Week 6 - CHT Analysis on a Graphics card

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

Objective:

• To understand the setup and requirements of a conjugate heat transfer problem.
• To understand and evaluate the characteristics of such problems.
• To evaluate the temperature profiles and heat transfer coefficient at the region of interest (the processor and fins).
• To understand and evaluate the specific region(s) of interest (such as hotspots).

Introduction:

The term conjugate heat transfer relates to the study of thermal interactions occurring at the interface between solid(s) and fluid(s) which includes convective (in fluids) and conductive (in solids as well as fluids) heat transfer where the convection can be either free (or natural) or forced. The analysis is governed by the set of equations that are conformal with the physics that separates the solid and fluid domains. The equations utilised are solid heat conduction (based on Fourier's law) and Navier-Stokes equations (Reynolds' Averaged (RANS) in case of turbulent flows). The applications of CHT analysis, although not limited, are profound in mechanical, aerospace, chemical, electronics, and nuclear engineering.

Computers today frequently employ processors and graphics devices. As computer functionality continues to advance, more and more power is needed to keep up with computing capabilities. The CPU, the processor base, and the cooling system are the three main parts of a graphics card. In this project, the processor will have a set of fins added to the top to absorb heat through conduction, and the graphics card configuration will be exposed to a cross-flow of air to absorb heat through convection. Convection generates more heat transfer from the graphics card to the surroundings and away from the heat source, hence the addition of airflow improves the cooling of the graphics card.

Workflow:

1. Preparing the geometry
2. Generate mesh for the geometry
3. Setting up the solver
4. Setting up the solution monitors
5. Post-processing to observe the results

Ansys Fluent:

In the Ansys workbench project window, Fluid Flow (Fluent) is dropped into the workspace.

Preparing the geometry:

The computational domain, as shown (provided as a STEP file) is imported in SpaceClaim. The cleanup work begins with checking for any redundancies. As a procedure of the CAD cleanup, the topology of all the components is shared, by the options available under the workbench tab. The common edges and interfaces are properly shared between the components to ensure the formation of a conformal mesh.

Generating Mesh:

Meshing is the process, wherein the geometry is divided into smaller parts which are known as elements so that the iterative solver can solve the governing equations across the domain. In the meshing utility, a randomly provided element size is used which is consecutively revised (up to 20mm). Based on this baseline mesh, further revisions in the mesh were done. The relative coarseness of the mesh in the free flow region in the enclosure is good enough for checking in the air flow happening in the enclosure. In this case, the events taking place (heat transfer and temperature rise) at the processor and regions surrounding it are essential to be captured, and hence, the following changes were made. A body sizing with an element size of 10 mm was inserted over the graphic card body so as to refine the mesh over the entire graphic card. Similar treatment was given to the fins and the processor with an element size of 0.5mm. This ensured that the cell count is sufficiently high near the regions of importance and the data such as velocity, temperature and heat transfer coefficient are captured accurately. This treatment gave a total cell count of 370865 elements. The mesh and mesh metrics are displayed below:

Orthogonal Quality	Aspect Ratio

For implementing the boundary conditions, the "boundaries" must be clearly defined. To do so, the mesh utility itself has a tool; named selections. By selecting the respective faces and surfaces, the velocity inlet, the pressure outlet, and the symmetry condition for the graphic card enclosure were provided.

Setting up the solver:

An important part of a CFD simulation, the setup for solver is done to obtain the solution. Firstly, the units of important parameters are set to desired units (preferably SI units). In the models tab, in the viscous models, $k-\omega$ (shear stress transport - SST) is applied. The materials defined for solving this problem are tabulated as:

The pressure-velocity coupling is set to SIMPLE and the momentum and pressure are set to second-order discretisation. Report definitions are set to capture the coefficient of drag and lift over the car body surface. The corresponding plots along with the residual plots are saved. Hybrid initialization is used to initiate the problem, thereafter, 500 iterations are provided for the solver.

Post-processing (Results):

		Case: 1	Case: 2	Case: 3
		Velocity: $1m/s$ Avg. Temp.:$331.61K$ Max. Temp.: $331.49K$ Heat Transfer Coefficient: $13.603W/m^{2}K$	Velocity: $3m/s$ Avg. Temp.:$315.01K$ Max. Temp.: $315.97K$ Heat Transfer Coefficient: $22.667W/m^{2}K$	Velocity: $5m/s$ Avg. Temp.:$310.95K$ Max. Temp.: $311.07K$ Heat Transfer Coefficient: $26.905W/m^{2}K$
Average Temperature
Maximum Temperature
Heat Transfer Coefficient
Scaled Residuals
Pressure
Velocity
Velocity Vector
Temperature
	Graphics Card
	Graphics Card Base
	Fin
	Processor

 Conclusion:

• The surface heat transfer coefficient is a function that is dependent on various factors such as flow conditions (velocity, fluid properties, etc.), geometry of the body, surface conditions, etc. It is evident from the contours obtained that the surface heat transfer coefficient raised with an increase in velocity. 
• This increase in surface heat transfer coefficient also leads to a more enhanced and uniform cooling, the difference between average and maximum temperature attained by the processor reduces. It can be observed that this improvement also helps in reducing the temperature gain of the PCB (graphic card base).
• The hotspot regions are encountered in the fin and processor, and the regions of high temperature are highlighted by checking the local temperature of these bodies. The location of these hotspots, more or less, remained the same, only the temperature values changed. The refinement of mesh in this region helped in gathering the relevant insights with better accuracy.
• Fins cause the flow to diverge and separate and as a result, vortices are formed behind the fins. This rotating flow helps dissipate the heat conducted by the fins.
• Most electronic gadget failure is a result of poor (or restricted) heat dissipation. Hence, to understand and improvise this, a conjugate heat transfer analysis of graphics card was done.

References:

• Introduction to GUI based CFD using Ansys Fluent - Course provided by Skill-Lync
• Ansys Inc.
• Supporting Material: Comsol blog: Conjugate Heat Transfer
• Supporting Material: A-TO-Z GUIDE TO THERMODYNAMICS, HEAT & MASS TRANSFER, AND FLUIDS ENGINEERING
• Supporting Material: Application of Conjugate Heat Transfer Simulations for the Development of Ventilation and Cooling Systems for Large Hydro Generators
• Supporting Material: ASME: Specific Applications of Conjugate Heat Transfer Models
• Supporting Material: Fluid Mechanics - Frank M White
• Supporting Material: Heat Transfer - Y Cengel and A Ghajar