Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
0% found this document useful (0 votes)
356 views

Notes For Finite Element Method Thermal Analysis

The document provides notes on using the finite element method for thermal analysis. It discusses the governing differential equations for 1D and 2D heat conduction, as well as approaches for modeling temperature-dependent material properties and resulting thermal stresses. Key steps of the finite element method are outlined, including defining shape functions to approximate temperature over an element, assembling the global conductivity matrix, and applying boundary conditions. Example problems are given to illustrate applying the method to simple 1D heat conduction problems.
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
356 views

Notes For Finite Element Method Thermal Analysis

The document provides notes on using the finite element method for thermal analysis. It discusses the governing differential equations for 1D and 2D heat conduction, as well as approaches for modeling temperature-dependent material properties and resulting thermal stresses. Key steps of the finite element method are outlined, including defining shape functions to approximate temperature over an element, assembling the global conductivity matrix, and applying boundary conditions. Example problems are given to illustrate applying the method to simple 1D heat conduction problems.
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 8

M.

Vable

Notes for finite element method: Thermal Analysis

One-dimensional heat conduction


Axial Member Differential Equation d du EA = p x dx d x EA= Axial Rigidity px= force per unit length Heat Conduction d dT k = qx d x d x k= Thermal conductivity qx= Heat flow (source) per unit length. fx=heat flux per unit area= qxA T= Temperature dT dx Positive: Flow into body Heat Capacity: Heat flow: q = k 1 dT 2 - U T = -- k dx 2 dx
L L m

Primary variable Secondary variable Energy

u= Displacement in x-direction Internal axial force: N = EA Strain Energy: 1 du 2 U = -- EA dx d x 2


L

du dx

Work Potential:
L m

W A = p x u dx +
0

Fq u ( xq )
q=1

WT =

qx T dx + Qq T ( xq )
0 q=1

Functional

Potential Energy A = UA WA
n

T = UT WT
n

Rayleigh-Ritz u( x) = Matrix K jk Right Hand Side Vector


L

Ci fi ( x )
i=1

T(x) =

Ci fi ( x )
i=1 L

df j df k = EA dx d x d x
0 L m

K jk =
L

k d x d x dx
0 m

df j df k

Rj =

px fj
0

dx +

Fq fj ( xq )
q=1

R j = q x f j dx +
0

Qq fj ( xq )
q=1

Other Applications:
Flow through pipes; Flow through porous media; Electrostatics By non-dimensionalizing the problem a software can be used to solve all the above applications.

M. Vable

Notes for finite element method: Thermal Analysis

Linear Elements:
T = T1 L1 ( x ) + T2 L2 ( x ) [K
(e) (e) (e)

Q1
(e)

(e)

Q2
(e)

(e)

k ] = --------) 1 1 (e L 1 1
(e) (e)

{R

(e) (e) (e) qx L 1 Q1 } = ----------------- + 2 1 (e) Q 2 (e) (e) (e) fx L 1 Q1 } = ---------------- + A 2 1 Q( e ) 2


(e)

[K

(e)

k A ] = ------------------ 1 1 (e) 1 1 L

{R

(e)

Quadratic Elements:
T = T1 L1 ( x ) + T2 L2 ( x ) + T3 L3 ( x ) 7 8 1 k (e) [ K ] = ------------ 8 16 8 (e) 3L 1 8 7
(e) (e) (e) (e)

Q1

(e) (e)

Q3
(e) (e) 1 qx L

Q1 (e) { R } = ----------------- 4 + 0 6 (e) 1 Q3

Class Problem 1
Heat qo is being added at point B at a constant rate as shown in Fig. 1. Using two linear elements determine the temperature at point B and the heat flowing out at A and C in terms of k, L, and qo for the following two cases: (a) The ends of the bar are maintained at a constant zero temperature. (b) The ends of the bars are maintained at a constant temperature To.
qo A L B C 2L A L 3qo B L qo C L D

Fig .1

Fig .2

Home Problem 1
Heat 3qo is being added at point B and qo is being taken out at a constant rate at point C as shown in Fig. 2. Using three linear elements determine the temperature at points B and C and the heat flowing out at A and D in terms of k, L, and qo for the following two cases: (a) The ends of the bar are maintained at a constant zero temperature. (b) The ends of the bars A and D are maintained at a temperature To and 2To, respectively. ANS: TB = ( 5qo L ) 3k T C = ( q o L ) 3k q A = ( 5q o 3 )

M. Vable

Notes for finite element method: Thermal Analysis

Thermal Stresses
Stress-strain curve with temperature effects
Normal Stress

O T

O1

-E

Normal Strain

= -- + T = -- + o E E
where, = linear coefficient of thermal expansion. o = Initial strain= Thermal strain unconstrained body due to uniform temperature changes.

No thermal stresses are produced in a homogenous, isotropic,


Axial Problem
We assume that the thermal problem and stress analysis problem can be solved independently.
xx = E ( xx o )
L

U AT

1 = -- xx ( xx o ) dV = 2
V L L

-- E ( xx o ) 2
0 L

A dx or
L L

U AT

1 1 du 2 du 1 2 2 = -- EA xx dx EA xx o dx + -- EA o dx = -- EA dx EA o dx + U o d x d x 2 2 2
0 L 0 m 0 0 0

