MODERN INSTITUTE OF TECHNOLOGY AND RESEARCH CENTRE, ALWAR

Name of faculty:Kalyan Singh

Subject & Subject Code: FEM, 7ME4-01
Semester: VII Session: 2024-25(Odd Sem.)
Branch: ME Batch: A
Unit No.:5 Date of Submission:13/11/2024

LEC. Topics Covered Total Page

No. Page Number

33 Heat conduction through a wall, 5 1-5

34 Heat transfer through a fin, 4 6-9

35 3 10-12
2-D heat Conduction in a Plate,

36 Torsion problems. 3 13-15

37 3 16-18
Bandwidth, Elimination method

38 method of factorization for solving simultaneous algebraic equations 3 19-21

39 Features of software packages, 3 22-24

40 Sources of error. 3 25-27

References:
[1] Seshu P., “Text Book of Finite Element Analysis” PHI Learning Private Limited, tenth
edition, 2003.
[2] Tirupathi.R. Chandrapatha and Ashok D. Belegundu, "Introduction to Finite Elements in
Engineering", Prentice Hall India, 2012
[3] Reddy J.N., "An Introduction to Finite Element Method", McGraw Hill, third edition, 2005.
CHAP 4 FINITE ELEMENTS FOR HEAT
TRANSFER PROBLEMS

1
 HEAT CONDUCTION ANALYSIS
• Analogy between Stress and Heat Conduction Analysis

Structural problem Heat transfer problem

Displacement Temperature (scalar)
Stress/strain Heat flux (vector)
Displacement B.C. Fixed temperature B.C.
Surface traction force Heat flux B.C.
Body force Internal heat generation
Young’s modulus Thermal conductivity

– In finite element viewpoint, two problems are identical if a proper

interpretation is given.
• More Complex Problems
– Coupled structural-thermal problems (thermal strain).
– Radiation problem
2
 THERMAL PROBLEM
• Goals:

[KT ]{T }  {Q}

Thermal load

Nodal temperature
Conductivity matrix

– Solve for temperature distribution for a given thermal load.

• Boundary Conditions
– Essential BC: Specified temperature
– Natural BC: Specified heat flux

3
4.2. FOURIER HEAT
CONDUCTION EQUATION

4
 STEADY-STATE HEAT TRANSFER PROBLEM
• Fourier Heat Conduction Equation:
– Heat flow from high temperature to low temperature

dT
q x = -kA
dx
Thermal conductivity (W/m/C )
Heat flux (Watts)

• Examples of 1D heat conduction problems

Thigh

Thigh Tlow
qx qx

Tlow
5
 GOVERNING DIFFERENTIAL EQUATION
• Conservation of Energy
– Energy In + Energy Generated = Energy Out + Energy Increase

Ein + Egenerated = Eout + DU

• Two modes of heat transfer through the boundary

– Prescribed surface heat flow Qs per unit area
– Convective heat transfer Qh = h (T ¥ - T )
– h: convection coefficient (W/m2/C )

T¥
Qs

dqx
qx Qg qx + Dx
dx
A

x
6
 GOVERNING DIFFERENTIAL EQUATION cont.
• Conservation of Energy at Steady State
– No change in internal energy (U = 0)
æ dqx ö÷
qx + Qs PDx + h (T - T ) PDx + Qg ADx = ç qx +
¥ ç Dx ÷÷
  è
dx ø
Ein Egen
Eout
– P: perimeter of the cross-section
dqx
= Qg A + hP (T ¥ - T ) + Qs P, 0£x £L
dx
• Apply Fourier Law

d æ dT ö÷
çç kA
dx è ÷
÷
dx ø
+ Qg A + hP ( T ¥
- T ) + Qs P = 0, 0£x £L

– Rate of change of heat flux is equal to the sum of heat generated and
heat transferred

7
 GOVERNING DIFFERENTIAL EQUATION cont.
• Boundary conditions
– Temperature at the boundary is prescribed (essential BC)
– Heat flux is prescribed (natural BC)
– Example: essential BC at x = 0, and natural BC at x = L:
T(0)  T0

 dT
kA dx  qL
 x L

8
4.3. FINITE ELEMENT
ANALYSIS – DIRECT METHOD

9
 Heat Conduction in a Long Wire
• Similar to direct method in 1D bar
• No need to work with differential equation
• Use conservation of energy

• Heat transfer on the boundary

• Internal heat generation
• No radiation
T¥

T(0) = T0 Qg x
qx(L) = qL
A

Qs
10
 DIRECT METHOD
• Follow the same procedure with 1D bar element
– No need to use differential equation
• Element conduction equation
– Heat can enter the system only through the nodes
– Qi: heat enters at node i (Watts)
– Divide the solid into a number of elements
– Each element has two nodes and two DOFs (Ti and Tj)
– For each element, heat entering the element is positive
1 2 N

Q1 Q2 Q3 QN
i j
e
qi(e ) q (je )
Ti L(e) Tj
xi
xj
11
 ELEMENT EQUATION
• Fourier law of heat conduction
(e ) dT (T j - Ti )
qi = -kA = -kA
dx L( e )

• From the conservation of energy for the element

(T j - Ti )
qi(e ) + q(j e ) =0 q (j e ) = +kA
L( e )
• Combine the two equation
ìï q (e ) üï kA é 1 -1ù ìï Ti üï
ï i ï= ê ú ï ïý Element conductance matrix
í (e ) ý í
ïï q j ïï L(e ) êë -1 1 úû ïîïT j ïþï
î þ

– Similar to 1D bar element (k = E, T = u, q = f)

12
 ASSEMBLY
• Assembly using heat conservation at nodes
– Remember that heat flow into the element is positive
– Equilibrium of heat flow:
ìï T1 üï ì ï Q1 üï
ï ï
ïï T ïï ï ï ï
ï
Ni
ï ï ï
ï Q ï
ï
Qi = å qi (e )
[KT ] í ý = í 2 2
ý
e =1 ( N´N ) ï ï  ï
ï ï
ï  ï
ï
ïï ïï ï ï ï
ï QN ï
îïTN þï î ï
þ
– Same assembly procedure with 1D bar elements
• Applying BC
– Striking-the-rows works, but not striking-the-columns because
prescribed temperatures are not usually zero
Q2

1 3
Element 1 q2(1) 2 q2(2) Element 2
13
 EXAMPLE
• Calculate nodal temperatures of four elements
– A = 1m2, L = 1m, k = 10W/m/C

