Calculating heat transfer in heat sinks by natural convection

[dcruzKA46W]

Hi everyone,

This is my first publication here in the forum and I would to know if somebody could help me with this simulation.

I´m trying to calculate the existing gradient of temperatures in a heatsink with the different heat fluxes in the semiconductors.

To avoid use a fan, the heat transfer will be produce by natural convection, between the vertical fins and the ambient air at 50 celsius degree. The way of get the final results has been by means of the nolinear steady state heat transfer analysis with the different heat focus.

I have many doubts that the analysis is correct because the temperatures in the base are very high, around 180 and 240 celsius degree. These values produce the destruction of the components!! In the convection loads I only can select the temperature of the ambient, the heat transfer area and the heat transfer coeffiient, which I calculated using a correlation of the fluid mechanics for natural convection. In my opinion, the positon of the heatsink is very important in this kind of natural convection calculating and I don´t know how the program know this. So how can I select in the program the vertical position of the heatsink? Is it correct the way of calculate this heatsink?

Somebody calculated any heatsink in natural convection? Any example?

I attach too the file simulation for future comments.

Thanks in advance.

[John_Holtz]

Hi @dcruzKA46W . Welcome to the Inventor Nastran forum.

If you calculate the heat sink temperature by hand, what result do you get?

• Total heat input = 3.93 W/m^2/C*(area of fins)*(T-50 C).

I think the temperature is hot and close to what Nastran is calculating. Are you sure your heat input is correct? Are you sure your convection coefficient is correct? (With a temperature gradient this large, the convection coefficient will be larger.)

I do not fully understand your question about the position of the heat sink and how the software knows it. The software knows the value that you enter for the convection coefficient, and that is all it needs to know. You need to calculate the convection coefficient based on the orientation of the heat sink and the other factors that are in the calculation of the convection coefficient. 

Also, what version of Inventor Nastran are you using?

[dcruzKA46W]

Hi John,

First of all, thank you for your fast response.

By other hand, you have reason with the calculating by hand, there is a similar result with the simulation.

I reviewed all input values again, all are corrects (numbers and units). The heat inputs were obtained with a electrical software simulator and we think that they are going to seem with the real test. The convection coefficient was obtained using the theorical Yovanovich´s correlation and I´m not sure that this value is correct because I don´t have the way to verify it yet.

Do you think that the simulation proposal by me it´s correct? ¿Is it necessary any change like the boundary conditions?

If I would like to simulate this case in forced convection, ¿are there any way to select a fan like a boundary condition or the simulation will be the same that last case with the change of the convection coefficient value?

¿Is there any way to introduce the thermal resistance value of the heat sink or is nastran who calculates it in function of temperature and thermal conductivity material?

My version of Autodesk Inventor Nastran is 2022.1.0.143.

Thanks in advance for your time.

Best Regards.

[John_Holtz]

Hi

"Do you think that the simulation proposal by me it´s correct? ¿Is it necessary any change like the boundary conditions?" The method is correct. The accuracy of the results depends on the accuracy of the input (and the mesh size).

"If I would like to simulate this case in forced convection, ¿are there any way to select a fan like a boundary condition or the simulation will be the same that last case with the change of the convection coefficient value?" Nastran is a general FEA program which does not include fluids. It sounds like you want to use Autodesk CFD and let the analysis calculate the effects of the fluid motion. Otherwise, you calculate the convection coefficient for forced air and enter that value into Nastran.

"¿Is there any way to introduce the thermal resistance value of the heat sink or is nastran who calculates it in function of temperature and thermal conductivity material?" I do not understand what you mean by thermal resistance of the heat sink. If the thermal resistance is related to the overall all heat transfer that includes the effects of contact resistance with the heat source, thermal conductivity of the material, and convection coefficient to the ambient air, then the answer is that is what you calculate my modeling the heat sink with conductivity and boundary conditions.