Stacked multi-cuboid geometries with PDE Toolbox for thermal analysis

With the PDE Toolbox, I know that I can use the multicuboid function to create a stack of cuboids for thermal analysis, each with their own thermal characteristics. However, it appears that each of the resulting cuboids must have the same width and height as all other cuboids in the stack. To more accurately represent a power semiconductor that I am analyzing, each cuboid would ideally have a different size (thickness, width, and depth). Is it possible to create such a geometry?

Answer 1

[Alan Weiss]

No, you can currently create a stack of cuboids only with the same width and depth. They can have different thicknesses, as in this example.

Alan Weiss

MATLAB mathematical toolbox documentation

[David Naviaux]

Thank you for your quick response. Is it possible to create a stack of cuboids with different dimensions and thermal characteristics in some other way than using the multicuboid() function?

[Alan Weiss]

Sure. Export STL geometry from a CAD program, then import it to your model.

But if you are asking if there is another way in PDE Toolbox, the only thing I can think of is to make 3-D geometry from a point cloud, though I am not 100% sure that this will be satisfactory.

Alan Weiss

MATLAB mathematical toolbox documentation

[David Naviaux]

The main problem that I am having is that I have not been able to create a geometry for thermal analysis that consists of a group of geometric shapes each with their own thermal characteristics. It seems like the PDE Toolbox thermal analysis is designed to allow that by allowing a geometric object to contain multiple cells. None of the examples I found showed importing an STL file that had multiple cells defined. Since it could be time consuming to get up to speed on using a 3D CAD software package, I just wanted to make sure that this was possible first.

Answer 2

[Svetlana Pease]

Hi David,

You can define a multi-cellular geometry by importing it from an STL file only if the cells are not connected. The last example here shows a multi-cellular geometry imported from STL. In all other cases importGeometry creates one cell.

Regards,

Svetlana Pease

MathWorks Documentation Group

[David Naviaux]

Thank you! After some experimentation using TinkerCad, it appears that if the cuboids touch, when imported they are combined into one cell (as you explained). Since leaving a gap between the cuboids would be problematic for thermal analysis, It doesn't look like I can solve my problem by importing an STL file. Can I create a multi-cellular geometry by combining multiple geometries via a sequence of Matlab commands in someway or by defining the thermal characteristics as functions of position within the stack of cuboids that have been imported as a single cell?

[Giulia Ulpiani]

Hey David, I have the same issue in my current project. Did you solve it somehow?

Thank you,

Giulia

[jin thomas]

hi David, Did you solve it somehow?

Thanks，

zhangjin

Answer 3

[David Naviaux]

Hi jin thomas,

I was not able to figure out how I could create the model and I never got an answer. I had to move on.

If you come up with a way to handle the modeling, I would appreciate any input you have. I would like to be able to model the thermal characteristics of a semiconductor. For example, of interest would be to create a more complete thermal model that I could use MATLAB to evaluate. The model would be of the semiconductor heat flow from the molding compound, the die, the solder, the lead frame, the thermal paste and heatsink. I can do this roughly without MATLAB, but using a complete thermal model would provide so much more information.

Regards,

David

[Erebros]

Good morning,

I am having the same issue than you have. I reduced a 3D process to a 2D model due to the symmetry in the z-axis (cylindrical coordinates). Unfortunately I have multiple areas which are directly in contact with each other. If I define the first geometry everything works just fine. But by adding the other geometries the program adds the cross-sectional areas of the two geometry in such a way, that I end up with only one structure containing the combined surface of both geometries. This is unfortunately not acceptable due to the differences in material properties.

Is there any chance that you resolved the problem in another way?

If I find a solution/workaround to this problem I will let you know.

Best regards,

Erebros

[David Naviaux]

Erebros,

I just noticed your queston. I realize that it has been almost 2 years ago, but I thought I should let you know that I did not find a work around and have had to move on.

Please let me know if you found a way around this limitation.

Thank you,

David