectangular element, the aspect ratio is conveniently defined as length
to breadth ratio. Aspect ratio closer to unity yields better results.
This is the last step in a finite element analysis. Results obtained in Analysis after the pre-
processing are usually in the form of raw data and difficult to interpret. In post analysis, a CAD
program is utilized to manipulate the data for generating deflected shape of the structure, creating
stress plots, animation, etc. A graphical representation of the results is very useful in
understanding the behaviour of the structure.
i) Heat transfer
ii) Fluid mechanics (Two dimensional flow).
iii) Solid mechanics.
iv) Boeing 747 aircraft.
v) Nuclear reaction vessel.
vi) Bio-mechanics
vii) Reinforced concrete beam.
Advantages:
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
12
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Disadvantages:-
i) There are many types of problems where some other method of analysis may probe
efficient then the finite element method.
ii) Cost involved in the solution of the problem.
iii) For vibration and stability problems in many cases the cost of analysis by finite element
method may be prohibitive.
24. Use the Gaussian elimination method to solve the simultaneous equations.
2a + b + 2c – 3d = 0
2a – 2b + c – 4d = 5
a +2c – 3d = -4
4a + 4b – 4c + d = -6
Matrix from is
2 1 2 3 a 0
2 2 1 4 b 5
1 0 2 3 c 4
4 4 4 1 d 6
1 0.5 1 3 a 0
0 1 0.33 0.33 b 1.66
0 0 1 1.14 c 4.14
0 0 0 1 d 10.80
The term stiffness matrix originates from structural analysis. The matrix relation between
temperature and heat flux is called the stiffness matrix.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
13
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
PART – B
3 1 4
1. A= -1 4 2
-2 2 2
3 1 4 : 1 0 0
-1 4 2 : 0 1 0
-2 2 -2 : 0 0 1
3 1 4 : 1 0 0 R2 3R2 + R1
0 13 10 : 1 3 0 R3 3R3 + 2R2
0 8 2 : 2 0 3
3 1 4 : 1 0 0
0 13 10 : 1 3 0 R 3 13R3 8R2
0 0 -54 : 18 -24 39
39 0 42 12 3 0
0 13 10 1 3 0 R1 13R1 8R2
0 0 54 18 24 39
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
14
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Solution:-
1
v T dv v uT fdv s uT Tds uiTpi
2 i
can be used for finding an approximate solution The Rayleigh – Ri+z method involves the
construction of an assumed displacement field, i.e.,
u aii(x,y,z) i 1 to l
v a ji (x,y,z) j l 1 to m
w akk (x,y,z) k m 1 to n
nml
Where r = number of independent unknowns Now, the extremum with respect to ai, (I = 1 to r).
yields the set of r equations
0, i 1,2,...,r.
ai
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
15
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The following figure contains a body B with boundary S. The boundary is divided into two
regions su with essential (Dirichlet) boundary conditions and a region sf with natural (Neumann)
boundary conditions.
The essential boundary conditions are specifications of the solution on the boundary (for
example, known boundary displacement), while the natural boundary conditions are
specifications of derivatives of the solution (for e.g. surface tractions). All points on the boundary
must have one or the other type of specified boundary conditions.
General body with boundary. The basic step in weighted residual methods is to assume a solution
of the form:
n
un a j j
j1
In that aj value should be find out and that gives a best approximation to the exact solution.
FEA solution of engineering problems, such as finding deflections stresses in a structure, requires
three steps:
2. Analysis
3. Post processing
The structure is modeled using a CAD program that either comes with the FEA software or
provided by another software vendor. The final FEA model consists of several elements that
collectively represent the entire structure. The elements not only represent segments of the
structure, they also simulate its mechanical behaviour and properties. Regions where geometry is
complex (curve, notches, holes, etc.) require increased number of elements to accurately represent
the shape; where as, the regions with simple geometry can be represented by coarser mesh (or
fewer elements). The selection of proper elements requires prior experience with FEA, knowledge
of structure’s behaviour, available elements in the software and their characteristics, etc. The
elements are joined at the nodes, or common post.
In the pre-processing phase, along with the geometry of the structure, the constraints, loads
and mechanical properties of the structure are defined. Thus, in pre-processing, the entire
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
16
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
structure is completely defined by the geometric model. The structure represented by nodes and
elements is called ‚mesh‛.
Step 2: Analysis
In this step, the geometry, constraints, mechanical properties and loads are applied to
generate matrix equations for each element, which are then assembled to generate a global matrix
equation of the structure. The from of the individual equations, as well as the structural equation
is always,
Where
The equation is then solved for deflections. Using the deflection values, strain, stress, and
reactions are calculated. All the results are stored and can be used to graphic plots and charts in
the post analysis.
This is the last step in a finite element analysis. Results obtained in step 2 are usually in the
form of raw data and difficult to interpret. In post analysis, a CAD program is utilized to
manipulate the data for generating deflected shape of the structure, creating stress plots,
animation, etc. A graphical representation of the results is very useful in understanding
behaviour of the structure.
In this method the approximating functions must satisfy the boundary conditions and
should be easy to use. Polynomials are normally used. Some times sine-cosine terms are also
used.
Results can be obtained for complex problems. But for complex problem it is difficult to
say whether the results obtained are accurate enough to use. The doubt will arise due to the
following two reasons
(i) Whether this is the only function which can be used
(ii) How many terms in the function are to be used.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
17
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The best way to ensure the accuracy is to get result using a certain number of terms and
then use additional terms to get the results. If the difference is negligible, we can conclude that the
satisfactory result is obtained. However it may be noted that the lowest terms in the series should
not be omitted in the approximating functions.
Let us first demonstrate how weighted residuals work using a bar subjected to body and end
loads (Figure).
F x 0 or:
A
x x. x f B
(x) 0
x
d
A f B (x) 0 (1)
dx
E (2)
du
(3)
dx
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
18
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
d du B
f (x) 0
dx dx
A E
d2u
AE b 0 for 0 x L (4)
dx 2
du
EA x L P (6)
dx
Equation (4), along with the boundary conditions (5) and (6), forms the differential equation for
the problem at hand. They can be solved, by direct integration, for the exact solution.
For the weight residual formulation, we first choose a weighting function w(x), multiply (4) by the
weighting function:
d2u
w EA 2 b 0 (7)
dx
This is called the weighted residual formulation. It is called this because if we assume an
approximate solution Un (that satisfies all boundary conditions) then,
d2un
AE 2 b R(x) 0 (9)
dx
Instead, we have an error (residual) that is a function of x. Thus (8) is really a weighting of the
residual over the body:
L
wRdx 0
0
(10)
We have taken the error (residual), multiplied by a weighting function and set the weighted
integral to zero.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
19
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
3.b. Using Rayleigh – Ri+z method determine the expression for deflection and bending moments
in a simply supported beam subjected to uniformly distributed load over entire span. Find the
deflection and moment at miss pan and compare with exact solutions.
Solution:
m x
The following figure shows the typical beam. The Fourier series y a
m 1,3
i sin
l
is the ideal
2
dy
function for simply supported beams since y = o and M = EI 0 at x = 0 and x = l are satisfied.
x 2
For the simplicity
Figure.
x 3 x
y a1 sin a2 sin <(a)
l l
2
EI d2 y
l l
2 dx wy dx <(b)
0
2 dx 0
EI 2 x 3 x 3 x
2 l
1
l l
2 1
a sin 9a2 sin dx w a1 a2 cos
2 l 0 l l 3 l 0
EI 4 x 3 x 3 x wl 2a
l
4 1
a sin2 18a1a2 sin 81a22 sin2 dx 2a1 2
2 l 0 l l l 3
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
20
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
l
x 1
l
2 x 1
Nothing that sin2 dx 1 cos dx
0
l 0
2 l 2
l
x 3 x 2 x
l
4 x
0 sin l sin l dx 0 cos l cos l dx 0
3 x 1 6 x
l l
1
and sin2 cos dx
0
l 0
2 l 2
EI 4 2 1 1 2wl a
y 4 1
a 81 a22 a1 2
2 l 2 2 3
we get,
EI 4 2wl a
4 l 4
a 12 81a22
a1 2
3
to be minimum, 0 and 0.
a1 a2
EI 4 2wl
i.e., 2a1 0
4l3
4wl4
or a1
3 5
EI 4 2wl
and 81 x 2a2 0
4l3
3
4wl4
or a2
243EI 5
4wl4 x 4wl4 3 x
y sin sin
EI 5
l 243EI 5
l
l
Max, deflection which occurs at x is
2
5 wl4 wl4
ymax
384 EI 76.8EI
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
21
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
d2 y 2 x 9 2 3 x
Now, Mx EI 2
EI
1 2
a sin a 2 2
sin
dx l l l l
4wl2 x 4wl2 x9 3 x
= EI sin sin
EI 243EI
3 3
l l
wl2
we know the exact value is .
8
1. h-elements and,
2. P-elements
H-element is the original and ‚classic‛ element. The name is derived from the field of
numerical analysis, where the letter ‘h’ is used for the step size, to achieve convergence in the
analysis. The h-element is always of low order, usually, linear or quadratic. When a finite
element mesh is refined to achieve convergence, the procedure is called h-convergence. For h-
elements, convergence is accomplished regions require a very fine mesh, thereby increasing the
number of elements. Finite elements used by commercial programs in the 1970s and 80s, well h-
elements. However, with improvement in computer power and efficiency, a much more useful, p-
elements were developed.
P-elements are relatively new, developed in late 1980s and offer not only the traditional static
analysis, they provides option of optimizing a structure. In Pro/M, P-elements can have edge-
ploynomial as high as 9th order, unlike the low order polynomials of h-elements. The high
polynomial edge order of p-elements makes it possible to model a curved edge of a structure with
accuracy. Therefore, fewer elements can be used to achieve convergence. When p-elements are
used, the number of elements in the mesh usually remains fixed; convergence is achieved by
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
22
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
increasing the polynomial order of the p-elements, rather than refinement of the mesh. For
optimization, as the dimensions of the structure being analyzed are changed, the number of
elements remains constant. Only the polynomial order of the elements is changed as needed.
In FEA, a top-down procedure refers to certain of FEA mesh by first building a solid
model, using a 3-D CAD program, and then dividing the model into nodes and elements. Thus,
the top-down method requires building of geometric model of the structure, which is then used to
create an FEA mesh. The advantages of the top-down approach are obvious; we don’t have to
define the geometry of individual elements in the structure, which can be very time-consuming.
Obviously, a 3-D model requires high-end computer hardware, along with familiarity with the
modeling software.
4.b. Using Rayleigh – Ri+z method, determine the expressions for displacement and stress in a
fixed bar subjected to axial force P as shown in the below fig. Draw the displacement and stress
variation diagram. Take 3 terms in displacement function.
Solution: Let the displacement at load point be u1. Then the strain energy of the bar
2
du
l
1
U EA dx
20 dx
And potential energy due to external forces = Pu1
2
du
l
1
EA dx Pu1
20 dx
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
23
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
u = a1 + a2x + a3 x2
0 = a1 (1)
a2 = -a3 l (3)
u a3lx a3 x2 a3 lx x 2
1
At x
2
l l2
u u1 a3 l
2 4
a3l2
i.e., u1 (4)
4
du
Now a3 l 2x
dx
l
1 l2
2 0
2
EA a 2
3 l 2x dx Pa3
4
l
l2
1
2
2 2
EA a 3 l 4lx 4x dx Pa3
2
4
0
l
1 2 2 4x 3 l2
EA a3 l x 2lx
2
Pa3
2 3 0 4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
24
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 l2 l2
EA a23 Pa3
2 3 4
d l3 l2
0 EAa3 P 0
da3 3 4
3P
a3
4EAl
3P
u lx x 2
4EAl
a3l2 3pl
u1
4 16
du
E Ea3 ( l 2x)
dx
3p 3p
E [ l 2x] [l 2x]
4EAl 4Al
3p
0 x 0
4A
1 l 0
x
2
3p
2 x l
4A
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
25
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
5.a. In the given spring structure, k1 = 20 lb/in, k2 = 25 lb/in, K3 = 30 ib/in, F = 5 lb. Determine
deflection at all the nodes.
Solution:
k e k e
K (e)
k e k e
1 2
20 20 1
Element1: k (1)
20 20 2
2 3
25 25 2
Element1: k (2)
25 25 3
Element 3 : k (3) 3 4
30 30 3
30 30 4
Assembling the terms according to their row and column position. We get
K g 1 2 3 4
20 20 0 0 1
20 20 25 25 0 2
0 25 25 30 30 3
0 0 30 30 4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
26
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Or, by simplifying
20 20 0 0
20 45 25 0
K g
0 25 55 30
0 0 30 30
First, applying the boundary conditions u1 = 0, the first column will drop out. Net, F1
=F2=F3=0, and F4 = 5 lb. The final form of the equation becomes.
0 45 25 0 u2
0 25 55 30 u3
5 0 30 30 u
4
This is the final structural matrix with all the boundary conditions being applied. Since the
size of the final matrices is small, deflections can be calculated by hand. It should be noted that in
a real structure the size of a stiffness matrix is rather large and can only be solved with the help of
a computer. Solving the above matrix equation by hand we get,
0 = 45 u2 – 25 u3
u2 0.2500
0 = -25 u2 + 55 u3 – 30 u4 Or u3 0.4500
u 0.6167
4
5 = -30 u3 + 30 u4
5. b. In the spring structure shown, k1 = 10 lb/in, k2 = 15 lb/in, k3 = 20 lb/in, P = 5 lb. Determine the
deflections at node 2 and 3.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
27
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 2
Element 1 : 10 10 1
k (1)
10 10 2
2 3
Element 2 : 15 15 2
k (2)
15 15 3
3 4
Element 3 : 20 20 3
k (3)
20 20 4
1 2 3 4
1 10 10 0 0 10 10 0 0
2 10 10 15 15 0 10 25 15 0
3 0 15 15 20 20 0 15 35 20
4 0 0 20 20 0 0 20 20
The known boundary conditions are : u1= u4 = 0, F1= P = 31b. Thus, rows and columns 1 and 4 will
out, resulting in the following matrix equation,
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
28
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Solution :
1 4
Element 1 : 10 10 1
k (1)
10 10 4
1 2
Element 2 : 15 15 1
k(2)
15 15 2
2 3
Element 3 : 20 20 2
k (3)
20 20 3
2 3
Element 4 : 25 25 2
k (4)
25 25 3
2 4
Element 5 : 30 30 2
k (5)
30 30 4
3 4
Element 6 : 35 35 3
k (6)
35 35 4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
29
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 2 3 4
10 15 15 0 10 1
15 15 20 25 30 20 25 30 2
k g
0 20 25 20 25 35 35 3
10 30 35 10 30 35 4
25 15 0 10
15 90 45 30
k g
0 15 80 35
10 30 35 75
Now, apply the boundary conditions, u1 = U4 = 0, F2 = 100N. The is carried out by deleting
the rows 1 and 4, columns 1 and 4, and replacing F2 by 100N. The final matrix equation is,
Which gives
u2 1.5459
u3 0.8696
Deflections:
Spring 1 : u4 – u1 = 0
Spring 2 : u2 – u1 = 1.54590
Spring 3 : u3 – u2 = 0.6763
Spring 4 : u3 – u2 = 0. 6763
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
30
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Spring 5 : u4 – u2 = 1.5459
Spring 6 : u4 – u3 = 0.8696.
Gaussian Elimination:
x1 2x 2 6x3 0 ( )
2x1 2x 2 3x 3 3 ( ) (1)
x1 3x 2 2 ( )
The equations are labeled as ,, and . Now, we wish to eliminate x1 from and. We
have, from Eq. , x1 = + 2x2 – 6x3. Substituting for x1 into Eqs. and yields
x1 2x 2 6x 3 0 ( )
0 6x 2 9x 3 3 ((1) ) (2)
0 x 2 6x 2 2 ( ) (1)
It is important to realize that Eq. Can also be obtained from Eq. by row operations.
Specifically, in Eq., to eliminate x1 from II, we subtract 2 times I from II, and to eliminate x1 from III
we subtract – 1 times I from III. The result is Eq. . Notice the zeroes blow the main diagonal in
column 1, representing the fact that x1 has been eliminated from Eqs. II and III. The superscript (1)
on the labels in Eqs. Denotes the fact that the equations have been modified once.
We now proceed to eliminate x2 from III in Eqs. For this, we subtract 1/6 times II from III. The
resulting system is
x1 2x 2 6x 3 0 ( )
0 6x 2 9x 3 3 ( )
(1)
(3)
15 3 ( )
(2)
0 0 x3
2 2
The coefficient matrix on the left side of Eqs. Is upper triangular. The solution now is
virtually complete, since the last equation yields x3 = 1/5, which, upon substitution into the second
equation, yields x2 = 4/5, and then x1 = 1/5, from the first equation. This process of obtaining the
unknowns in reverse order is called back-substitution
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
31
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
These operations can be expressed more concisely in matrix form as follows: Working with
the augmented matrix [A,B], the Gaussian elimination process is
1 2 6 0 1 2 6 0 1 2 6 0
2 2 3 3 0 6 9 3 0 6 9 3 (4)
1 3 0 2 0 1 6 2 0 0 15 / 2 3 / 2
1 4 2
x3 x2 x1 . (5)
5 5 5
x1 + 2x2 + x3 =2
-4x1 - 2x2 – x4 = 0
Solution:
4 2 2 8 x1 4
1
2 1 0 x 2 2
(a)
0.5 1 4 4 x3 10
4 2 0 1 x 4 0
Divide row 1 by 4. Subtract the new row 1 from row. Multiply the new row 1 by 0.5 and
subtract it from row 3. Multiply row 1 by – 4 and subtract it from row 4. The result is
Divide row 2 by 1.5. Multiply the new row 2 by – 1.25 and subtract it from row 3. A zero
already appears in row 4, and no modification is required. The result is
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
32
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 0.5 0.5 2 x1 1
0 1
1 1.3333 x 2 0.6667
0 0 5.5 6.6667 x3 10.3333
0 0 2 7
x 4 4
Divide row 3 by .5.5. Multiply the new row 3 by -2 and subtract it from row 4:
1 0.5 0.5 2 X1 1
0 1
1 1.3333 X2 0.6667
0 0 1 1.2121 X3 1.8788
0 0 0 4.5758
X4 7.7576
8.a. Explain the term convergence & briefly explain about its types.
Convergence refers to this process, where we optimize the mesh to arrive at the desired
results. In general, there are three main types of convergences:
i) Von-Misses Stress (VMS) convergence: Mesh is refined until the percentage variation
in VMS is less than 1,5,10 or any given value selected by the user. VMS convergence
should be avoided if there are stress concentrations points, convergence will be difficult
to achieve.
ii) Strain Energy Convergence: Mesh is refined until the percentage variation in the
average strain of elements is less then a chosen value. Strain convergence is a better
criterion for optimizing an FEA mesh. Stress concentrations points do not significantly
influence the average strain energy of elements and variation in strain energy is
influenced by mesh size or polynomial order of the elements only.
iii) Deflection Convergence: It is similar to the above convergences, except, node deflection
values are used for the convergence criterion.
The following steps can summarize FEA procedure that works inside software:
i) Using the user’s input, the given structure is graphically divided into small elements
(sections or regions) so that every element’s mechanical behaviour can be defined by as
set of differential equations.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
33
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
ii) The differential equations are converted into algebraic equation, and then into matrix
equations, suitable for a computer-aided solution.
iii) The element equations are combined and a global structure equation is obtained.
iv) Appropriate load and boundary conditions, supplied by the user, are incorporated into
the structure matrix.
v) The structure matrix is solved and deflections of all the nodes are calculated.
vi) A node can be shared by several elements and the deflection at the shared node
represents deflection of the sharing elements at the location of the node.
vii) Deflection at any other point in the element is calculated by interpolation of all the node
points in the elements.
The individual element matrix equations are assembled into a combined structure equation, {F} =
[k] {u}.
As defined earlier
[k] = Stiffness matrix of the structure, which is always a symmetric matrix. This matrix is
analogues to n equivalent spring constant of several of connected springs.
{u} = Column matrix representing the deflection of all the node points, that results when the load
{F} is applied.
Liner elastic materials, the stress-strain relations come from the generalized Hooke’s for
isotropic materials, the two material properties are Young’s modulus (or models of elasticity) E
and Poisson’s ratio v. Considering an elemental cube inside the, Hooke’s law gives
x y z
Ex v v
E E E
x y z
E y v v
E E E
x y z
Ez v v (1)
E E E
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
34
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Txz
yz
G
T
xz xz
G
T
xy xy
G
E
G (2)
2(1 v)
(1 2v)
Ex Ey Ez ( x y z ) (3)
E
DE (4)
1 v v v 0 0 0
v 1 v v 0 0 0
E v v 1 v 0 0 0
D (5)
(1 v)(1 2v) 0 0 0 0.5 v 0 0
0 0 0 0 0.5 v 0
0 0 0 0 0 0.5 v
Special Cases
One dimension. In one dimension, we have normal stress along x and the corresponding
normal strain. Stress-strain relations (Eq.4) are simply
= E (6)
Two dimensions. In two dimensions, the problems are modeled as plane stress and plane strain.
Plane Stress. A thin planar body subjected to in-plane loading on its edge surface is said to be in
plane stress. A ring press fitted on a shaft, Fig. a, an example. Here stresses z, Tyz are set as zero.
The Hooke’s law relations (Eq.1) then give us
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
35
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
x y
x v
E E
x y
y v
E E (7)
2(1 v)
xy Txy
E
v
z ( x y )
E
x 1 v 0 x
y v 1 0 y (8)
1 v xy
Txy 0 0
2
Which is used as D .
Plane strain. If a long body of uniform cross section is subjected to transverse banding along its
length, a small thickness in the loaded area, as shown in Fig. b, can are treated as subjected to
plane strain. Here z, zx, yz are taken as zero. Stress may not be zero in this case. The stress-
strain relations can be obtained directly from Eqs., and , :
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
36
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
x 1 v v 0 x
E
y v 1 v 0 y (9)
(1 v)(1 2v)
xy v xy
1
0 0
2
Anisotropic bodies, with uniform orientation, can be considered by using the appropriate D
matrix for the material.
10.
1 1 1 1
k1 12 k 2 2 2 k 3 3 2 k 4 4 2 F1q1 F3 q3
2 2 2 2
1 1 1 1
k1(q1 q2 )2 k 2q2 2 k 3 (q3 q2 )2 k 4 q3 2 F1q1 F3 q3
2 2 2 2
where q1,q2 and q3 are the displacements of nodes 1,2 and 3, respectively.
Figure. A.