Q4 = –200W
T1 T2 T3 T4 T5
200 C x
1 2 3 4

Q1 Q2 = 500W Q3 = 0 Q5 = 0

• Element conduction equation

é 1 -1ù ìï T1 üï ì ïï q1(1) üïï é 1 -1ù ìïT2 üï ì ïï q2(2) ü
ïï
10 ê úí ý = í ý 10 ê úí ý = í ý
êë -1 1 úû ïï
îT2 ïïþ ïîï q2 ïþï
(1) êë -1 1 úû ïîïT3 ïþï ïï q (2) ï
î 3 ï þ
é 1 -1ù ì ü ì
ï
ï 3ï ï 3 ï
T q (3) ü
ï é 1 -1ù ìT ü ì
ï q (4) ü
ï 4 ï ï 4 ïï
10 ê úí ý = í ý 10 ê úí ý = í ý
êë -1 1 úû ï
ïîT4 ï ï q ïï
ïþ ï (3)
ë ï
ê -1 1 ûú î
ï T ï
ï
þ ï
ï q (4) ïï
î 4 þ 5 î 5 þ

14
 EXAMPLE cont.
• Assembly
ìï q (1) üï
ìï Q1 üï ï 1 ï é 1 -1 0 0 0 ù ìï T1 üï
ïï ïï ï (1) ï (2) ï
ï ê ú ïï ïï
ïï Q2 ïï ïï q2 + q2 ïï ê -1 2 -1 0 0 ú ïïT2 ïï
ïï ïï ïï (2) ïï ê ú ïï ïï
í Q3 ý = í q3 + q3 ý = 10 ê 0 -1 2 -1 0 ú íT3 ý
(3)
ïï ïï ïï ï ê úï ï
ïïQ4 ïï ïï q (3) + q (4) ïï ï ê 0 0 -1 2 -1ú ïïïT4 ïïï
ïï ïï ïï 4 4
ïï ê úï ï
îï Q5 þï ï q (4) êë 0 0 0 -1 1úû ïîïT5 ïïþ
îï 5
ï
ï
þ
• Boundary conditions (T1 = 200 oC, Q1 is unknown)
é 1 -1 0 0 0 ùì ï 200 ü
ï ì Q1 ïü
ï
ê ú ïï ï ïï
ï ï
ï
ê -1 2 -1 0 0 ú ïï T2 ïï ïï 500 ïï
ê úï ï ï ï
10 0 -1 2 -1 0 í T3 ïý = ïí 0 ïý
ê ú ï
ê úï ï ï ï
ê 0 ú
0 - 1 2 -1 ïï T ï
ï ï
ï - 200 ï
ï
ê úïï
4 ï
ï ï
ï ï
ï
êë 0 0 0 -1 1 úû ï ï ï
îï 5 þï ïî
T 0 ï
ïþ

15
 EXAMPLE cont.
• Boundary conditions
– Strike the first row
ì
ï 200 ü
ï
é -1 2 -1 0 0 ïù ï ï ì
ï 500 ü
ï
ê ú ï T2 ïï ï ï
ê 0 -1 2 -1 0 ú ïï ïï ï 0 ï
ï ïï
10 ê ú ï ï ï
T3 ý = í
ê 0 0 -1 2 -1ú í ï ï ï -200 ï
ý
ê úïï T ï
ï ï
ï ï
ï
ê 0 0 0 - 1 1 ú ï 4 ï ï 0 ï
ë ï
ûï T ï ï ï
î ï
þ
ï 5 þ
î ï
– Instead of striking the first column, multiply the first column with
T1 = 200 oC and move to RHS

é 2 -1 0 0 ùìï T2 ü
ï ïïì 500 ïïü ï ì 2000 ü
ïï
ê ï
úï ï ï ï ïï ï ï ïï
ê -1 2 -1 0 ú ï T ï ï 0 ï 0
10 ê ú ïí 3 ïý = ïí ïý + ïí ïý
ê 0 -1 2 -1ú ï T4 ï ï -200 ï ï 0 ïï
ê úï ï ï
ï ï
ï ï
ï ï
ï ïï
êë 0 ú
0 -1 1 û îï
ïT5 ïï ï
þ ïî 0 þ ï
ï î ï
ï 0 þ ï
– Now, the global matrix is positive-definite and can be solved for nodal
temperatures
16
 EXAMPLE cont.
• Nodal temperatures

{T } T = { 200 230 210 190 190 } C

• How much heat input is required to maintain T1 = 200oC?

– Use the deleted first row with known nodal temperatures

Q1 = 10T1 - 10T2 + 0T3 + 0T4 + 0T5 = -300 W

• Other example 100W 3

4
2

1 2 4 5
1 3 5
50 C x

200W Q=0 Q=0

17
 Exercise
• Calculate nodal temperatures of five elements in the figure
– A = 1m2, L = 1m, k = 10W/m/C

100W 3

4
2

1 2 4 5
1 3 5
50 C x

200W Q=0 Q=0

18
4.4. GALERKIN METHOD FOR
HEAT CONDUCTION

19
 GALERKIN METHOD FOR HEAT CONDUCTION
• Direct method is limited for nodal heat input
• Need more advanced method for heat generation and
convection heat transfer
• Galerkin method in Chapter 2 can be used for this purpose
• Consider element (e) (e )
i
e
j
qi q (je )
• Interpolation T L(e) T xi
T( x ) = Ti Ni ( x ) + T j N j ( x )
i j
xj
æ x - xi ÷÷ö, N j ( x ) = x - xi
ç
Ni ( x ) = ç1- (e )
è L ÷ø L(e )
ìïïTi üïï

T ( x ) = êë Núû {T} = êë Ni ( x ) N j ( x ) úû í ý Temperature varies linearly
ïïîT j ïïþ
• Heat flux
dT ê 1 1 ú
= ê - (e ) ú {T } = êë B úû {T }
(e ) ú Heat flow is constant
dx êë L L û
20
 GALERKIN METHOD cont.
• Differential equation with heat generation
d æ dT ö÷
çç kA ÷÷ø + Qg A = 0, 0 £ x £ L
dx è dx
• Substitute approximate solution
d æç dT ö÷
÷ + AQg = R ( x ) Residual
dx ççè
kA
dx ø÷

• Integrate the residual with Ni(x) as a weight

xj
æ d æ dT ÷ö ÷ö
çç çç kA
ò çè dx çè dx ÷÷ø + AQg ÷÷øNi ( x )dx = 0
xi
• Integrate by parts
xj xj

dT
xj

dT dNi
kA N ( x ) - ò kA dx = -ò AQg Ni ( x )dx
dx i x
d x d x
i xi xi

21
 GALERKIN METHOD cont.
• Substitute interpolation relation
xj xj
æ dNi dN j ÷ö dNi
ç
ò kA çTi
çè dx
+ Tj ÷
dx ÷ø dx
dx = ò AQg Ni ( x )dx - q( x j )Ni ( x j ) + q( xi )Ni ( xi )
xi xi

• Perform integration
xj
kA
( Ti - T j ) = Q (e )
+ q (e )
Qi( e ) = ò AQg Ni ( x )dx
L( e )
i i
xi

