Documente Academic
Documente Profesional
Documente Cultură
EME 165
ABSTRACT
This report shows the optimization of a heat dissipation system for a microprocessor, and the effects of the changes in the
heat sink dimensions, fluid velocity and the heat conductor materials.
sink cross section, only one symmetric portion of the heat sink is
used in order to reduce significantly the amount of time needed
to do the simulation.
q00 p
Adiabatic
Isotherm
Microprocesor
Thermal Interfase
Heat Sink
Fluid
(25oC)
hf
Adiabatic
Isotherm
height
width
PROBLEM DESCRIPTION
Heat from a computer microprocessor is transfered trough a
thermal interface to a heat sink and then to a liquid as is shown in
Figure 1. The microprocessor generates 500[W ] as heat, and has
20[mm]x20[mm] in size. The thermal interface has a resistivity
of R00 = 2x10 5[oCm2 /W ]
Adiabatic
Surface
wf
(a) Thermal
(b) Dimensions
Insulator
Liquid Channel
The parameters that will vary to get the optimum heat sink
dissipator are:
-Fin width w f and fin height h f (Figure 2b).
-Liquid channel velocity u.
-Material Thermal properties: Density (), Specific heat capacity (c p ), Thermal conductivity (k).
SOLUTION APPROACH
Because of the simulation time needed when a parameter
is changed, the system will be analyzed as an one dimensional
steady state heat transfer problem.
Another relevant point is the maximum temperature at the
surface of the chip must be 75[oC], with the thermal interface
resistance given (Rti00 = 2x10 5[oCm2 /W ]), we can get the temperature drop in this material:
T p Ts = q p
Rti00
= 25[oC]
A p
Simulation method
Once the optimal approximated dimensions (OAD) are
known, the same procedure as the approximation method will be
followed, but with the finite difference model, and only for points
near to the OAD, in order to validate the approximation method
and get the fin dimensions that give us the minimum temperature,
then materials and fluid speeds will be tested to get a better understand of their influence in the heat sink performance in order
to select the optimal configuration of these variables.
(1)
RESULTS
Approximation method
Simplified model
Assuming the steady state in the dotted symmetric section
from the heatsink in Figure 1, it can be modeled as a thermal
conductor with a uniform cross section fin , with boundary conditions of the case B (Insulated Tip) from the table 3.4 in the
textbook [1]. The simplified system is shown in Figure 3, that
is modeled as one dimensional thermal circuit, where heat is applied to the surface of the conductive material, modeled as Rcond ,
at the base surface heat is transfered by the fin as q f and by convection where the base surface is in contact with the fluid Rconv .
45
Temp [C]
qsection
50
qsection
Ts
Rcond
40
Rconv
qf
Tb
0
1
35
0
2
0.5
1
1.5
2
1
Rconv
!
(2)
[C]
Ts = T + qsection Rcond +
hf [mm]
1
+ Mtanh(mL)
wf [mm]
T T
min
Aproximation method
For the approximation to the optimal dimensions, the heat
sink material will be aluminum (k = 237[W /mK]) and a fluid
velocity in the channel will be v = 4.65[m/s]. Ts will be calculated for different widths and heights of the fin and then it will be
plotted in a 3D surface .
Once having the best dimensions of the fin, the geometry
will remain constant and the material thermal conductivity will
vary and the temperature values for different k values will be
plotted. The same procedure will be done with the variation of
the fluid rate.
6
4
2
500
1000
1500
2000
k [W/(m K)]
2500
3000
Figure 5 shows that the heat sink material isnt really important before k = 500[W /mK] the temperature improvement would
be less than 5[oC]. The only considerable change could be change
2
100
min
80
60
TsTs
100
40
Ts [C]
20
50
0
10
12
14
16
18
v [m/s]
0
v [mm/s]
x 10
x 10
3.5
60
60
60
59
59
59
x 10
36
58
58
57
57
57
56
56
56
55
55
55
54
54
54
53
53
53
52
52
52
51
51
51
50
50
50
49
49
49
33
35
35
29
28
27
30
34
34
26
33
29
30
33
2.5 30
2.5
47
46
48
48
47
47
height (m)
27
28
31
28
30
1.5
1.5
26
48
height (m)
height (m)
1.5
32
29
29
2
440
442
43
41
39
38
36
35
34
33
32
31
30
29
36
36
31
31
31
3.5
28
42
41
40
3933
38
37
36
35
34
32
31
30
29
37
37
32
32
32
2.5
39
38
33
33
3.5
x 10
34
34
34
58
35
35
35
32
31
29 30
30
27
46
46
45
4443
42
45
4443
42
27
29
45
Simulation method
29
44
0.5
0.5
0.5
26
1
width (m)
1.5
2
3
x 10
min
10
TsTs
0.5
43
0.5
1
width (m)
1.5
0.5
x 10
1
width (m)
1.5
2
3
x 10
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
wf [mm]
CONFIGURATIONS
( h f = 3.9mm, v = 4.65m/s,Aluminum)
convection will be predominant at the top and there will be a
small gradient of temperature added to a small amount of heat
transfer by convection at the bottom.