For equilibrium of this three degrees of freedom system, we need to minimize with respect to
q1,q2, and q3. The three equations are given by
0 i 1,2,3 (1)
qi
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
37
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Which are
k1(q1 q2 ) F1 0
q1
k1 q1 q2 k 2q2 k 3 (q3 q2 ) 0
q2
k 3 (q3 q2 ) k 4q4 F3 0
q3
If , on the other hand, we proceed to write the equilibrium equations of the system by considering
the equilibrium of each separate node, as shown in Fig. b, we can write.
11.
The potential energy for the elastic one-dimensional rod (Fig), with body force neglected, is
2
du
L
1
EA dx 2u1
20 dx
U = a1 + a2x + a3x2
0 = a1
0 = a1 + 2a2 + 4a3
Hence,
A2 = -2a3
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
38
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Figure. E
12. Give the brief explanation of FEA for a stress analysis problem.
The steps involved in finite elements analysis are clarified by taking the stress analysis of a
tension strip with fillets (refer Fig.). In this problem stress concentration is to be studies in the
fillet zone. Since the problem is having symmetry about both x and y axes, only one quarter of the
tension strip may be considered as shown in Fig. . About the symmetric axes, transverse
displacements of all nodes are to be made zero. The various steps involved in the finite element
analysis of this problem are discussed below:
Step 1 : Four nodded isoparametric element (refer Fig) is selected for the analysis (However note
that 8 noded isoparametric element is ideal for this analysis). The four noded isoparametric
element can take quadrilateral shape also as required for elements 12,15,18, etc. As there is no
bending of strip, only displacement continuity is be ensured but not the slope continuity. Hence
displacement of nodes in x and y directions are taken as basic unknowns in the problem.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
39
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Step 2 : The portion to be analyzed is to be discretised. Fig. shows discretised portion. For this
33 elements have been used. There are 48 nodes. At each node unknowns are x and y
components of displacements. Hence in this problem total unknowns (displacements) to be
determined are 48 x 2 = 96.
Step 3 : The displacement of any point inside the element is approximated by suitable functions in
term of the nodal displacements of the elements. For the typical element (Fig.b), displacements at
P are
u Nu
i i N1u1 N2u2 N3u3 N4u4
(1)
and v Nu
i i N1u1 N2u2 N3u3 N4u4
Step 4 : Now the stiffness characters and consistent loads are to be found for each element. There
are four nodes and at each node degree of freedom is 2. Hence degree of freedom in each element
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
40
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
is 4 x 2 = 8. The relationship between the nodal displacements and nodal forces is called elements
stiffness characteristics. It is of the form
For the element under consideration, ke is 8 x 8 matrix and Fe are vectors of 8 values. In
solid mechanics element stiffness matrix is assembled using variational approach i.e. by
minimizing potential energy. If the load is acting is the body of element or on the surface of
element, its equivalent at nodal points are to be found using variational approach, so that right
hand side of the above expression is assembled. This process is called finding consistent loads.
96 x 96 96 x 1 96 x 1
Each element stiffness matrix is to be placed in the global stiffness matrix appropriately.
This process is called assembling global stiffness matrix. In this problem force vector F is zero at
all nodes except at nodes 45, 46, 47 and 48 in x direction. For the given loading nodal equivalent
forces are found and the force vector F is assembled.
Step 6 : In this problem, due to symmetry transverse displacements along AB and BC are zero.
The system equation [k] {} = {F} is modified to see that the solution for {} comes out with the
above values. This modification of system equation is called imposing the boundary conditions.
Step 7: The above 96 simultaneous equations are solved using the standard numerical procedures
like Gausselimination or Choleski’s decomposition techniques to get the 96 nodal displacements.
Step 8: Now the interest of the analysis to study the stresses at various points. In solid mechanics
the relationship between the displacements and stresses are well established. The stresses at
various points of interest may be found by using shape functions and the nodal displacements and
then stresses calculated. The stress concentrations may be studies by comparing the values at
various points in the fillet zone with the values at uniform zone, far away from the filet (which is
equal to P/b2 t).
13.
Determine the displacement and stress in a bar of uniform cross section due to self weight
only when held as shown in Fig. Use (i) two terms (ii) three terms, for approximating polynomial.
Verify the expression for total extension with the exact value.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
41
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Figure.
Solution: Let ‘p’ be unit weight and E Young’s modulus of the material of the bar. If A is the
cross section of the bar then,
1 T
U dV
v
2
T
1 du du
l
E A dx (1)
0
2 dx dx
2
du
l
1
EA dx
20 dx
v
and l (2)
u A dx
0
At X = 0, u = 0
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
42
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
du
u a1x a1
dx
2
du
l l
1 1 1
U EA dx EA a12dx EA a12l
20 dx 20 2
l
Wp u A dx
0
l
l
x2
a1x A dx Aa1
0 2 0
l2
Aa1
2
1 l2
EA a1 l A a1
2
2 2
d l2
0 i.e.,0 EAa1l A
da1 2
l
or a1 (4)
2E
l
u x (5)
2E
du l
and E (6)
dx 2E
(a)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
43
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
du
u a1x a2 x 2 a1 2a2 x
dx
U Wp
2
du
l l
1
EA dx u A dx
20 dx 0
l l
a x a x dx
1
EA a1 2a2 x dx A
2 2
1 2
2 0 0
l l
1 x2 x3 a x2 a x3
EA a12 4a1a2 4a2 2 A 1 2
2 2 3 0 2 3 0
1 2 4 2 3 l2 l3
EA a1 l 2a1a2l a2 l A a1 a 2
2
2 3 2 3
d 1 1
EA 2a1l 2a2l2 A 0
da1 2 2
l
i.e., a1 a2l (8)
2E
d 1 2 8 3 l3
0 EA 2a1l a2l A 0
da2 2 3 3
4 l
i.e., a1 a2l (9)
3 3E
From equation 8 and 9 we get,
1 l l l
a2l
3 3E 2E 6E
a2
2E
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
44
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
l l
a1 l
2E 2E E
l x2
u a1x a2 x 2 x x2 lx
E 2E E 2
du
E l x
dx
The variations of displacements and stress in this case also are shown in Fig.
2 l2 l2
Extension of the bar u1 u0 l
E 2 2E
Figure.
l
l
Px dx l Axdx x 2 l2
AE 0 AE
0
E 2 0 2E
Thus total extension of the bar obtained is exact in both the cases.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
45
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
UNIT – II
PART – A
1. Use the lagrangian interpolation formula for deriving one-dimensional three-node shape
functions for the element illustrated in the below fig specialize the shape function for an
element of length L with a note at its center.
(x x j )(x xk )
Li Ni
(xi x j )(xi xk )
(x xi )(x xk )
L j Nj
(x j xi )(x j xk )
(x xi )(x x j )
Lk Nk
(xk xi )(xk x j )
Let xi=0, xj=L/2 and xk=L and let corresponding node numbers ,12, and 3:
(x L / 2)(x L) x(x L)
N1 ;N2
( L / 2)( L) (L / 2)(L / 2 L)
x(x L / 2)
N3
L(L L / 2)
In the finite element analysis aims is to find the field variables at nodal points by rigorous
analysis, assuming at any point inside the element basic variable is a function of values at nodal
points of the element. This function which relates the field variable at any point within the
element to the field variables of nodal points is called shape functions.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
46
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
3. Derive a one-dimensional linear interpolation formula for a function u=u(x) that is valid in
the range u1 through u2 as shown in fig.
The function u(x) is shown in above figure. A linear equation that would approximate u(x)
between u1 and u2 is assumed as
u= A+Bx (a)
where A and B are constants, substituting the boundary conditions u(x 1)= u1and u(x2)=u2 gives two
equations that can be solved for A and B.
u1=A+Bx1
u2=A+Bx2
Solving for A and BB substituting into Eq. (a) gives the interpolating polynomial:
u1x 2 u2 x1 u2 u1
A B=
x3 x1 x 2 x1
(b)
x2 x x x1
u u1 u2
x 2 x1 x 2 x1
u1x 2 u2 x1
The results of A=
x 2 x1
u2 u1
B
x 2 x1
x2 x x x1
u u1 u2
x 2 x1 x 2 x1
Can be used to derive the shape function at node 1 is the coefficient of u 1 in the above said
equation or N1=(x2-x)/(x2-x1). Similarly the shape function for node 2 is N2=(x-x1). Note that N1=1
for x =x1 and N1=0 for x=x2. N2 is zero at node 1 and 1 at node 2.
5. Derive a linear one-dimensional interpolation formula in terms of shape functions and nodal
point variables and write the result as a matrix equation.
u1x 2 u2 x1
The interpolation formula A=
x 2 x1
u2 u1
B ;
x 2 x1
x2 x x x1
u u1 u2
x 2 x1 x 2 x1
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
47
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
N1=(x2-x)/(x2-x1)
N2=(x-x1)/(x2-x1)
u=N1u1+N2u2
[N]=[N1 N2]
{u}=[u1 u2]T
u=[N]{u}
Element division:
The first step is to model the bar as stepped shaft, consisting of a discrete number of
elements, each having a uniform cross section.
Specifically, let us mode the bar using four finite elements. A simple scheme for doing this is to
ivied the bar into four regions, as shown in figure. the average cross-sectional area within each
region is evaluated and then used to define an element with uniform cross section. The resulting
four-element, five node finite element model is shown in fig. In the finite element mode, every
element connects to two nodes. In fig the element numbers are circled to distinguish them from
one numbers. In additional to the cross section, traction and body forces are also (normally0
treated at constant within each element. However, cross-sectional area, traction and body forces
can differs in magnitude from element to element. Better approximation are obtained by
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
48
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
increasing the number of elements. It is convenient to define a node at each locations where a
point load is applied.
For easy implementation, an orderly numbering scheme for the model has to be adopted.
Each element has two nodes; therefore the element connectivity information can be
conveniently represented as shown in fig. further the element connectivity table is also given. In
the connectivity table, the headings 1 and 2 refer to local node numbers of an element, and the
corresponding node number on the body are called global numbers. Connectivity thus establishes
the local –global correspondence. In this simple example, the connectivity can be easily generated
since local node 1 is the same as the element number e, and local node 2 is e+1. Other ways of
numbering nodes or more complex geometries suggest the need for a connectivity table. The
connectivity is introduced in the program using the array NOC.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
49
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The concepts off dof, nodal displacement, nodal loads and element connectivity are central
to the finite element method and should be clearly understood.
1
T Adx uT f Adx uT Tdx ui Pi
2 L
...(1)
L L
i
The quantities ,u,f and T in equation (1). In the last term, Pi represents a force acting at
point I, and ui is the x displacement at that point. The summation on I gives the potential energy
the potential energy due to all point load.
Since the continuum has been discretized into finite elements, the expression for becomes
1
A dx u fA dx- u T dx- QiPi
T T T
e 2 e
e
e
e
e
i
...(2)
the last term in Eq. assumes that point loads Pi are applied at the nodes. This assumption makes
the present derivation simpler with respect to notation and is also a common modeling practice,
Equation can be written as
1 T
2
Ue = A dx
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
50
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
d
( ) .... (2)
dx
T
( )Adx T fAdx T Tdx Pi 0 ...(3a)
L L L
i
this equation should hold for every consistent with the boundary condition. The first term
represent the internal virtual work, while the load terms represents the external virtual work.
e
e
T
E ( )A dx T fAdx T Tdx Pi 0
e
e
e
e
i
...(3b)
Note that is the strain due to the actual loads in the problem, while is the strain due to
the actual loads in the problem, while () is a virtual strain. Similar to the interpolation steps in
Eqs. We express
N
<(4)
( ) B
Where =[ 1, 2]T represents the arbitrary nodal displacement of element e. Also, the global
virtual displacements at the nodes are represented by
Consider the first term, representing internal virtual work in the equation.
e
e
T
E ( )Adx T fAdx T dx iPi 0
e
e
e
e
i
..(1)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
51
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
substituting the values of =N & ()=B in the above equation & nothing that =Bq then we get,
in the finite element model, the cross sectional area of element e, denoted by Ae, is constant. Also,
B is a constant matrix, Further, dx =( ( e / 2)d. thus,
e T 1
E ( )A dx=qT [Ee A e B B d ]
T
...(3a)
e 2 1
=qT k e ...(3b)
= k qT e
Ke E Ae eBTB ...(4)
Ee A e 1 1
ke
e 1 1
12. How the global stiffness matrix can be assembled in one dimensional problem:
1
qT k e q qT f e qT f e qT T e PQ
i i <(1)
e 2 e e e i
1 T
= Q KQ QTF
2
by taking element connectivity into account. This step involves assembling K and F from
element stiffness and force matrices. The assembly of the structural stiffness matrix K from
element stiffness matrices ke+ will first be shown here.
Referring to the finite element model in fig. (0.1b) let us consider the strain energy in say, element
3. We have
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
52
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 T 3
U3 = qk q (0.1 a)
2
1 T E3 A 3 1 1
U3 = q q (o.1b)
2 3 1 1
0 0 0 0 0
0 Q1
0 0 0 0
Q
E3 A 3 E3 A 3 2
1 0 0 0 Q3
U3 Q1,Q2 ,Q3 ,Q 4 ,Q5 3 3 (0.3)
2
E3 A 3 E3 A 3
0 0 0 Q4
3 3
Q5
0
0 0 0 0
From the previous equations, we see that elements of the matrix k3 occupy the third and
forth rows and columns of the K matrix. Consequently, when adding element-strain energies, the
elements of ke are placed in the appropriate locations of the global K matrix, based on the element
connectivity; overlapping elements are simply added. We can denote this assembly symbolically
as
K ke (0.4a)
e
Similarly, the global load vector F is assembled from element-force vectors and point loads are
F (f e + T e ) P (0.4b)
e
1. The dimension of the global stiffness K is (N X N), where N is the number of nodes. This
follows from the fact that each node has only one degree of freedom.
2. K is a symmetric.
3. K is a banded matrix. That is, all elements outside of the band are zero. This can be seen in
Example, just considered. In this example, K can be compactly represented in banded form
as
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
53
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
A1 A1
1
1
A1 A 2 A2
1 2 2
A A A
Kbanded = E 2 3 3
2 3 3
A A A
3 4 4
3 4 4
A 4
0
4
Elimination Approach:
Qp 1 = a1, Qp 2 , = a2 , ...., Qp r = ar
Step 1: Store the p1th, p2th,<<and prth rows of the global stiffness matrix K and force vector F.
These rows will be used subsequently.
Step 2: Delete the p1th row and column, the p2th row and column, <<.,and the prth row and
column from the K matrix. The resulting stiffness matrix K is of dimensions (N – r, N – r).
Similarly, the corresponding load vector F is of dimension (N – r, 1). Modify each load component
as
Step 3: For each element, extract the element displacement vector q from the Q vector, using
element connectivity, and determine element stresses.
Step 4: Using the information stored in step 1, evaluate the reaction forces at each support dof
from
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
54
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Penalty Approach:
Qp 1 = a1, Qp 2 , = a2 , ...., Qp r = ar
Step 1: Modify the structural stiffness K by adding a large number C to each of the p1th, p2th,
<<, prth diagonal elements of K. Also, modify the global load vector F by adding Ca1 to Fp , Ca2 1
to Fp , <<., and Car to Fp , . Solve KQ = F for the displacement Q, where K and F are the modified
2 r
Of all possible displacements that satisfy the boundary conditions of a structural system,
those corresponding to equilibrium configurations make the total potential energy assume a
minimum value.
1
i. A shape function has a value of either 1 or 0 at a nodal point : Ni j ij =
ie i = j
0 if i j
ii. The sum of all shape functions at any point equals 1= Ni () = 1.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
55
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The cross-sectional dimensions and elastic properties of each member are constant along its
length.
The element may interconnect in a 2-D or 3-D configuration in space.
The element is mechanically equivalent to a spring, since it has no stiffness against applied
loads except those acting along the axis of the member.
However, unlike a spring element, a truss element can be oriented in any direction in a
plane, and the same element is capable of tension as well as compression.
19. Write down the characteristic features of global stiffness matrix of a truss element.
For a truss element, loads can be applied on a node only. If loads are distributed on a
structure, they must be converted to the equivalent loads can be applied at nodes. Loads can be
applied in any direction at the node, however, the element can resist only the axial component,
and the component perpendicular to the axis merely causes free rotation at the joint.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
56
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
PART – B
Let us consider the truss member, shown in figure. The element is inclined at an angle , in
a counter clock wise direction. The local deflections are 1 and 2. The global deflections are: u1, u2,
u3, and u4. We wish to establish a relationship between these deflections in terms of the given
trigonometric relations.
u1x
1 c s 0 0 u1y
=
1 0 0 c s u2x
u
2y
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
57
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Along with equation (3, 2), we also need an equation that relates the local and global forces.
By using trigonometric relations similar to the previous section, we can derive the desired
relationship between local and global forces. However, it will be easier to use the work-energy
concept for this purpose. The forces in local coordinates are : R1 and R2, and in global coordinates:
f1, f2, f3, and f4, see figure for their directions.
Since, work done is independent of a coordinate system, it will be the same whether we use
a local coordinate system or a global one. Thus, work done in the two systems is equal and given
as,
W = T R = uT f, or in an expanded form,
f1
R
W= 1 2 1 = u1x u1y u2x u2y f2
R2 f3
f4
= R = u f
T T
T u R
T
Substituting = T u in the above equation, we get, u T R u f .dividing by u
T T T T
T R = f
T
Equation can be used to convert local forces into global forces and vice versa.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
58
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Now we will derive the finite element equation in local coordinates system. This equation
will be converted to global coordinate system, which can be used to generate a global structural
equation for the given structure. Note that, we can not use the element equations in their local
coordinate form; they must be converted to a common coordinate system, the global coordinate
system.
Consider the element shown below, with nodes 1 and 2, spring constant k, deflections 1
and 2 and forces R1 and R2. As established earlier, the finite element equation in local coordinates
is given as,
R1 k -k 1
=
R2 k k 1
1, R1 1 k 2
2, R2
Using the relationship between local and global deflections and forces, we can convert an
element equation from a local coordinate system to a global system.
We want a similar equation, but in global coordinates. We can replace the local force R with
the global force f derived earlier and given by the relation:
{ f } = [TT] {R}
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
59
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
{ f } = [kg] {u}
c 0
s 0 AE / L AE / L C S 0 0
[Kg]
0 c AE / L AE / L 0 0 C S
0 s
c2 cs c 2 cs
cs s2 cs s2
K g 2 (AE / L)
c cs c 2 cs
cs s s2
2
cs
2. Example
find displacements of joints 2 and 3. find stress, strain & internal force in each member.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
60
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Solution:
Using Shigley’s Machine Design book for yield strength values, we have
Element (1)
=0
c=cos =1, c2=1
s=sin=0, s2=0
CS=0
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
61
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
c2 cs c 2 cs
cs s2 cs s2
(AE / L)x 2
(1)
K g
c cs c 2 cs
cs s s2
2
cs
1 0 1 0
0 0 0 0
(53.1)x
(1)
K g
1 0 1 0
0 0 0 0
Element 2:
=90
c=cos 90 =cos 0=1,s2=1
cs=0
EA/L=69 x 200(1/150)=92 kN/mm
Element 3
=30
c=cos 30 =0.866,c2=0.75
s=cos 60 =.5,s2=0.25
cs=0.433
EA/L=207 x 100 (1/300)=69 kN/mm
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
62
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Since there are 6 deflections (or DOF), u1 through u6, the matrix is 6 x6. Now, we will place
the individual matrix element from the element stiffness matrices into the global matrices into the
global matrix according to their position of row and column members.
Element [1]
Element [3]
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
63
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Assembling all the terms for elements [1],[2] and [3], we get the complex matrix equation of the
structure.
Boundary conditions:
Node 1 is fixed in both x and y directions, where as, node 2 is fixed in x-direction only and free to
move in the y-direction. Thus,
U1x=u1y=u2x=0
Therefore, all the columns and rows containing these elements should be set to zero. The reduced
matrix i:
29.9u2y -29.9u3x-17.2u3y= 0
-29.9u2y+104 u3x+29.9u3y=0
-17.2u2y+ 29.9 u3x+17.2 u3y = -0.4
solving we get
u2y= 0.0043
u3x =0.0131
u3y= 0.0502
Element (1)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
64
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Note that u1x, u1y, u2x etc are not coordinates, they are actual displacements.
L u3x 0.0131
L / L 0.0131111/ 260 5.02 10-5mm / mm
E 69 5.02 105 0.00347kN/ mm2
reaction R= A 0.00347 kN
L u3x 0.0043
L / L 0.00043 /150 2.87 10 -5mm / mm
E 69 2.87 10 5 1.9803kN/ mm2
Reaction R= A (1.9803 10 3 )(200)=0.396 kN
Element (3)
Since element (30 is at angle 30, the change into the length is found by adding the
displacement components of nodes 2 and 3 along the element (at 30). Thus,
=L/L=-0.0116/300=-3.87 x 10 -5 3.87x10 5 mm / mm
Factor of Safety:
Factor of safety ‘n’ is the ration of yield strength to the actual stress found in the part.
Sy 0.0375
Element(1) n= 10.8
0.00347
Sy 0.0375
Element(2) n= 18.9
0.00198
Sy 0.0586
Element(3) n= 7.325
0.0080
The lowest factor of safety is found in element (3), and therefore the steel bar is the most
likely to fail before the aluminum bar does.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
65
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
3.
Given;
Required :
Solution:
AE 1.5 10 10h6 lb
K1 7.5 10s
L in
AE 1 10 10 6
lb
K2 2.5 105
L 40 in
AE 30 10 10 6
lb
K3 6.0x105
L 50 in
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
66
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Element (1)
Denoting the Spring constant for element (1) by k1, and the stiffness matrix by K(1), the stiffness
matrix in global coordinates is given as,
cs s2 cs s2 u1y
K1 2
(1)
K g
c cs c 2 cs u2x
cs s s2 u2y
2
cs
Element (2)
C=cos =0 , c2=0
S=sin =1, s2=1
Cs2=0
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
67
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Element 3
u 4x u 4y u2x u2y
.36 .48 .36 .48 u4x
.48 .64 .48 .64 u
K (3) K 3 4y
.36 .48 .36 .48 u2x
.48 .64 48 .64 u2y
Following the procedure for assembly described earlier, the assembled matrix is,
1 k1 0 K1 0 0 0 0 0
0
2 0 0 0 0 0 0 0
3 K1 0 K1 .36K 3 .48K 3 0 0 .36K 3 .48K 3
4 0 0 .48K 3 K 2 .64K 3 0 K 2 .48K 3 .64K 3
K g
5 0 0 0 0 0 0 0 0
6 0 0 0 K 2 0 K2 0 0
7 0 0 .36K 3 .48K 3 0 0 .36K 3 .48k 3
8 0 0 .48k 3 .64K 3 0 00 .48K 3 .64K 3
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
68
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
u1x=u1y=u3x=u3y=u4x=u4y=0
We will suppress the corresponding rows and columns. The reduced matrix is a 2 x2, given below.