• Repeat with Nj(x) as a weight xj

kA Q(j e ) = ò AQg N j ( x )dx
( T j - Ti ) = Q (e )
+ q (e )
L( e )
j j
xi

22
 GALERKIN METHOD cont.
• Combine the two equations
ì
ï (e ) (e ) ü
é 1 -1ù ì ü
ïï i ïï ï i Q + q i ï T ]{T }  {Q
[k (e) }  {q(e) }
T (e)
kA
ê ú = ï
í ý í (e ) ý
(e ) ê
-1 1 ú ï ï ï (e ) ï
îQ j + q j ïþ
L ë T
û îï j þï ï Similar to 1D bar element

– {Q(e)}: thermal load corresponding to the heat source

– {q(e)}: vector of nodal heat flows across the cross-section
• Uniform heat source
xj
é Ni ( x ) ù AQg L( e ) ìï1üï
{Q ( e ) } = ò AQg ê ú dx =
êë N j ( x ) úû 2
í ý
ïïî1ïïþ
xi

– Equally divided to the two nodes

• Temperature varies linearly in the element, and the heat
flux is constant

23
 EXAMPLE Insulated
No heat flow

• Heat chamber
Wall
Wall temperature = 200 C
Uniform heat source inside 200 C
the wall Q = 400 W/m3.
Thermal conductivity of the x
wall is k = 25 W/mC.
1m
Use four elements through
the thickness (unit area) No heat flow
Boundary Condition:
x
T1 = 200, qx=1 = 0. 200 C

T1 T2 T3 T4 T5
1 2 3 4
24
 EXAMPLE cont.
• Element Matrix Equation
– All elements are identical

 1 1  T1  50  q1(1) 

100          (1) 
 1 1  T2  50  q2 

– Assembly

 Q1   q1(1)  1 1 0 0 0   T1   50 
Q   (1) (2)   1 2 1 0 0  T2  100 
q
 2   2  q2      
(3)  
Q3   q3  q3   100  0 1 2 1 0  T3   100 
(2)

Q  q(3)  q(4)   
0 0 1 2 1 T4  100 
 4   4 (4) 4     
Q5   q5   0 0 0 1 1  T5   50 

25
 EXAMPLE cont.
• Boundary Conditions
– At node 1, the temperature is given (T1 = 200). Thus, the heat flux at
node 1 (Q1) should be unknown.
– At node 5, the insulation condition required that the heat flux (Q5)
should be zero. Thus, the temperature at node 5 should be unknown.
– At nodes 2 – 4, the temperature is unknown (T2, T3, T4). Thus the heat
flux should be known.
1 2 3 4 5
Q1 Q5

1 1 0 00  200  50  Q1 
 1 2 1 0 0   T2   100 
    
100  0 1 2 1 0   T3    100 
 
0 0 1 2 1  T4   100 
   
 0 0 0 1 1   T5   50 

26
 EXAMPLE cont.
• Imposing Boundary Conditions
– Remove first row because it contains unknown Q1.
– Cannot remove first column because T1 is not zero.
200 
 1 2 1 0 0    100 
 0 1 2 1 0   T2  100 
100    T3    

 0 0 1 2 1    100 
   T4  
 0 0 0 1 1   50 
 T5 
100( 1 200  2  T2  1 T3 )  100
– Instead, move the first column
100(2  T2  1 T3 )  100  20000
to the right.

 2 1 0 0  T2  100  20000  20100 

 1 2 1 0  T3  100   0   100 
100        
 0 1 2 1 T4  100   0   100 
 
0 0 1 1  T5   50   0   50 
27
 EXAMPLE cont.
• Solution
T1  200 C, T2  203.5 C, T3  206 C, T4  207.5 C, T5  208 C
208

207
FEM
206
Exact
205

204
T

203

202

201

200
0 0.2 0.4 x 0.6 0.8 1

• Discussion
– In order to maintain 200 degree at node 1, we need to remove heat

Q1  50  100T1  100T2  350

 Q1  400 W

28
 Exercise
• Consider a heat conduction problem described in the figure.
Inside of the domain, heat is generated from a uniform heat
source Qg = 10 W/m3, and the conductivity of the domain is k
= 0.1 W/m/oC. The cross-sectional area A = 1 m2. When the
temperatures at both ends are fixed at 0oC, calculate the
temperature distribution using (a) two equal-length elements
and (b) three equal-length elements. Plot the temperature
distribution along with the exact temperature distribution

Insulated

T = 0oC x Qg = 10 W/m3 T = 0oC

Insulated
1m
29
4.5. CONVECTION BOUNDARY
CONDITION

30
 CONVECTION BC
• Convection Boundary Condition
– Happens when a structure is surrounded by fluid
qh Wall
– Does not exist in structural problems
– BC includes unknown temperature (mixed BC) T T

qh = hS (T ¥ - T )
Fluid Temperature
Convection Coefficient
– Heat flow is not prescribed. Rather, it is a function of temperature on
the boundary, which is unknown
• 1D Finite Element
– When both Nodes 1 and 2 are convection boundary

ìï q1 = hAT1¥ - hAT1
ïí
ïïî q2 = hAT2¥ - hAT2 T1 T2
T1 T2
31
 EXAMPLE (CONVECTION ON THE BOUNDARY)
• Element T1 T2 T3

equation T1¥ T3¥

h1 1 2 h3

ï ì kA é 1 -1ù ìïT2 üï ìïï q2 üïï

(1) ü (2)
kA é 1 -1ù ì ï T1 ü ï
ï q1 ï ïý
ê úí ý = í ê úí ý = í ý
L êë -1 1 úû ï
ïT2 ï
î ï ïï
þ (1) ï L ëê -1 1 úû îï
ïT3 þï ïî q3 ïþï
ï ï (2)
î 2 ïþ
q

• Balance of heat flow


– Node 1: q1  h1A(T1  T1 )
(1)

– Node 2: q2  q2  0
(1) (2)


– Node 3: q3  h3 A(T3  T3 )
(2)

• Global matrix equation

é 1 -1 0 ù ìï T1 üï ìïï h1A(T1¥ - T1 ) ü ï
ïï
kA ê ú ï
ï ï
ï ï
ê -1 2 -1ú íT2 ý = íï 0 ýï
L ê ú ïï ï ï ï
êë 0 -1 1 ûú ïîT3 ïþï ïï h3 A(T3¥ - T3 ) ïï
îï ï
þ
32
 EXAMPLE cont.
• Move unknown nodal temperatures to LHS
é kA kA ù
ê + h1A - 0 ú
ê L L ú ìï T üï ìï h AT ¥ üï
ê 2kA kA úú ïï ïï ïïï
1 1 1 ï
ïï
ê - kA - íT2 ý = í 0 ý
ê ú
ê L L L ú ïïïT ïïï ïïï h AT ¥ ïïï
ê kA kA ú î 3 þ ïî 3 3 ïþ
ê 0 - + h3 A ú
êë L L úû

