Conjugate Heat Transfer analysis of a Graphics card
Introduction:
The need for miniaturization and increased performance of the electronics components has contributed to the increased power consumption and heat generation within the system and an innovative cooling system is required to remove the heat which helps in achieving higher performance and a better life for the product. There are several techniques for cooling including various types of heat sinks, thermoelectric coolers, forced air systems, and heat pipes. The type of cooling system depends on the electronic component and the amount of heat generated by it.
The use of simulation tools helps us to understand the heat transfer phenomenon on the component and in the selection of a suitable cooling system. A conjugate heat transfer (CHT) model enables us to simulate the heat transfer at the interface between the solid body and fluid. The physical processes and solutions of the governing equations are considered separately for each domain. Matching conditions for these solutions at the interface provide the distribution of temperature and heat flux along the body–flow interface, eliminating the need for a traditional heat transfer coefficient surrounding it.
A conjugate heat transfer analysis was performed on a graphics card with the help of ANSYS Fluent, a commercially available CFD software. The aim of this study was to determine the maximum temperature on the surface of the graphics card. In the subsequent sections, the analysis of the graphics card is discussed from case description to post-processing of the results.
Case Description:
The graphics card is cooled using the phenomenon of forced convection. The graphics card is placed on a PCB base and the heat produced by it is conducted through the heat sink placed directly on it. The heat is then removed by the air that is flowing over the fins.
Model Preparation:
The PCB base, the graphics card, and the cooling fins are modeled in SOLIDWORKS and they are assembled as shown in figure 1. The model is then imported into Space Claim where it is prepared for simulation. An enclosure was constructed around the model for the purpose of simulating air flow around the component. The topology was shared in order to create a conformal mesh between the various components which is essential to avoid interpolation problems at the interface.
The model is meshed using the ANSYS meshing tool. First, a base case is simulated using a coarse mesh and in the second case, a finer mesh is used.
Base case: Coarse Mesh
Element size: 0.003 m
Number of Elements: 204414
Number of nodes: 38375
Figure 2 shows a picture of the sliced plane across the body to display the mesh and figure 3 shows the quality of the mesh elements.
A steady-state simulation is performed in the solver. The boundary conditions used in the simulation are summarized in table 1.
The materials assigned to the appropriate solid bodies and their properties are listed in table 2.
The processor is modeled as a constant heat source. The value of heat source was considered as 1e8 W/m3.
The simulation converged at 210th iteration which can be observed form the following figure 4. The temperature of the air at the outlet was 300.5 K.
Second Case:
The mesh is refined by reducing the size of the element to 0.0018m. A local sizing of 0.0008m was added to the faces of cooling fins with the help of face sizing option. The details of the mesh can be visualized from figure 5 and the corresponding quality of the mesh elements is shown in figure 6.
Element size: 0.0018 m
Number of Elements: 463298
Number of nodes: 86419
Results and Discussion:
The simulation results were post-processed in CFD-Post. The temperature profile on the surface of the graphics card processor, the cooling fins and the PCB base are displayed in figure 8, figure 9 and figure 10 respectively. It can be observed from the figures that the maximum temperature occurs on the surface of the graphics card processor because of the heat generation. The maximum temperature on the processor, cooling fins and the base is 368.42 K, 368.4 K, and 354 K respectively.
The second case is simulated with the same boundary conditions after refining the mesh. The temperature distribution on the surface of the processor, the cooling fins and the base are shown in figure 13, figure 14 and figure 15 respectively.
Refining of the mesh improved the quality of the results. Now, the maximum temperature on the processor is 366.8 K, 366.8 K, and 353 K respectively
Due to the low thermal conductivity of the PCB material, the heat generated is predominantly carried through the fins to be cooled by air, which can be visualized in the temperature profile around the whole component as shown in figure 11 and figure 16.