K1 48K 3 u2x
K g u
.48K 3 K 2 .64K 3 2y
u2x 0.0000438 in
u2y 0.01241 in .
Solution:
(a) Using Eq: the coordinate of point P is given by
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
69
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
2
p= (24 20) 1
16
=-.0.5
up
=0.75(0.003)+0.25(-00.0005)
=0.001 in.
du
dx
Upon using the chain rule of differentiation, we obtain
du d
(1)
d dx
d 2
(2)
dx x 2 x1
Also since,
1 1
U=N1q1+N2q2= q2
2 2
We have
du q1 q2
(3)
d 2
thus, Eq yield
1
( q1 q2 ) (4)
x2 x2
=Bq (5)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
70
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1
B [ 1 1] (6)
x 2 x1
Consider a typical finite element in figure. In the local number scheme, the first node will
be numbered 1 and the second node. 2. The notation x1=x-coordinate of node 1,x2=x-coordinate of
one 2 is used. We define a Natural or intrinsic coordinate system, denote by , as
2
(x x1 ) 1 (1)
x 2 x1
From fig, we see that =1 at node 2. the length of an element is covered when changes from -1 to
1. We use this system of coordinates in defining shape functions, which are used in interpolating
the displacement field.
Now the unknown displacement field within an element will be interpolated by a linear
distribution (fig). this approximation becomes increasingly accurate as more element are
considered in the model. To implement this linear interpolation, linear shape functions will be
introduced as.
1
N1( ) (2)
2
1
N2 ( ) (3)
2
The shape function N1 and N2 are shown in fig a and b, respectively. Then graph of the
shape function N1 in fig (2a) is obtained from equation (2) by noting that N11 at =-1, N1=0 at =1,
and N1 is a straight line between he two points. Similarly, the graph of N2 in fig (4b) is obtained
from equation (3) . Once the shape function are defined the linear displacement filed within the
element can be written in terms of the nodal displacement q1 and q2 as
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
71
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Figure :4(a) Shape function N1, 9b) shape function N2, and (c) linear interpolation using N1 and
N2
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
72
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
It may be noted that the transformation from x to in equation (1) can be written in terms
of N1 and N2 as
X=N1x1+N2x2 (6)
Comparing E qs.(4a) and (6) , we see that both the displacement u and the coordinate x are
interpolated within the element using the same shape functions N 1 and N2. this is referred to as
isoparametric formation in the literature.
Though linear shape function have been used previously, other choices are possible. Shape
functions need to satisfy the following:
1. First derivatives must be finite within an element
2. Displacements must be continuous across the element boundary.
Rigid body motion should not introduce any stress in the element.
6. Consider the bar as shown in fig (1). For each element i,Ai and I are the cross-sectional area and
length, respectively. Each element i is subjected to a traction force T i per unit length and a body
force f per unit volume. The units of Ti,f, Ai and so on are assumed:- be consistent. The Young’s
modulus of the material is E. A concentrated load P2 is applied at node 2. The structural stiffness
matrix and nodal load vector will bow be assembled.
The element stiffness matrix for each element I is obtained from Equation as
EAi 1 1
[K(1) ]
i 1 1
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
73
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Element 1 2
1 1 2
2 2 3
3 3 4
4 4 5
The element stiffness matrices can be ‚expanded’ using the connectivity table and then
summed (or assembled) to obtain the structural stiffness matrix as follows:*
1 1 0 0 0 0 0 0 0 0
1 1 0 0 0 0 1 1 0 0
EA1 EA 2
K 0 0 0 0 0 0 1 1 0 0
1 2
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0
EA 3 EA 4
+ 0 0 1 1 0 0 0 0 0 0
3 4
0 0 1 1 0 0 0 0 1 1
0 0 0 0 0 0 0 0 1 1
Which gives
A1 A 1
0 0 0
1
1
A A1 A 2 A2
1 0 0
1 1 2 2
A 2 A2 A3 A 3
K E 0 0
2
2 3 3
A3 A3 A 4 A4
0 0
3 3 4 4
A 4 A4
0 0 0
4 4
*This ‚expansion‛ of element stiffness matrices as shown in Examples is merely for illustration
purposes and is never explicitly carried out in the computer. Since storing zeroes is inefficient.
Instead , K is assembled directly from k’ using connectivity table.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
74
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
A11f 1T1
0
2 2
A11f T A 2 2 f 2 T2
2 1 1
2 P2
2 2
A 2 2 f 2 T2 A 3 3 f 3 T3
F
2 2 2 2 0
A 3 3 f T A f T
3 3 4 4 4 4
2 2 2 2 0
A f T
4 4 4 4
2 2 0
7. Consider the thin (thin) plate in fig.(1a). the plate has a uniform thickness t=1 in. Young’s
modulus E=30 x 106 psi, and weight density =100 lb at its midpoint.
Solution:
(a) Using two element each of 12 in, in length, we obtain the finite element model in fig. Nodes
and elements are numbered as shown. Note that the area at the midpoint of the plate in fig (1a) is
4.5 in2. consequently, the average and of element 1 is A1 =(6+4.5)/2 =5.25 in2, and the average area
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
75
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
of element 2 is A2=(4.5+3)/2 =3.75 in2. the boundary condition for this model is Q1=0.
(b) From Eq. we can write down expressions for the element stiffness matrices of the two element
as
1 2 Global dof
30 106 5.25 1 1 1
K1 1 1 2
12
and
2 3
30 106 3.75 1 1 2
K2 1 1 3
12
global dof
5.25 12 0.2836 1 1
f 1=
2 1 2
and
3.75 12 0.2836 1 2
f2=
2 1 3
1 2 3
5.25 5.25 0 1
30 106 2
K 5.25 9.00 3.75
12
0 3.75 3.75 3
the externally applied global load vector f is assembled from f1,f2, and the point load P=100 lb; as
8.9334
F 15.3144 100
6.3810
(d) In the elimination approach, the stiffness matrix K is obtained by deleting rows and columns
corresponding to fixed dofs. In this problem , dof 1 is fixed. Thus, K is obtained by deleting the first row
and column of the original f. the resulting equations are
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
76
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
2 3
30 10 9.00 3.75 Q2 115.3144
6
12 3.75 3.75 Q3 6.3810
solution of these equations yields
Q2=0.9272 x105 in
Q3=0.9953 x 10-5 in
(e) using Eqs, 3.15 and 3.16 , we obtain the stress in each element
1 0
30 106 {1 1} -5
12 0.9272 10
= 23.18 psi
and (f) the reaction force R1 at node 1 is obtained from
1 0.9272 10
-5
30 106 {1 1} -5
12 0.9953 10
= 1.70 psi
Eq. This calculation require the first row of K from part ©. Also , from part ©, note that the
externally applied load 9due to the self-weight) at note 1 is F1= 8.9334 lb. thus,
8.9334
0
30 10
6
R1 [5.25 -5.25 0] 0.9272 10-5
12 0.9953 10 5
=-130.6 lb
8. An axial load P= 300 x 103 N is applied at 20C to the rod as shown in fig . the temperature is then raised
to 60 C.
(a) Assemble the K and F matrices.
(b) Determine the nodal displacement and elements stresses.
Solution:
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
77
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
315 315 0
K 10 315 1115 800 N / mm
3
0 800 800
Now, in assembling F, both temperature and point load effects have to be considered. The element
temperature forces due to T=40C are obtained from Eq. as
Global dof
1 1
1 70 103 900 23 10 6 40 N
1 1
and
1 2
2 200 103 1200 11.7 10 6 40 N
1 3
57.96
f=103 57.96 112.32 300
112.32
or
F 103 [57.96,245.64,112.32]T N
(b) the elimination approach will now be used to sole for the displacements. Since dofs 1 and 3 are
fixed , the first and third rows and columns to K, together with the first and third components
of F, are deleted. This results in the scalar equation.
103[115]Q2=103 x 245.64
yielding
Q2=0.220 mm
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
78
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Thus,
Q=[0,0.220, 0]Tmm
70 102 0
1 [ 11] 6
70 10 23 10 40
3
200 0.220
=12.60 MPa
and
200 103 0.220
2 [1 1] 6
200 10 11.7 10 40
3
300 0
=-240.27 MPa
9. Consider the bar shown in figure. an axial load P=2500 x 103 N . using the penalty approach for
handling boundary conditions, do the following.
Figure
Solution:
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
79
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 2 3
0.56 -0.56 0
K 10 -0.56 0.86 -0.30
6
0 -0.30 0.30
8600.56 0.56 0
K 10 0.56
6
0.86 0.30
0 0.30 8600.30
8600.56 0.56 0 Q1 0
3
10 0.56
6
0.86 0.30 Q 2 200 10
0 0.30 8600.30 Q3 0
1 15.1432 10-6
1 70 103 (1 1)
300 0.23257
=54.27 MPa
1 0.23257
2 200 103 ( 1 1) -6
400 8.1127 10
=-116.27 MPa
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
80
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
R1 CQ1
=-[0.86 1010 ] 15.1432 106
=-130.23 103
Also,
R3 CQ1
=-[0.86 1010 ] 8.1127 106
=-69.77 103N
10. In the following fig , a load P= 60 x 103 N is applied as shown Determine the displacement field
stress, and support reactions in the body. Take E= 20 x103 N/mm2.
Solution:
In this problem, we should first determine whether contact occurs between the bar and the
wall , B. to do this, assume that the wall does not exist. Then, the solution to the problem can be
verified to be.
Mm
QB=1.8 mm
Where QB is the displacement of point B’ . from this result, we see that contact does occur.
The problem has to be re-solved, since the boundary conditions are now different: the
displacement at B’ is specified to be 1.2 mm. Consider the two –element finite element model in
figure. the boundary conditions are Q1=0 and Q3=1.2 mm. The structural stiffness matrix K is
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
81
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 1 0
20 103 250
K 1 2 1
150
0 1 1
In the penalty approach, the boundary conditions Q1=0 and Q3=1.2 imply the following
modification: A large number c chosen here as C= (2/3)x 1010 , is added on to the 1st and 3rd
diagonal elements of K. also, the number (C x 1.2) gets added on to the 3rd component of F. thus,
the modified equation are.
20001 1 0 Q1 0
105 3
1 2 1 2
Q 60.0 10
3
0 1 20001 Q3 80.0 107
the solution is
Q=[7.49985 x 10-5, 1.500045, 1.200015]Tmm
The results obtained from the penalty approach have a small approximation error due to
flexibility of the support introduced. In fact, the reader may verify that the elimination approach
for handling boundary conditions yields the exact reactions, R1=-50.0 x 103 N and R3 =-10.0 x 103 N.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
82
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
So far, the unknown displacement field was interpolated by linear shape functions within
each element. In some problems, however, use of quadratic interpolation leads to far more
accurate results. In this section, quadratic shape functions will be introduced, and the
corresponding element stiffness matrix and load vectors will be derived. The reader should note
that the basic procedure is the same as that used in the linear one-dimensional element earlier.
Consider a typical three-node quadratic element, as shown in figure (a). In the local
numbering scheme, the left node will be numbered 1, the right node 2, and the midpoint 3. Node 3
has been introduced for the purposes of passing a quadratic fit and is called an internal node. The
notation xi = x – coordinate of node i, I = 1, 2, 3, is used. Further, q = [q1, q2, q3]T, where q1, q2, and q3
are the displacements of nodes 1, 2 and 3, respectively. The x-coordinate system is mapped onto a
-coordinate system, which is given by the transformation.
2 x x3
.......(1)
x2 x1
From Equation 1, we see that = -1, 0 and +1 at nodes 1, 3 and 2 (Fig. (b)).
Now, in -coordinates, quadratic shape functions N1, N2, and N3 will be introduced as
1
N1 1 ....(2a)
2
1
N2 1 .....(2b)
2
N3 1 1 .....(2c)
The shape function N1 is equal to unity at node 1 and zero at nodes 2 and 3. Similarly, N2
equals unity at node 2 and equals zero at the other two nodes; N 3 equals unity at node 3 and
equals zero at nodes 1 and 2. The shape functions N 1, N2 and N3 are graphed in figure. The
expressions for these shape functions can be written down by inspection. For example, since N1 = 0
at = 0 and N1 = 0 at = 1, we know that N1 must contain the product (1 - ). That is, N1 is of the
form
N1 = c(1 - ) <.(3)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
83
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1
The constant c is now obtained from the condition N1 = 1 at = -1, which yields c = , resulting
2
in the formula given in Equation 2(a). These shape functions are called Lagrange shape functions.
Now the displacement field within the element is written in terms of the nodal
displacements as
where N = [N1, N2, N3] is a (1 3) vector of shape functions and q = [q1, q2, q3]T is the (3 1)
element displacement vector. At node 1, we see that N1 = 1, N2 = N3 = 0, and hence u = q1. Similarly,
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
84
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
du
(strain-displacement relation)
dx
du d
= (chain rule)
d dx
2 du
= (using Eq.1) ......(5)
x 2 x1 d
2 dN1 dN2 dN3
= , , .q (using Eq. 4)
x 2 x1 d d d
We have,
2 1 2 1 2
, , 2 q ......(6)
x2 x1 2 2
which is of the form
= Bq <..(7)
where B is given by
2 1 2 1+2
B , , 2 <..(8)
x2 x1 2 2
E Bq <(9)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
85
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Note that since Ni are quadratic shape functions, B in Eq.8 is linear in . This means that the
strain and stress can vary linearly within the element. Recall that when using linear shape
functions, the strain and stress came out to be constant within the element.
We now have expressions for u, , and in Eqs. 4(b), 7, and 9, respectively. Also, we have
dx e / 2 d from Eq.1.
Again, in the finite element model considered here, it will be assumed that cross-sectional
area Ae, body force F, and traction force T are constant within the element. Substituting for u, , ,
and dx into the potential-energy expression yields.
1
A dx - uT fA dx - u T dx - QiPi
T
T
e 2 e e
e
e
e
i
1 e
2q BT B d q qT Ae e f N T d
1 1
T
= Ee Ae 2
e 1
e 2 1
- qT e T N T d Q i Pi
1
<.(10)
e 2 1 i
1
qT k eq qT f e qTT e Qi Pi
e 2 e e i
yields
E A 1
k e e e e BT B d <(11a)
2 1
1 2 3 local dof
7 1 8 1
<(11b)
K e e e 1 7 8 2
E A
3 e
8 8 16 3
the element body force vector fe is given by
Ae ef 1
fe
2 1
NT d <<(12a)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
86
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
local dof
1/ 6 1
<.. (12b)
f e Ae ef 1/ 6 2
2 / 3 3
eT 1
Te
2
1
NT d <..(13a)
which results in
local dof
1/ 6 1
<..(13b)
T ef 1/ 6 2
e
2 / 3 3
1 T
The total potential energy is again of the form Q KQ QT F , where the structural
2
stiffness matrix k and nodal load vector f are assembled from element stiffness matrices and load
vectors, respectively.
12. Consider the structure shown in figure. A rigid bar of negligible mass, pinned at one end, is
supported by a steel rod and an aluminium rod. A load P = 30 103 N is applied as shown.
(a) Model the structure using two finite elements. What are the boundary conditions for your
model?
(b) Develop the modified stiffness matrix and modified load vector. Solve the equations for Q.
Then determine element stresses.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
87
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Solution
(a) The problem is modeled using two elements as shown in the following connectivity table:
CONNECTIVITY TABLE
The boundary conditions at nodes 3 and 4 are obvious: Q3 = 0 and Q4 = 0. Now, since the
rigid bar has to remain straight, Q1, Q2, and Q5 are related as shown in Figure(b). The multipoint
constraints due to the rigid bar configuration are given by
Q1 0.333 Q5 0
Q2 0.833 Q5 0
3 1
53.33 3
200 103 1200 1 1 3 53.33
k1
1 1 10
4500
-53.33 53.33 1
and
4 2
70 10 900 1 1
3
21 21 4
k2 103
3000 1 1 -21 21 2
1 2 3 4 5
53.33 0 53.33 0 0 1
0 21 0 21 0 2
K 10 53.33
3
0 53.33 0 0 3
0 21 0 21 0 4
0 0 0 0 0 5
The K matrix is modified as follows: a number C = 53.33 103 104 , large in comparison
to the stiffness values, is chosen. Since Q3 = Q4 = 0, C is added on to the (3, 3) and (4, 4) locations of
K. Next, multipoint constraints given in part (a) are considered. For the first constraint, Q 1 -
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
88
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
0.333Q5 = 0, we note that 0 = 0, 1 = 1, and 2 = -0.333. The addition to the stiffness matrix is
obtained from Equation as
1 5
C 12 C 1 2
7 53.33 17.77 1
2
10
C 1 2 C 2
17.77 5.925926 5
The force addition is zero since 0 = 0. Similarly, the consideration of the second multipoint
constraint Q2 – 0.833Q5 = 0 yields the stiffness addition
2 5
53.33 44.44 2
107
44.44 37.037037 5
On addition of all the preceding stiffnesses, we obtain the final modified equations as
The solution, obtained from a computer program that solves matrix equations, is
2 28.35 MPa
13. Consider the rod (a robot arm) in figure. Which is rotating at constant angular velocity = 30
rad/s. Determine the axial stress distribution in the rod, using two quadratic elements. Consider
only the centrifugal force. Ignore bending of the rod.
Solution:
A finite element model of the rod, with two quadratic elements, is shown in figure. The
model has a total of five degrees of freedom. The elements stiffness matrices are (from equation).
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
89
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 3 2 Global dof
7 1 8 1
107 0.6
K
1
1 7 8 3
3 21
8 8 16 2
Figure
And
3 5 4
7 1 8 3
107 0.6
K
2
1 7 8 5
3 21
8 8 16 4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
90
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Thus,
1 2 3 4 5
7 8 1 0 0 1
8 16 8 0 0 2
107 0.6
K 1 8 14 8 1 3
3 21
0 0 8 16 8 4
0 0 1 8 7 5
r 2
f Ib / in3
g
Global dof
1 1
6
1
f 0.6 21 f1 3
1
6
2
3 2
and
Global dof
1 3
6
1
f 1 0.6 21 f2 5
6
2
3 4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
91
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Which yields
There stress can now be evaluated from equations. The element connectivity table is as follows:
Thus,
q = [Q1, Q3, Q2] T
Q1
2 1 2 1 2
1 10
7
, , 2 Q3
21 2 2
Q2
The stress at node 1 in element 1 is obtained by substituting = -1 into the previous equation,
which results in
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
92
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
0
1 1 10 10 1.5, 0.5, 2.0 1.07076
7 2 3
21 .5735
= 583 psi
The stress at node 2 (which corresponds to the midpoint of element 1) is obtained by substituting
for = 0:
0
1 3 10 10 0.5,0.5,0 1.07076
2
7 3
21 .5735
= 510 psi
Similarly, we obtain
1 2 2 1 437psi 2 3 218psi 2 2 0
The axial distribution is shown in figure. The stresses obtained from the finite element model can
be compared with the exact solution, given by
2
exact (x) (L2 x 2 )
2g
14.
Consider the four bar truss shown in figure. It is given that E = 29.5 x 10 6 psi and Ae = l in 2.for all
elements. Complete the following:
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
93
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Figure
Solution:
a) It is recommended that tabular form be used for representing nodal coordinate data are as
follows:
Node X Y
0
1 0
0
2 40
30
3 40
30
4 0
Element 1 2
1 2
1
2 2
3
3 3
1
4 3
4
Note that the user has a choice in defining element connectivity. For example, the
connectivity of element 2 can be defined as 2 – 3 instead of 3 – 2 as in the previous table.
However, calculations of the direction cosines will be consistent with the adopted connectivity
scheme. Using formulas in equations and, together with the nodal coordinate data and the given
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
94
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Element e M
1 40 1 0
2 30 0 -1
3 50 0.8 0.6
4 40 1 0
For example, the direction cosines of elements 3 are obtained as = (x3 – x1)/ e = (40 – 0)/50 = 0.8
and m = (y3 – y1)/ e = (30 – 0)/50 = 0.6. Now, using equation, the element stiffness matrices for
element 1 can be written as
1 2 3 4 Global dof
1 0 1 0 1
0
29.5 10 0
6 0 0 2
K1
40 1 0 1 0 3
0 0 0 0 4
The global dofs associated with elements 1, which is connected between nodes 1 and 2, are
indicated in K1 earlier. These global dofs are shown in figure and assist in assembling the various
element stiffness matrices.
5 6 3 4
0 0 0 0 5
0 1 6
29.5 10 0 1
6
K2
50 0 0 0 0 3
0 1 0 1 4
1 2 5 6
.64 .48 .64 .48 1
.48 .36
29.5 106 .48 .36 2
K
3
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
95
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
7 8 5 6
1 0 -1 0 7
0 0 0 8
29.5 106 0
K
4
40 -1 0 1 0 5
0 0 0 0 6
b) The structural stiffness matrix K is now assembled from the element stiffness matrices. By
adding the element stiffness contributions, noting the element connectivity, we get
1 2 3 4 5 6 7 8
22.68 5.76 15.0 0 7.68 5.76 0 0 1
5.76 4.32 0 0 5.76 4.32 0 0 2
15.0 0 15.0 0 0 0 0 0 3
29.5 106 0 0 0 20.0 0 20 0 0 4
K
600 7.68 5.76 0 0 22.68 5.76 15.0 0 5
5.76 4.32 0 20.0 5.76 24.32 0 0 6
0 0 0 0 15.0 0 15.0 0 7
0 0 0 0 0 0 0 0 8
c) The structural stiffness matrix K given above needs to be modified to account for the
boundary conditions. The elimination approach will be used here. The rows and columns
corresponding to dots 1, 2, 4, 7, and 8, which correspond to fixed supports, are deleted form
the K matrix. The reduced finite element equations are given as
The nodal displacement vector for the entire structure can therefore be written as
D) The stress in each element can now be determined form equation, as shown below.