• The above matrix is P.D. because of additional positive terms

in diagonal
• How much heat flow through convection boundary?
– After solving for nodal temperature, use
q1(1) = h1A(T1¥ - T1 )

• This is convection at the end of an element

33
 EXAMPLE: FURNACE WALL
• Firebrick
Firebrick Insulating
k1=1.2W/m/oC brick
hi=12W/m2/oC Tf = 1,500 Ta = 20 C
• Insulating brick C
k2=0.2W/m/oC hi x
ho
ho=2.0W/m2/oC
0.25 m 0.12 m
16.8 4.8 0   T1  18,000 
 4.8 6.47 1.67  T    0 
  2  
 0 1.67 3.67  T3   40 
{T } T  {1,411 1,190 552} C Convection Convection
No heat flow
boundary boundary

1,500 C x 20 C

q(2)
3  h0 (Ta  T3 )  1054 W/m 2

T1 T2 T3
Tf Ta
hi 1 2 ho
34
 CONVECTION ALONG A ROD

• Long rod is submerged into a fluid

• Convection occurs across the entire surface
• Governing differential equation
d  dT 

dx 
kA 
dx 
 AQ g  hP T 
 
 T  0, 0  x  L P  2(b  h)

Convection
Fluid T 

b
h
qi(e ) i
j q (je )

xi Convection
xj
35
 CONVECTION ALONG A ROD cont.
• DE with approximate temperature
d  dT 

dx 
kA
dx 
  AQ g  hP T

 T  R(x)

• Minimize the residual with interpolation function Ni(x)

dT 
xj
 d  

x  dx  kA dx   AQg  hP(T  T) Ni (x)dx  0


i  

• Integration by parts
xj

dT
xj

dT dNi
xj xj xj

Ni (x)   kA  dx   AQ N (x)dx  hPT Ndx

dx   hPTN
kA
dx xi
dx dx xi
i 
xi
g i 
xi
i
xi

36
 CONVECTION ALONG A ROD cont.
• Substitute interpolation scheme and rearrange
xj x
 dNi dNj  dNi j

x kA  Ti dx  Tj dx  dx dx  x hP(TN
i i  TN
j j )Ni dx
i i

xj

  (AQg  hPT  )Ni dx  q(x j )Ni (x j )  q(xi )Ni (xi )

xi

• Perform integration and simplify

(e)  Ti
kA Tj 
L(e)
 Ti  Tj   hpL 
3

6
  Q(e)
i  q(e)
i
 
xj

 g

Q(e)
i  (AQ  hPT )Ni (x)dx
xi

• Repeat the same procedure with interpolation function Nj(x)

37
 CONVECTION ALONG A ROD cont.
• Finite element equation with convection along the rod
 kA  1 1 hPL(e) 2 1   Ti  Qi  qi 
(e) (e)

 (e)  1 1   6  1 2  T   Q(e)  q(e) 

L       j   j j  

[k (e)
T ]  [k (e)
h ] 
 T  {Q (e)
}  {q(e)
}

• Equivalent conductance matrix due to convection

hPL(e)  2 1
k   6
(e)
h  1 2
 
• Thermal load vector
(e) 
 Qi  AQ gL(e)
 hPL T 1
{Q }    
(e)

Q j  2 1

38
 EXAMPLE: HEAT FLOW IN A COOLING FIN
• k = 0.2 W/mm/C, h = 2104 W/mm2/C
• Element conductance matrix
0.2  200  1 1 2  104  320  40 2 1
[k ]  [k ] 
(e)
T
(e)
h     1 2
40  1 1  6  
• Thermal load vector
Convection
2  104  320  40  30 1 T  = 30 C
{Q } 
(e)

2 1
160 mm
1.25 mm
• Element 1

330 C x Insulated
120 mm

T1 T2 T3 T4
1 2 3
39
 EXAMPLE: HEAT FLOW IN A COOLING FIN cont.
• Element conduction equation
 1.8533 0.5733   T1  38.4  q1(1) 
       (1) 
– Element 1  0.5733 1.8533  T2  38.4  q2 

 1.8533 0.5733  T2  38.4  q(2) 

     (2) 
2
– Element 2  
 0.5733 1.8533  T3  38.4  q3 
 1.8533 0.5733  T3  38.4  q(3) 
– Element 3        (3) 
3

 0.5733 1.8533  T4  38.4  q4 

• Balance of heat flow
– Node 1 q1(1)  Q1

– Node 2 q(1)
2  q2 0
(2)

– Node 3 q(2)
3  q3  0
(2)


4  hA(T  T4 )
q(3 )
– Node 4

40
 EXAMPLE: HEAT FLOW IN A COOLING FIN cont.
• Assembly
1.853 .573 0 0   T1   38.4  Q1 
 .573 3.706 .573 0  T2   76.8 

     
 0 .573 3.706 .573  T3   76.8 
   
 0 0 .573 1.853  T4  38.4  hA(T   T4 )

• Move T4 to LHS and apply known T1 = 330

1.853 .573 0 0  330  38.4  Q1 
 .573 3.706 .573 0   T2   76.8 
   
 0 .573 3.706 .573   T3   76.8 
 
 0 0  .573 1.893   T4   39.6 
• Move the first column to RHS after multiplying with T1=330
3.706 .573 0  T2  265.89 
 .573 3.706 .573  T    76.8 
  3  
 0 .573 1.893  T4   39.6 

41
 EXAMPLE: HEAT FLOW IN A COOLING FIN cont.
• Solve for temperature
T1  330C, T2  77.57C, T3  37.72C, T4  32.34C

350
300
250
200
150
100
50
0
0 40 80 120

42
 Exercise
• Determine the temperature distribution (nodal temperatures)
of the structure shown in the figure using two equal–length,
linear finite elements with the cross-sectional area of 1 m2.
The thermal conductivity is 10 W/m/ºC. The left side is
maintained at 300 ºC. The right side is subjected to heat loss
by convection with h = 1 W/m2/ºC and Tf = 30 ºC. All other
sides are insulated.
Insulated
T = 300 ºC Tf = 30 ºC

20 m

T1 T2 T3

43
4.6. 2D HEAT TRANSFER

44
 2D Heat Transfer Problem
• 2D Heat transfer: Temperature remains constant through z-
coordinate
– No heat flow in z-dir
– ST: prescribed temperature
– SQ: prescribed heat flux
– Sh: convection boundary T = T0 S
T
SQ q n = q0

• Temperature distribution T(x,y)

on a thin plate
y
• Temperature distribution Sh
over the cross-section of x
a nuclear reactor T¥

