Conjugate Heat Transfer Analysis on a graphics card

Aim: To Perform a steady-state conjugate heat transfer analysis on a model of a graphics card.

Objectives:

• To perform a steady-state conjugate heat transfer analysis on a model of a graphics card.
• To understand the effect of mesh in various regions using both coarse and refined mesh.
• To study the effects on the graphic card by varying the velocities from 1m/s to 5m/s.
• To determine the maximum temperature and heat transfer coefficient obtained by the processor.
• To prove that the simulations are converged by showing plots and contours.

Description:

• Heat transfer is known as the passage of thermal energy from a hot to the cold material.
• The passage generally occurs in three different modes i.e. conduction, convection and radiation.
• Conjugate Heat Transfer (CHT) is used to describe processes which involve the variations of the temperature with solids and fluids.
• The CHT analysis is used to predict the surface temperature and heat exchanging capabilities of the engineering devices.
• The mode of heat transfer in solids is conduction and in case of liquids it is convection.
• Bascially heat transfer in solids is calculated by Fouriers law. The exchange of thermal energy between the two physical bodies is called as the study of heat transfer.
• The rate of transfered heat is directly praportional to the temperature difference between the bodies.
• The conjugate heat transfer corresponds with the combination of heat transfer in solids and in fluids.
• Forced convection is the most common way to achieve high heat transfer rate. Also provides the velocity and pressure distribution of the fluid inside the solid.
• The conjugate heat transfer analysis helps in measuring the heat transfer through two different regions. Helps to locate/ predicit the hot spot regions i.e. region that have maxximum temperatures during the flow.
• It plays a vital role in design optimization by improving the heat transdfer rate and provinding sufficient cooling system to it.
• With help of CHT analysis visualization of flow patterns and also trimming of geometry can be done based on the temperaturerequirement. It can be performed to improve the cooling performance of the water jacket and increase engine life.
• This analysis is used for Motor and Battery to get optimized design for high voltage systems.

CHT analysis on a graphic card:

Geometry:

• Initially the 3D model of the graphic card is generated using the cadd software and then it is imported into the SpaceClaim. Once the geometry is importedclearly check if all the parts are imported.

Creating the Enclosure :

• Using the enclosure option, enclose the imported graphic card geometry.

Meshing:

• Open Meshing Interface.
• Initially we start with the coarse mesh(baseline mesh). As we know that mesh with larger size cannot predict the accurate solution/results.This baseline mesh helps to compare other cases to compare and hence we can decide which cases is more effective and accurate.
• Non-conformal mesh means that the vertex should be linked each other and faces also should be connected to each other. We have a smooth heat transfer only when these are connected properly. For a CHT simulations we need a perfect overlap of mesh.
• To see that mesh is overlapped correctly, go to geometry > click on FFF > in propeties > select Share Topology = Share. Now both the interfaces are shared. Go to Workbench in the top toolbar select Share. This helps to create interfaces between two bodies and helps to see interfaces where information is shared. We can see pink circles appear on the shared surfaces.
• Go to Meshing interface and fist create the name the necessary surfaces and generate a basic mesh to visualize it.
• For the baseline mesh the element size is comsidered to be 5.e-002m. Number of inflation layers used around the body is 5 for the growth rate of 1.2. For this case, select Inflation option > Smooth Transition.
• Now generate the mesh.
• Also check the number of elements we got for the basile mesh. As we know that for acedamic licence we can go upto half a million cell count. To check the Element Quality go to Details of Mesh > select Quality > click on Mesh Metric. Plot is created to check the quality.

Setup:

• Go to Mesh > click on Check Mesh. This helps to check the quality of the mesh.
• Go to Setting up Physics > select Solver=Pressure-Based, Velocity Formulation=Absolute, Time=Steady State.
• Go to Setting up Physics > select Energy > also go to Models > go to Viscous > Select k-Epsilon Mode.
• Go to Materials

Material Type = Fluid

Fluent Fluid Materials = Air

Material Type = Solid

Fluent solid Materials = 1. Aluminium (graphic card base)

2. Copper (fins)

3. Gold (Processor)

In this case we consider the processor is the heat source , hence we should provide the source term . Go to cell zones > assign materials as mentioned above > for the processor add source term i.e. go to Source Terms > Energy (edit) > select Number of Energy Sources = 1 > enter the value i.e. 20000000w/m^3(considered for this problem) .