The connectivity of element 1 is 1 – 2. Consequently, the nodal displacement vector for element 1
is given by q = [0, 0, 27.12 x 10-3, 0]T, and equation yields.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
96
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
0
29.5 10 6
0
1 1 0 1 0 3
40 27.12 10
0
= 20 000.0 psi
5.65 10 3
3
29.5 106 22.25 10
2 0 1 0 1 3
30 27.12 10
0
=-21 880.0 psi
3 = 5208.0 psi
4 = 4167.0 psi
d) The final step is to determine the support reactions. We need to determine the reaction
forces along dofs 1, 2, 4, 7, and 8, which correspond to fixed supports. These are obtained
by substituting for Q into the original finite element equation R = KQ – F. In this
substitution, only those rows of K corresponding to the support dofs are needed, and F = O
for theses dofs. Thus, we have
0
0
22.68 5.76 15.0 0 7.68 5.76 0 0
R1 27.12 10 3
Which results in
R1 15833.0
R
2 3126.0
R 4 21879.0
R 4167.0
7
R8 0
A free body diagram of the truss with reaction forces and applied loads is shown in figure.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
97
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
15. The four bar truss of Example is considered here, but the loading is different. Take E = 29.5 x
106 psi and = 1/150 000 per F.
a) There is an increase in temperature of 50F in bars 2 and 3 only (figure). There are no other
loads on the structure. Determine the nodal displacements and element stresses as a result
of this temperature increase. Use the elimination approach.
b) A support settlement effect is considered here. Node 2 settles by 0.12 in. Vertically down,
and in addition, two point loads are applied on the structure (figure). Write down (without
solving) the equilibrium equations KQ = F, where K and F are the modified structural
stiffness matrix and load vector, respectively. Use the penalty approach.
c) Use the program TRUSS2 to obtain the solution to part (b).
Solution:
a) The stiffness matrix for the truss structure has already been developed in example. Only
the load vector needs to be assembled due to the temperature increase. Using equation, the
temperature loads as a result of temperature increases in elements 2 and 3 are, respectively.
Figure
Global dof
0 5
29.5 106 50 1 6
2
150,000 0 3
1
4
And
0.8 1
29.5 10 50 0.6 2
6
3
150,000 0.8 5
0.6 6
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
98
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The 2 and 3 vectors contribute to the global load vector F. Using the elimination approach;
we can delete all rows and columns corresponding to support dofs in K and F. The resulting finite
element equations are
15.0 0 0 Q3 0
29.5 106
0 22.68 5.76 Q5 7866.7
600
0 5.76 24.32 Q6 15733.3
Which yield
Q3 0
Q5 0.003951 in.
Q 0.01222
6
The element stresses can now be obtained from equation. For example, the stress in element 2 is
given as
0.003951
0.01222
29.5 106 29.5 106 50
2
30
0 1 0 1
0
150,000
0
= -8631.7 psi
1 0
2 2183
psi
3 3643
4 2914
b) Support 2 settles by 0.12 in. vertically down, and two concentrated forces are applied
(figure). In the penalty approach for handling boundary conditions, recall that a large
spring constant C is added to the diagonal elements in the structural stiffness matrix at
those dofs where the displacements are specified. Typically, C may be chosen 104 times the
largest diagonal element of the unmodified stiffness matrix (see equation). Further, a force
Ca is added to the force vector, where a is the specified displacement. In this example, for
dof 4, a = -0.12 in., and consequently, a force equal to -0.12C gets added to the fourth
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
99
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
location in the force vector. Consequently, the modified finite element equations are given
by
c) Obviously, the equations in (b) are too large for hand calculations. In the program TRUSS,
that is provided, these equations are automatically generated and solved from the user’s
input data. The output from the program is
Q3 0.0271200
Q
4 0.1200145
in.
Q
5 0.0323242
Q6 0.1272606
And
1 20 000.0
2 -7 125.3
psi
3 -29 791.7
4 23 833.3
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
100
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
UNIT - III
PART - A
T
u v u v
, ,
x y y x
The points where the cameras of the triangles meet are called nodes, and each triangle
formed by three nodes and three series called an element.
ui
Vi
u j
d
v j
um
v
m
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
101
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The displacement inside the element are now written using the shape functions and the
nodal values of the unknown displacement field. We have
u q1 q5 q3 q5 qs
u q2 q6 q4 q6 q6 5.12b
The relations 5.12a can be expressed in a matrix form by defining a shape functions matrix.
N 0 N2 0 N3 0
N 1 (5.13)
0 N1 0 N2 0 N3
and
u=Nq 5.14
for the triangular element, the coordinates x, y can also be represented interms of nodal
coordinates using the same shape functions. This isoparametric representation.
Element Stiffness
We now substitute for the strain from the element strain – displacement relationship in Eq.
into the element strain energy Ue in Eq.,b to obtain
1
Ue
2 e
T D tdA (5.29a)
1
qTBTDBqt A
2 e
Taking the element thickness te as constant over the element and remembering that all
terms in the D and B matrices are constants, we have
Ue
1
2 e
qTBTDBt e dA q
e
5.29b
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
102
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 T
Ue q t e A eBTDBq 5.29c
2
or
1 T e
Ue qK q 5.29d
2
For plane stress or plane strain, the element stiffness matrix can be obtained by taking the
appropriate material property matrix D defined 1 and carrying out the previous multiplication on
the computer. We note that ke is symmetric since D is symmetric. The element connectivity as
established in Table is now used to add the element stiffness values in K e into the corresponding
global locations in the global stiffness matrix K, so that
1
U qTk e q 5.31
e 2
1
= QTKQ
2
A traction force is a distributed load acting on the surface of the body. Such a force acts on edges
connecting boundary nodes. A traction force acting on the edge of an element contributes to the
global load vector F. This contribution can be determined by considering the traction force term
e uT Ttdl.
5.46
T
x , y
and
T
x , y , x y 5.47
x y y x
Where is an arbitrary (virtual) displacement vector, consistent with the boundary conditions The
variation form is given by
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
103
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
e T tdA T ftdA T Ttdl iTPi =0(5.48)
A A
i
where the first term represents the internal virtual work. The expression in parentheses represents
the external virtual work.
Since strains are constant in a constant – strain triangle (CST) element, the corresponding
stresses are constant. The stress values need to be calculated for each element. Using the stress –
strain relations in Eq. and element strain-displacement relations in Eq. we have
=DBq.
If the distribution of the change in temperature T(x,y) is known, the strain due to this
change in temperature can be treated as an initial strain 0. From the theory of mechanics of solid,
0 can be represented by
0 T.T.0 5.61
T
The effect of temperature can be accounted for by considering the strain energy term. We have
1
U 0 D 0 tdA
T
2
T D 2 T D 0 0T D 0 tdA 5.64
1
2
The first term in the previous expansion gives the stiffness matrix derived earlier. The last
term is a constant, which has no effect on the minimization process. The middle term. Which
yields the temperature load, is now considered in detail. Using the strain displacement
relationship =Bq.
A
T
D 0 tdA qT BTD 0 t e A e 5.65
e
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
104
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
This step is directly obtained in the Galerkin approach where T will be T() and qT will be T.
It is convenient to designate the element temperature load as
e=teAeBTD0 (5.66)
e=[1, 2, 3, 4, 5, 6 ]T (5.67)
The vector 0 is the strain in Eq. 5.61 or 5.62 due to the average temperature change in the
element. e represents the element nodal load contributions that must be added to the global force
vector using the connectivity.
The stresses in an element are then obtained by using Eq. 5.63 in the form
=D(Bq-)
10. Consider the two-dimensional loaded plate showing Fig. E5.4. In addition to the conditions
defined in Example 5.5, there is an increase in temperature of the plate of 80 oF. The coefficient of
linear expansion of the material is 10-6 /oF. Determine the additional displacement due to
temperature Also, calculate the stresses in element 1.
T 5.6
0 T 10 5.6
4
0 0
Thickness <<<<..the area of the element A is 3 in 2. The element temperature loads are
=tA(DB)T0
Picking the forces for dofs ,3, and 4 <<the previous equations, we have
FT=[F1F3F4]=[16800 11206 16800]
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
105
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
On solving KQ = F, we get
1=(DB1)Tq1 -D0
Substituting for the terms on the right – hand side, we get
11. Write down the general equation of the stiffness matrix for 2-D element.
Line Element:
D is replaced by E
dV is replaced by dL
[B]=[-I/L I/L]
Plane Element
dV=tdA,
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
106
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
[k]=tA[B]T[D][B] (6.2.52)
12. Find the strain – nodal displacement matrices Be for the elements shown in fig. E5.3. Use local
numbers given at the corners.
Solution We have
y 23 0 y 31 0 y12 0
1
B
1
0 x 32 0 x13 0 x 21
det J
x 32 y 23 x13 y 31 x 21 y12
2 0 0 0 2 0
1
= 0 0 0
6
3 2 0 0 2
where det J is obtain x13y23 – x23y13 = (3) (2) - (3)(0)=6. Using the local numbers at the corners., B2 It
can be written using the relationship as
2 0 0 0 2 0
B 0 3 0 3 0 0
2 1
6
3 2 3 0 0 2
13. A CST elements is shown in Fig. E5.5. The element is subjected to a body force f x = x2 N/m3.
Determine the nodal force vector fe. Take element thickness =1m.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
107
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The work potential is - e fTudV, where fT=[fx, 0]. Substituting for u=Nq. We obtain the work
potential in the form =qTfe. where fe = e NTtdV, where N is given in Eq. 5.13. All y components of fe
are zero. The x components at nodes 1,2,3 are given, respectively, by
We now make the following substitutions: f=x=x2,x=x1+x2+(1--)x3=4, dV= det J dd, det J =
2Ae, and Ae =6. Now, integration over a triangle is illustrated in Fig. 5.6. Thus,
1 1
f dV
x 162 12 dd 3.2N
e 0 0
Fe=[3.2,0,9.6,0,3.2,0]TN
The modified equations are solved using Gaussian elimination and the strains may be computed
as
B
u
v
CB
u
v
A portion is assigned to each node and load on that region as the nodal load. This method
is called as lumped load method.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
108
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
PART - B
1. A) Explain in Detail
i) Constant –strain triangle (CST)
The displacements at points inside an element need to be represented in terms of the nodal
displacements of the element. As discussed earlier. The finite element method uses the concept of
shape functions in systematically developing these interpolations. For the constant strain triangle,
the shape functions are linear over the element. The three shape functions, N1,N2 and N3
corresponding to nodes 1,2, and 3, respectively, are shown in fig. 5.4 shape function N1 is 1 at node
1 and linearity reduces to 0 at nodes 2 and 3, respectively, and dropping to 0 at the opposite edges.
Any linear combination of these shape functions also represents a plane surface. In particular,
N1+N2 +N3 represents a plane at a height of at nodes ,2,and 3, and, thus, it is parallel to the
triangle 123. Consequently for every N1,N2, and N3.
N1+N2+N3= (5.9)
N1, N2, and N3 are therefore not linearity independent; only two of these are independent. The
independent shape functions are conveniently represented by the pair ,as
Where, are natural coordinates (Fig.) At this stage, the similarity with the one dimensional
element should be noted: in the one – dimensional problem the x-coordinates were mapped onto
the coordinates, and shape functions were defined as functions of . Here, in the two-
dimensional problem, the x-,y-coordinates are mapped onto the -,-coordinates, and shape
functions are defined as functions of and .
The shape functions can be physically represented by area coordinates. A point (x,y) in a triangle
divides it into three areas, A,A2, and A3, as shown in Fig. The shape functions N1, N2, and N3 are
precisely represented by
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
109
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
A1 A2 A
N1 N2 N3 3 5.11
A A A
where A is the area of the element. Clearly, N1+N2+N3=1 at very point inside the triangle.
The displacements inside the element are now written using the shape functions and the
nodal values of the unknown displacement field. We have
The relations 5.12a can be expressed in a matrix form by defining a shape function matrix
N 0 N2 0 N3 0
N 1 5.13
0 N1 0 N2 0 N3
and
u=Nq (5.14)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
110
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
For the triangular element, the coordinates x,y can also be represented in terms of nodal
coordinates using the same shape functions. This is isoparametric representation. This approach
lends to simplicity of development and retains the uniformly with other complex elements. We
have
X=(x1-x3)+( x2-x3)+x3
Y=(y1-y3)+( y2-y3)+y3 (5.15b)
Using the notation, xij =xi and yij=yi-yi, we can write Eq. 5.15b as
X=x13+x23+x3
Y=y13+y23+y3 (5.15c)
This equation relates x- and y- coordinates to the - and - coordinates. Equation 5.12
express u and v as functions of and .
b) Evaluate the shape functions N1, N2 and N3 at the interior point P for the triangular element.
3.85=1.5N1+7N2+4N3=2.5+3+4
4.8=2N1+3.5N2+7N3=5-3.5+7
2.5-3=0.15
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
111
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
5+3.5=2.2
Solving the equations, we obtain =0.3 and =0.2, which implies that
In evaluating the strains, partial derivatives of u and v are to be taken with respect to x and
y. From Eqs. 5.12 and 5.15, we see that u, v and x,y are functions of and . That is, u=u(x( ,),y(
,)) and similarly v=v (x( ,),y( ,)). Using the chain rule for partial derivatives of u, we have
u u x u y
x y
u u x u y
x y
u x y u
x
(5.16)
u x y u
y
where the (22) square matrix is denoted as the Jacobian of the transformation. J:
x y
J (5.17)
x y
Some additional properties of the Jacobian are given in the appendix. On taking he derivative of x
and y.
x y13
J 13
y 23
(5.18)
x 23
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
112
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
u u
x
1
J (5.19)
u u
y
1 y 23 y13
J1 x13
(5.20)
det J x 23
detJ=x13 y 23 x 23 y13 (5.21)
From the knowledge of the area of the triangle, it can be seen that the magnitude of det J is
twice the area of the triangle. If the points 1,2, and 3 are ordered in a counter clockwise manner,
det J is positive in sign. We have
1
A | det J | (5.22)
2
where | | represents the magnitude. Most computer codes use a counter clockwise order
for the nodes and use det J for evaluating the area.
2. Determine the Jacobian of the transformation J for the triangular element shown in fig. E5.1.
Solution: We have
Thus, detJ =23.75 units. This is twice the area of the triangle. If 1,2,3 are in a clockwise order,
then det J will be negative.
u u u
y 23 y13
x 1
(5.23a)
u det J x u x13
u
y 23
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
113
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
v v v
y 23 y13
x 1
(5.23b)
v det J x v v
x13
y 23
Using the strain-dis-placement relations (5.5) and Eqs. 5.12b and 5.23, we get
u
x
v
y
u v
y x
y 23 q1 q5 y13 q3 q5
1
x 23 q2 q6 x13 q4 q6 (5.24a)
det J
x
23 1 5 q q x13 q3 q5 y 23 q2 q6 y 13 q 4 q
6
From the definition of xij and yij we can write y31 =y13 and y12 =y13 –y23, and so on. The foregoing
equation can be written in the form
y 23 q1 y 31q3 y12q5
1
x32q2 x13 q4 x 21q6 (5.24b)
det J
x32q1 y 23 q2 x13q3 y 31q4 x 21q5 y12q6
where B is a (36) element strain – displacement matrix relating the three strains to the six
nodal displacements and is given by
y 23 0 y13 0 y12 0
1
B 0 x 32 0 x13 0 x 21 (5.26)
det J
x32 y 23 x13 y 31 x 21 y12
It may be noted that all the elements of the B matrix are constants expressed in terms of the nodal
coordinates.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
114
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
=[x,y]T (5.46)
and
x , y , x , y (5.47)
x y y x
where is an arbitrary (virtual) displacement vector, consistent with the boundary conditions. The
variational form is given by
T tdA T ftdA T Ttdl iTPi 0 5.48
A
A A
i
where the first term represents the internal virtual work. The expression in parentheses represents
the external virtual work. On the discretized region, the previous equation becomes.
e
e
T
D tdA
e
e
T
ftdA T Ttdl iTPi 0
L
i
5.49
where
Represents the arbitrary nodal displacements of element e. The global nodal displacement
variations are represented by
T
D tdA qTBTDBtdA
e e
Noting that all terms of B and D are constant and denoting t e and Ae as thickness and area of
element, respectively, we find that
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
115
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
T
D tdA qTBTDBt e dA
e e
=q t e A eB DB
T T
(5.54)
=qTK e
Ke=teAeBTDB (5.55)
The material property matrix D is symmetric, and, hence, the element stiffness matrix is
also symmetric. The element connectivity as presented in table 5.1 is used in adding the stiffness
values of ke to the global locations. Thus,
e
e
T
D tdA qTk e TK eq
e e
= KQT
5.56
The global stiffness matrix K is symmetric and banded. The treatment of external virtual
work terms follows the steps involved in the treatment of force terms in the potential energy
formulation, where u is replaced by . Thus,
ftdA
T T e
f (5.57)
T
which follows from Eq. 5.33, with fe given by Eq. 5.36. Similarly, the traction and point load
treatment follows from Eqs. 5.38 and 5.43. The terms in the variation form are given by
The stiffness and force matrices are modified to use the full size (all degrees of freedom). From the
Galerkin form (Eq. 5.49). the arbitrariness of gives
KQ=F (5.59)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
116
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
4. For the two – dimensional loaded plate shown in fig. E5.6, determine the displacement of
nodes 1 and 2 and the element stresses using plane stress conditions. Body force may be
neglected in comparison with the external forces.
Solution: For plane stress conditions, the material property matrix is given by
1 v 0 3.2 107 0.8 107 0
E
D 2
v 1 0 0.8 107 3.2 10 7
0
1 v
1 v 0 0 7
1.2 10
0 0
2
Using the local numbering pattern used in fig. E5.3, we establish the connectivity as follows:
Nodes
Element No.
1 2 3
1 1 2 4
2 3 4 2
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
117
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
These two relationships will be used later in calculating stresses using e=DBeq. The multiplication
teAeBeT DBe gives the element stiffness matrices.
1 2 3 4 7 8 Global of
0.983 0.5 0.45 0.2 0.533 0.3
1.4 0.3 1.2 0.2 0.2
0.45 0 0 0.3
k1 107
1.2 0.2 0
Symmetric 0.533 0
0.2
5 6 7 8 3 4 Global of
0.983 0.5 0.45 0.2 0.533 0.3
1.4 0.3 1.2 0.2 0.2
0.45 0 0 0.3
k 2 107
1.2 0.2 0
Symmetric 0.533 0
0.2
In the previous element matrices, the global dof association is shown on top. In the problem
under consideration, Q2,Q5, Q6, Q7, and Q8, are all zero. It is now sufficient to consider the stiffness
associated with the degrees of freedom Q1,Q3 and Q4. Since the body forces are neglected, the first
vector has the component f4 =-1000Ib. the set of equations is given by the matrix representation
Q1=10-5[1.913,0,0.875,-7436,0,0]T
Similarly,
Q2=10-5[0,0,0,0,0.875,-7.436]T
2= [93.4,23.4-297.4]Tpsi
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
118
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Turner was first to suggest it and that was real starting point of FEM. Consider the typical
element shown in fig. It is subjected to constant stresses along its all the three edges. Let the
constant stresses be x, y, xy, =yx . Assembling stiffness matrix means finding nodal equivalent set
of forces which are statically equivalent to the constant stress field acting at the edges of the
elements.
The equivalent nodal forces to be found are F1, F2, F3,<F6 as shown in Fig. we have six
unknown nodal forces, but only three equilibrium. Hence it is not possible to determine F1, F2 <..
F6 in terms of x,y,xy mathematically. Turner resolved the uniform stress distribution into an
equivalent force system at midsides as shown in Fig. Note side I is the side opposite to node. With
this notation,
Fm1x x y3 y2 t xy x2 x3 t
Fm1y y x3 x 2 t xy y 3 y 2 t
Fm2x x y3 y1 t xy x 3 x1 t
Fm2y y x3 x1 t xy y 3 y1 t ...(7.1)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
119
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Fm3x x y 2 y1 t xy x 2 x1 t
Fm3y y x 2 x1 t xy y 2 y1 t
After this Turner transferred half of mid side forces to nodes at the end of sides to get equivalent
nodal forces. Thus he got
1
F1 Fm2x Fm3x
2
1
= x y 3 y1 t xy x 3 x1 t x y 2 y1 t xy x 2 x1 t
2
1
= x y 3 y1 y 2 y1 xy x 3 x1 x 2 x1
2
1
= x y 2 y 3 xy x 3 x 2
2
1
F2 Fm1x Fm3x
2
1
= x y 3 y 2 t xy x 2 x 3 x y 2 y1 xy x 2 x1
2
1
= x y 3 y1 xy x1 x 3
2
1
F3 Fm1x Fm2x
2
1
= x y 3 y 2 xy x 2 x 3 x y 3 y1 xy x 3 x1
2
1
= x y1 y 2 xy x 2 x1
2
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
120
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1
F4 Fm2x Fm3x
2
1
= y x 3 x1 xy y 3 y1 y x 2 x1 xy y 2 y1
2
1
= y x 3 x 2 xy y 2 y 3
2
F5
1
2
Fm1y Fm3y
1
= y x 2 x 3 xy y 3 y 2 y x 2 x1 xy y 2 y1
2
1
= y x1 x 3 xy y 3 y1
2
F6
1
2
Fm1y Fm2y
1
= y x 2 x 3 xy y 3 y 2 y x 3 x1 xy y 3 y1
2
1
= y x1 x 3 xy y 3 y1
2
b1 0 c1
b 0 c 2
2
b 0 c3
2A B
T
But, 3 ....(7.3)
0 c1 b1
0 c 2 b2
0 c 3 b3
F k e
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
121
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
In the stiffness equation [k] []=[F] refers to the nodal forces. Generally, while subdividing a
structure, nodal locations are selected to as to coincide with the external forces applied. This can
be easily done in case of concentrated loads acting on the structure. But in case of distributed loads
like self weight, uniformly distributed load. Uniformly varying load, we need a technique of
transferring the load as nodal load. There are two procedures for it. Namely direct procedure and
variational approach.
This procedure was first to be used in the finite element method. In this procedure classical
structural analysis background is utilized or a portion is assigned to each node and load on that
region as the nodal load. The latter method is called as lumped load method.
Consider the self weight of the uniform bar element shown in fig. Half the bar length may be
Assumed to contribute to each node. Hence at each node vertical downward load is
Al
2
where is unit weight of the material. A the cross sectional area and l is the length of the element.
In case of beam element subject to uniformly distributed load.
(i) Lumped load procedure: half the region is assigned to each node as shown in fig. (b) Its
wl
equivalent is taken at the centre of gravity of the element as shown in fig. (c). Then at the node,
2
wl wl2
the force are and . Hence Lumped load vector is
2 8
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
122
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
wl wl2 wl wl2
F
T
2 8 2 8
(ii) Classical Structural Analysis Approach: In case of beam the end reactions, for a fixed beam
wl wl2
and
2 12
wl wl2 wl wl2
2 12 2 12
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
123
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1
(iii) In case of a CST element area may be assigned to each node and hence equivalent
3rd
1
nodal force the self weight as shown in fig.
