Thermal simulation

[bili]

Hi,
I need any information about "thermal simulation".

Thank you in advance.

bili,

 

[therm0]

Hey Bili,

I took a course at the University of Manitoba last term that covered some of this type of material. Generally I believe Finite Difference Method is used. Simply put you divide the area you want to analyze into a finite mesh, and then establish Boundary conditions around the edges of the mesh, and initial conditions at each node on the mesh. Each 'cell' of the mesh has 4 sides, and is called an element, (I suppose it could have any number of sides). Next an energy balance equation is written for each element, and you end up with a system of equations that can be solved, usually by writing a computer program.

The course only covered the methods for solids, I assume because in fluids all sorts of fluid mechanic properties play a role, and make it more complicated. I have heard of commercial programs that do thermofluid simulations though.

Hope that helps.

- Mike

 

[bdiamond]

You should check out CFD (computational fluid dynamics) programs such as "Coolit" or "Flotherm". They are very good simulation programs.

 

What do you need to know?

 

[Haf]

BJ asked a good question ("What do you need to know?&quot. The subject of thermal simulation has been studied for decades. Any heat transfer text book would cover numerical simulation, and there have been thousands of articles written over the years. There are several commercial codes available, and, in the case of relatively simple problems (with simple geometries), simple computer codes can be written relatively quickly.

If you want a simple description of thermal simulation, here goes:

Thermal simulation generally refers to the numerical modeling of a thermal system. It is generally used when the governing equation is too difficult (or even impossible) to solve in closed form. The governing equations are solved numerically by dividing the domain into small segments and employing some pre-described numercial method to solve the system. There are literally dozens (maybe even hundreds) of ways to numerically solve a system. These methods vary in accuracy and simplicity (generally, the more simple, the less accurate).

Haf