45
 Heat Balance Equation
• Conservation of energy qy y+
dy
2
Ein + Egenerated = Eout

(q )t dy
qx x-
dx qx x+
dx
x x-dx - qx x+
dx z
2
dy Qg 2
2 2
æ ö÷ dx
+çç qy dy - qy y +dy ÷÷ t zdx + Qg tzdxdy = 0
è y-
2 2 ø
qy y-
dy
2
é ¶qx æ dx ÷öù é ¶qx æ dx ÷öù ¶qx
qx x-dx - qx x +dx = ê qx + ç - ÷ ú - ê qx + ç ÷ ú =- dx
2 2 êë x ¶x è 2 ÷ø úû êë
ç x ç ÷
¶x è 2 ø úû ¶x

¶qy
qy dy - qy dy =- dy
y-
2
y+
2 ¶y

¶qx ¶qy
+ = Qg
¶x ¶y
46
 Constitutive Equation (Fourier’s Law)
• Fourier’s law: heat flux is proportional to the negative of
temperature gradient
¶T ¶T
q x = -k xx - k xy
¶x ¶y
¶T ¶T
qy = -k xy - k yy
¶x ¶y

– kxx, kxy, kyy: Thermal conductivities

– isotropic material: kxx = kyy = k and kxy = 0
ìï ¶T üï
ï ï
ìï q x ïüï
ï é k xx k xy ù ïï ¶x ïï
í ý = -[ k ] {T } = - ê úí ý ì
ï
ï qx ü
ï
ïïî q y ïïþ êë k yx ú
k yy û ïï ¶ T ï í ï ý = -k {T }
ïï ¶y ïïï ï
ï qy ï
î ï
þ
î þ
Conductivity matrix For isotropic material
We will only consider
isotropic material case
47
 Governing Differential Equation
• Fourier’s law  Conservation of energy
¶æ ¶T ¶T ö÷ ¶ æ ¶T ¶T ö÷
ç k xx + k xy ÷ ç
+ ç kyx + kyy ÷ + Qg = 0
ç
¶x è ¶x ÷
¶y ø ¶y è ¶x ¶y ø÷

 ⋅ ( [ k ] {T } ) + Qg = 0 2nd-order differential equation

• Boundary conditions
– Normal inward heat flux: qn = q ⋅ (-n) = ( k {T } ) ⋅ n

– Temperature boundary ST: T = T0

– Heat flux boundary SQ: q n = q0

– Convection boundary Sh: qn = h(T ¥ - T )

In 2D heat transfer, we need temperature T(x, y)

48
 2D Finite Element Interpolation
• Let 2D domain is discretized by 2D elements
• Commonly either triangular or rectangular shapes
• Then, FE interpolation is (x, y) coordinates
N
T( x, y ) = å Nk ( x, y )TK
k =1 y

– Tk: nodal temperature x

– Nk(x, y): interpolation function
y
3 T3

T1
1
T2
2
x 49
 2D Triangular Element
• Interpolation is a function of x and y coordinates
• Interpolation function is a three term polynomial in x and y
– Three nodal temperatures (T1, T2, T3) are available
y
T( x, y ) = a1+ a2 x + a3 y 3 T3

• Goal is to express in terms of T1

nodal temperature 1
T2
T( x, y ) = N1( x, y )T1 + N2 ( x, y )T2 + N3 ( x, y )T3
2
x
• Interpolation requirement (property)
ì
ïT( x , y ) º T = a + a x + a y
ï
ï
1 1 1 1 2 1 3 1
ï
íT( x2 , y 2 ) º T2 = a1 + a2 x2 + a3 y 2
ï
ï
ïT( x , y ) º T = a + a x + a y
ï
î 3 3 3 1 2 3 3 3

50
 Interpolation of Triangular Element
• In matrix notation
ì
ï T1 ü
ï é 1 x1 y1 ù ïì a1 ïü
ï
ïT ï ï ê ú ïï ïï
í 2ý = ê 1 x2 y 2 ú í a2 ý Is the matrix singular? When?
ï
ï ï
ï ê ú ïï ïï
ï
îT3 ïþ ëê 1 x3 y 3 ûú îï a3 ïþ
-1
ìï a1 üï é 1 x1 y1 ù ìï T1 üï é f1 f2 f3 ù ìï T1 üï
ïï ïï ê ú ïï ïï 1 ê ú ïï ïï
í a2 ý = ê 1 x2 y 2 ú íT2 ý = ê b1 b2 b3 ú íT2 ý
ï
ï ïï ê ú ïï ïï 2 A ê ú ïï ïï
ï
î a3 ïþ ëê 1 x3 y 3 úû ïîT3 ïþ êë c1 c2 c3 ûú ïîT3 ïþ

ì
ï f1 = x2 y 3 - x3 y 2 , b1 = y 2 - y 3 , c1 = x3 - x2
ï
ïf = x y - x y , b = y - y , c = x - x
í2 3 1 1 3 2 3 1 2 1 3
ï
ï
î f3 = x1y 2 - x2 y1, b3 = y1 - y 2 , c3 = x2 - x1
ï

1 x1 y1
1
A= det 1 x2 y2 Area
2
1 x3 y3
51
 Interpolation of Triangular Element
1
• Coefficients a1 = (f1T1 + f2T2 + f3T3 )
2A
1
a2 = (b1T1 + b2T2 + b3T3 )
2A
1
a3 = (c1T1 + c2T2 + c3T3 )
2A
• Interpolation equation
T( x, y ) = a1 + a2 x + a3 y
1
= [ (f1T1 + f2T2 + f3T3 ) + (b1T1 + b2T2 + b3T3 )x + (c1T1 + c2T2 + c3T3 )y ]
2A
1 N1(x,y)
y T
= (f1 + b1x + c1y ) T1 N1(x,y) 3 3
2A
1
+ (f + b2 x + c2 y ) T2
2A 2 N2(x,y)
1 T1
+ (f3 + b3 x + c3 y ) T3 1
2A N3(x,y) T2
2 x
52
 Interpolation of Triangular Element
• Temperature interpolation of triangular element
ìï T1 üï
3 ïï ïï
T( x, y ) = å Nk ( x, y )Tk = [N1 N2 N3 ] íT2 ý = êë N úû {T }
k =1 ïï ïï
ïîT3 ïþ

ìï 1
ïï N1( x, y ) = (f1 + b1x + c1y )
ïï 2A
ïï 1
í N2 ( x, y ) = (f2 + b2 x + c2 y )
ïï 2A Shape Function
ïï 1
ïï N3 ( x, y ) = (f + b3 x + c3 y )
ïî 2A 3

• N1, N2, and N3 are linear functions of x- and y-coordinates.