3rd
For complex loading and elements this method may not be of much use. The distribution obtained
in lumped load approach may be one of the possible distribution.
7. Derive a local stiffness matrix for heat transfer using shape functions for a four –mode
quadrilateral element.
The variational function is written in matrix format. As in Eq. (f) of Prob. 3,4 and will be
minimized with respect to the unknown variable that in this case is {T}. It follows that
The first three matrices on the left-hand side of Eq. (a) are multiplied together to give a 44
local stiffness matrix. The local stiffness matrix is defined as
The derivatives of the shape functions are required before an attempt is made to evaluate
the integral and are listed below for ready reference. Refer to Eq. (e) of prob3.1.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
124
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
N1 b y N1 b x
x ab y ab
N2 b y N2 x
x ab y ab
N3 y N3 x
(d)
x ab y ab
N4 y N4 a x
x ab y ab
The proper derivatives are substituted into Eq. © to <<<<<<<<<<..in the stiffness matrix
t
2
v v k y a x k y dxdy
a b
K11
2 2
0 0 ab 2
K11
b k 2
x
a 2k y t
3ab
The remaining terms in the stiffness matrix are evaluated in a similar manner. The local
stiffness matrix appears as follows and the reader may verify the results as an exercise:
The local stiffness matrix is symmetric, and some terms are repeated in the matrix. For
instance, all diagonal terms are the same. The term on the right-hand side of Eq. (a) represents the
distribution of the heat – source term.
a x b y ab / 4
ab / 4
a b t x b y
0 0 ab Qdxdy tQ (f)
xy ab / 4
y a x ab / 4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
125
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Unlike Spring and Beam elements. There is no deflection equation available for CST
element.
The displacement equation is derived by assuming an equation and then boundary
conditions are applied to solve the equation.
The displacement function is assumed to be a linear equation given by:
U(x,y)=a1+a2x+a3y
V(x,y)=a4+a5x+a6y (6.2.2)
Ui=u(xi,yi)=a1+a2xi+a3yi
Uj=u(xj,yj)=a1+a2xj+a3yj
Um=u(xm,ym)=a1+a2xm+a3ym
vi=v(xi,yi)=a4+a5xi+a6yi
vj=v(xj,yj)=a4+a5xj+a6yj
vm=v(xm,ym)=a4+a5xm+a6ym
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
126
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
ui 1 xi yi a1
u j 1 x j y j a2
u 1 x ym a3
m m
and
v i 1 xi y i a 4
v j 1 x j y j a5
v 1 x ym a6
m m
{a}=[x]-1{u}
i j m
x
1
1/(2A) i j m
i j m
1 xi yi
where 2A= 1 x j yj
1 xm ym
The values of ,,I are found using the given nodal coordinates (x,y).
Now,. The coefficient values can be found in terms of the nodal coordinates and the boundary
conditions.
a1 i j m ui
a 1/(2A) j
m u j
2 i (6.2.11)
a3
i j m um
and
a 4 i j m v i
a 1/(2A) j
m v j
5 i (6.2.12)
a6
i j m v m
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
127
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
i j m ui
u 1/(2A)1 x y i j m u j (6.2.14)
i j m um
and similarly,
i j m v i
v 1/(2A)1 x y i j m v j
i j m v m
9. Explain the general procedure when CST elements are in the usage.
Since plane stress and plane strain problems are two dimensional problems, we need two
dimensional elements. Any one from the family or triangular elements (CST/LST/QST) are ideally
suited for these problems. Any one element from the family of two dimensional isoparametric
elements also may be used. In these elements there are two degree of freedom at each node i.e. the
displacement in x direction and displacement in y direction. Hence total degree of freedom in
For a CST element shown in Fig. the displacement vector may be taken as
or as
In most of the programs the order shown in equation (b) is selected. Hence the
displacement vector {} is used in the form of equation (b) Then the x and y displacements of the
node in global system are referred as 2n-1th and 2nth displacements.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
128
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Step2: Discritization
Discritization of the structure should be made keeping in mind all the points listed. For all
nodes x and y coordinates are to be supplied/ generated. Then nodal connectivity details is to be
supplied. For the dam analysis problem shown in fig. the nodal connectivity detail is of the form
shown in Table.
Table Nodal connectivity
a1 b1x c1y a b2 x c 2 y a b3 x c 3 y
N1 ,N2 2 andN3 3
2A 2A 2A
where a1 x 2 y 3 x 3 y 2 a2 x 3 y1 x1y 3 a3 x1y 2 x 2 y1
b1 y 2 y 3 b 2 y 3 y1 b3 y1 y 2
c1 x 3 x 2 c 2 x1 x 3 c 3 x 2 x1
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
129
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 x1 y1
and 2A= 1 x 2 y2
1 x3 y3
ux,y
N1 0 N2 0 N3 0
ux,y e (12.3)
v x,y
0 N1 0 N2 0 N3
u
x x
v
y
y
2 u v
y x
and nodal displacement vector is in the form 12.3, the strain displacement vector ({}=[B]{}),[B] is
given by
b1 0 b2 0 b3 0
1
[B] 0 c1 0 c2 0 c 3 (12.4)
2A
c1 0 c 2 0 c3 0
[k]e=[B]T[D][B]v (12.5)
where V=At
This is exactly same as equation which was obtained by Turner by the direct approach. In
equation [D] is the elasticity matrix, In case of isotropic materials, for plane stress case,
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
130
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 0
E
[D] 2
1 0 (12.6)
1
1
0 0
2
1 0
E
[D] 1 0 (12.7)
1 1 2 1 2
0 0
2
Consistent Loads
If there are nodal forces they are to be added directly to the vector {F}e
Using nodal connectivity details the exact position of every term of stiffness matrix and
nodal vector must be identified and placed in global stiffness matrix.
Since in most of the problems in plane stress and plane strain degree of freedom is quite
high, the computers are to be used. These problems are not suitable for hand calculations. When
computer programs are to be developed, imposition of boundary condition is conveniently done
by penalty method.
After getting nodal displacements stresses and strains in each element is assembled using the
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
131
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
relations
Be
and DBe
The calculated value of stress for an element is constant. It is assumed to represent the
value at the centroid of the element. As a designer is normally interested in the principal stresses,
for each element these values also may be calculated.
10. Find the nodal displacements and element stresses in the propped beam shown in fig.
Idealize the beam into two CST elements as shown in the figure. Assume plane stress
condition. Take = 0.25, E= 210=5 N/mm2, Thickness = 15mm.
Solution: For element (1), global nodal numbers are 1,3,4. Local numbers 1,2,3 selected are
indicated in Fig. Selecting node 4 as the origin of global coordinate system.
1(0,0),2(750,500)and 3(0,500)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
132
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 0 0
2A 1 750 500 750 500 0 750 500
1 0 500
0 0 500 0 500 0
[B]
1 0 750 0 0 0 750
750 750
750 0 0 500 750 500
0 0 1 0 1 0
1
= 0 15 0 0 0 15
750
15 0 0 1 15 1
1 0
E
[D] 1
0
1 1 2
1 2
0 0
2
0.75 0.25 0 3 1 0
2 105
0.25 0.75 0 0.2 10 1 3 0
5
1.25 0.5
0 0 0.25 0 0 1
3 1 0 0 0 1 0 1 0
5 15 0 0 0 15
1
[D][B] 0.2 10 1 3 0 0
75
0 0 1 15 0 0 1 15 1
0 15 3 0 3 15
2 105
0 45 1 0 1 45
750
15 0 0 1 15 1
0 0 15
0 15 0
0 15 3 0 3 15
15 750 500 1 1 0 0 0.2 105 0 45 1 0 1 45
2 750 0 0 1 750
15 0 0 1 15 1
1 0 15
0 15 1
u1 v1 u3 v3 u4 v 4 Global
225 0 0 15 2.25 150 u1
0 6.75 15 0 15 6.75 v1
0 15 3.0 0 3.0 15 u3
100000
15 0 0 1 15 1.0 v 3
225 15 3 15 525 3.0 u4
15 6.75 15 1 3.0 7.75 v 4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
133
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1(0,0),2(750,0)3(750,500)
b1 y 2 y 3 =-500 b 2 y 3 y1=-500 b 3 y1 y 2 0
c1 x 3 x 2 =0 c 2 x1 x 3 =-750 c 3 x 2 x1 750
1 0 0
2A 1 750 0 1 750 500 750 500
1 750 500
500 0 500 0 0 0
[B]
1 0 0 0 750 0 750
750 500
0 500 750 500 750 0
1.0 0 1.0 0 0 0
1
= 0 0 0 750 0 750
750
0 500 750 500 750 0
3 1 0
[D] 0.2 10 1 3 0 same as for element 1.
5
0 0 1
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
134
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
3 1 0 1.0 0 1.0 0 0 0
5 1.5 0 15
1
[D][B] 0.2 10 1 3 0 0 0 0
750
0 0 1 0 1.0 0 1.5 1.0 1.5 0
10 0 0
0 0 1.0
3.0 0 3.0 1.5 0 1.5
750 500 1 1.0 0 1.5 0.2 105 1.0 0
[k]2 15 1.0 4.5 0 4.5
2 750 0 1.5 1.0 750
0 1.0 1.5 1.0 1.5 0
1 0 1.5
0 1.5 0
u1 v1 u2 v2 u3 v3
3.0 0 3.0 1.5 0 1.5 u1
0 1.0 1.5 1.0 1.5 0 v1
3.0 1.5 5.25 3.0 2.25 1.5 u2
10000
1.5 1.0 3.0 7.75 1.5 6.75 v 2
0 1.5 2.25 1.1 2.25 0 u3
1.5 0 1.5 6.75 0 6.25 v 3
[k]=100 000
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
135
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
{F}T=[0 0 0 0 50000 0 0 0]
The equation is
[k][]={F}
1 2 4 7 8 0
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
136
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
6 0.010459
5 0.118236
3 0.053661
1 DB
0
0
0 1.5 3 0 3 1.5 5.584
2 105 0.053661
0 4.5 1 0 1 4.5 2.977
750 0
1.5 0 1 1.5 1.0 5.000
0.118236
0
0.010459
0
0
3.0 0 3.0 1.5 0 1.5 9.877
2 105 0.118236
2 1.0 0 1.0 4.5 0 4.5 4.408
750 0.010459
0 1.0 1.5 1.0 1.5 0 5.008
0
0
11. (i) Derive the expression for consistent load vector due to self weight in a CST element.
Solution: The general expression for the consistent load in any element due to the body force is
0
For self weight x b
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
137
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
L 0 L 2 0 L3 0
N 01 L1 0 L2 0 L3
u
When nodal vector selected is in the order 1
v1
L1 0 0
0 L L
1 1
L 0 0 0
[F]e 2 hdA hdA
A
0 L 2 L
A 2
L3 0 0
0 L3 L3
p!q!r!
L L L dA p q r 2 2A
p q r
1 2 3
A
1!0!0! 1 hA
we get L hdA 1 0 0 2 h2A 2 3 2A
A
1
3
hA
Similarly L 2dA
A
3
hA
and L3dA
A
3
0
1
hA 0
Fe Answer
3 1
0
1
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
138
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
11. (ii) Find the expression for nodal vector in a CST element subject to pressure Px1, Py1 on side
1. Px2, Py2 on side 2 and Px3, Py3 on side 3 as shown in fig.
Solution: Let us first consider nodal vector due to pressure Pxt and Pyt only.
[N]=[L]
Along side 1-L2
i.e., L3=1-L2
P
xs Px1
y1
Hence the line integral form exists for nodal force vector as given below:
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
139
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
0 0 0
0 0 0
1 L P
t
L 2 0 Px1
t ds t L P ds
2 x1
o 2 y1
o
0 L 2 Py1
L 3 0 L P
3 x1
0 L3 L3Py1
Nothing that the standard integration form for natural coordinate system is
p!q!
L L ds p q 1! l
p q
1 2
1!0! l
we get L 2x1ds l1Px1 1 Px1
1 0 1! 2
l1
L P2 y1 ds
2
Py1
0!1! l
L3Px1ds 0 1 1 l1Px1 21 Py1
l
and L P
3 x1ds 1 Py1
2
tl1
Fe 0 0 Px1 Py1 Px1 Py1
T
2
l P l3 Px3
Fe l1 Px1 l3 Py3
1 y1
l P l2 Px2
1 x1
l1 Py1 l2 Py2
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
140
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
12. A three-node triangular element is defined in fig. in an x,y coordinate system with nodes
1,2, and 3 located at (x1,y1), (x2,y2), and (x3,y3), respectively, in the global system. Derive shape
functions in terms of global coordinates.
=C1+C2x+C3y (a)
C1
1 x y C2 or in matrix format =[]{c} (b)
C
3
Substitute Eqs. © into Eq. (a) to obtain three equations that can be solved for C 1,C2, and C3:
1 C1 C2 x1 C3 y1 1 1 x1 y1 C1
2 C1 C2 x 2 C3 y 2 or 2 1 x 2 y 2 C2 (d)
3 C1 C2 x3 C3 y3 1 x y3 C3
3 3
{}=[x]{c} (e)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
141
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
{c}[x]-1{} (f)
The shape functions are the product of the first two matrics on the right- hand side of Eq. (g)
N1 x 2 y 3 x 3 y 2 x y 2 y 3 y x 3 x 2 / 2A
N2 x 3 y1 x1y 3 x y 3 y1 y x1 x 3 / 2A
N3 x1y 2 x 2 y1 x y1 y 2 y x 2 x1 / 2A (i)
1 x1 y1
A= det 1 x 2 y 2 (j)
1
where
2
1 x 3 y 3
13. Derive a local stiffness matrix for heat transfer using the three-node triangular element
defined in Fig.
T
[k] [B] [B]t dx dy (a)
The temperature is defined in terms of the unknown temperature at each node and is
similar to Eq.(b) of Prob. Except there are only three nodal values:
The shape function given by Eqs. (i) of Prob. And, for convenience, are usually written in the
following form
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
142
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
a1 b1x c1y
N1
2A
a b2 x c 2 y
N2 2 (c)
2A
a b3 x c 3 y
N3 3
2A
a1 x 2 y3 x3 y 2 b1 y 2 y 3 c1 x 3 x 2
where a2 x3 y1 x1y3 b2 y3 y1 c 2 x1 x 3
a3 x1y 2 x 2 y1 b3 y1 y 2 c 3 x 2 x1
Define the shape function matrix of Eq. (b), using Eqs. (c) as
a1 a2 a3
[N] [1 x y] b1 b2 b3 2A
c1 c 2 c 3
The [B] matrix is formed as [L] [N] where the operator matrix [L] is given by Eq. (e) of Prob.
For the heat-transfer problem or any physical problem governed by an equation of the form of Eq.
when = 0:
a1 a2 a3
/ x 0 1 0
L N / y N 0 0 1 b1 b2 b3 2A
c c c
1 2 3
b1 b2 b3
or [B] 2A
c1 c 2 c 3
The material matrix [k] of Eq. (a) is a 2 x 2 matrix and is the same as that in Prob. The area
integration is dx dy = A, the area of the triangle, since all terms are constant. Substituting into
Eq. (a) gives the symmetric matrix
Equation (d) can be used for numerical computation, however, for computer
implementation it is probably more practical to formulate the individual matrices and use a matrix
multiplication routine to compute [k].
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
143
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
14. Derive the shape functions for the nine-node rectangular element.
Figure.
The nine-node element has midside nodes located at the midpoint of each side of the
element, and the ninth node located at the center of the element. The node numbering shown is
standard, with midside nodes numbered 5 through 8, but the nodes can be numbered in any
sequence. In this case the dimensions of the element (the local coordinate system) are assumed as
a and b. Other dimensions, such as 2a x 2b could also be used. A local coordinate system similar
to –a and –b will be introduced.
The shape function can be derived using the results of Prob. And fig. The global x
coordinates of Fig. are reproduced in Fig. along with corresponding y coordinates. The results of
Prob. Are repeated for reference and given an extra subscript to indicate the x direction in the
global system:
In the local system let node 1 correspond to x1 = 0 a/2, and x3 = a. The shape function , in the
x direction only, along nodes1,5 and 2 is obtained by substituting into Eq.(a):
N1x
x (a / 2)(x a) (2x a)(x a) (b)
0 (a / 2)(0 a) a2
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
144
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Similarly, N5x for the nine-node element can be constructed using N5x of Eqs.(a):
(x 0)(x a) x(x a)
N5x (c)
(a / 2) 0(a / 2) a (a2 / 4)
Visualize that the shape function of Prob. Could have been derived along the y axis of the
global system using y1,y2, and y3 of Fig. and the result would be Eqs.(a) with all x value replaced
by y values. A shape function for node 1 of the nine-node element would be identical to Eq.(b)
with x and a replaced by y and b, respectively. The two-dimensional shape function for node 1 is
the product of the two one-dimensional shape functions for node 1:
substituting local coordinates corresponding to node 1(x = 0, y = 0) into Ew.(d) will give N1
= 1, and the value of N1 all other nodes will be zero.
The y contribution of the two-dimensional shape function for node 5 is constructed using
the first Eq.(a) and is the same as that used in Eq.(d) or, corresponding to Fig. is written as
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
145
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
15. Derive: Two dimensional shape function for a four –model quadrilateral element.
Consider the general quadrilateral element shown in Fig. The local nodes are numbered as
1,2,3, and 4 in a counter clock wise fashion as shown, and (xi,yi) are the coordinates of node i. The
vector q=[q1, q2,<q8]T Denotes the element displacement vector. The displacement of an interior
point P located at (x,y) is represented as u= [u(x,y),v(x,y)]T.
Shape Functions
We first develop the shape functions on a master element. Shown in fig. the master element
is defined in -,-coordinates (or natural coordinates) and is square shaped. The Lagrange shape
functions where i=1,2,3. and 4, are defined such that N1 is equal to unity at node I and is zero at
other nodes. In particular, consider the definition of Ni:
N1 =1 at node 1
=0 at nodes 2,3, and 4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
146
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Now, the requirement that N1 =0 at nodes 2,3, and 4 is equivalent to requiring that N1=0 along
edges =+and =+1 Thus, N1 has to be of the form
N1=c(1-)(1-) (7.2)
1
N1 1 (1 ) (7.4)
4
1
N1 1 (1 )
4
1
N2 1 (1 )
4
1
N3 1 (1 )
4
1
N4 1 (1 ) (7.5)
4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
147
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1
N1 1 1 (1 1 ) (7.6)
4
We now express the displacement field within the element in terms of the nodal values.
Thus, if u=[u,v]T represents the displacement components of a point located at (, ), and q,
dimension (8x1), is the element displacement vector, then
u N1q1 N2q3 N3 q5 N4 q7
v N1q2 N2q4 N3 q6 N4 q8 (7.7a)
u=Nq (7.7b)
where
N 0 N2 0 N3 0 N4 0
N 1 (7.8)
0 N1 0 N1 0 N3 0 N4
In the isoparametric formulation, we use the same shape functions N 1 to also express the
coordinates of a point within the element in terms of nodal coordinates. Thus,
x N1x1 N2 x 2 N3 x 3 N4 x 4
y N1y1 N2 y 2 N3 y 3 N4 y 4 (7.9)
f f x f y
x y
f f x f y
x y
or
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
148
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
f f
x
J (7.11)
f f
y
x y
J (7.12)
x y
1 1 x1 1 x 2 1 x 3 1 x 4 1 y1 1 y 2 1 y 3 1 y 4
J
4 1 x1 1 x 2 1 x 3 1 x 4 1 y1 1 y 2 1 y 3 1 y 4
J J
11 12
21 22
J J
f f
x
J (7.14a)
f f
y
or
f f
x 1 J22 J12
J (7.14b)
f det J 21 J11 f
y
These expression will be used in the derivation of the element stiffness matrix. An
additional result that will be needed is the relation.
Dxdy=detJ dd.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
149
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
UNIT – IV
PART – A
2 2 2
u Trdld u P
1
rdAd u frdAd
T T T T
i i
2 0 A 0 A o i
Where rdld is the elemental surface area and the point load Pi represents a line load distributed
around a circle.
2 u fr dA 2 (uf fz )r dA
T
e e
Once again, approximating the variable quantities by their values at the centroid of the triangle,
we get
2 uT fr dA = qT f e
e
2 rA e
fe [f r ,f z ,f r ,f z ,f r ,f z ,]T
3
The bar on the f terms indicates that they are evaluated at the centroid. Where body force is
the primary load, greater accuracy may be obtained by substituting r = N 1r1 + N1r2 + N1r3 into
equation and integrating to get nodal loads.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
150
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The two dimensional region defined by the revolving area is divided into triangular
elements. Though each element is completely represented by the area in the rz plane, in reality. It
is a ring shaped solid of revolution obtained by revolving the triangle about the z- axis.
u= Nq
u= [u, w] T
N 0 N2 0 N3 0
N 1
0 N1 0 N2 0 N3
U = q1 + q3 + (1 - - q6)
U = q2 + q4 + (1 - - q6)
r r
J 13 13
r23 r23
det J r13 z23 r23 z13
Let us consider a rotating flywheel with is axis in the z direction. We consider the flywheel
to be stationary and apply the equivalent radial centrifugal (inertial) force per unit volume of pr2,
where is the density (mass per unit volume), and the angular velocity in rad/s. In addition, if
gravity acts along the negative z-axis, then
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
151
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
f r r 2 ,f z g
For more precise results with coarse meshes, we need to use r = N1r1 + N1r2 + N1r3 and integrate.
For a uniformly distributed load with components Tr and Tz, shown in figure, on the edge
connecting nodes 1 and 2, we get
2 uT Trd qT T e
e
Where,
q = [q1, q2 q3, q4,] T
In this derivation, r is expressed as N1r1 + N1r2 and then integrated. When the line 1-2 is pa
rallel to the z-axis, we have r1 = r2, which gives a = b = 0.5r1.
2 T ( )r dA
A
q K e
T
T
K e 2 rA e B DB
We note that Ke is symmetric. Using the connectivity of the elements, the internal virtual work can
be expressed in the form
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
152
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
q K
T e T
K eq
Internal virtual work = e
= TKQ
Where K is the global stiffness matrix. The external virtual work terms in equation
involving body forces, surface tractions, and point loads can be treated in the same way as in the
potential energy approach, by replacing q with . The summation of all the force terms over the
elements then yields.
From the set of nodal displacements the elements nodal displacements q can be found
using the connectivity. Then, using stress strain relation in = d and strain displacement relation
in the equation = Bq we get DBq , where B is b and it is evaluated at the centroid of the
element.
