Can initial temperature values be assigned to nodes?

[ArdavanY]

Hello,

I generated a cubic part and partitioned it to 8 smaller cubes. Using predefined field, I assigned a different temperature to each cube. The objective is to observe how heat is transferred from warmer areas to colder areas. I meshed the part to eight cubic elements, which coincided with the 8 partitions (I prefer not to use finer mesh). The problem is that the nodes have double or more values; for example, the node at the middle has 8 values. Is there a way to assign initial temperatures directly to the nodes? What is the best way to solve the heat transfer problem that I described?

Thank you,
Ardavan.

 

[corus]

You can do this in Abaqus by meshing each region as separate parts and using the predefined field to apply the initial temperatuers. For contacting surfaces (lines if it's 2D) you can then tie the surfaces together or use the *equation to link freedom 11 across.

Tata

 

[ArdavanY]

Tata,

Thank you for your response. Actually, I had done a similar thing before: Generated small cubes, meshed them separately and gave initial temp to each. Then I used the interaction module to connect them. There I had to creat an interaction property. The type of interaction was "contact". Then I had to define a clearance dependent "thermal conductance". I used the same conductivity number value as that of the cubes for zero clearance, and used some distance value for conductivity zero. But this method didn't work well and when I ran the program I could obviously see that the interface was not as conductive as the parts. Have I done what you meant? Actually I am not sure what you meant by "tie the surfaces together or use the *equation to link freedom 11 across" and how to do that. I would appreciate some explanations.

Thank you,
Ardavan.

 

[corus]

Your method is the correct way to do this. With unequal temperatures between two contacting surfaces you'll find that the temperature change at the interface will be rapid and so a finer mesh will be required there to capture the temperature gradient there.
I don't see how the contact interface can be less conductive than the parts. If there was an error in the way you have set the job up then most likely you'd see no heat flow at the interface. How the temperatures change in the parts will depend on the density, specific heat, as well as the conductivity.

Tata