• Interpolated temperature changes linearly along the each
coordinate direction.
53
 Interpolation and Heat Flux
• Derivatives of temperature
¶ æç ö÷
3 3 3
¶T ¶N k bk
= çç å Nk ( x, y )Tk ÷÷ = å ¶x Tk = å 2 A Tk
¶x ¶x çè k =1 ÷ø k =1 k =1

¶ æç ö÷
3 3 3
¶T ¶N k ck
= çç å Nk ( x, y )Tk ÷÷ = å ¶y Tk = å 2 A Tk
¶y ¶y çè k =1 ÷ø k =1 k =1

ì
ï T1 ü
ï
ì
ï ¶T / ¶ x üï 1 é b1 b2 b3 ù ï ï
ï
{ T } = í ý= ê ú íT2 ïý = [B]{T }
ïîï ¶T / ¶y ïïþ 2 A êë c1 c2 c3 úû ïï ïï
ï
îT2 ïþ
• Heat flux

ìïï q x üïï
í ý = -k {T } = - k[B]{T }
ïîï q y ïþï

– Heat flux is constant within an element

54
 Exercise
• Build shape functions of the triangular element shown in the
figure and express functional expression of temperature and
heat flux (qx, qy). Use T1 = 50, T2 = 80, T3 = 30 oC, t = 1 m, k =
10 W/m/oC. y

(0,1)
3

x
1 (0,0) 2 (1,0)

55
 Galerkin Method for Heat Conduction
• 2D weighted residual form:

òòA R( x, y )W ( x, y )t dA = 0
é æ ¶2T ¶2T ö÷ ù
òòA êêë k çççè ¶x 2 + ¶y 2 ÷÷ø + Qg úúû fk t dA = 0 k: trial function
k = 1,, N

òòA T + Qg )fk t dA = 0
( k 2

• Green’s theorem (integration by parts) S

A
òòA (2u )v dA = òS (u ⋅ n)v dS - òòA (u ⋅ v )dA
n
unit normal vector
56
 Galerkin Method for Heat Conduction
• Apply Green’s theorem to the Galerkin’s method

òòA T + Qg )fk t dA = 0
( k 2

òS k (T ⋅ n)fk t dS - òò
A
( k T ⋅ fk ) t dA + òò Qg fk t dA = 0
A

– Apply boundary heat flux and rearrange

òòA ( kT ⋅ fk ) t dA = òòA Qg fk t dA + òS qnfk t dS

– Divide the boundary into S = ST È SQ È Sh

òS qnfk t dS = òS T
qnfk t dS + ò qnfk t dS + ò qnfk t dS
SQ Sh

= òS
Q
q0fk t dS + ò h(T ¥ - T )fk t dS
Sh

òòA ( kT ⋅ fk ) t dA = òòA Qg fk t dA

+ò q0fk t dS + ò h(T ¥ - T )fk t dS
S Q S h 57
 Finite Element Formulation
• Consider triangular element A(e) and shape functions Nk(x, y)
• LHS (conductivity part) with k = 1, 2, 3
æ é N1 ù ÷ ö
çç ê ú ÷÷ é ù {T }
òòA(e ) çç ê 2 ú
ç k ê N ú ⋅ T ÷
÷÷ t dA = êë òòA( e ) ( k [B]T
[B] ) t dA úû
çè ê N ú ÷ø
ë 3û constant
[k(e)] element stiffness matrix

[k( e ) ] = ktA[B]T [B]

é b1 c1 ù
1 ê ú 1 é b1 b2 b3 ù
= ktA ê b2 c2 ú ⋅ ê ú
2A ê ú 2 A êë c1 c2 c3 úû
êë b3 c3 úû
é b12 + c12 b1b2 + c1c2 b1b3 + c1c3 ù
kt êê 2 2
ú
ú
= b b + c c b + c b b + c c
4A ê 1 2 1 2 2 2 2 3 2 3 ú
êb b + c c b b + c c b32 + c32 úúû
êë 1 3 1 3 2 3 2 3

58
 Finite Element Formulation
• Load due to distributed heat source: k = 1, 2, 3
– Ex) heat generated due to electrical current flow
ì
ï N1 ü
ï ì
ï Qg1 üï
ï ï ï
ï ï
ï
ï ï
òòA(e ) g ï 2 ï
Q í N ý t d A = í
ï
Qg2 ý
ï
= {Q (e )
g }
ï
ï ï ï ï
î N3 ï
þ ï
ï
î Qg 3 ïï
þ

1 (e )
• Uniform heat generation: Qg = constant òA( e ) Nk dA = 3 A
ì
ï N1 ü
ï ì
ï 1ü
ï
ï
ï ï
ï 1 ï
(e ) ï ï
ï
{Q g } = òò ( e ) Qg í N2 ý t dA = Qg A t í1ý
(e )
A ï
ï ï
ï 3 ï
ï ï
ï
ï
î N3 ï
þ î1ï
ï þ

• This nodal heat sources should be assembled to the global

heat source vector
59
 Finite Element Formulation
• Load due to distributed heat flux on SQ
– If element (e) is within a domain, there is no SQ boundary
– Only elements on the boundary may have a part of SQ boundary
– Let’s assume edge 1-2 of element (e) is on the heat flux boundary
– N3 = 0 along edge 1-2
ìï N1 ü ì 1 üï
ï (e ) ï
ï ï q0tL ï ïï ï 1 (e )
{QQ } = ò ( e ) q0 í N2 ï
(e) ï
ý t dS
(e)
{QQ } = í 1ý òS( e ) k
N d S = L
SQ ïï ï 2 ïï ïï 2
ïî 0 ï
ï
þ ïî 0 ïþ
k = 1, 2
Uniform heat flux
L(e): length of edge 1-2
q0 y
3 T3

T1 Element (e)
1
T2
q0 2
x 60
 Finite Element Formulation
• Convection boundary condition
– Let’s assume edge 1-2 of element (e) is on the convection boundary
– N3 = 0 along edge 1-2
– Same as 1D convection, stiffness and heat load terms

ì
ï N1 ü
ï ì 1ü
ï ï (e ) ¥ ï ï
ï ï htL T ï ï ï ï
{Q h } = htT ò í N2 ý dS =
(e ) ¥
í 1ý
Sh ï ï 2 ï ï
ï ï ï ï
î0 ï
ï þ î0ï
ï þ

ì
ï N1 ü
ï ì
ï N1 ü
ï
ï ï ï ï
ht ò ( e ) T í N2 ý dS = ht ò ( e ) í N2 ï
ï ï ï
ý êë N1 N2 0 úû dS { T }
Sh ï
ï ï
ï Sh ï
ï ï
ï
î0 ï
ï þ î0ï
ï þ

é2 1 0ù
(e )
htL ê ú Add to the heat conduction matrix
[kT(e ) ] = ê 1 2 0ú
6 ê ú
êë 0 0 0 úû
61
 Finite Element Equation for Heat Transfer
• Element equation
– With constant heat source, heat flux on Edge 1-2, and convection on
Edge 1-2