On substitution into the strain energy, this yields an additional term of -TD0 in the
potential energy II. Using the element strain displacement relation in equation we find that
2 T D 0 r dA q (2 rA B D
T T
e 0
A e
The consideration of the temperature effect in the Galerkin approach is rather simple. The
term T in equation is replaced by T ().
The expression in parentheses gives elements nodal contributions. The vector 0 is the initial
strain evaluated at the centroid, representing the average temperature rise of the element. We
have
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
153
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
T
e 2 rA e B D0
e = [1, 2, 3, 4, 5, 6]T
The modeling of a cylinder of infinitive length subjected to external pressure is shown below.
In that the length dimensions are assumed to remain constant. This plane strain condition
is modeled by considering a unit length.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
154
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
PART – B
Considering the elemental volume shown figure, the potential energy can be written in the
form
1 2 2 2
T r dA d - uT fr dA d uTTr d d - uiT Pi (1)
2 0 A 0 A 0 L
i
where r d d is the elemental surface area and the point load Pi represents a line load
distributed around a circle, as shown in figure.
1
2 T r dA - uTfr dA uTTr d ui T Pi (2)
2 A A L
i
where
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
155
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
T
u w u w u
= , , , .....(6)
r z z r r
r , z ,Trz ,
T
....(7)
D ...(8)
Elemental volume
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
156
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Where the (4 4) matrix D can be written by dropping the appropriate terms from the three-
dimensional matrix, as
v v
1 1 v
0
1 v
v 1 0
v
E 1 v 1 v 1 v
D ...(9)
1 v (1 2v ) 0 0
1 2v
0
2(1 v )
v v
0 1
1 v 1 v
A
2 T ( )r dA- 2 T fr dA + 2
A Tr d +
L
T
i
T
Pi 0 ...(10)
where
r , z
T
.....(11)
T
( ) = r , z , r , z , r .....(12)
r z z r r
The two-dimensional region defined by the revolving area is divided into triangular
elements, as shown in figure. Though each element is completely represented by the area in the rz
plane, in reality, it is a ring-shaped slid of revolution obtained by revolving the triangle about the
z-axis. A typical element is shown in figure.
The definition of connectivity of elements and the nodal coordinates follow the steps
involved in the CST element discussed in Section. We note here that the r-and z-coordinates,
respectively, replace x and y.
U = Nq <<(1)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
157
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
N 0 N2 0 N3 0
N 1
N3
.....(2)
0 N1 0 N2 0
q = q1, q2 , q3 , q4 , q5 , q6 .....(3)
T
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
158
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
r r1 r2 1 r3
z z1 z2 1 z3 .....(5)
u u
r
J ....(6)
u u
z
and
w w
r
J ....(7)
w w
z
r z13
J 13
z23
.....(8)
r23
In the definition of J earlier, we have used the notation rij = ri – ri and zij = zi – zj.
The determinant of J is
Recall that |det J| = 2Ae. That is, the absolute value of the determinant of J equals twice the area of
the element. The inverse relations for Eqs. 6 and 7 are given by
u u w w
r
1 r 1
J and J .....(10)
u u w w
z z
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
159
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
where
1 z23 z13
J 1 z13
.....(11)
det J r23
Introducing these transformation relationships into the strain-displacement relations in Eq. (4), we
get
z23 q1 q5 z13 q3 q5
det J
r23 q2 q6 r13 q4 q6
det J
r23 q1 q5 r13 (q3 q5 ) z23 (q2 q6 ) z13 q4 q6
det J
N1q1 N2q3 N3 q5
r
Bq <(12)
3. (a)
An axisymmetric body with a linearly distributed load on the conical surface is shown in
figure. Determine the equivalent point loads at nodes 2, 4, and 6.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
160
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Solution: We approximate the linearly distributed load by the average uniformly distributed
loads on the edges 6-4 and 4-2 as shown in figure. Relationships for more precise modeling of a
linearly distributed load are provided in Problem. We now consider the two edges 6-4 and 4-2
separately and then merge them.
r1 r2 z1 z2 25mm
2 2
1-2
z 2 z1 r1 r2
c= 0.6, s= 0.8
1 2 1 2
Tr pc 0.21, Tz ps 0.28
2r1 r2 r 2r2
a= 26.67, b = 1 23.33
6 6
T
T1 2 1 2 aTr , aTz,bTr , bTz
r1 r2 z1 z2 25mm
2 2
1-2
z 2 z1 r1 r2
c= 0.6, s= 0.8
1 2 1 2
Tr pc 0.15, Tz ps 0.2
2r1 r2 r 2r2
a= 16.67, b = 1 13.33
6 6
T
T1 2 1 2 aTr , aTz,bTr , bTz
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
161
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
These loads add to F1, F8, F3, and F4, respectively. Thus,
[F3 F4 F7 F8 F11 F12]
Potential-Energy Approach
1
e 2
e
2 T D r dA 2 uT fr dA - 2
e
u Tr d
e
T
- uiT Pi ....(1)
The element strain energy Ue given by the first term can be written as
Ue
1 T
2
q 2 BT DBr dA q
e
.....(2)
k e 2 BT DBr dA ....(3)
e
The fourth row in B has terms of the type Ni/r. Further, this integral also has an additional r
in it. As a simple approximation, B and r can be evaluated at the centroid of the triangle and used
as representative values for the triangle. At the centroid of the triangle,
1
N1 = N2 = N3 = <.(4)
3
And
r1 r2 r3
r =
3
where r is the radius of the centroid. Denoting B as the element strain-displacement matrix B
evaluated at the centroid, we get
k e 2 r BTDB dA
e
or
k e 2 r A eBT DB ...(5)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
162
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
We note here that 2 r A e is the volume of the ring-shaped element shown in figure. Also, Ae is
given by
1
Ae det J ....(6)
2
4. In figure, a long cylinder of inside diameter 80 mm and outside diameter 120 mm snugly fits in
a hole over its full length. The cylinder is then subjected to an internal pressure of 2 MPa. Using
two elements on the 10-mm length shown, find the displacements at the inner radius.
Solution
We will use the units of millimeters for length, newtons for force, and megapascals for
stress and E. These units are consistent. On substituting E = 200 000 MPa and v = 0.3, we have
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
163
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
for both elements, det J = 200mm2 and Ae = 100mm2. forces F1 and F3 are given by
The B matrices relating element strains to nodal displacements are obtained first. For element 1,
1
r 40 40 60 46.67 mm and
3
0.05 0 0 0 0.05 0
0 0.1 0 0.1 0 0
B
1
0.1 0.05 0.1 0 0 0.05
0.0071 0 0.0071 0 0.0071 0
1
For element 2, r= 40 60 60 53.33mm and
3
0.05 0 0.05 0 0 0
0 0 0 0.1 0 0.1
B2
0 0.05 0.1 0.05 0.1 0
0.00625 0 0.00625 0 0.00625 0
The stiffness matrices are obtained by fixing 2 rAe BT DB for each element:
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
164
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Global of 1 2 3 4 7 8
4.03 -2.58 -2.34 1.45 -1.932 1.13
8.45 1.37 -7.89 1.93 -0.565
2.30 -0.24 0.16 -1.13
k1 107
7.89 -1.93 0
Symmetric 2.25 0
0.565
Global of 3 4 5 6 7 8
2.05 0 -2.22 1.69 -0.085 -1.69
0.645 1.29 -0.645 -1.29 0
5.11 -3.46 -2.42 2.17
k 2 107
9.66 1.05 -9.01
Symmetric 2.62 0.241
9.01
Using the elimination approach, on assembling the matrices with reference to the degrees of
freedom 1 and 3, we get
Q1 0.014 102 mm
Q3 0.0133 102 mm.
6. Derive the shape functions and the element Stiffness matrix for an axisymmetric annular
ring element as shown in figure.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
165
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1
U 1 2 r = < 1 r > .....(1)
2
U1 1 r1 1 r1
1
U1
or a
1 r2
U2 1 r2 U2
U / r r
U / r
0 1 U1
1 r2 r1
= 1 ....(4)
1 r2 r1 1 1
r U2
U1
= B
U2
K = B TC B dv ......(6)
r2
= 2 t BT C B r dr. ......(7)
r1
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
166
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
7. A solid element in the form of a right circular cone is under the action of self weight
along its axis. It is proposed to be analyzed by using single axisymmetric ring element of
triangular cross section.
(b) Derive the components of vectors of nodal forces. Assume the height, radius, weight density to
be equal to 1.
1
U 1 r z 0 0 0
: ....(1)
V 0 0 0 1 1 z
6
U1
1 0 0 0 0 0
:
1 1 0 0 0 0
U3 1 0 1 0
0 0
= ....(2)
V1 0 0 0 1 0 0
0 0 0 1 1 0
:
0 0 0 1 0 1
V3
1 0 0 0 0 0
1 1 0 0 0 0
1 0 1 0 0 0 U
...(3)
0 0 0 1 0 0 V
0 0 0 1 1 0
0 0 0 1 0 1
so,
U 1 r z r z 0 0 0 U
1 r z r z V
....(4)
V 0 0 0
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
167
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 0 0 0
0 1 1
0 U U
or B .....(5)
0 1 1 0 V V
1 0 0 0
1 1
z 0 r 0
Qx 0
1 1
y
Q
2
z 0 r 0
NT r dr dz ......(7)
pg
Qx
The integration of the elements of [K] and the consistent load vector are to be carried out
Qy
using Table.
I4 z dr dz
z i z j zk
A
3
kij k 5jk k ki5
5
where
zi z j zi 3r j ri
z dr dz kij5
I5 4 ri r j z j 3ri r j
r
1 ri z j r j zi
2
log ri / r j
2 r r 2
i j
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
168
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
I6
z2
dz dr
+ 2 zi z j 2.5 ri2 11ri r j 2.5r j2
r
+ zi2 11rj2 7ri r j 2ri 2
1 ri z j r j zi
3
r
- log i
3 r r 3
r j
i j
A = Area of Note: If ri = rj in an element some of the integrals become infinity. In such cases special
the triangle techniques must be used, if ri = rj for an element the logarithm term becomes very
ijk large and loss of accuracy may result. In such cases r i = rj = (ri + rj) /2.
If the loads are not axisymmetric but are symmetric about a plane containing the axis of
symmetry, such as static equivalent wind load shown in figure. There are three displacements
which can be expanded in Fourier series as
Pr Prn (r , z )cos n
P P n n, z cos n ...(1)
Pz Pzn r , z cos n
U un (r , z )cos n
V v n n, z cos n ...(2)
W w n r , z cos n
When the amplitude in the three displacements functions are assumed to be linear function in rz
plane within each element.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
169
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
When n = 0 all values are independent of 1 V = 0. The strain displacement relation is given by
U
rn cos n
r
W
z zn cos n
z
1 V U
= n cos n
r r
U V V
v r v r n sin n
r r r
V 1 W
v z v z n sin n .....(4)
z r
U W
v rz v rzn cos n
z r
0 1 0 0 0 0 0 0 0
1
rn 0 0 0 0 0 0 0 0
1
zn
z n nz
1 n 0 0 0 a1n
n r r r r
n nz 1 z : <(5)
v r n r n 0 0 0 0 a
v r r r
9n
z n n nz
0 0 0 0 0 1 n
v rzn r r
0 0 1 0 0 0 0 1 0
or
Hence
n Gn an and an A qn <<.(6)
1
Hence
n Gn A qn Bn qn
1
.....(7)
K B C B dv
T
.....(8)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
170
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
0
p cos m = a
0 n=0
n cos n cos m dv <(10)
cos m
0
cos n d =2 for m=n=0
2
1
a0
2 p( ) d
0
.....(12a)
2
1
an
p( )cos n
0
d .....(12b)
The boundary condition of the problem is u = 0 on . Solve the problem using finite element
method.
A problem possesses symmetry of the solution about a line only when there is symmetry of
(a) the geometry, (b) the material properties, (c) the source variation, and (d) the boundary
conditions about the line.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
171
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Finite element analysis of the Poisson equation in a square region: (a) geometry and
computational domain, and boundary conditions of the problem; (b) a coarse finite element mesh
of linear triangular elements; (c) a refined mesh of linear triangular elements.
The problem at hand has symmetry about the x = 0 and y = 0 axes; it is also symmetric
about the diagonal line x = y. We can exploit such symmetries in modeling the problem. Thus, we
can use a quadrant for meshes of rectangular elements and an octant for meshes of triangular
elements of the domain to analyze the problem. Of course, it is possible to mix triangular and
rectangular elements to represent the domain as well as the solution.
Solution by linear triangular elements. Owing to the symmetry along the diagonal x = y,
we model the triangular domain shown in figure(a). As a first choice, we use a uniform mesh of
four linear triangular elements to represent the domain (fig.(b)), and then a refined mesh (fig. (c))
to compare the solutions. In the present case, there is no discretization error involved in the
problem because the geometry is exactly represented.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
172
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The element matrix in (3a) is valid for the Laplace operator -2 on any right-angled triangle
with sides a and b in which the right-angle is at node 2, and the diagonal line of the triangle
connects node 3 to node 1. Note that the off-diagonal coefficient associated with the nodes on the
diagonal line is zero for a right-angled triangle. These observations can be used to write the
element matrix associated with the Laplace operator on any right-angled triangle, i.e., for any
element-node numbering system. For example, if the right-angled corner is numbered as node 1,
and the diagonal-line nodes are numbered as 2 and 3 (following the counter-clockwise numbering
scheme), we have (note that a denotes the length of side connecting nodes 1 and 2)
a 2 b 2 b 2 a 2
1
K e b 2 b2 0 ....(3b)
2ab
a 2
0 a
2
1 1 0
1
K
e
1 2 1 <(4)
2
0 1 1
The assembled coefficient matrix for the finite element mesh is 6 6, because there are six
global nodes, with one unknown per node. The assembled matrix can be obtained directly by
using the correspondence between the global nodes and the local nodes, expressed through the
connectivity matrix
1 2 3
5 3 2
B <..(5)
2 4 5
3 5 6
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
173
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 1 0 0 0 0 U1 1 Q11
1 4 2 1 0 0 U 3 Q + Q2 + Q3
1
2 2 3 1
4
1 0 2 4 0 2 0 U3 f0 3 Q3 + Q2 + Q1
1 2
(6)
2 0 1 0 2 1 0 U4 24 1 Q23
0 0 2 1 4 1 U 3 Q 2 + Q 3 + Q 4
5 1 3 2
0 0 0 0 1 1 U6 1
Q3 4
ˆ
Q12 + Q33 + Q24 = Q 5
The specified boundary conditions on the primary degrees of freedom of the problem are
U4 = U5 = U6 = 0 <(8)
Qˆ1 0, Q
ˆ 0, Q
2
ˆ 0 <.(9)
3
Since U4, U5, and U6 are known, the secondary variables at these nodes, i.e., Qˆ 4 ,Qˆ 5 , and Q
ˆ , are
6
Since the only unknown primary variables are (U1, U2 and U3), and (U4, U5, and U6) are
specified to be zero, the condensed equations for the primary unknowns can be obtained by
deleting rows and columns 4, 5 and 6 from the system. In retrospect, it would have been sufficient
to assemble the element coefficients associated with the global nodes 1, 2 and 3 i.e., writing out
equations 1, 2 and 3:
K11
1 1
K12 1
K13 U1 f11 0
1
K 21 K 22 K33 K11 K 23 K32 U2 f21 + f + f 0
1 2 3 1 2 2 3
3 1
4 1
K31
1 1
K32 K 23
2 1
K 33 K 22
2
K11 U3 f3 + f + f
2
2 1
4
0
<<(10)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
174
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Qˆ 4 f23 0 3
K 21 0 U1
ˆ 4
5
Q 1
f 2
f3
3
f 4
2 0 K 2
13 K 3
31 K 2
12 K 21 U2 <.(11)
ˆ f34 0 0 4
K 31
Q6 U3
where
u u u
q
3
nx ny 0 nx 0, 0
x y 12 y
n 1 2
u u u
q
3
nx ny nx 1, ny 0
x y 23 x
n 23
23
2 3
1
y
h23
, 23
13
0
Thus,
h 23 u y
Qˆ 4 Q22
3
1 dy
0 x h23
u 3 j3
u 3j
x j 1 2A3
We obtain
h 23 a, 13 a,2A3 a2 , U4 U5 0
3
ˆ h23
Q4
4 A3
j 1
u 3j j3 0.5U2 ...(12b)
Using the numerical values of the coefficients Kije and fie (with f0 =1), we write the
condensed equations for U1, U2 and U3 as
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
175
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Q22 3
1 0 0.5 0 U1 -0.197917
3 4 1
Q32 Q22 3 0 0 1 U2 = -0.302083 <(15)
Q4 24 1 0
32 0 0 U3
-0.041667
3 3
By interpolation, Q22 , for example, is equal to -0.5u2, and it differs from Q22 computed from
1
equilibrium by the amount f23 = .
24
10. Explain in detail. The steps involved in finite element solution of the axisymmetric
problem.
The finite element solution of the problem is given by the following steps.
Step 1: Replace the solid body of revolution by an assembly of triangular ring elements as shown
in figure.
Step 2: We use a natural coordinate system and assume linear variation of temperature inside an
element e so that the temperature T(e) can be expressed as
e
T N q
e
....(1)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
176
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
and
Ti
e
q T j ......(3)
Tk
The natural coordinates L1, L2 and L3 are related to the global cylindrical coordinates (r, z) of nodes
I, j and k as
1 1 1 1 L1
r = ri rj rk L2 .....(4)
z z zk L
i zj 3
or, equivalently,
L1 a1 b1 c1 1
1 a b c r ......(5)
L2 e 2 2 2
L 2A a3 b3 c3 z
3
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
177
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
where
a1 r j zk rk z j
a2 rk zi ri zk
a3 ri z j r j zi
b1 z j zk
b2 zk zi ....(6)
b3 zi z j
c1 rk r j
c2 ri rk
c3 r j ri
ri z j zk r j zk zi rk zi z j ....(7)
1
A
e
2
Step 3: The element matrices and vectors can be derived using Eqs. as follows.
rk 0
D 0r rkz
......(8)
and
Ni N j Nk
r r 1 b b b
B N Nr N e c1 c2 c3 .......(9)
i j k 2A 1 2 3
z z z
and by writing dV as 2r. dA where dA is the differential area of the triangle ijk. Eq. gives
(e)
K1e 2
r B D B dA
T
A
e
b2b3
c12 c1c2 c1c3
2 K z
+ e 2 c1c2 c22 c 2c3 r
2
dA .....(10)
4A c1c3 c32 A
e
c 2c3
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
178
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The radial distance r can be expressed in terms of the natural coordinates L1, L2 and L3 as
2 1 1 ri
R r dA
2 1
12
ri r j rk 1 2 1
2
r j ......(13)
A
e
1 1 2
rk
and hence
b12 c12
2
b1b2 c1c2 b1b3 c1c3
K1e kR b1b2 c1c2 2 2
b 2
c 2
b b c c ...(15)
2A e 2 3 2 3
b b c c
1 3 1 3 b2 b3 c2c3 b32 c32
To evaluate the surface integral of equation, we assume that the edge ij lies on the surface S 3 from
which heat convection takes place. Along this edge, L3 = 0 and dS3 = 2r ds so that equation gives
sj
p! q !
s sI
L1p Lq2ds s ji
p q 1!
....(17)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
179
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
3ri r j r r 0
hs ji
i j
K 2 e ri r j r 3r 0 ......(18)
6
i j
0 0 0
To evaluate the volume integral for P1e as
e
P1 rq N
T
dV <<(19)
v
e
We use the approximation rq rc q = constant where rc = (ri + rj + rk)/3 and the relation dV = 2r.dA
to obtain
rL1
e
P1 2 rc q e rL2 dA <..(20)
A
rL3
2ri r j rk
e
e rcqA
P1 ri 2r j rk ...(21)
6
ri r j 2rk
e
The surface integral involved in the definition of P3 can be evaluated as in the case of equation
(18). Thus if the edge ij lies on the surface S3,
2ri r j
rL1
hT S ji
sj
e
P3 hT N dS3 2 hT rL2 ds ri 2r j (22)
T
e s si 3
0
S3
0
e
Similarly expressions for P2 can be obtained as
2ri r j
e qS ji
P2 q N ri 2r j
T
dS2 if the edge
i j lies on S2 .....(23)
e
S2
3
0
Step 4: Once the element matrices and vectors are available, the overall or system equations can be
derived as
K T P <.(24)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
180
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
where
E
K K1e K2e <<(25)
e 1
and
E e e e
P P1 P2 P3
e 1
<<..(26)
Step 5: The solution of the problem can be obtained by solving Equation (24) after the
incorporation of the known boundary conditions.
11. Derive the element matrices and vectors for the elements shown in figure.
Solution:
From the data shown ............... required element properties can be computed as
b1 = zj -.........
b2 = zk – zi
b3 = zi - zj
c1 = rk - rj
c2 = ri - rk
c3 = rj - ri
Figure: 16.2
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
181
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
9175 9175 0
[KI(e) 9175 14330 5160 from equation .......... (1)
0 5160 5160
Since convection occurs along the two edges i j and jk, the [KI(e) ] matrix can be written as
(3ri rj ) (ri rj ) 0 0 0 0
(h)ij s ji (h) jk Skj
[K ]
(e)
2 (ri rj ) (ri 3rj ) 0 0 (3rj rk ) (rj rk )
6 0 6
0 0 0 (rj rk ) (rj 3rk )
(12 7) (4 7) 0 0 0 0
(15)(3) (10)(4) 0 (21 7) (7 7)
= (4 7) (4 21) 0
6 6
0 0 0 0 (7 7) (7 21)
447.7 259.2 0
= 259.2 1176.0 293.2
0 293.2 586.4
Equation gives
8 7 7 20734.5
(e) (60)(50)(6)
PI 4 14 7 23561.9
6 4 7 14 23561.9
(e)
As no boundary heat flux is specified, PI 0 . From equation and a similar equation for the edge jk,
we obtain
(2r r ) 0
(e) (hT )ij S ji i j (hT) jk skj
P3 (ri 2rj ) (2rj rk )
3 3
0 (ri 2rk )
(8 7) 0
(15)(40)(3) (10)(40)(4)
= (4 14) (14 7)
3 0 3 (7 14)
28274.3
= 69115.0
35185.8
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
182
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Thus,
9622.7 8915.8 0
[K ] [k ] [K ] 8915.8 15506.0 4866.8
(e) (e)
I
(e)
2
0 4866.8 5746.4
and
49008.8
(e) (e) (e) (e)
P PI P2 P3 92676.9
58747.7
12. Derive the three dimensional heat transfer equation for axisymmetric problems.
The governing differential equation for the steady state heat conduction in a solid body is
given by equation with the right hand side term zero and the boundary conditions by equation.