WA =

px ( x )u ( x ) dx + Fq u ( xq )
0 L q=1

A = U AT W A

1 du 2 du = -- EA dx EA o dx + U o d x d x 2
0 0

px ( x )u ( x ) dx + Fq u ( xq )
0 q=1

M. Vable

Notes for finite element method: Thermal Analysis

1 du 2 = -- EA dx + U o d x 2
0 L m

px ( x )u ( x ) dx + Fq u ( xq ) + EAo d x dx
du
0 L q=1 0

= U A W AT

W AT =
L

px ( x )u ( x ) dx + Fq u ( xq ) + EAo d x dx
du
0 q=1 L L 0 0

du I = EA o dx = EA o u ( x ) d x
0

d o EA u dx = EATu ( x ) dx
0 L

L L 0

d EA T u dx dx
0

EA dT I = EAT L u ( L ) EAT 0 u ( 0 ) + ----------- q x u dx , where q x = k k dx


0 L m

W AT =

[ px ( x ) + pxT ( x ) ]u ( x ) dx + Fq u ( xq ) + FT2 u ( L ) + FT1 u ( 0 )


0 q=1

where,

F T1 EAT 1 = F T2 EAT 2 qx = 0

EA p xT = ----------- q x k

Special case: Constant Temperature Change


T = Cons tan t

Thermal loads are added only at the element ends. Linear Element:
[K
(e)

EA ] = ------- 1 1 L 1 1

{R

(e)

(e) (e) p o L 1 p To L 1 F 1 F T1 } = -------- + ----------- + + 2 1 2 1 (e) (e) F 2 F T2

Quadratic Element:
(e)

[K

7 8 1 EA ] = ------- 8 16 8 3L 1 8 7

F(e ) 1 1 1 p o L p To L { R } = -------- 4 + ----------- 4 + 0 6 6 1 1 F(e ) 3

F( e ) T1 0 + F( e ) T3

M. Vable

Notes for finite element method: Thermal Analysis

Class Problem 2
Heat qo is being added at point B at a constant rate as shown in Fig. 3. Using two quadratic elements determine the displacement of node B, reaction force at A and the axial stress just before B in terms of E, A, , k, L, and qo. Assume that the entire bar was at zero temperature and the ends of the bars are maintained at zero temperature.
qo A L B C 2L A L 3qo B L qo C L D

Fig .3

Fig .4

Home Problem 2
Heat 3qo is being added at point B and qo is being taken out at a constant rate at point C as shown in Fig. 4. Using three linear elements determine the displacements of points B and C, the reaction force at A, and the axial stress just before B in terms of E, A, , k, L, and qo. Assume that the entire bar was at zero temperature and the ends of the bars are maintained at zero temperature. ANS: u B
5q o L = -----------------18k
2

uC

7q o L = -----------------18k

5EAq o L R A = -----------------------9k

M. Vable

Notes for finite element method: Thermal Analysis

2-D Steady State Thermal Analysis


Differential Equation:
T T k 2 + 2 = qv x y
2 2

or

k T = q v

Functional (Stored Heat):


UT = T T k x + y
2 2

t dx dy

where t= thickness of the body.


T T x x U T = [ k ] t dx dy T A T y y
T

Isotropic material:

[k] = k

---scaler quantity.

Element Approximation:
(e) T1 n T(e ) 2 (e) T ( x ) = T i f i ( x, y ) = f 1 f 2 f n i=1 (e) Tn

fi(x,y) are Lagrange Polynomials.


T f 1 f 2 x x x = T f 1 f 2 y y y f 3 f n x x f 3 f n y y (e) T1 (e) T2 (e) Tn (e) (e) = [B ]{d }

UT =

-- { d 2

(e) T

} [B

(e) T

] [k

(e)

][B

(e)

]{d

(e)

1 (e) T (e) (e) }t dx dy = -- { d } [ K T ] { d } 2

M. Vable

Notes for finite element method: Thermal Analysis

Element Conductivity Matrix


[ KT ] =
(e)

[B
A

(e) T

] [k

(e)

][B

(e)

]t dx dy

Heat Conduction Boundary Conditions


k T T T = k n x + n y = q n x n y

where,

n= direction of the unit normal to the boundary. nx, ny = direction cosines of the unit normal to the boundary. qn = specified heat flow in the n-direction on the boundary.

Right Hand Side Vector:


(e) { R } = qv (e) A f1 f2 t dx dy + q n (e) fn f1 f2 t ds fn

where,

= the boundary of the element. s= tangential coordinate along the element boundary.
T = h ( Tf T ) n

(e)

Convection Boundary Conditions


k

where,

h = convection heat transfer coefficients Tf = Temperature of the surrounding fluid h depends upon many factors: velocity of fluid, viscosity of fluid, density of fluid, and other properties of fluid. It also depends upon the surface roughness and surface geometry. Addition to Element Matrix: Addition to Element Right Hand Side Vector:
K ij = Ri
(e) (e)

(e)

f i f j t ds f i T f t ds

(e)

(e)

(e)

If

is an element boundary in the interior, then there is no convection there and hence no addition to the matrix or the element RHS vector.

(e)

M. Vable

Notes for finite element method: Thermal Analysis

Radiation Boundary Condition:


Heat radiated is proportional to the difference in the fourth power of temperature between the radiating bodies.
k T 4 4 = BC ( Tr T ) n

where,

BC is the proportionality constant. Tr is the temperature of the other radiating body. Temperatures Tr and T are in absolute degrees i.e., oK.

For two infinite parallel black bodies (planes) it is called the Boltzmann
constant. For regular bodies BC depends upon the emissivity of the bodies, the geometry and other factors. Radiation boundary conditions lead to non-linear thermal problem. A general approach is:
k T 2 2 = BC ( Tr + T ) ( Tr + T ) ( Tr T ) = hr ( Tr T ) n

In the iteration process, at each step treat the radiation boundary condition like a convection term with coefficient dependent upon the temperature at a particular step.

You might also like