é é b12 + c12 b1b2 + c1c2 b1b3 + c1c3 ù é 2 1 0 ù ù ïì T1 ïü

ê kt ê ú htL( e ) ê ú úï ï
ê ê b1b2 + c1c2 b22 + c22 ú
b2b3 + c2c3 ú + ê 1 2 0 ú íT2 ï
ú ï
ý
ê 4A ê 6 ê ú ú ï ï
ê êb b + c c b2b3 + c2c3 2 2
b3 + c3 úûú ê 0 0 0 ú ú ï
ïT ï
ï
êë êë 1 3 1 3 ë û úû î þ3

ìï1üï ì 1ü ì 1ü
ï ï (e ) ï ï (e ) ¥ ï ï
1 (e ) ï ï q0tL ï ï htL T ï ï
ï ï ï ï
= Qg A t í1ý + í 1ý + í 1ý
3 ïï ïï 2 ïï ï ï 2 ï
ï ï
ï
ïî1ïþ ïî 0 ï
þ ï 0
î ï þ

62
 Ex) Heat Transfer along a Conducting Block
• Determine temperature 10mm
distribution using
h
triangular elements
q0
5mm Qg
T¥
• Thickness = 5 mm,
k = 0.2 W/mm/oC,
Qg = 0.06 W/mm3 Insulated
Heat flux boundary Convection boundary
• q0 = 0.04 W/mm2, (0, 5) (5, 5) (10, 5)
1
h = 0.012 W/mm2/oC, 3 5
T∞ = 25oC

4 6
2
(0, 0) (5, 0) (10, 0)
63
 Ex) Heat Transfer along a Conducting Block
Element LN1 LN2 LN3
• Element connectivity table
1 1 2 4
2 1 4 3
• Element 1 3 3 4 6
b1 = y 2 - y 3 = 0, c1 = x3 - x2 = 5 4 3 6 5
b2 = y 3 - y1 = -5, c2 = x1 - x3 = -5
b3 = y1 - y 2 = 5, c3 = x2 - x1 = 0

1 x1 y1
1 25 Heat flux boundary Convection boundary
(e )
A = det 1 x2 y2 =
2 2 (0, 5) (5, 5) (10, 5)
1 x3 y3 1
3 5

E2 E4
1 é b1 b2 b3 ù
[B] = ê ú
2 A ë c1 c2 c3 úû
(e ) ê
E1 E3
1 é 0 -5 5 ù 4 6
= ê ú 2
25 êë 5 -5 0 úû (0, 0) (5, 0) (10, 0)
64
 Ex) Heat Transfer along a Conducting Block
• Element stiffness matrix (E1 and E3)
é 1 -1 0 ù
ê ú
[ K1 ] = [ K3 ] = ktA [B] [B] = 0.5 ê -1 2 -1ú
(e ) T
ê ú
êë 0 -1 1 úû

• Element stiffness matrix (E2 and E4)

é 1 0 -1ù
ê ú
[ K2 ] = [ K 4 ] = ktA [B] [B] = 0.5 ê 0 1 -1ú
(e ) T
ê ú
êë -1 -1 2 úû

• Heat generation (for all elements)

ìï1ü
ï ì
ï 1ü
ï ì
ï 1ü
ï
1 ï ï 1 æ 25 ÷ö ï ï ï ï
{Q g } = Qg A t í1ý = ⋅ ( 0.06 ) ⋅ çç ÷÷ ⋅ ( 5 )í1ý = 1.25 í1ýï
(e ) (e ) ï ï ï ï ï
3 ïï ï ï 3 è 2 ø ï
ï ï
ï ï
ï ï
ï
ïî1ïþ î1ï
ï þ î1ï
ï þ

65
 Ex) Heat Transfer along a Conducting Block
• Heat flux (along Edge 1-2 of Element 1)
ì 1 üï ì 1ü ì 1 üï
(1) ï ï ï ï
(1) q0tL ï ïï 0.04 ⋅ 5 ⋅ 5 ï
ï ï ï
ï ï
ï ïï
{QQ } = í 1ý = í 1 ý = 0.5 í 1 ý
2 ïï ïï 2 ï
ï ï
ï ï
ï ïï
ïî 0 ïþ î0ï
ï þ î 0 ïþ
ï

• Convection (along Edge 5-6 of Element 4)

ì
ï 0ü
ï ì
ï 0ü
ï ìï 0 üï
htL T (4) ¥ ï
ï 1ï
ï = 0.012 ⋅ 5 ⋅ 5 ⋅ 25 ï
ï 1ï
ï = ïï 3.75 ïï
{Q(4)
h } = í ý í ý í ý
2 ï
ï ï
ï 2 ï
ï ï
ï ï
ï ïï
î 1ï
ï þ î 1ï
ï þ ïî 3.75 ïþ
é0 0 0ù é0 0 0ù é0 0 0ù
(4)
htL ê ú 0.012 ⋅ 5 ⋅ 5 ê ú ê ú
[kT(e ) ] = ê 0 2 1ú = ê 0 2 1 ú = 0.05 ê 0 2 1 ú
6 ê ú 6 ê ú ê ú
êë 0 1 2 úû êë 0 1 2 úû êë 0 1 2 úû

66
 Ex) Heat Transfer along a Conducting Block
• Assembly
é 1 + 1 -1 -1 0 0 0 ùìï T1 ü
ï ìï1 + 1 + 0.4 ü
ï
ê ï
úï ï ï ïï ï
ï
ê -1 2 0 -1 0 0 ú ïT2 ï ï 1 + 0.4 ï
ê úï ï ï
ï ï
ï ï
ï
ê -1 0 2 + 1 + 1 -1 - 1 -1 0 ú ïïT3 ïï ï
ï 1 + 1 + 1 ï
ï
0.5 ê ú í ý = 1.25 í ý
ê 0 -1 -1 - 1 1 + 2 + 1 0 -1 ú ïT4 ï ï 1 + 1 + 1 ï
ê úïï ï
ï ï
ï ï
ï
ê 0 0 - 1 0 2 + 0.2 - 1 + 0.1 ú ï
ï T ï
ï ï
ï 1 + 3 ï
ï
ê úï ï 5
ï ï
êë 0 0 0 -1 ú ï
-1 + 0.1 1 + 1 + 0.2 û ï ï
ï ï
ï 1 + 1 + 3 ï
ï
T
î 6þ î þ

• E1: E2: E3: E4:

• Solution: (no essential BC)
{T1 T2 T3 T4 T5 T6 } = {95.75 94.92 90.59 90.58 77.93 78.75}
• Heat flux Edge 1-2 (constant for Element 1)
ì
ï T1 ü
ï ì
ï 95.75 ü
ï
ìïï q x üïï -k é b1 b2 ù ï ï
b3 ï ï -0.2 é 0 -5 5 ù ï ï ïï ìï 0.174 ïü
q=í ý= ê ú íT2 ý = ê ú í 94.92 ý = í ý
ïîï qy ïþï 2 Ae ëê c1 c2 ï
c3 ûú ï ï ï 25 ï
ê 5 -5 0 ûú ï
ë ï
ï ï -0.0332 ïïþ
ï
î
ï
îT4 ïþ î 90.58 ï
ï þ 67
 Exercise
• When temperature T1 = 20oC and heat generation Qg = 12
W/m3, calculate the temperature T2 and T3 as well as heat flux
at Node 1. Use k = 0.1 W/m/oC and t = 0.1 m.