The finite element solution of these equations can be obtained by using the following procedure.
Step 2: We use a natural coordinate system and assume linear variation of temperature inside an
element e so that the temperature T(e) can be expressed as
(e )
T(e) = [N] q ............(1)
Where
[N] = [Ni Nj Nk Nl] = [L1 L2 L3 L4] ......... (2)
And
Ti
(e) T
j
q ........... (3)
Tk
T
l
The natural coordinates L1, L2, L3, and L4 are related to the global Cartesian coordinates of the
nodes i, j, l, and l by equation.
Step 3: The element matrices and vectors can be derived using equations as follows:
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
183
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
k x 0 0
[D] 0 ky 0 .......... (4)
0 k z
0
(b12 c12 d12 ) (b1b2 c1c 2 d1d2 ) (b1b3 c1c 3 d1d3 ) (b1b 4 c1c 4 d1d4 )
(b22 c 22 d22 ) (b2b3 c 2c 3 d2 d3 ) (b2b 4 c 2c 4 d2 d4 )
k
[KI(e) ]
36V (e) (b32 c 32 d32 ) (b3b 4 c 3c 4 d3 d4 )
symmetric (b24 c 24 d24
Ni2 NN
i j NN
i k NNi l
2
Nj NjNk NjNl
2 ] h
[K (e) ......... (6)
s(3e ) Nk2 Nk Nl
symmetric Nl2
If the face ijk of the element experiences convection, Nl = 0 along this face and hence equation
gives
2 1 1 0
hA ijk 1 2 1 0
[K (e) ............. (7)
12 1 0
2
1 2
0 0 0 0
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
184
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Where Aijk is the surface area of the face ijk. There are three other forms of equation, one for
each of the other faces jkl, kli, and lij. In each case the value of the diagonal terms will be two and
the values of the nonzero off-diagonal terms will be one. The coefficients in the row and the
column associated with the node not lying on the surface will be zero.
Ni
N 1
(e) j q0 V (e)
PI ( e ) q Nk dV 4 1 ......... (8)
V
N 1
l
If the face ijk lies on the surface S2 on which heat flux is pacified,
Ni L1 1
N L
(e) j 2 qA ijk 1
P2 N q dS 2 q ( e ) L3
dS 2 .......... (9)
S(2e ) k s2 3 1
N 0 0
l
Ni L1 1
(e) N L
j 2 hTAijk 1
P3 ( e ) hT N dS3 hT ( e ) L3 dS3 3 1 ....... (10)
S3 k S3
N 0 0
l
There are three other forms of equation (9) and (10). In these equations, the zero
coefficients will be located in the row corresponding to the node not lying on the face.
13. Derive the element equations for the elements shown if figure.
Solution:
From the given data, the required element properties can be computed as follows:
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
185
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 0 1 2
11 4 3 1 5
V (e) ,
6 1 1 0 0 6
1 2 2 4
1 3 1 4 1 1
b1 1 0 1 10, c1 1 1 0 17,
2 2 1 2 1 4
4 3 1
d1 1 0 1 1
2 2 1
1 0 0 1 1 0
b2 1 2 4 0, c 2 2 1 4 2,
1 1 2 0 1 2
1 0 1
d2 2 2 1 1
0 1 1
1 2 4 2 1 4
b3 1 1 2 5, c 3 0 1 2 10,
1 3 1 4 1 1
2 2 1
d3 0 1 1 0
4 3 1
1 1 2 0 1 2
b4 1 3 1 5, c 4 4 1 1 11
1 0 0 1 1 0
0 1 1
d4 4 3 1
1 0 1
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
186
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
= (25+9+1)1/2 = 5.916
= (9+4+16)(1/2) = 5.385
= (4+1+9)(1/2) = 3.742
1 1
s ( ) (5.916 5.385 3.742) 7.522
2 2
A jkl [7.522(7.522 5.916)(7.522 5.385)(7.522 3.742)]
1/ 2
= 9.877
Equation gives
0 0 0 0 0 0 0 0
h.A jkl 0 2 1 1 (10)(9.877) 0 2
1 1
2 ]
[K (e)
12 0 1 2 1 12 0 1 2 1
0 1 1 2 0 1 1 2
0 0 0 0
0 16.462 8.231 8.231
=
0 8.231 16.462 8.231
0 8.231 8.231 16.462
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
187
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 10.42
(e) 20 5
PI 1 10.42
64
1 10.42
0
(e)
P2 0 since no boundary heat fluxis specified
0
0 0
(e) 10 40 9.877
1 1316.92
P3
3 1 1316.92
1
1316.92
[K (e) ] [K1(e) ] [K (e)
2 ] (E1 )
10.42
(e) (e) (e) (e) 1327.34
P P1 P2 P3
1327.34
1327.34
(e) (e)
[K (e) ]q P
(e)
where [K (e) ] and P are given by equation (E1 ) and (E 2 ) result and
Ti
(e) T
j
q
Tk
T
l
2 2u
u2 fo (or) - u2 2 fo in
x y
in a square region the boundary condition of the problem is u = 0 on .
Solution by linear rectangular elements. Note that we cannot exploit the symmetry along the
diagonal x = y to our advantage when we use a mesh of rectangular elements. Therefore, we use 2
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
188
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
2 uniform mesh of four linear rectangular elements to discretize a quadrant of the domain. Once
again, no discretization error is introduced in the present case.
Since all elements in the mesh are identical, we shall compute the element matrices for only
one element, say element1. We have
1 1 2x 1 2y , 2 2x 1 2y , 3 4 xy , 4 1 2 x 2y
0.5 0.5 i j i j
K ije dx dy .......(1)
x x y y
0 0
0.5 0.5
fie fo i dx dy
0 0
Finite element discretization of the domain of Example 8.3 by linear rectangular elements.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
189
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
UNIT – V
PART – A
If the shape functions defining the boundary and displacements are the same, the element
is called as isoparametric element.
The element in which more number of nodes are used to define geometry compared to the
number of nodes used to define displacement are known as subparametric element.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
190
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Sub parametric element in which less number of nodes are used to define geometry
compared to the number of nodes used for defining the displacements. Such elements can be used
advantageously in case of geometry being simple but stress gradient high.
Theorem I:
If two adjacent elements are generated using shape functions, then there is continuity at the
common edge.
Theorem II:
It states that if the shape functions used are such that continuity of displacement is
represented in the parent co-ordinates, then the continuity requirement will be satisfied in the
isoparametric elements.
Theorem III:
The constant derivative conditions and condition for rigid body are satisfied for all isoparametric
elements.
The Lagrange polynomial formula is an interpolation formula that is useful for generating
shape functions and is defined as
n
x xm ( x x0 )( x x1 )...( x xk 1 )( x xk 1 )...( x x n )
Lk
m 0 x k x m ( xk x0 )( xk x01 )...( xk xk 1 )( xk xk 1 )...( xk xn )
k m
and represents the product of all terms. When x = x k, the product becomes unity. However, when
x = xm with m k, the product becomes zero. It follows that Lk of Equation has properties similar
to a shape function.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
191
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The shape function can be derived in terms of area coordinates using the formula
n
Fm (L1, L2 , L3 )
Nk
m 1 Fm|L1,L2,L3
where n is the order of the triangle and is equal to 1 less than the number of nodes along a
side. The function Fm is obtained from the equations of n lines that pass through all the nodes
except the one of interest. The denominator is the value of Fm when evaluated at the coordinates of
node k.
( x x j )( x xk ) ( x xi )( x xk ) ( x xi )( x x j )
Li Ni Lj N j Lk Nk
( xi x j )( xi xk ) ( x j xi )( x j xk ) ( xk xi )( xk x j )
( x L / 2)( x L) x( x L ) x( x L / 2)
N1 N2 N3
(L / 2)(L) (L / 2)(L / 2 L) L(L L / 2)
Figure :
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
192
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Use the Lagrange polynomial formula and Figure. The interpolation function for node 1 is
the product of shape functions along the lines = -1 and = -1. Note that the interpolation
function is identical to the shape function and that the notation for the shape function is used:
1 1 (1 )(1 )
N1 L1 L1
1 1 1 1 4
9. Discuss the evaluation of the jacobian matrix for a four-node isoparametric finite element.
The interpolation functions are defined in Pro. 6.4, and the jacobian matrix is defined by
Equation, (g) of Prob. 6.6. The jacobian matrix can be written as a product of two matrices:
(Ni / )
J E [ xi y i ]
(Ni / )
Or
x1 y1
x2
1 (1 ) (1 ) (1 ) (1 ) x2
J
4 (1 ) (1 ) (1 ) (1 ) x3 x3
x4 x4
This definition of J can be extended to any interpolation function. The matrix multiplication
would be a [2 x n ] [n x2], where n is the number of nodes to be used for geometry transformation.
Equation (a) can be used to evaluate J at any , location for an element defined in the x, y system.
In finite element analysis Equation (a) is used in conjunction with the numerical integration and
the evaluation of the local stiffness matrix.
10. Discuss the derivation of the [B] matrix for axisymmetric elasticity.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
193
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
rr / r 0
1/ r 0 N1 0 N2 0 N3 0 N4 0
u ...(a)
zz 0 / r 0 N1 0 N2 0 N3 0 N4
rz / r / r
The terms containing partial derivatives are obtained from Equation (b) of problem, and
substituted into equation (b) above, with r and z replacing x and y, respectively. The terms
containing the shape function divided by r are computed directly for each node (shape function).
For instance, let the x coordinate correspond to the radial coordinate r: Equation (b) of problem is
used to compute the r in equ. (b) above. The , and of prob. correspond to the coordinates of the
integration point in the , system. The shape functions are evaluated by substituting the
coordinates of the integration point (Gauss point) into the corresponding shape function equation
(see prob. for a four-node quadrilateral).
11. Discuss higher-order isoparametric element in terms of number of nodes versus the
complete polynomial representation that satisfies the Pascal triangle requirement. In Particular,
compare the six-node quadratic triangular element, the eight-node serendipity element, and the
nine-node Lagrangian element.
All these elements can be classified as quadratic elements because they have three nodes
along each side. The triangular element was discussed in Problem, and the corresponding
interpolation function contains all possible quadratic terms and none higher than quadratic.
A study of the shape functions for the eight-node serendipity element shows that all six
constant, linear, and quadratic terms are represented with the addition of two cubic terms. It
follows that an eight-node element must have eigth terms in its corresponding interpolation
function, and the cubic terms in this case are 2 and 2. The nine-node Lagrangian element
shape functions are given in problem, and a study of these shape functions indicates that there are
two additional third-order terms and one fourth-order term that can be shown to be 2 , 2, and
22.
It can be concluded that quadrilateral elements do not satisfy the Pascal triangle
requirement that the shape function be represented by a complete polynomial. However, the idea
is to satisfy the Pascal triangle (completeness) requirement in the best possible way. In higher-
order elements the excessive interior nodes in the Lagrange family of elements can cause difficulty
with convergence and should be avoided. For additional study see Burnett (1987) or Zienkiewicz
and Taylor (1989).
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
194
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
12. Use area coordinates to determine the shape function N4 for both the quadratic and cubic
triangular finite elements of Figure.
The quadratic element is shown in figure. The order of the triangle is n = 3 -1, and it follows
that two lines should be sufficient to pass through all nodes except node 4. These lines are shown
in the figure as L1 = 0 passing through nodes 2, 5, and 3 and L2 = 0 passing through nodes 1, 6, and
1 1
3. The location of node 4 is defined using all three area coordinates as L1 , L2 , L3 0.
2 2
L1 0 L2 0
N4 1 4L1L2
1
2 2
where the numerator in each term is an equation of a line and the denominator is the equation of
the line evaluated using the coordinates of node 4.
PART – B
1. (a) A four-element model of a plane area is shown in Figure. Use the interpolation function
for a four-node quadrilateral derived in problem and for element III shown that coordinate
location (x = 7.0, y = 6.0) corresponds to point (1, 1) in the generalized space. Also, for = 0.5 and
= - 0.5, determine the corresponding point in the global system.
Figure :
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
195
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The interpolation functions can be written as follows using the notation of problem.
4 4
x Ni x i y Ni y i (a)
i 1 i 1
1 1 1 1
or x= (1 )(1 )x1 (1 )(1 )x2 (1 )(1 )x3 (1 )(1 )x4 (b)
4 4 4 4
1 1 1 1
y= (1 )(1 )y1 (1 )(1 )y 2 (1 )(1 )y 3 (1 )(1 )y 4 (c)
4 4 4 4
1 1 1 1
x= (1 )(1 )(2) (1 )(1 )(5.5) (1 )(1 )(7) (1 )(1 )(4) (d)
4 4 4 4
1 1 1 1
x= (1 )(1 )(3) (1 )(1 )(3) (1 )(1 )(6) (1 )(1 )(6) (e)
4 4 4 4
Substitute = 1 and = 1 into Equations (d) and (e) and compute the corresponding and x
and y. Note that all terms in Equations (d) are zero except the third term and that it corresponds to
node 3. Similarly, all terms in equation (e) are zero except the third that corresponds to node 3. It
k can be seen that the interpolation function is similar to the shape function. For = 1 and =1, N3
= 1 with N1 = N2 = N3 = 0.
Substituting = 0.5 and = -0.5 into Eqs. (d) and (e) gives x = 5.0313 and y = 3.75. It is easily
verified that these solutions represent a linear approximation for coordinate locations along the
element boundaries and on the interior of the element.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
196
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
x = x (Ni) = x (, ) y = y (Ni) = y (, ) (the Ni are the interpolation functions) (b)
The derivatives of the shape functions can be written as follows, using the chain rule:
The first matrix on the right-hand side is defined as J and is referred to as the Jacobian matrix:
x y
J (e)
x y
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
197
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Multiplying Equ. 9d) by J-1 gives the form of the equation that can be used to compute derivatives
of the shape functions in the x, y system:
1
Ni Ni x y Ni
x
= J-1 = (f)
Ni Ni x y Ni
y
Substituting Equations (a) of Problem into (f) gives the final form
1
Ni Ni x N Ni
x i i y i
= (g)
Ni Ni Ni Ni
y xi y i
Note that the Ni within the Jacobian matrix are coordinate interpolation functions, whereas
the Ni within the column matrices are shape functions. The formulation of Eq. (g) has been related
to a four-node quadrilateral finite element, and that dictates that both the interpolation function
and the shape function be linear and that the result be an isoparametric element. A subparametric
finite element is formulated in exactly the same manner except that the Ni within the jacobian
matrix can be linear (a four-node coordinate approximation) and the Ni in the column matrices can
correspond to any higher node shape function approximation (such as the nine-node element of
Problem).
In addition, for the purpose of performing area integration, an infinitesimal element is defined as
where |det J| is the determinant of the jacobian matrix and is often merely referred to as the
jacobian.
2. A quadrilateral element is shown in Figure. Evaluate the stiffness matrix for heat transfer
using the definition given by Eq. of problem. Assume the thermal conductivity as k x=ky = k=1
Btu/(hr. in.F). Assume a unit thickness for the element.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
198
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
A 2 x 2 gaussian quadrature will be used, and the parent element is shown in Figure. The
numerical integration procedure is similar to that illustrated in Eq. (a) of Problem. The [B] matrix,
in the form given by Eq. (b) of problem, must be evaluated for each Gauss point of Figure. Each
time the [B] matrix is computed, the contribution to the stiffness matrix is computed as [B] T [k] [B],
and the final stiffness matrix is the sum of the four contributions. In what follows, the
computations will be shown in detail, and the reader should keep in mind that they are being
done using a computer code with a nested DO loop where I = 1 TO 2 as j = 1 TO 2, where I and J
correspond to the 2 x 2 gaussian integration. In order to evaluate [B] at each Gauss point, the
jacobian of the transformation for that Gauss point must be computed. Equation (a) of Prob. is
used to compute the jacobian in matrix format. The matrix that defines the element in the x, y
coordinate system corresponds to the second matrix of that equation and is evaluated using
Figure:
2 1
5 2
(a)
4 6
1 4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
199
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Gauss point 1:
The location of the Gauss point is obtained from Table and is shown in Figure. Let i = 1 = -
0.57735 and j = i = -0.57735. Use Eq. (a) of Prob. to evaluate the jacobian matrix:
1
J11
4
(1 )x1 (1 )x2 (1 )x3 (1 )x4
1
J11 (1 0.57735)2 (1 0.57735)(5) (1 0.57735)(4) (1 0.57735)(1)
4
1
3.15470 7.88675 1.69060 0.42265 = 1.50 Similarly,
4
1
J12 (1 0.57735)(1) (1 0.57735)(2) (1 0.57735)(6) (1 0.57735)(4)
4
1
= 1.57735 3.15470 2.53590 1.69060 = 0.60566
4
1
J21 (1 0.57735)(2) (1 0.57735)(5) (1 0.57735)(4) (1 0.57735)(1)
4
1
= 3.15470 2.11325 1.69060 1.57735 = -0.5
4
1
J22 (1 0.57735)(1) (1 0.57735)(2) (1 0.57735)(6) (1 0.57735)(4)
4
1
= 1.57735 0.84530 2.53590 6.3094 = 1.60566
4
1.5 0.60566
J=
-0.5 1.60566
J22 j12
J J11
J 1 = 21 (b)
| det EJ |
where |det J| = J11J22 J21J21 (c)
Substituting into Eqs. (c) and (b) gives |det J| = 2.71133 and
0.59221 -0.22338
J-1 = (d)
0.18441 0.55324
Let dA=dx dy = |det J| d d = |det J| wi = 1 wj = 1 = 2.71133, and the weights have been included in
this calculation. Note that the weight functions for 2 x 2 integration are equal to 1.0.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
200
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The [B] matrix of Eq. (b) of Problem is evaluated using the derivatives of the shape functions in the
, system. These numbers are the same as those used above to compute J.
The contribution to the stiffness matrix for Gauss point 1 can be written as
The computation for the remaining Gauss points will be shown but with less detail.
Gauss point 2:
i 1 0.57735 and j=2 0.57735
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
201
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Gauss point 3:
det J 3.14434
Gauss point 4:
det J 3.28868
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
202
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
3. Assume the eight-node quadrilateral isoparametric element shown in Figure is loaded with a
uniform pressure loading, px = 1.0, acting in the x direction along the side defined by nodes 1, 8,
and 4. Compute the distribution of the pressure loading to each node. Use a three-point
gaussian quadrature.
[N]
T
{T } dS (a)
s
where {T} is the surface traction matrix and [N] is the shape function matrix and is similar
to Eq. (b) or Prob. except that there are 16 rows rather than 8. The uniform pressure loading along
side 1-8-4 corresponds to the isoparametric coordinate = 1, and all shape functions will compute
as zero except at nodes 1, 8, and 4. Formally, Eq. (a) can be written as
T
N1 0 0 0 0 0 N4 0 0 0 0 0 0 0 N8 0 px 1.0
1
1 0 N1 0 0 0 0 0 N4 0 0 0 0 0 0 0 N8 px 0 d (b)
1
The term 1
N1px d is evaluated numerically with = -1 and and the weight functions taken
from Table. The shape functions are given in Problem.
(1 + 1) (1 - )(1 - - 1) 2
N1 (with = -1) =
4 2
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
203
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Similarly, the integral involving N4 becomes 0.333333px, and the integral representing the
contribution to the center node 8 becomes 0.6666667 px. The total length of the side of the
isoparametric element is 2, and the results are interpreted to mean that each corner node is
assigned one-sixth of the uniform load and the center node is assigned two-thirds of the uniform
load. Note that the result given by Eq. (c) could have been obtained by direct integration.
While the preceding computations illustrate distributing the force to each node, the analyst
should take advantage of the isoparametric formulation to compute the node loading. Consider
the plane element of Figure (b). In the isoparametric formulation the term dS should be expressed
in terms of the is , system using an interpolation function. Assume a unit thickness (t = 1) or
carry t through the derivation and write dS = t dL, where dL defines a curve corresponding to the
boundary of the plane area. In the Cartesian coordinate system,
x x x x
dx d d dy= d d (e)
The boundary of the element will always correspond to = 1 or = 1. Let = -1,
corresponding to the element of Figure, and it follows that / = 0. Use Eq. (a) of Prob. and
rewrite Eq. (e):
x Ni
dx d xi d (i = 1 to 8) (f)
y Ni
dy d y i d (i = 1 to 8) (g)
For illustration, use the element of Figure in x, y coordinates. Let = -1 and = -0.774597,
then substitute into Equations (f) and (g) to obtain dx = 0 and dy = 0.5 and by Eq. (d), dL = 0.5. The
first contribution to the pressure loading is computed as follows.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
204
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Node1:
(N1)(px)(dL)(w1)=0.687298px(0.5)(0.555555)=0.190916px
Node 4:
(N4)(px)(dL)(w1) = (-0.087298)px (0.5)(0.555555) =
- 0.02424px
Node 8:
(Ng)(px)(dL)(w1)=(0.400)px(0.5)(0.555555)=0.111111px
Similarly, let = -1 and = 0.0, dx = 0.0 and dy = 0.5, and dL = 0.5. The second contribution to the
pressure loading is (note that w2 = 0.888888) as follows.
Node 1 : N1 = 0.0
Node 4: N4 = 0.0
Node 8: (1.0)px= (0.5) (0.88888) = 0.44444px
The third integration point corresponds to = -1 and = +0.774597 with dx = 0.0, dy = 0.5, dL = 0.5,
and w3 = 0.55555, and the contribution to the pressure loading is computed as follows.
Node 1: -0.024249px
Node 4: 0.190916px
Node 8: 0.111111px
The total loading is the sum of the three contributions and is computed as 0.166667p x at nodes 1
and 4, and as 0.666666px at node 8.
4. (a) Discuss node placement and area coordinates for linear and higher-order triangular
isoparametric elements.
Linear, quadratic, and cubic elements are shown in Figure. The node numbering sequence is
arbitrary, but very often the corner nodes are numbered 1, 2, and 3, with the intermediate nodes
numbered in sequence starting with 4 along the side between nodes 1 and 2. The areas coordinates
then correspond to Figure for triangular elements of any order. For instance, area coordinate L 2
emanates from the side opposite node 2. The intermediate nodes are placed equidistant from the
end nodes, and for a cubic element the side nodes are placed at the one-third points along the side.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
205
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Figure:
The derivation for the linear three-node triangular finite element discussed in Prob. began by
assuming an interpolation function that was linear in terms of the x, y coordinates, or
It follows that a quadratic triangular finite element must contain all possible linear and quadratic
coordinate functions, or
This complete polynomial representation for interpolation functions corresponds to the Pascal
triangle shown in Figure (d). The cubic element should contain all linear, quadratic, and cubic
coordinate terms, and it can be seen from the Pascal triangle that there are 10 such terms. The
cubic element, Figure (c), must have 10 corresponding nodes. Therefore, the tenth node is located
at the centroid of the triangular element. The derivation of shape functions using interpolation
functions and x, y coordinates can become a tedious algebraic task. The use of area coordinates
and Eq. simplifies the derivation and formulation of a stiffness matrix.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
206
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
(b). Use area coordinates to determine the shape function N4 for both the quadratic and cubic
triangular finite elements of Figure.