In the results of Table 2 in the Appendix B, the temperature variation for different heights and weights of the fin shows a
similar behavior as the one shown in Figure 4, with a difference
of temperatures of 7[oC] along w f and a temperature variation
greater than 50[oC] changing h f .
TsTs
min
w f = 1.5[mm]
h f = 3.5[mm]
Material: Copper Pure
v=19.05 [m/s] w/ pump MCP655 (20 lt per min)
4
2
0
500
1000
1500
2000
2500
k [W/(m K)]
34
33
32
Temperature (C)
In the Table 1 (Appendix B) is shown that with different materials the temperature improvement is small as the approximated
model suggested it, with a maximum temperature difference of
about 6[oC], from Aluminium to Diamond.
Figure 10 shows in a small chanel there will be a predominant conduction phenomena rather than convection, in a thin
channel there will be a conductive phenomena but, with a smaller
gradient in the vertical gradient, finally, with a tall channel the
31
30
29
28
27
26
25
500
1000
1500
2000
Time steps ( 5e07 seconds)
2500
3000
x 10
33
33
32
32
32
3.5
31
30
29
28
31
31
30
27
3
30
29
28
30
2.5
29
height (m)
29
2
28
28
27
1.5
REFERENCES
[1] Fundamentals of Heat & Mass Transfer, TL Bergman, AS
Lavine, FP Incropera & DP Dewitt, 7th Ed., John Wiley &
Sons, Inc. (2011)
27
0.5
0.5
1
width (m)
1.5
2
3
x 10
q00 p =
(5)
(3)
q p
qSection
=
A p
ASection
qsection = q f +
(4)
Tb T
Rconv
(6)
and
Ts Tb
Rcond
(7)
1
h(Asection A f in )
(8)
qsection =
where
DISCUSSION
As is shown in the results, the predominant variable to enhance the heat transfer in the system is the fluid velocity and the
fin height. The fin height enhances the heat transfer increasing
the temperature gradient in the horizontal direction, resulting in a
Rconv =
4
Rcond =
L
kAsection
(9)
(10)
h f [mm]
w f [mm]
Tsmax [o ]
Tsmin [o ]
3.9
0.1
48.670
39.485
3.9
0.5
47.04
37.46
3.9
44.66
39.043
3.9
1.5
41.73
40.242
3.9
1.9
41.05
38.5212
3.5
1.9
43.716
42.29
3.0
1.9
46.9
46.22
2.0
1.9
55.33
55.14
0.1
3.9
>90.79
>90.78
with
M=
r
m=
(11)
hPkAc
hP
kAc
(12)
Solving for Ts
Ts = T + qsection Rcond +
1
Rconv
1
+ Mtanh(mL)
!
(13)
Ac and P are the cross sectional area and perimeter of the fin
respectively.