(0,1)
3

x
1 (0,0) 2 (1,0)

68
The Elimination Method in Finite Element Method (FEM)

The elimination method in the Finite Element Method (FEM) is an approach to simplify and solve

systems of equations that arise from discretizing a continuum problem. In FEM, physical problems

(like structural analysis, fluid dynamics, heat transfer, etc.) are broken down into smaller, simpler

parts called finite elements. After assembling these elements, we obtain a global system of linear

equations. The elimination method helps in reducing and solving this system by systematically

removing or eliminating variables, often in cases with boundary conditions or sparse matrices.

Key Concepts

1. Finite Element Discretization:

- Divide the domain of the problem into finite elements (e.g., triangles, quadrilaterals for 2D

problems or tetrahedrons, hexahedrons for 3D).

- Formulate local stiffness matrices for each element based on material properties and geometry.

- Assemble these local matrices into a global stiffness matrix representing the entire domain.

2. Formulating the Global System:

- Once the elements are assembled, we have a large system of equations that represent the

relationships between nodes, usually expressed as [K]{u} = {F}, where:

- [K] is the global stiffness matrix.

- {u} is the displacement (or other field variable) vector.

- {F} is the force (or load) vector.

3. Boundary Conditions:
 - Boundary conditions (Dirichlet or Neumann) are applied to specific nodes or edges of the

system.

- For Dirichlet conditions, known values (e.g., fixed displacements) are directly applied to nodes,

which changes the system by eliminating certain degrees of freedom.

- Neumann conditions (e.g., known forces) are incorporated into the {F} vector.

4. Elimination Process:

- In the elimination method, unknowns associated with the boundary conditions are removed from

the system of equations to reduce the size and complexity of the matrix.

- For Dirichlet boundary conditions, the rows and columns corresponding to fixed nodes can be

eliminated or set to represent the constraints, effectively reducing the matrix size.

5. Solving the Reduced System:

- After elimination, a reduced system of equations is obtained.

- This smaller matrix system can then be solved using numerical methods like Gaussian

elimination, LU decomposition, or iterative solvers (e.g., conjugate gradient method for sparse

matrices).

6. Back Substitution (if necessary):

- In some cases, especially with Neumann boundary conditions, the solution obtained from the

reduced system may be substituted back to compute reactions or other quantities of interest at the

boundary nodes.

Applications

The elimination method is especially useful in problems with large sparse matrices (like those in
FEM) where storage and computation can be optimized by reducing the matrix size. This method is

commonly used in structural analysis, thermal analysis, and other applications where FEM is

applied, making the solution process more efficient and manageable.

Advantages and Limitations

- Advantages:

- Reduces computational cost by working with a smaller system.

- Helps manage large systems typical of FEM applications by handling boundary conditions

directly.

- Limitations:

- Not always ideal for very large or complex systems where iterative solvers might be more

efficient.

- Elimination might lead to loss of sparsity in the matrix, depending on the problem's constraints.

In summary, the elimination method is an effective approach to simplifying FEM problems with

specific boundary conditions, enhancing computational efficiency and making it easier to solve large

systems of equations encountered in finite element analysis.

The Method of Factorization for Solving Simultaneous Algebraic Equations

The method of factorization (or LU decomposition) is a technique used to solve systems of

simultaneous algebraic equations,

particularly useful for linear equations in matrix form. It breaks down a complex matrix equation into

simpler parts, which

can be solved more efficiently, especially in large-scale computations. Factorization methods are

commonly used in engineering

and scientific applications, such as finite element analysis, where large systems of equations often

need to be solved.

Basic Concept

In matrix terms, solving a system of linear equations means finding a solution vector x for the

equation:

Ax=b

where:

- A is a square coefficient matrix of size n x n.

- x is the unknown vector of size n x 1.

- b is a known vector of size n x 1.

Factorization helps in breaking down the matrix A into simpler matrices. Once decomposed, it

becomes easier to solve for x.

LU Decomposition

The most common type of factorization in solving algebraic equations is LU decomposition, where

the matrix A is factored into two matrices, L and U:

A = LU

where:

- L is a lower triangular matrix with ones on the diagonal (only elements below the diagonal are

non-zero).

- U is an upper triangular matrix (only elements on and above the diagonal are non-zero).

Steps to Solve Using LU Decomposition

1. Decompose the Matrix A into L and U:

- Find matrices L and U such that A = LU.

- This step involves manipulating the rows of A to zero out elements below the diagonal, forming L

and U.

2. Set Up Two Sub-Systems:

- Substitute A = LU in the original equation:

LU x = b

- Let y = U x. This changes the equation to:

Ly=b

- We now have two systems of equations to solve:

 - L y = b (forward substitution).

- U x = y (back substitution).

3. Solve for y Using Forward Substitution:

- Solve L y = b by forward substitution, starting from the top row, since L is lower triangular.

4. Solve for x Using Back Substitution:

- Once y is determined, substitute it into U x = y and solve for x using back substitution, starting

from the last row, since U is upper triangular.

Example

Consider the system of equations:

2x + 3y + z = 5

4x + y + 2z = 6

3x + 2y + 4z = 7

The coefficient matrix A and vector b are:

A = [ [2, 3, 1], [4, 1, 2], [3, 2, 4] ], b = [5, 6, 7]

Decompose A into L and U, then solve for y and x as described.

Advantages of Factorization
- Efficiency: Once L and U are determined, the system can be solved multiple times for different b

vectors, making it efficient for iterative problems.

- Simplicity: Breaking down the problem into forward and backward substitution steps simplifies the

solution process, especially for large matrices.

Applications

The factorization method is widely used in engineering fields, where multiple systems of linear

equations must be solved,

such as in structural analysis, fluid mechanics, and electrical circuit analysis.

Summary

The factorization method is a powerful approach for solving simultaneous algebraic equations. By

decomposing the coefficient

matrix into lower and upper triangular matrices, it simplifies solving large systems of equations,

making it a cornerstone method in computational mathematics and engineering.