CHT analysis on Graphics-card

CONJUGATE HEAT TRANSFER ANALYSIS ON A GRAPHICS CARD:

OBJECTIVE:

To perform a conjugate heat transfer analysis on a graphics card and to obtain the following result:

1. To find out the velocity of the cooling air at which the there is a maximum heat transfer coefficient (h.t.c)
2. To find out the maximum temperature attained by the processor in every cases simulations performed
3. To identify the potential hotspots within the graphics card.

DESCRIPTION:

Computational Fluid Dynamics (CFD) and more specifically conjugate heat transfer (CHT) analysis can accurately predict heat transfer by simulataneous solving all the relevant solid and fluid heat transfer preocesses,like that of the conduction through solids ,free and forced convection in the gases/fluids and thermal radiation.The CHT approach in particular has more advantage over the Finite Element thermal analysis in that the wall heat teansfer coefficients and their local variations on surfaces are calculted directly within the model rather than based on simplified empirical calculations.The CHT approach therefore has benefits for those applications where heat trasnsfer is either non-uniform or difficult to calculate empirically.

CHT analysis is performed on the graphics is basically to predict the cooling characteristics that can be obtained by the processor during the working of the component by forced convection.The graphics card is a basically a hardware on the computer which produces images that we are able to see on our screen.The graphics card is responsible for rendering an image to monitor.Thus in this process the processor gets heated up and large amount of heat is being generated within the processor of the graphics card and inorder to achieve optimum performance from the graphics card without being overheated,it must be cooled down with a proper system.In normal cases a fan is being used where a cool air being circulated over the graphics card and as a result of this ,forced convection heat transfer take place and the heat is dissipated to the incoming air and cooling of the component takeplace.Thus,velcoity places a very important role in the forced convection heat transfer.Hence by this simualtion we are going to determine the velocity where the efficient cooling take place.

METHODOLOGY:

A Steady state simulation was performed on the geometry of the graphics card and the inlet velocity of air was varied from 1 m/s to 10 m/s and the results obatined were analysed.For the simulation the thermal design power of the chip was assumed to be 20 W.

Thermal Design Power(TDP) ,sometimes called as the thermal design point is the maximum amount of heat that geneated by a computer chip that the cooling system in a computer is designed to dissipitate under any workload.TDP is not the largest amount of heat the CPU could generate,such as by running a power virus,but rather the maximum amount of heat that it would generate while running 'real applications'.This ensures that the computer will be able to handle essentially all applications without exceeding its thermal envelope.

Thus the TDP for the current problem is taken to be 20W

The volume of the processer is 64`mm^3`.Therefore the volumetric heat generation in the processor is 312,500,000 `W/m^3`.

NUMERICAL SIMULATION:

1.GEOMETRY:

The geometry above is of the graphics card that were are simulating.Since the flow is taking place over the graphics card,an enclosure is provided and it is treated as the fial geometry for the simulation of the flow.

2.MESHING:

The meshing for the problem has been done for two cases,one is having a base line mesh and other is having a more refined meshing.

CASE-A : BASELINE MESH

• Element Size : 10 mm
• Type of Mesh Elements : Tetrahedral
• Number of Elements : 84,805

CASE-B: REFINED MESHING:

The meshing has been more refined in the areas where there is more heat dissipation.Fins and the processor areas have been more refined.

• Mesh Element size: 2 mm
• Maximum size of the elements : 2 mm
• Orthogonal Quality : 0.773

Sizing : - Face sizing : Element size 0.5 mm

Body sizing : Element size: 1 mm

• Number of Nodes : 1,82,657
• Number of Elements: 10,40,776

3.SETUP:

Solver - Type : Pressure-Based

- Velocity Formulation : Absolute

- Time : Steady State

Model - k-omega SST

Material - Fluid : Air

- Solid : Silicon

Copper

Aluminium

Cell Zone Conditions - Processor

• Type: Solid
• Material : Silicon
• Source Term : ON

Number of source term = 1

Heat generation = 312,500,000 `W/m^3`

- Fins

• Type: Solid
• Material : Copper
• Source Term :OFF

- Base

• Type: Solid
• Material : Aluminium
• Source Term: OFF

NOTE: The case has been simulated for various inet velocities from 1 m/s to 10 m/s and rest of all conditions remains unchanged

Boundary Conditions - Inlet :

• Velocity-inlet: CASE 1. 1 m/s

CASE 2. 2 m/s

CASE 3. 5 m/s

CASE 4. 8 m/s

CASE 5. 9 m/s

CASE 6. 10 m/s

• Thermal - Temperature 300 K

- Oulet:

• Pressure-outlet ( zero guage pressure )

- Bottom-wall

• No slip condition

- Symmetry

• Symmetry boundary condition

Solver - Pressure-velocity Coupling

Method : Coupling

4.SOLUTION & RESULTS:

CASE A: BASELINE MESHING

• CASE 1. INLET VELOCITY 1 m/s

1.Residual Plot:

• The solution converged at 50 iterations.