TABLE 3: MAX AND MIN SURFACE TEMPERATURES FOR
Temperature Range
h[W/m.K]
Tsmax [o ]
Tsmin [o ]
0.1
1455 (L)
>124.737
>115.58
17853
>65.97
>62.50
MATERIALS
Temperature Range
Material
k[W /m.K]
Tsmax [o ]
Tsmin [o ]
31085
52.48
49.9
Aluminium
273
41.212
39.49
42995.6
46.699
44.658
Copper
401
39.46
38.05
74859.7
39.53
38.4294
Silver
429
39.29
37.95
12
130338
36.02
34.27
Diamond
2300
35.38
34.95
17.1
172982
34.7839
32.6642
%% Clear Workspace/Command/Figures
clear all; clc; close all;
3
4
5
6
7
8
9
10
11
12
%% Define Geometry
width=2e-3;
%width of symmetry element (m)
height=4e-3;
%height of symmetry element (m)
wf=1.5e-3;
%width of fin in symmetry element (m)
hf=3.5e-3;
%height of fin in symmetry element=channel height (m)
wc=width-wf;
%channel width (m)
hs=height-hf;
%thickness of base (where there is no fin) (m)
dx=0.1e-3;
%node size delta x (m)
dy=0.1e-3;
%node size delta y (m)
13
14
15
16
17
Ncolumn=int16(width/dx)+1;
%number of rows;
Nrow=int16(height/dy)+1;
%number of columns;
fluid_yloc=int16(hs/dy)+1; %row number location of fluid y contact
fluid_xloc=int16(wf/dx)+1; %column number location of fluid x contact
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
%% Liquid properties
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
else
h=3.61*kfh/Dh; %convective heat transfer coeff. to the fluid (W/m2.K)
end
53
54
55
56
57
58
59
60
61
%% Initialize indexing and other variables for Euler forward time stepping method
ti=0;
% initialize time index to use for time stepping
err=1;
% initial error start the while loop
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
while err>tol
% loop while error is larger than preset tolerance
ti=ti+1;
% move time step by one
Temp(:,:,ti)=Tnew;
T=Tnew;
% use new temperature array from the previous step to determine change in tem
incr=zeros(Nrow,Ncolumn);
% reset "incr" array, which tracks temperature change at this tim
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
%% "fin" area, where the heat sink extends to the bottom insulation
for n=2:fluid_xloc-1
7
102
103
104
105
106
107
108
%
%
%
%
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
end
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
end
198
199
200
201
202
203
204
205
k*dy/2*(T(m,n-1)-T(m,n))/dx );
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
%% Update temperature array (at time step ti+1) with change in temperature calculated by temp
Tnew=T+incr;
err=abs(max(incr));
% Determine when to stop solution
if ti>SetStepLimit
disp(Tolerance not met within set iteration limit);
break;
end
223
224
225
226
227
228
229
230
end
231
232
233
234
235
236
237
238
MaxSurfaceTemp=max(Tnew(1,:));
MinSurfaceTemp=min(Tnew(1,:));
TempDifference=MaxSurfaceTemp-MinSurfaceTemp;
TimePassed=ti*dt;
%
%
%
%
[X,Y]=meshgrid(0:dx:width, 0:dy:height);
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
F1=figure(1);
set(F1,units,inches,OuterPosition,[0 0 10*width/height 10]); %set plot size
MinChanTemp=min(min(Tnew(fluid_yloc,:)),min(Tnew(:,fluid_xloc))); %minimum surface temperature o
ContourSpacing=1;
%contour line spacing (degC)
ContourRange=ceil(MinChanTemp:ContourSpacing:MaxSurfaceTemp);
contour(X,height-Y,Tnew,ShowText,on,LevelList,ContourRange);
boundary1=line([wf wf],[0 hf],linewidth,2);
% draw boundary lines where fluid and hea
boundary2=line([wf width],[hf hf],linewidth,2);
set(boundary1,Color,k);
set(boundary2,Color,k);
rectangle(Position,[wf,0,wc,hf],FaceColor,b);
% "white out" channel area from plot
xlabel(width (m));
%label axis
ylabel(height (m));
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
F2=figure(2);
Teo=Temp(1,:,:); %Surface temperature
Teo=squeeze(Teo); %
Teo=Teo(:,1:100:length(squeeze(Teo(1,:))));
surf(Teo,EdgeColor,none,LineStyle,none,FaceLighting,phong);
xlabel(Time x 2 (\mu s));
%label axis
ylabel(width (m));
zlabel(Temperature ( C ))
system=toc
.2
1
2
3
4
5
6
7
8
9
10
11
12
ApproxMicrochip.m code
%% Clear Workspace/Command/Figures
clear; close all; clc;
Tf=25;
%Temperature of the fluid (degC)
CPUWatts=500;
%CPU heat dissipation (W)
width=2e-3;
%width of symmetry element (m)
height=4e-3;
%height of symmetry element (m)
Z=0.02;
%dimension of one side of CPU (m)
k=237;
%material conductive heat transfer coeff. (W/m.K)
rhoh=1000
%Fluid density (kg/m3)
muh=118e-6
%Fluid viscosity (N s/ m2)
kfh=0.642
% Fluid conductive heat transfer coeff. (W/m.K)
Prh=0.86
%Prandlt Number
13
14
15
16
17
18
19
20
21
22
23
L=h_f(j);
24
25
Ach=2*h_f(i)*(width-w_f(i));
v=0.02/(5*Ach*60)
Ph=2*h_f(i)+4*(width-w_f(i));
Dh=4*Ach/Ph;
Reh=rhoh*v*Dh/muh;
26
27
28
29
30
%Channel area
%Fluid velocity
%Channel perimeter
%Hydraulic diameter
%Reynolds number
31
if Reh>2300
h=0.023*Reh0.8*Prh0.3*kfh/Dh; %Laminar convection coeff.
