Fill out the form to download

Required field
Required field
Not a valid email address
Required field
Required field

Documentation

Validation Case: RF Electronics Package Cooling

This study aims at validating the CHT v2 solver of SimScale GmbH. A peer-reviewed publication of R.Boukhanouf\(^1\) focussing on thermal analyses and cooling of a Radio Frequency (RF) electronics package has been used as the basis for this validation case. Reasonable assumptions and approximations have been made to bridge uncertainty in the publication.

The validation study involves qualitative and quantitative comparisons of the

  • temperature, and
  • velocity vectors

between SimScale’s CHTv2 solver and the commercial CFD code: FloTHERM.

Geometry

The battery pack was modeled using the CAD tool Onshape. Necessary assumptions were made to overcome missing/inconsistent geometric data:

the model from the publication was reversed engineered to be used in this validation case
Figure 1: A 1:1 model (right image) reverse-engineered from measurements and images provided in the publication (left image).

The model contains the following parts:

  • DC Shelf
    • Upper Copper mount for DC component
  • DC Shelf components
    • DC component with 15 \(W\) Heat Rating
    • Thermal via circuit built into RF4 PCB
    • Solder layer
    • Thermal insulating material (TIM)
  • RF Shelf
    • Upper Copper mount for RF components
  • RF components
    • Three RF components with 60 \(W\) , 0.5 \(W\), 0.5 \(W\) Heat Ratings
    • Aluminium RF4 PCB
    • Dielectric substrate to insulate components from each other
    • Solder layer
    • Thermal insulating material (TIM)
  • Heat sink: Finned Heat Exchanger  
  • Aluminium Enclosure Box
  • Fan: Air delivery 11 \(m^3 \over \ h\) at 40 \(^o C\)

Analysis Type and Mesh

Tool Type: OpenFOAMⓇ

Analysis Type: Incompressible, steady-state analysis with the Conjugate Heat Transfer v2 (CHT v2) solver.

Mesh and Element Types:

The Standard mesher algorithm with tetrahedral and hexahedral cells was used to generate the mesh:

the resulting mesh with the standard meshing algorithm
Figure 2: The three-dimensional unstructured mesh containing tetrahedral and hexahedral elements was created with the Standard algorithm.

A mesh sensitivity analysis has been carried out to determine the dependence of the CHTv2 solver temperature predictions on the mesh:

Mesh CountMesh TypeTC@DC Component \([°C]\)TC@RF1 Component
\([°C]\)
TC@RF2 Component
\([°C]\)
TC@RF3 Component
\([°C]\)
Mesh 13.4M cells, 1M nodesStandard106.72148.8588.2588.25
Mesh 24.4M cells, 1.2M nodesStandard105.95148.6587.7587.75
Absolute TC deviation (Mesh2-Mesh1)0.770.20.50.5
% TC deviation (Mesh2-Mesh1)0.73%0.13%0.57%0.57%
Table 1: The results of the mesh sensitivity analysis after the area-averaged temperatures of the package components have been compared.

TC stands for the Temperature in Celsius degrees. The values in the table are area-averaged values over the respective components.

It is indicated that the temperature deviation between Mesh 1 and Mesh 2 was found to be within 1% for all the package components:

deviation between meshes to decide the most appropriate set for the validation case
Figure 3: Based on the mesh sensitivity analysis, the fine mesh (Mesh 2) has been used for all comparisons.

Simulation Setup

Fluid Material:

  • Air
    • Dynamic viscosity \((\mu)\) = 1.83e-5 \(m^2 \over\ s\)
    • Specific heat = 1004 \(J \over\ (kg \times\ K)\)

Solid Materials:
The table highlights the materials and thermal conductivity values used for each component of the RF electronics package:

ComponentMaterialThermal conductivity \(W \over \ (m \times \ K) \)
Enclosure BoxAluminum180
RF & DC shelvesCopper385
Solder layerTin50
Thermal via circuit(Derived)22 (Derived)
TIMSil Pad® and Gap Pad®2
Dielectric SubstrateDielectric material0.6
PCB Aluminum 180
Table 2: Properties of the solid materials

The RF components have been modeled as using the thermal resistance network and therefore are not assigned any material properties.

The Heat Ratings and Thermal Resistance values used are respectively as follows:

  • RF1: 15 \(W\), 0.7 \(^o C \over \ W\) 
  • RF2: 0.5 \(W\), 0.8 \(^o C \over \ W\)  
  • RF3: 0.5 \(W\), 0.8 \(^o C \over \ W\) 
  • DC: 60 \(W\), 0.78 \(^o C \over \ W\)

Boundary Conditions:

  • Natural convection inlet-outlet with an ambient temperature of 40 \(^o C\)
  • Fan as a momentum source with a velocity of 1.85 \(m \over \ s\)
  • Thermal Resistance Network: Star Network Resistance Model
  • DC Component:
    • Resistance: 0.78 \(K \over \ W\)
    • Power Source: 15 \(W\)
  • RF1 Component:
    • Resistance: 0.7 \(K \over \ W\)
    • Power Source: 60 \(W\)
  • RF2 Component:
    • Resistance: 0.8 \(K \over \ W\)
    • Power Source: 0.5 \(W\)
  • RF2 Component:
    • Resistance: 0.8 \(K \over \ W\)
    • Power Source: 0.5 \(W\)
  • No-slip walls
  • Coupled contact interfaces

Result Comparison

Convergence below 1e-3 has been achieved. Calculated physical quantities such as inlet pressure, outlet velocity, and cell average temperatures have also been allowed to converge to stable values.

The validation of the SimScale’s CHTv2 has been carried out by qualitatively comparing temperatures and velocities with the reference results\(^1\). The reference results have been produced with the commercial CFD code Flotherm. All post-processing was done in SimScale’s online post-processor:

temperature distribution on slice for qualitative comparison of results between the validation and SimScale
Figure 4: Qualitative comparison of the temperature distribution between the reference and SimScale results

Another section was compared between the two:

temperature distribution on package mid-section for qualitative comparison of results between the validation and SimScale
Figure 5: This comparison shows temperature distribution at the package mid-section between the FloTHERM and SimScale solvers.

A quantitative comparison of the velocity field shows a good match between the two solvers:

the velocity vectors where compared between the two solvers
Figure 5: FloTHERM and SimScale showcase consistency when it comes to the velocity vectors, to both direction and magnitude.

Except from the qualitative comparison, a quantitative study was performed:

ComponentFloTHERM
\([°C]\)
CHT v2 SimScale
\([°C]\)
TC deviation (SimScale – FloTHERM) \([°C]\) % TC deviation (SimScale – FloTHERM)
TC@DC Component101105.954.954.9
TC@RF1 Component 137.85148.6510.87.83
Table 3: A qualitative comparison of area-averaged temperatures on the package components has been performed.
comparison of temperature between literature and SimScale results
Figure 6: It is indicated that the component temperatures are over-predicted by the CHTv2 solver

From Table 3 and Figure 6 above, the DC and RF1 component temperatures are within 5% & 8% of the reference FloTHERM results respectively.

This deviation could be traced down to two possible factors:

  • The publication does not clarify how the thermal resistance has been implemented (top/board/side). 
  • Temperature measurement type in the paper.

Overall, The study shows a moderately good agreement between the temperature predictions from the two CFD solvers.

Note

If you still encounter problems validating you simulation, then please post the issue on our forum or contact us.

Last updated: November 17th, 2021

What's Next

part of: Validation Case: RF Electronics Package Cooling

Contents
Data Privacy