Validation Case: LED Thermal Management – Chip on Plate (COP)
This case validates the thermal management inside an LED against the CFD and experimental results obtained in the conference paper, “Thermal Analyses and Measurements of Low-Cost COP Package for High-Power LED”\(^1\), using the Conjugate Heat Transfer v2.0 analysis type in SimScale.
The geometry consists of a cylindrical LED module constructed using the images from the conference paper\(^1\) with certain thickness assumptions for the Silicone encapsulant and the Aluminum Pkg substrate.
Figure 1: Detailed LED module CAD geometry with associated parts used for the validation case
A cubic domain, with seven times the size of the LED module, is also constructed to model the air region surrounding the module in each direction.
Mesh and Element Types: The mesh is generated using the standard meshing algorithm in SimScale.
Mesh
Number of cells
Element type
F0 (Coarse)
153000
Tetrahedral and hexahedral
F5 (Moderate)
658000
Tetrahedral and hexahedral
F8 (Fine)
3500000
Tetrahedral and hexahedral
Table 1: Standard mesh characteristics for three different fineness levels: 0, 5, and 8
Figure 2: LED module meshed with a fineness of 5 with tet and hex elements using the standard algorithm
Simulation Setup
Material
Solid: The materials constituting the LED are tabulated with their properties as follows:
Materials
Thermal conductivity \([W/m\ °C]\)
Mass density \([kg/m^3]\)
Specific heat \([J/kg\ °C]\)
Silicone (Encapsulant)
0.3
1200
2000
LED Chip
42
2330
712
Sapphire (Chip substrate)
35
3980
761
Ag epoxy (die attach)
8
2300
671
Aluminum (Pkg substrate)
220
2702
910
Thermal Grease
3.6
1180
1044
Copper (Heat sink)
380
8800
380
Table 2: Material properties for the solid parts in the LED module
Fluid: Air
Viscosity model: Newtonian
Kinematic viscosity \((\nu)\): 1.469e-5 \(m²/s\)
Density \((\rho)\): 1 \(kg/m^3\)
Boundary Conditions
As in the reference paper all the faces of the LED module are modeled as no-slip walls with the temperature being zero-gradient. All the faces of the air domain are also modeled as no-slip walls with a fixed temperature value of 24 \(°C\) except for the top face which is open for natural convection.
Figure 3: Mostly the wall boundary conditions are used for the solid and the fluid domain. The top face of the fluid domain is open for natural convection.
The LED module is placed in the air domain such that the Silicone encapsulant faces the direction in which gravity is modeled (+z axis).
Power Source
The experiments mentioned in the conference paper\(^1\) were performed at three different input powers: 0.96 \(W\), 1.22 \(W\), and 1.47 \(W\). The paper assumes that only 78% of the input power is transferred into heat, which is uniformly distributed over the entire chip volume. Thus, the LED Chip is modeled as an absolute power source with the values of 0.749 \(W\), 0.952 \(W\), and 1.147 \(W\) respective to the power inputs used in the experiment\(^1\).
You can read more about power sources in SimScale in the following documentation:
The simulation results from SimScale are compared against the CFD results and experimental results from [1]. The results for the temperature of the surface-junction between LED Chip and Chip Substrate in \(°C\), for various power inputs, are tabulated as follows:
Power input \([W]\)
SimScale \([°C]\)
CFDesign 8.0 \([°C]\)
Experiment \([°C]\)
% Deviation SimScale
% Deviation CFDesign 8.0
0.96
52.12
52
55.1
5.41
5.63
1.22
58.48
57
61.4
4.76
7.17
1.47
64.39
62
69.3
7.09
10.53
Table 3: Result comparison for the temperature of the surface-junction between LED Chip and Chip Substrate
Figure 4: Result comparison for the temperature of the surface-junction between LED Chip and Chip Substrate
Mesh Sensitivity Study
For the power input of 1.47 \(W\), a mesh sensitivity study was also performed for meshes with a fineness of 0 (F0), 5 (F5), and 8 (F8). This study is plotted below:
Figure 5: Mesh sensitivity study performed for the power input of 1.47 \(W\) for three different fineness 0, 5, and 8 shows that the results are reliable
Conclusion
The discrepancy is likely due to the difference between simulation setup and physical testing environment. Unfortunately, full experimental details were not provided in the conference paper.
The following conclusion can be made from this LED thermal management validation case:
Reasonable agreement with experimental trends
Good agreement with alternate CFD software CFDesign 8.0
Outperformed alternate CFD in % result deviation
Figure 6: Temperature distribution around the LED module as viewed in SimScale’s online post-processor
This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.
Strictly Necessary Cookies
Strictly Necessary Cookie should be enabled at all times so that we can save your preferences for cookie settings.
If you disable this cookie, we will not be able to save your preferences. This means that every time you visit this website you will need to enable or disable cookies again.