The quadratic element is shown in Figure (a). The order of the triangle is n = 3 -1, and it follows
that two lines should be sufficient to pass through all nodes except node 4. These lines are shown
in the figure as L1 = 0 passing through nodes 2, 5, and 3 L2 = 0 passing through nodes 1, 6, and 3.
The location of node 4 is defined using all three area coordinates as L1 = ½ , L2 = ½ , L3 = 0. There
are two terms in Equation:
L 0 L 0
N4 1 1 2 1 = 4L1L2
2 2
where the numerator in each term is an equation of a line and the denominator is the equation of
the line evaluated using the coordinates of node 4.
Figure:
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
207
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
5. (a). The isoparametric formulation is advantageous for higher-order triangular elements. Like
the quadrilateral element, the distorted element in x, y coordinates can be mapped into a parent
element, and standard integration routines can be used to evaluate the stiffness integrals.
Discuss the coordinate transformation for isoparametric triangular elements using area
coordinates.
The method used by Zienkiewicz and Taylor (1989) may help to avoid some confusion in the
derivation. There are three area coordinates for a two-dimensional element, but the original
coordinates is 2-space (x and y). It follows that the parent element should be 2-space ( and for
the quadrilateral). Recall from Prob. that there are only two independent area coordinates, or L 1
+L2 + L3 = 1. Identify
and then Eqs. (c) – (f) and (h) of Prob. are valid. The shape functions are written in terms of L1, L2,
and L3, or
Hold , defined in Eq. (a), independent of L1 in Eq. (b) and write the partial derivative
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
208
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
(b). (i). Assume a five-element model as shown in Figure to solve the long cylinder problem
described in Problems. The cylinder has an inside radius of 1 in and outside radius of 2 in.
Assume an axisymmetric internal pressure loading of 1000 psi. Show the results for the [B] 1
matrix and the local stiffness matrix for element I. Compare results for radial displacement
with the exact solution. Assume E = 1.0 psi and = 0.3.
(ii) Compare the results for the 5 – element model with those for a 10-element model .
(a) The five-element model of Figure is constructed of square elements 0.2 x 0.2 in. In the
derivation for the axisymmetric finite element the differential area used was 2r dr dz, and
that requires that the internal pressure be distributed around the inside circumference of
the cylinder. The pressure loading is converted to nodal point loading as (1000 psi) (2) (1
in) (0.2 in) = 1265.64 lb, and one-half of that is applied to nodes 1 and 2.
five-element model
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
209
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The [B]I matrix for element I is a 4 x 8 matrix corresponding to Eq. (b) of Prob. and is computed for
Gauss point 1 using i = 1 = -0.57735 and j = 1 = -0.57735;
Note that the complete stiffness matrix is made up of four parts as in Prob. There is a separate B
matrix corresponding to each integration point, and the stiffness matrix for element I is the 8 x 8
symmetric matrix
The exact solution is computed using the results of Prob. Results for displacement are given in
Table. Note that all displacements in the z direction are zero for this problem and were entered
into the computer solution as computer solution as zero displacement boundary conditions.
(b) The 10 – element solution is computed using square elements 0.1 x 0.1, and the nodal point
loading is computed as 314.16 lb at nodes 1 and 2. The results are given in Table. The
results for displacement tabulated in Table are for illustration purposes, and the large
values are a result of assuming the material constant E = 1 psi.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
210
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
6. A triangular element is shown in Figure. Use the isoparametric formulation to compute the
terms in the stiffness matrix corresponding to convection as discussed in Probs. Integration
results for triangles are given in Figure (b). Illustrate the use of the both linear and quadratic
integration formulas.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
211
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The function to be integrated is Eq. (b) of Prob. Assume a unit value of t: for reference, the
function is
N1 N1
ux 0 N1 / x N2 / x N3 / x
A N2 N2
uy N1 / y N3 / y
dA (a)
0 N2 / y
N3 N3
Recall that there is an equality between the linear shape functions and area coordinates:
y = L1 + 3L2 + 5L3
There are partial derivatives in Eq. (a), and that will require evaluating the inverse of the jacobian
given by Eq. (f) of Problem. Refer to Eqs. (d) and (e) of Prob. it follows that
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
212
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
0 4 1 1
and J J-1 41 4
det J 8 2A ©
2 2 4
0
The jacobian for the coordinate transformation is complete, and now the right-hand matrix of Eq.
9a) must be evaluated. The computation is analogous to Eq. (g) of Problem. Using Eqs. (b) and (c)
above,
Combining Eq. (d) with Eq. (a) gives the final from of the equation that is to be integrated
numerically.
Numerical integration for the stiffness matrix is accomplished using the formulas of Figure (b)
corresponding to area coordinates. The limits of integration are written in terms of area
coordinates, and the weighted integral approximation follows as
n
1 1L2 1
F dL1 dL2 w i Fi (L1i L2i L3i ) (f)
i 1 2
0 0
where wi are the weights and L1i, L2i, L3i are the sampling points. Keep in mind that the function
of Eq. (a) is defined in the x, y system, dA = dx dy, and in the new system, dA = |det J| dL1 dL2.
Consider the first term of Eq. (e) and a linear order integration formula. For each L i, substitute 31
and note that wi = 1. Of course, only L1 appears in the first term, or
1 1 ux uy
K11 (1) 3 (u x u y )(8)
2 4 3
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
213
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Similarly,
1 11 2u
K 22 (1) ux (8) x
2 23 3
The remaining terms can be computed in the same manner and should be verified by the reader.
The correctness of the linear integration can be checked using the exact solution given in Problem.
u x b1 uy c1 )
K11 b1 y 2 y1 2 c11 xl 3 x2 2
6
u x uy
K11
3
The linear integration gives an exact result for this term. Similarly,
ux b2 uy c2 )
K 22 b2 y 3 y1 4 c 2 x1 x3 0
6
2ux
K 22
3
The quadratic integration formula will be illustrated for the K 11 term. Referring to Figure (b), L1
takes the values 21 , 21 and 0. and wi = 31 .
1 1 1 1 1 1 1 1 1
K11 (ux uy ) (ux uy ) (0)(ux uy )
2 4 2 3 4 2 3 4 3
u x uy
K11 (8)
3
Again, the exact result is obtained. The remaining terms are computed in a similar manner and
will not be recorded. Note that linear integration was sufficient to give exact results for the linear
triangular finite element. However, the reader should study Prob., where shape functions (no
derivatives appear in that problem) are integrated, and discover that linear integration is not
sufficient to give exact results in that case.
Assembling element stiffness matrix is a major part in finite element analysis. Since it involves
coordinate transformation from natural local coordinate system to Cartesian global system,
isoperimetric elements need special treatment. In this article assembling of element stiffness
matrix for 4 noded quadrilateral elements is explained in detail. The procedure can be easily
extended to higher order elements by using suitable functions and noting the increased number of
nodes.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
214
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Figure shows the typical parent element and isoparametric quadrilateral element.
(1 i )(1 )i
Ni ......... (1)
4
(1 )(1 ) (1 )(1 )
N1 ,N2
4 4
(1 )(1 ) (1 )(1 )
N3 ,and N4
4 4
We use the above functions for defining the displacement as well as for defining the
geometry of any point within the element in terms of nodal values.
x1
y
1
x2
x N1 0 N2 0 N3 0 N4 0
y2
N4
........ (2)
y 0 N1 0 N2 0 N3 0
x 4
y 4
The above relation helps to determine the (x, y) coordinates of any point in the element
when the corresponding natural coordinates and are given.
We are also using the same functions for defining the displacement at any point in the
element.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
215
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
u1
v
1
u2
u N1 0 N2 0 N3 0 N4 0
v 2
N4
........ (3)
v 0 N1 0 N2 0 N3 0
u 4
v 4
In assembling the stiffness matrix we need the derivatives of displacements with respect to
global x, y system. It is easy to find derivatives with respect to local coordinates and but it
needs suitable assembly to get the derivatives w.r.t. to global Cartesian system.
The relationship between the coordinates can be computed using chain rule of partial
differentiation.
Thus,
x y
x y
x y
x y
i.e.,
x y
x x
[J ] ........ (4)
x y
y y
x y
Where, [J ] ........ (5)
x y
The matrix [J] shown above is called Jacobian matrix. It relates derivative of the function in
local coordinate system to derivative in global coordinate system. In case of three dimensional it
is given by
x y z
x y z
[J ]
x y z
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
216
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Let
J J
[J] 11 12
J21 J22
Where
x y
J11 J12
x y
J21 J22
4
We know, x Ni xi N1x1 N2 x2 N3 x3 N4 x 4
i1
Then we get
4 Ni 4
Ni
xi yi
J 4
i 1 i 1
...... (6)
Ni 4
Ni
xi xi
i1 i 1
For any specified point the above matrix can be assembled. Now,
x
[J ]
y
x
-1
[J ]
y
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
217
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
J *
J
*
11
12
...... (7)
J
*
21 J
*
22
* * *
Where J11 ,J12 ,J21, and J*22 , are the element of Jacobian inverse matrix. Since for a given point
Jacobian matrix is known its inverse can be calculated and jacobian matrix is assembled. With this
transformation relation known, we can express derivatives of the displacements as shown below:
u u
x
*
*
u J11 J12 0 0 u
y J
*
J* 0 0
21 22 * ............. (8)
v 0 v
*
0 J11 J12
x 0 J*21
*
v
0 J22
v
y
u
x
x 1 0 0 0 u
y
y 0 0 0 1
0 1 1 0 v
xy x
v
y
u
u
J11
* *
J12 0 0
*
0 0 J*21 J22 ................ (9)
J*21 J*22 * * v
J11 J12
v
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
218
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
4 4
But u Nu
i i and v= Ni Vi
i1 i1
K B DB dv
T
In this case,
K t B DB x y
T
........... (12)
Where t is the thickness.
It can be shown that elemental area in Cartesian coordinates (x, y) can be expresses in terms of the
area in local coordinates (, ) as
x y = J ............... (13)
K t B DB J
T
........ (14)
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
219
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Integration is to be performed so as to cover entire area i.e. the limit of integration is from is from
– 1 to 1 and is also form – 1 to 1. It is difficult to carryout all the multiplications in equation (14)
and then the integration. It is convenient to go for numerical integration.
8. Assemble Jacobian matrix and strain displacement matrix corresponding to the Gauss point
(0.57735, 0.57735) for the element shown in figure. Then indicate how do you proceed to assemble
element stiffness matrix.
Solution : The coordinates of node points in Cartesian system are (0,0), (60, 0), (65.7735, 10) and
(5.7735, 10).
1
Ni 1 i 1 i
4
Ni 1
i 1 i
4
Ni 1
and i 1 i
4
N1 1 1
1 1 0.57735 0.10566
4 4
N2 1 1
1 1 0.57735 0.10566
4 4
N3 1 1
1 1 0.57735 0.39438
4 4
N4 1 1
1 1 0.57735 0.39438
4 4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
220
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Similarly,
N1 1 1
1 1 0.57735 0.10566
4 4
N2 1 1
1 1 0.57735 0.39438
4 4
N3 1 1
1 1 0.57735 0.39438
4 4
N4 1 1
1 1 0.57735 0.10566
4 4
Ni Ni
xi yi
J N Ni
xi y i
i
Ni
J11 xi
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
221
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
0.033333 0.019246 0 0
B 0 0 0 0.166667
0 0.166667 0.033333 0.019246
9. Determine the Cartesian coordinate of the point P( = 0.5, = 0.6) shown in Fig.
Figure.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
222
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
9.b. In the element shown in Fig. P is the point (6.5). On this point the load components in x and y
directions are 8 kn and 12 kn respectively. Determine its nodal equivalent forces.
Solution: We have to first determine the local natural coordinates of point P. We know for the
quadrilateral element
1
6 (1 )(1 )2 (1 )(1 )8 (1 )(1 )7 (1 )(1 )3
4
24 2(1 ) 8(1 ) 7(1 ) 3(1 )
20 10 0 2
4 10 2
or 2 5 n
y Ni y i gives
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
223
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1
2 5
2
or 4 10 (1 ) 9 2
i.e. 2 9 4 0
9 92 4x4
0.42443
2
1 0.42443
0.28779
2
X
F N
T
Y
For Point P.
(1 0.42443)(1 0.28779)
N1 0.10248
4
(1 0.42443)(1 0.28779)
N2 0.25362
4
(1 0.42443)(1 0.28779)
N3 0.45859
4
(1 0.42443)(1 0.28779
N4 0.18530
4
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
224
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
10. The quadrilateral element shown in Fig. is 20 mm thick and is subjected to surface forces T x
and Ty. Determine expressions for its equivalent nodal forces. If Tx = 10N/mm2 Ty = 15N/mm2,
determine the numerical values of the nodal forces.
Figure:
Solution: The element is subjected to load edge 3-4. We know along edge 3-4, = 1
1
N1 (1 )(1 ) 0
4
1
N2 (1 )(1 ) 0
4
1 1
N3 (1 )(1 ) (1 )
4 2
1 1
N4 (1 )(1 ) (1 )
4 2
We know,
2 2
l x y
l (x) (y)
2 2
and
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
225
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
1 1
x00 (1 )x 3 (1 )x 4
2 2
In limiting case,
dx x 1
(x 3 x 4 )
d y 2
1 1
Similarly, y 0 0 (1 )y 3 (1 )y 4
2 2
In limiting case
dy 1
(y 3 y 4 )
d 2
2 2
dl l 1 1 1
(x 3 x 4 ) (y 3 y 4 ) l34
d 2 2 2
1
dt l34 d
2
{Fx } t [N]2 {Tx }dl
0
0 0
l 1
0
t 1 (1 ) {Tx } l34d t 34
1 l d
1 2
2 4 1 (1 )Tx
1 (1 )Tx
2 (1 )
For uniformly distributed load, Tx is constant,
Fx1 0
Fx2 tl 0
1
34 2T Since (1 ) d 2
Fx3 4 x 1
F 2Ty
x4
0
tl34 0
Similarly
2 Tx
Ty
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
226
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Fy1 0
0
Fy2 tl34
[N] {Ty }ds
T
T
Fy3 2 x
F Ty
y4
In this problem,
t 20 mm
Tx 10 N/ mm2
Fx1 0
Fx2 0
Fx3 44721.36
Fx4 44721.36
Fy1 0
Fy2 0
Fy3 67082.04
F 67082.04
y4
Theorem 1: If two adjacent elements are generated using shape functions, then there is continuity
at the common edge.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
227
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
It may be observed that in the parent element , for any point on edge AB, shape functions
Ni =0 for nodes not on the edge and N1 exists for node on the edge. Hence the final function is the
same for the common edge AB in any two adjacent element, when we give the same coordinate
values for the nodes on common edge. Hence edge AB is contiguous in the adjacent elements.
Theorem II: It states, if the shape functions used are such that continuity of displacement is
represented in the parent coordinates, then the continuity requirement, will be satisfied in the
isoparametric elements also.
Theorem III: the constant derivative conditions and conditions for rigid body are satisfied for all
isoparametric elements if,
N i 1
u 1 2 x y y 4 z <..(1)
ui 1 2 xi 3 yi 4 zi
In finite element analysis we define nodal displacement at any point in the element in terms
of nodal displacement as
u Nu
i i
u Ni (1 2 xi 3 yi 4 zi )
1 Ni 2 Ni xi 3 Ni yi 4 Ni zi
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
228
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
N x i i x
N y i i y
<(2)
N z i i z
u= i Ni 2 x 3 y 4 z
N i 1
The shape function developed in natural systems satisfy this requirement. Hence they can
be safely used for isoparametric representation. This theorem is known as convergence criteria for
isoparametric elements.
COORDINATE TRANSFORMATION:
So far we have used the shape function for defining deflection at any point interms of the
nodal displacement Tag[1] suggested use of shape function for coordinate transformation form
natural local coordinate system to global Cartesian system and successfully achieved in mapping
parent element to required shape in global system. Thus the Cartesian coordinate of a point in an
element may be expressed as
x n1x1 N2 x 2 .... Nn xn
y n1y1 N2 y 2 .... Nn yn
z n1z1 N2 z 2 .... Nn zn
or in matrix form
x Nxe
where N are shape functions and (x)e are the coordinates of nodal points of the elements.
The shape functions are to be expressed in natural coordinate system.
For example consider mapping of a rectangular parent element into quadrilateral element
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
229
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
The parent rectangular element shown in fig (a) has nodes 1,2,3 and 4 and their coordinate
are (-1,-1),(-1,1),(1,1) and (1,-1). The shape functions of this element are
1 1 1 1
N1 ,N2
4 4
1 1 1 1
N3 and N1
4 4
P is a point with coordinate (,). In global system, the coordinates of the nodal points are
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
230
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
SEVENTH SEMESTER
MECHANICAL ENGINEERING
ENGINEERING)
(REGULATION 2004)
PART – A
7. Why polynomial type of interpolation functions are preferred over trigonometric functions?
PART – B
sin x
11. (a) Compute the value of central deflection (Figure 1) by assuming y .
L
The beam is uniform throughout and carries a central point load P.
Figure 1
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
231
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Or
12. (a) (i) Derive the shape function for a 2D beam element.
(ii) Derive the shape functions for 2D truss element.
Or
(b) Why higher order elements are needed? Determine the shape functions of an eight
nodded rectangular element.
13. (a) For the constant Strain Triangular element shown in Figure 2, assemble strain-displacement
matrix. Take t = 20 mm and E = 2 x 105 N/mm2.
Figure 2
Or
(b) The temperature at the four corners of a four-noded rectangle are T1, T2, T3 and T4.
Determine the consistent load vector for a 2-D analysis, aimed to determine the thermal stresses.
14. (a) Derive the expression for the element stiffness matrix for an axisymmetric shell element.
Or
(b) (i) Explain the terms ‚Plane stress‛ and ‚Plane strain‛ problems. Give constitutive laws for
these cases.
(ii) Derive the equations of equilibrium in case of a three dimensional system.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
232
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
15. (a) Establish the strain-displacement matrix for the linear quadrilateral element as shown in
figure 3 at Gauss point r = 0.57735 and s= -57735.
Figure 3
(b) Derive element stiffness matrix for a Linear Isoperimetric Quadrilateral element.
SEVENTH SEMESTER
MECHANICAL ENGINEERING
PART – A
2. Write the potential energy for beam of span ‘L’ simply supported at ends, subjected to a
concentrated load ‘P’ at mid span. Assume EI constant.
5. What is the governing differential equation for a one dimensional heat transfer?
8. Give one example each for plane stress and plane strain problems.
9. What do you mean by Constitutive Law and give Constitutive Law for axi-symmetric
problems?
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
233
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
PART – B
11. (a) Explain the Gaussian elimination method for solving of simultaneous linear
algebraic equations with an example.
Or
(b) A cantilever beam of length L is loaded with a point load at the free end. Find the
maximum deflection and maximum bending moment using Rayleigh-Ritz method
using the function Y = A (1 – cos (x / 2L)}. Given: EI is constant.
12. (a) (i) Derive the shape functions for a 2D beam element.
(ii) Derive the shape functions for 2D truss element.
Or
(b) Each of the five bars of the pin jointed truss shown in figure (b) has a cross sectional
area 20 Sq.cm. and E = 200 GPa.
(i) Form the equation F = KU where K is the assembled stiffness matrix of the
structure.
(ii) Find the forces in all the five members.
13. (a) Find the temperature at a point P(1, 1.5) inside the triangular element shown with
the nodal temperatures given as TI = 40C, TJ = 34C, and Tk = 46C. Also determine
the location of the 42C contour line for the triangular element shown in figure (a).
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
234
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Or
(b) Calculate the element stiffness matrix and thermal force vector for the plane stress
element shown in figure (b). The element experiences a rise of 10C.
14. (a) Derive the constant-strain triangular element’s stiffness matrix and equations.
Or
(b) Derive the Linear-Strain triangular element’s stiffness matrix and equations.
15. (a) Integrate f(x) = 10 + (20x) – (3x2 / 10) + (4x3 / 100) – (-5x4 / 1000) + (6x5 / 10000) between
8 and 12. Use Gaussian Quadrature Rule.
Or
(b) Derive element stiffness matrix for a Linear Isoparametric Quadrilateral element.
***********
SEVENTH SEMESTER
MECHANICAL ENGINEERING
PART – A
1. Write the potential energy for beam of span ‘L’ simply supported at ends, subjected to a
concentrated load ‘P’ at mid span. Assume EI constant.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
235
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
3. What is Constitutive Law and give constitutive law for axi-symmetric problems?
5. Give one example each for plane stress and plane strain problems.
6. Write the stiffness matrix for the simple beam element given below.
10. Write the natural co-ordinates for the point ‘P’ of the triangular element. The point ‘P’ is the
C.G. of the triangle.
PART – B
11. (a) Determine the expression for deflection and bending moment in a simple supported
beam subjected to uniformly distributed load over entire span. Find the deflection
and moment at midspan and compare with exact solution using Rayleigh-Ritz
method.
Or
(b) Derive the equation of equilibrium in case of a three dimensional stress system.
12. (a) Derive the shape function for a 2 noded beam element and a 3 noded bar element.
Or
(b) Write the mathematical formulation for a steady state heat transfer conduction
problem and derive the stiffness and force matrices for the same.
13. (a) Find the expression for nodal vector in a CST element shown in figure (a) subject to
pressures Px1 on side 1.
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
236
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
Or
(b) Determine the shape functions for a constant strain triangular (CST) element in
terms of natural coordinate system.
14. (a) For the CST element given below figure (a) assemble strain displacement matrix.
Take t = 20 mm, E = 2 105 N/mm2.
Or
(b) Derive the expression for constitutive stress-strain relationship and also reduce it to
the problem of plane stress and plane strain.
Or
b a
(b) (i) Use Gauss quadrature rule (n = 2) to numerically integrate xydxdy.
0 0
(ii) Using natural coordinates derive the shape function for a linear quadrilateral
element.
*****************
VEL TECH VEL TECH MULTI TECH VEL TECH HIGH TECH
237