Sei sulla pagina 1di 15

COMPUTER ASSIGNMENT #1

STEADY 2-D HEAT CONDUCTION






Submitted by:

CARA B. MCLAUGHLIN




AEROSPACE AND OCEAN ENGINEERING DEPARTMENT

VIRGINIA POLYTECHNIC INSTITUTE AND STATE UNIVERSITY

BLACKSBURG, VIRGINIA

16 SEPTEMBER 2014





AOE 3044: BOUNDARY LAYER THEORY

COURSE INSTRUCTOR: PROFESSOR JOSEPH SCHETZ






Honor Pledge:
By electronically submitting this report I pledge that I have neither given nor received
unauthorized assistance on this assignment.
______905603903______ ________09/16/14_____ __
Student Number Date



1. PROBLEM DESCRIPTION
The objective of this assignment is to analyze the relationship between the temperature
gradient and area during two-dimensional heat flow. A square body is subjected to various
temperatures on the edges, then divided into nodes in order to determine the temperature at any
position on the material. Figure 1 depicts the exact scenario being studied in this problem.





Figure 1. Four node problem from Figure 3-6 of Heat Transfer by J.P. Holman.

The accuracy of steady and unsteady methods are analyzed through the use of the Heat
Conduction Applets developed by William J. Devenport, Joseph A. Schetz, and Yu Wang in
1998. The code is specifically designed to be used by students to allow for easy grasping of the
concepts behind the calculations.
For the steady-state condition, assuming constant thermal conductivity, the Laplace
equation applies:

(1)
Where the solution, which gives the temperature in a two-dimensional body as a function of the
two independent space coordinates x and y, may be obtained using analytical (in the form of the
Fourier series), numerical, or graphing methods. The total heat flow at any point in the material
is the resultant of the heat flow in the x-direction and the y-direction at that specific point. Its
vector is perpendicular to the lines of constant temperature. For this problem, the Gauss-Seidel
method of numerical analysis iteration is used by the applet to determine the temperature
gradients. The square-shape is divided into equal increments in the x and y directions, and finite
differences are used to approximate differential increments in the temperature and space
coordinates. Since x = y, the net heat flow can be expressed as:



(2)
Where the heat flows are expressed in terms of temperature differentials due to constant thermal
conductivity.
For the unsteady-state condition, time must elapse before an equilibrium temperature is
obtained in the body. The analysis conducted for steady-state must be modified to take into
account the change in internal energy of the body with time, and the boundary conditions must
be adjusted to match the specific physical situation present. The heat flow is governed by the
differential equation:

(3)
Where rho, c, and k are all constant properties. For this problem, the Explicit Finite Difference
Method is used to solve the system of equations. It is conditionally stable, and must obey the
constraint:

()

(4)
For two-dimensions.

2. INPUT INFORMATION
For the steady-state analysis, the applet required that the coordinates of various points
along the boundary be entered along with their respective temperature (in a clockwise direction).
Two separate experiments were performed. When the grid size was x = y = 1, the values
displayed in Table 1 were input to represent Figure 1, along with an initial temperature of 100
C.

x y Temperature
(C)
0 0 100
0 1 100
0 2 100
0 3 500
1 3 500
2 3 500
3 3 500
3 2 100
3 1 100
3 0 100
2 0 100
1 0 100
0 0 100

When the grid size was x = y = 0.5, the values displayed in Table 2 were input to represent
Figure 1, along with an initial
temperature of 100 C.

x y Temperature
(C)
0 0 100
0 0.5 100
0 1 100
0 1.5 100
0 2 100
0 2.5 100
0 3 500
0.5 3 500
1 3 500
1.5 3 500
Table 1. Steady-state input values, x = y = 1.
Table 2. Steady-state input values, x = y = 0.5.
2 3 500
2.5 3 500
3 3 500
3 2.5 100
3 2 100
3 1.5 100
3 1 100
3 0.5 100
3 0 100
2.5 0 100
2 0 100
1.5 0 100
1 0 100
0.5 0 100
0 0 100

For the unsteady-state analysis, the same input coordinates and initial temperature are
required. However, the applet also requires the user to select a material. Based on Equation 3,
this analysis depends on the density, specific heat capacity, and thermal conductivity of the
bodyall of which depend on the material of the body. The unsteady analysis is based on the
time it takes heat to be conducted on the body, so heat distribution must be computed at each
time interval, whereas the steady analysis disregards these time intervals and assumes enough
time has passed that it has already reached equilibrium. A time step is also required to ensure that
Equation 4 is upheld for convergence. An error is returned by the applet if the time step is not
small enough. Two separate experiments were run. Each one employed the boundary
coordinates displayed in Table 3, along with an initial temperature of 100 C. When the selected
material was Iron, the values displayed in Table 3 were populated by the applet.
Material Iron
k [W/m*K] 73.0
Density [kg/m
3
] 7897
Cp [kJ/kg*K] 0.452
Time Step [sec] 0.2
Table 3. Unsteady-state input material: Iron.