2.Temperature Contour Plot:

3.Velocity Contour Plot:

4.Wall h.t.c Plot (Global):

• CASE 2. INLET VELOCITY 2 m/s

1. Residual Plot:

• The solution converged at 55 iterations.

2.Temperature Contour Plot:

3.Velocity Contour Plot:

4.wall h.t.c contour Plot:

• CASE 3. 5m/s

1.Residual Plot:

• The solution converged at 59 iterations.

2.Temperature Contour Plot:

3.Velocity Contour Plot:

• CASE 4. 8 m/s

1.Residual Plot:

• The solution converged at 68 iterations.

2.Temperature Contour plot:

3.Velocity Contour Plot:

4.Wall h.t.c Contour Plot:

• CASE 5. 9m/s

1.Residual Plot:

• The solution converged at about 76 iterations.

2.Temperature Contour Plot;

• The temperature distribution in different components of the graphics card is shown above.

3.Velocity Contour Plot:

4.Wall h.t.c Plot:

• CASE 6. 10 m/s

1.Residual Plot:

• The solution converged at 83 iterations.

2.Temperature Contour Plot;

3.Velocity Contour Plot:

The datas obtained from the contour plots and the simulation of the base line mesh can be tabulated as follows:

From the above table, it can be summarised that as the velocity of the inlet air is increased the total heat transfer rate increases from the processor, ie we can observe that the local h.t.c is increased.Further more when the temperature attained by the processor is observed there is a significant reduction in the temperature of the processor by almost 200 K when the velocity of the flow is increased from 1 to 10 m/s .Further when the mesh is refined more,there can again be significant variation in the results obatined as it will be more fitting to the practical problems.

CASE B- REFINED MESHING:

• CASE 1.INLET VELOCITY 1 m/s

1.Residual Plot:

• The solution has been converged at 21 iterations

2. Temperature Contour Plot:

3.Velocity Contour Plot:

4.Wall h.t.c Contour Plot:

• CASE 2. INLET VELOCITY 2 m/s

1.Residual Plot:

• The solution has converged at 25 iterations.

2.Temperature Contour Plot:

3.Velocity Contour Plot:

4.Wall h.t.c contour Plot:

• CASE 3 . 5 m/s

1.Residual Plot:

• The solution converged at 27 iterations.

2. Temperature contour Plot;

3.Velocity Contour Plot:

4.Wall h.t.c Contour Plot:

• CASE 4 . 8 m/s

1.Residual Plot;

• The solution converged at 28 iterations.

2.Temperature Contour Plot:

3.Velocity Contour Plot:

4.Wall h.t.c contour Plot:

• CASE 5. 9 m/s

1.Residual Plot:

• The solution converged at 28 iterations.

2.Temperature Contour Plot:

3.Velocity Contour Plot:

4.Wall h.t.c contour Plot:

• CASE 6. 10 m/s

1.Residual Plot:

• The solution converged at 28 iterations.

2.Temperature Contour Plot;

3. Velocity Contour Plot:

4.Wall h.t.c Contour Plot:

The data obtained from the simulation of the above can be tabulated as below:

• From the above table it can be summarised that:

1. As the velocity of the flow increases the wall h.t.c increases,which shows that more heat transfer take place

2. The temperature attained by the processor reduces almost by 200 K when the velocity is increased from 1 to 10 m/s.

3. As the mesh become more refined the total heat transfer rate increases which leds to efficient cooling of the graphics card.

• From the simulation carried out,it can be concluded that the potential hotspot in the system is the processor.

• In every cases the temperature attained by the processor is the highest in the system,thus it can be regarded as the potential hotspot.

• The velocity at which the maximum heat transfer coefficient is attained is 10 m/s.Thus,for inlet velocity of 10 m/s ,the comparison of parameters between the baseline mesh and refined mesh is given below:

CONCLUSIONS:

• The velocity at which the maximum heat transfer coefficient is obtained is 10 m/s
• The potential hotspot in the system is that of the Processor.
• The more refined the meshing especially along the components where the heat generation term is involved,the better the results will be as the problem will be more close towards the practical approach.
• As the velocity is increased ,the Reynolds number will be increased which inturn increase the Nusselt number and heat transfer coefficient and further leading to better heat transfer rate
• For the efficient cooling of the graphics card using the forced convection heat transfer, the velocity must be kept maximum within the practical limitations.

Thank You

Arjun Vs