Thermal simulation with strange results of temperature lower than ambient

[a.fabregat]

Hi,

I have done a thermal simulation and I don't understand the results. I have aplied an internal heat (1W) load as well as radiation (E=0.9) and convection (5 w/m^2K) loads to external surfaces with an ambient temperature of 20ºC. The results show a minimum temperature of 16.6ºC which I do not understand how it can be below ambient (20ºC). You can see the results in this image.

If I increase the internal heat power to 10W the problem disappears because the minimum temperature is much higher, but I would like to understand why the simulation with a 1W gives temperatures below ambient.

Thanks in advance.

[John_Holtz · Accepted solution]

Hi @a.fabregat

The short answer is the problem occurs because of mathematics. The effect is exaggerated when the elements are too large or distorted, or when the heat flow changes direction rapidly within an element. You can try a finer mesh in the areas of the problem, but it may not be possible to eliminate the problem entirely. (It depends on the model, thickness, materials, loads, etc.) As long as the result is not too extreme, the usual "solution" is to ignore it.

The longer answer is the following (from the article on the Knowledge Network, Temperature results of a simulation are out of range😞

The abnormal temperatures occur because

1. The heat flow through the elements is calculated based on the applied loads and the element shape.
2. The temperature of the nodes is then calculated based on the calculated heat flow through the element and the element shape.
3. Sometimes the temperature gradient necessary to support the calculated heat flow through the element can result in one of the nodes being colder (hotter) than it should be. Such conditions usually occur at a corner or edge of the model that has a load applied.

The abnormal temperature is less accurate (colder than the coldest applied load, or hotter than the hottest applied load) when the element is large, distorted, the applied load has a high magnitude (such as a large convection coefficient), or when the thermal conductivity has a large change from one part to another.

Let us know if you have any other questions.

[a.fabregat]

Thanks for the answer. I have applied a mesh control in the edges with lower temperature and results are more consistent.

By the way, one new feature is supposed to be able to generate XY plots along paths or with probe values, but I can not find it. Is it already implemented?

Thanks in advance

[John_Holtz]

Hi @a.fabregat

I am not aware of the ability to do path plots (plots along an edge), but graphs of the results at a probe can be done. Of course, this only applies to transient analyses. (It probably applies to analyses with multiple steps. A nonlinear static analysis is not really a "transient".) For reference, here is the documentation: Review Results.

A thermal analysis does not have transient results or multiple steps, so you cannot create a graph of the results at the probe locations.