When the selected material was Silver, the values displayed in Table 4 were populated by the
applet.





3. PLOTS
For the steady-state condition where x = y = 1, the computational grid displayed in
Figure 2 was established, and resulted in the isothermal pattern seen in Figure 3.


T0 [C] 100
Material Silver
k [W/m*K] 419.0
Density [kg/m
3
] 10524
Cp [kJ/kg*K] 0.234
Time Step [sec] 0.02
T0 [C] 100
Table 4. Unsteady-state input material: Silver.
Figure 2. x = y = 1 computational grid.


For the steady-state condition where x = y = 0.5, the computational grid displayed in Figure 4
was established, and resulted in the isothermal pattern seen in Figure 5.






For the unsteady-state condition where x = y = 0.5 and a material of Iron was used, the
computational grid displayed in Figure 4 was established, and resulted in the isothermal patterns
Figure 3. Steady-state x = y = 1 isothermal pattern.
Figure 4. x = y = 0.5 computational grid.

Figure 5. Steady-state x = y = 0.5 isothermal pattern.
seen in Figure 6 at various time increments. After 6.6 seconds, the isothermal pattern matched
the stead-state pattern seen in Figure 5.



For the unsteady-state condition where x = y = 0.5 and a material of Silver was used, the
computational grid displayed in Figure 4 was established, and resulted in the isothermal patterns
seen in Figure 7 at various time increments. After 6.6 seconds, the isothermal pattern matched
the stead-state pattern seen in Figure 5.
6.6 sec
0.8 sec
0 sec
5.6 sec 4.8 sec
4 sec 3.2 sec 2.4 sec
1.6 sec
Figure 6. Unsteady-state x = y = 0.5 isothermal patterns: Iron.


4. DISCUSSION
The results of each steady-state test are compared to the results detailed in Table 3-3 of
Holman, which were derived using the Gauss-Seidel method with relaxation, listed under
0 sec
0.8 sec 0.88 sec 0.72 sec
0.64 sec 0.56 sec 0.48 sec
0.4 sec 0.32 sec
0.16 sec
0.24 sec
0.08 sec
Figure 7. Unsteady-state x = y = 0.5 isothermal patterns: Silver.
Textbook in Table 5. The main interior nodes labeled in Figure 1 are analyzedthe rest of the
data is presented in Tables 7 and 8 of Section 5.
Table 5. Steady-state nodal temperature comparison.
Node Textbook x = y = 1 x = y = 0.5
(1, 1) T3 150 C 150 C 148.5 C
(1, 2) T1 250 C 250 C 251.5 C
(2, 2) T2 250 C 250 C 251.5 C
(2, 1) T4 150 C 150 C 148.5 C

The results produced from the applet for the coarse grid (x = y = 1) exactly match those
produced by the textbook. On the other hand, the use of a finer grid (x = y = 0.5) results in
values that were slightly differentthey are 1.5 C above or below the textbook data, which
equates to 1%. Based on the comparison of Figures 3 and 5, it can be seen that the highest
temperatures are always present at the top of the grid, regardless of the distribution. However,
using a finer grid with more nodes to evaluate at creates an overall higher temperature at the
higher y-valuesaveraging around 400 C compared to 200 C. The finer grid allows for a
gradual transition to these higher temperature values, whereas the coarse grid has a significant
jump to these values. These understandings indicate that the determination of temperature
gradients are not independent of grid type.
The results of each unsteady-state test are compared to the results obtained using the
applet with a finer grid and displayed in Table 6. The main interior nodes labeled in Figure 1 are
analyzedthe rest of the data is presented in Tables 9 and 10 of Section 5.
Table 6. Unsteady-state nodal temperature comparison.
Node Steady-state Iron (6.6 s) Silver (0.88 s)
(1, 1) T3 148.5 C 137.9 C 148.1 C
(1, 2) T1 251.5 C 241.8 C 251.1 C
(2, 2) T2 251.5 C 243 C 251.2 C
(2, 1) T4 148.5 C 139.1 C 148.1 C

