Conjugate Heat transfer on Graphics Card
CONJUGATE HEAT TRANSFER ANALYSIS ON GRAPHICS CARD
A graphics card is primarily designed to remove the graphical processing tasks from the processor or RAM. It includes a dedicated graphical processing unit (GPU) and a dedicated RAM that help it to process graphical data quickly. Like most processors, a graphics card also has a dedicated heat sink to keep the heat out of the GPU. A graphics card enables the display of 3-D images, image rasterization, higher pixel ration, a broader range of colors and more.
In this project, CHT analysis is done on graphics cards by CFD analysis.for the CFD simulation ANSYS Fluent is used. The geometry of a graphics card is created in space claim as shown in figure 1. There are main three parts of the graphics card. (1) base (2) fins (3) processor.
The processor acts as a heat source from where heat is generated. Fins are used to increase the heat transfer so heat can remove fastly.base is used to support the fins and processor.
In this geometry, air enclosure is created around the graphics card so that air can take that heat.
Figure 1. 3-D model
Figure 2. 3-D model
Figure 3. 3-D model with air enclousure
This model has meshed in ANSYS Mesher as shown in figure 4.Here The simulation is carried out for two types of mesh. The first type is baseline mesh and the second type is fined mesh.
CASE 1 :
In the baseline mesh, all the default value is taken for mesh generation.
Element size = 4 mm
skewness = 0.26
elements = 230539
Figure 4. Mesh generation (baseline mesh)
This model is imported in Ansys Fluent for the simulation. laminar flow is selected for the viscous model and boundary condition taken as inlet velocity of air and temperature of air that is 0.5 m/s and 26.85-degree celsius. heat generation from the processor taken as 25600000 w/m3.Convergent criteria for the simulation taken as 0.001. The transient simulation is carried out with time step size 0.025 and the number of time step 300.
RESULTS :
The simulation is carried out for 7.5 seconds. the figure below is shown the residuals graph for the simulation. The velocity and temperature contours are shown in the figures.
Figure 5. Residuals
Figure 6. velocity contour
Figure 7. Temperature contour
The maximum temperature attained by the processor was 29.53-degree celsius.
Figure 8. Temperature contour
The wall adjacent heat transfer coefficient of the fin walls is 16395.27 W/m2 K.
CASE 2 :
Fined mesh
element size = 2 mm
skewness = 0.26
elements = 402007
Figure 9. Mesh generation of air domain(fine mesh)
Figure 10. Mesh generation (fine mesh)
RESULTS :
The simulation is carried out for 7.5 seconds. the figure below is shown the residuals graph for the simulation. The velocity and temperature contours are shown in the figures.
Figure 11.Residuals
Figure 12. Velocity contour
The maximum temperature attained by the processor was 29.41-degree celsius.
Figure 13. Temperature contour
Figure 13. Temperature contour on fins
The wall adjacent heat transfer coefficient of the fin walls is 1604233.23 W/m2 K.
|
Sr no. |
Elements |
Processor Temperature (◦C) |
Wall heat transfer coefficient of fins (W/m2 K) |
|
1 |
230539 |
29.53 |
16395 |
|
2 |
402007 |
29.41 |
1604233.33 |
Conclusion :
From the results, it is concluded that for fine-meshed geometry results would be better than baseline meshed because there is large number of nodes in fine-meshed.
maximum processor temperature reached up to 29.41 ◦C because of air flowing around the fins which takes the heat from the fins.