TIME-DEPENDENT HEAT TRANSFER
Edward Tefft
ME 5180
12/2/2016
Outline of Presentation
I. Refresher on heat transfer concepts
II. Finite element application
III. Numerical time integration
IV. Example: MATLAB 1-D
V. Example: ANSYS 2-D
Concepts
Types of Heat Transfer
• What is heat flux again?𝑞 (W/m2) – heat transfer rate per perpendicular area
• Conduction:
𝑞" = −𝐾𝑑𝑇𝑑𝑥
• Convection:𝑞) = ℎ(𝑇, − 𝑇∞ )Heat transfer illustration [1]
𝜎" = 𝐸𝑑𝑢𝑑𝑥
analogous to:
1-D Heat Transfer Equation
• For the amount of energy in a volume:𝐸in + 𝐸generated = Δ𝑈 + 𝐸out
• For a system with: Heat IN: conduction, generationHeat OUT: convection
𝜕𝜕𝑥 𝐾
𝜕𝑇𝜕𝑥 + 𝑄 = 𝜌𝑐
𝜕𝑇𝜕𝑡 +
ℎ𝑃𝐴 (𝑇, − 𝑇∞ )
Change of energy stored in volume (time dependent!!)
Δ𝑈 = specificheat×mass×changeintemperature
Outline of Presentation
I. Refresher on heat transfer concepts
II. Finite element application
III. Numerical time integration
IV. Example: MATLAB 1-D
V. Example: ANSYS 2-D
1-D Finite Element Equation
𝐾 𝑇 + 𝐶 �̇� = 𝐹
Thermal conductivity
matrix
Heat capacity matrix
Thermal loads
Considered for an axial bar element undergoing conduction and convection.
𝑇 𝑥 = 𝑁R 𝑥 𝑇R + 𝑁S 𝑥 𝑇S
𝑁R = 1 −𝑥𝐿
𝑁S =𝑥𝐿
movingfluid
T1
x, T(x)
T2
h, TA, k, ρ
8
L
Thermal Conductivity Matrix
𝐾 𝑇 + 𝐶 �̇� = 𝐹
Thermal conductivity
matrix
𝐾 =𝐴𝑘𝐿
1 −1−1 1 +
ℎ𝑃𝐿6
2 11 2
Conduction Part of the convection equation
𝑞" = −𝐾𝑑𝑇𝑑𝑥
𝑞) = ℎ(𝑇, − 𝑇∞ )
Heat Capacity Matrix
𝐾 𝑇 + 𝐶 �̇� = 𝐹
Heat capacity matrix 𝐾 =
𝐴𝑘𝐿
1 −1−1 1 +
ℎ𝑃𝐿6
2 11 2
So far…
𝐶 =𝑐𝜌𝐴𝐿6
2 11 2
𝐶 =𝑐𝜌𝐴𝐿2
1 00 1
This is analogous to the mass matrix!
Consistent-heat capacity matrix• virtual work principle• uses shape functions
Lumped-heat capacity matrix• assumes all of the mass is
lumped at the nodes• generally not as accurate [2]
Thermal Loads
𝐾 𝑇 + 𝐶 �̇� = 𝐹
Thermal loads𝐾 =
𝐴𝑘𝐿
1 −1−1 1 +
ℎ𝑃𝐿6
2 11 2
So far…
𝐶 =𝑐𝜌𝐴𝐿6
2 11 2
𝐹 = 𝑄𝐴𝐿2
11 + 𝑞
𝑃𝐿2
11 + ℎ𝑇∞
𝑃𝐿2
11
Internal heat
generation
Heat flow into lateral
surface
Part of the convection equation
𝑞) = ℎ(𝑇, − 𝑇∞ )
Outline of Presentation
I. Refresher on heat transfer concepts
II. Finite element application
III. Numerical time integration
IV. Example: MATLAB 1-D
V. Example: ANSYS 2-D
Numerical Time Integration
Galerkin method:
1. Rewrite the equation to this form:
2. Givenaknown 𝑇_ at 𝑡 = 0 and a time step of Δ𝑡, solve for 𝑇Rat 𝑡 = Δ𝑡
3. Using 𝑇R , find 𝑇S at 𝑡 = 2(Δ𝑡) and so on
=1Δ𝑡
𝑀 −13𝐾 𝑇b +
13𝐹b +
23𝐹bcR
1Δ𝑡
𝑀 +23𝐾 𝑇bcR
𝐴 𝑇bcR = 𝐵 𝑇b + 𝐶
Outline of Presentation
I. Refresher on heat transfer concepts
II. Finite element application
III. Numerical time integration
IV. Example: MATLAB 1-D
V. Example: ANSYS 2-D
Example: MATLAB 1-D
Example 16.7from A First Course in the Finite Element Method by D.L. Logan
1. Find 2-element solution
2. Find 4-element solution
1 2 3 Insulated tip𝑇∞ = 25℃85℃
Results
Outline of Presentation
I. Refresher on heat transfer concepts
II. Finite element application
III. Numerical time integration
IV. Example: MATLAB 1-D
V. Example: ANSYS 2-D
Example: ANSYS 2-D
Example 13.6from A First Course in the Finite Element Method by D.L. Logan
Extended to include transients – assumed similar to lead
𝑇 = 100℉
4
2 ft
2 ft
ℎ = 20Btu
hftS℉
𝑇∞ = 50℉
1 2
3
5
x
y𝐾 = 25
Btuhft℉
𝑐j = 0.03Btulbm℉
𝜌 = 709lbmfto
Results
ANSYS results:𝑇S = 69.23℉𝑇o = 69.23℉𝑇p = 84.62℉
Book results:𝑇S = 69.33℉𝑇o = 69.33℉𝑇p = 84.62℉
References
[1] http://www.tutorvista.com/physics/example-of-convection-heat-transfer
[2] Archer, J.S., “Consistent Matrix Formulations for Structural Analysis Using Finite Element Techniques,” Journal of the American Institute of Aeronautics and Astronautics, Vol. 3, No. 10, pp. 1910-1918, 1965.