Initial analysis shows that use of Silver produced results that are extremely similar to the steady-
state data, only varying by about 0.3 - 0.4 C, or 0.02%. Iron, on the other hand, produced
results that are in the relative range to the steady state-data, but differ by a more significant
factor8.5 - 10.6 C, or about 5%. While Figures 6 and 7 seem to show very similar changes to
the isothermal patterns with respect to time, it is important to notice the scale. Iron also took 7.5
times longer than Silver to achieve equilibrium temperatures. Based on Tables 3 and 4, it can be
seen that Silver has a thermal conductivity (k) that is almost 6 times that of Iron, while its c
p
is
about half that of Ironthis results in it reaching steady-state faster.
Overall, the results prove that the type of grid used does play a factor into the precision of
the temperature gradients. Whether a steady or unsteady analysis is used, though, does not play
a factorthe results between the two methods proved to be very close in value. Additionally,
the material used during unsteady analysis will only affect how quickly the complete data is
collecteda higher thermal conductivity and lower heat capacity at constant pressure results in
reaching the equilibrium state faster, but does not guarantee more accurate temperature values.

5. TABULAR OUTPUT
Table 7. Steady-state x = y = 1 output data.

x y Temperature (C)
0 0 100
1 0 100
2 0 100
3 0 100
0 1 100
1 1 150
2 1 150
3 1 100
0 2 100
1 2 250
2 2 250
3 2 100
0 3 500
1 3 500
2 3 500
3 3 500


Table 8. Steady-state x = y = 0.5 output data.

x y Temperature (C)
0 0 100
0.5 0 100
1 0 100
1.5 0 100
2.0 0 100
2.5 0 100
3 0 100
0 0.5 100
0.5 0.5 112.5
1 0.5 121.4
1.5 0.5 124.5
2.0 0.5 121.4
2.5 0.5 112.5
3 0.5 100
0 1 100
0.5 1 128.7
1 1 148.5
1.5 1 155.4
2.0 1 148.5
2.5 1 128.7
3 1 100
0 1.5 100
0.5 1.5 153.8
1 1.5 188.4
1.5 1.5 200
2.0 1.5 188.4
2.5 1.5 153.8
3 1.5 100
0 2 100
0.5 2 18.2
1 2 251.5
1.5 2 267.7
2.0 2 251.5
2.5 2 198.2
3 2 100
0 2.5 287.5
0.5 2.5 287.5
1 2.5 351.7
1.5 2.5 367.8
2.0 2.5 351.7
2.5 2.5 287.5
3 2.5 100
0 3 500
0.5 3 500
1 3 500
1.5 3 500
2.0 3 500
2.5 3 500
3 3 500


Table 9. Unsteady-state x = y = 0.5 output data: Iron.

x y Temperature (C)
0 0 100
0.5 0 100
1 0 100
1.5 0 100
2 0 100
2.5 0 100
3 0 100
0 0.5 100
0.5 0.5 108.6
1 0.5 115
1.5 0.5 117.6
2 0.5 115.7
2.5 0.5 109.4
3 0.5 100
0 1 100
0.5 1 122.2
1 1 137.9
1.5 1 143.9
2 1 139.1
2.5 1 123.6
3 1 100
0 1.5 100
0.5 1.5 146.6
1 1.5 176.7
1.5 1.5 187.3
2 1.5 178.1
2.5 1.5 148.2
3 1.5 100
0 2 100
0.5 2 192.2
1 2 241.8
1.5 2 257.2
2 2 243
2.5 2 193.6
3 2 100
0 2.5 100
0.5 2.5 284.2
1 2.5 346.4
1.5 2.5 362
2 2.5 347
2.5 2.5 285
3 2.5 100
0 3 500
0.5 3 500
1 3 500
1.5 3 500
2 3 500
2.5 3 500
3 3 500


Table 10. Unsteady-state x = y = 0.5 output data: Silver.

x y Temperature (C)
0 0 100
0.5 0 100
1 0 100
1.5 0 100
2 0 100
2.5 0 100
3 0 100
0 0.5 100
0.5 0.5 112.4
1 0.5 121.1
1.5 0.5 124.3
2 0.5 121.2
2.5 0.5 112.4
3 0.5 100
0 1 100
0.5 1 128.5
1 1 148.1
1.5 1 154.9
2 1 148.1
2.5 1 128.5
3 1 100
0 1.5 100
0.5 1.5 153.6
1 1.5 188
1.5 1.5 199.5
2 1.5 188.1
2.5 1.5 153.6
3 1.5 100
0 2 100
0.5 2 198
1 2 251.1
1.5 2 267.3
2 2 251.2
2.5 2 198
3 2 100
0 2.5 100
0.5 2.5 287.4
1 2.5 351.5
1.5 2.5 367.5
2 2.5 351.5
2.5 2.5 287.4
3 2.5 100
0 3 500
0.5 3 500
1 3 500
1.5 3 500
2 3 500
2.5 3 500
3 3 500

Potrebbero piacerti anche