• After assigning the cell zones , go to Boundaries > change the values of inlet ( 3 cases- 1m/s, 3m/s, 5m/s). Also maintain temperature= 300K.
• Go to Solution > select Hybrid initiallization.
• Before calculating, set the plots and animation that are required for the problem.
• Enter the number of iterations and press Calculate.

Case 1: (Baseline Mesh)

Residual plot:

Area Weighted Average Temperature plot:

Temperature plot:

Velocity plot:

Pressure plot:

Velocity Vector Plot:

Wall heat transfer coefficient plot:

Hotspot:

Wall heat transfer coefficient -contour:

Mesh Refinement:

Mesh refinement is bascially done to find out the finest mesh that fits the body by providing the accurate results. It also depends on the software we are using as we have restrictions for the elements in student versions. Here in this case we refine the enclosure along with Fin, processor and the base of graphic card. Also standard inflation is considered in this case. For the academic version of software we can go upto half a million cells only, hence we adjust the mech accordingly to get the refined mesh. For this grid dependence test is performed to check the maximum elements that we can go upto.

Case-2: (Refined Mesh at Velocity = 1m/s)

Inputs-

Default Element size = 0.0025m

Body sizing (Fins) = 0.0005m

Body sizing (Processor) = 0.0005m

Velocity = 1m/s.

Number of nodes = 99155

Number of Elements = 485288

Residual plot:

Area Weighted Average Temperature plot:

Temperature plot:

Velocity plot:

Pressure plot:

Velocity Vector Plot:

Wall heat transfer coefficient plot:

Hotspot:

Wall heat transfer coefficient contour:

Case-3: (Refined Mesh at Velocity = 3m/s)

Residual plot:

Area Weighted Average Temperature plot:

Temperature plot:

Velocity plot:

Pressure plot:

Velocity Vector Plot:

Wall heat transfer coefficient plot:

Hotspot:

Wall heat transfer coefficient contour:

Case-4: (Refined Mesh at Velocity = 5m/s)

Residual plot:

Area Weighted Average Temperature plot:

Temperature plot:

Velocity plot:

Pressure plot:

Velocity Vector Plot:

Wall heat transfer coefficient plot:

Hotspot:

Wall heat transfer coefficient contour:

Results:

	Case-1 Base Mesh	Case-2	Case-3	Case-4
No. of Nodes	7051	99155	99155	99155
No.of Elements	30947	485288	485288	485288
Velocity	1	1	3	5
Max. Total Temperature (K)	318	315	308	306

• Here simulations are done for 3 different velocities for the refined case and we can see that we got converged sollution in all the cases.
• Initially for the baseline mesh we didnot get accurate results, where as after mesh refinement it is clearly oberved that the results are accurate. On the finer refinement of mesh we could see that temperature distribution is clear when compared to the baseline mesh.
• As the processor is in between the base and the fins , it doesnot cause major change in the processor side -heat transfer coefficient. Where as for fins heat transfer rate increases which transfers more from processor to the surrounding as we are increasing the velocity of air gradually. Hence, the temperature of the processor is decreasaing for the increase in the velocity of air. This is because of increase in the air fins and air base heat transfer rate.
• From the above plots its also clear that, hotspot region is formed near the processor(heat source) as they are directly connected to the base and fins.
• As the surface area of the fin is larger, it helps to absorb/ distribute the heat throught and maintains the proper cooling in the graphic card.

Conclusion:

A steady-state conjugate heat transfer analysis on a model of a graphics card is performed by varying the velocities for the refined mesh. Plots and contours that help to understand the maximum temperature and heat transfer coefficient are plotted . From the generated contours the hotspot regions are also identified i.e. processor, fins and the contact region where both processor and base coem in contact. Mesh refinement helped to get the accurate temperature distribution as well as the wall heat transfer coefficient. Hence it is clear that velocity change effects the heat transfer rate.