else
h=3.61*kfh/Dh;%Turbulent convection coeff.
end
m=sqrt((h*P)/(k*A_c));
M=sqrt(h*P*k*A_c);
R_cond=(height-h_f(j))/(k*Z*2*width);
R_conv=1/(h*Z*2*(width-w_f(i)));
q_f(i,j)=M*tanh(m*L);
T_ch(i,j)=Q*R_cond+Q/(1/R_conv+q_f(i,j))+Tf; %Temperature at base
if T_ch(i,j)>50
%discard values greater than max temperature
T_ch(i,j)=50;
end
end
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
end
48
49
50
51
52
53
54
figure
%Surface of Temperature(width,height)
surf(T_ch,EdgeColor,none,LineStyle,none,FaceLighting,phong);
xlabel(h_f [mm]);
ylabel(w_f [mm]);
zlabel(Temp [ C ]);
colormap(gray)
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
v=4.64511
rhoh=1000
muh=118e-6
kfh=0.642
Prh=0.86
%Fluid velocity
%Fluid density (kg/m3)
%Fluid viscosity (N s/ m2)
% Fluid conductive heat transfer coeff. (W/m.K)
%Prandlt Number
70
71
72
73
74
75
Ach=2*h_f*(width-w_f);
Ph=2*h_f+4*(width-w_f);
12
76
77
78
79
80
81
82
Dh=4*Ach/Ph;
Reh=rhoh*v*Dh/muh
if Reh>2300
h=0.023*Reh0.8*Prh0.3*kfh/Dh;
else
h=3.61*kfh/Dh;
end
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
for i=1:3000
Q=CPUWatts/Z2*(Z*2*width);
P=2*Z+2*(2*w_f);
A_c=(2*w_f)*(Z);
L=h_f;
m=sqrt((h*P)/(i*A_c));
M=sqrt(h*P*i*A_c);
R_cond=(height-h_f)/(i*Z*2*width);
R_conv=1/(h*Z*2*(width-w_f));
q_f(i)=M*tanh(m*L);
T_ch(i)=Q*R_cond+Q/(1/R_conv+q_f(i))+Tf;
end
figure
plot(T_ch-min(T_ch),k,linewidth,1.5)
ylabel(T_s-T_{min} [ C ])
xlabel(k [W/(m K)])
grid on
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
v=4.64511
rhoh=1000
muh=118e-6
kfh=0.642
Prh=0.86
116
117
118
w_f=1.9e-3;
h_f=3.9e-3;
119
120
121
122
123
124
125
126
127
13
128
129
130
131
132
if Reh>2300
h=0.023*Reh0.8*Prh0.3*kfh/Dh;
else
h=3.61*kfh/Dh;
end
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
Q=CPUWatts/Z2*(Z*2*width);
P=2*Z+2*(2*w_f);
A_c=(2*w_f)*(Z);
L=h_f;
m=sqrt((h*P)/(k*A_c));
M=sqrt(h*P*k*A_c);
R_cond=(height-h_f)/(k*Z*2*width);
R_conv=1/(h*Z*2*(width-w_f));
q_f(i)=M*tanh(m*L);
hh(i)=h;
T_ch(i)=Q*R_cond+Q/(1/R_conv+q_f(i))+Tf;
end
figure
plot(T_ch-min(T_ch),k,linewidth,1.5)
ylabel(T_s [ C ])
xlabel(v [mm/s])
grid on
14