[SOLVED] transient heating of 2D object

[chatpall]

Hi all.

I would like to ask you for some advices in preparation of transient heating example.

I am heating an 2D rectagle object with several holes in it. My boundary conditions are set for 60 ° (C or K, it does not matter for this example simulation) on lower boundary, side boundaries are set as isolants (heat flux is 0.0 W/Km2), upper boundary is set to heat transfer coefficient 10 W/mK with external temperature of 30 °(C or K). Please do not reffer to the physical justification of this example, it is just a way I am trying to understad how can Elmer help me to make some decision in my technical practice. I have attached the picture export of steady state simulation.

My question is? When I am trying to make transient simulation with such setup (for example 20 to 1 timing) I do not obtain any "real" results. I have attached picture for the first step of the resulting transient simulation - other steps visualization looks the same.

Thank you very much for any advice.

[mzenker]

Hi,

could you post your sif file?
What are the dimensions of your object?

Matthias

[annier]

Hi Chatpall,
I have come across such type of result when I perform transient simulation in case of such solvers.
There are two things which i would like to infer:
i. Scale of dimensions
ii. Time scale of simulation
A realistic simulation is attained with an accurate considerations of length and time scale.

Notes:
a. Attaching sif is always a clear way of stating our queries.
b. In transient simulation, for a small red colored region at the bottom, i guess that either a very small time scale is considered or a very large domain (than the heating power) is considered.
c. The thing i need to understand for transient simulation is "why is the temperature of the region becoming negative- it is not possible if it is in Kelvin?" [The same thing happened to concentration value(becoming negative) when i tried to do transient simulation in advection-diffusion solver].

Query from my side:
How did you manage to draw a wonderful geometry? Are they bubbles inside a liquid?


Yours
Anil Kunwar

[chatpall]

Hi,

could you post your sif file?
What are the dimensions of your object?

Matthias

I am attaching the sif file.

The dimmensions should be 80 x 40 mm. I have created geometry in Freecad, than imported it into salome meca in brep file and created mesh in salome mesh module. Maybe there is some missinterpretation in dimmensions, but I have never met with dimension importing problems.

[chatpall]

Anil, thank you for your answer.

As I wrote to Matthias, maybe there can be some dimmension importing issues with my brep to salome meca mesh file.

I will check the dimmensions and try to reproduce my simulation with new geometry.

Time scale of simulation was chosen in "proper" way, 20 seconds per 1 second. Or maybe not.

This geometry (holes in a plate) was used just to see how the temperature gradient changes with missing volumes of plate "material".

Hi Chatpall,
I have come across such type of result when I perform transient simulation in case of such solvers.
There are two things which i would like to infer:
i. Scale of dimensions
ii. Time scale of simulation
A realistic simulation is attained with an accurate considerations of length and time scale.

Notes:
a. Attaching sif is always a clear way of stating our queries.
b. In transient simulation, for a small red colored region at the bottom, i guess that either a very small time scale is considered or a very large domain(than the heating power) is considered.
c. The thing i need to understand for transient simulation is "why is the temperature of the region becoming negative- it is not possible if it is in Kelvin?" [The same thing happened to concentration value(becoming negative) when i tried to do transient simulation in advection-diffusion solver].

Query from my side:
How did you manage to draw a wonderful geometry? Are they bubbles inside a liquid?


Yours
Anil Kunwar

[mzenker]

Hi,

for a transient simulation, you should provide an initial condition - it is missing in your sif.

If you use ElmerGUI, you can check your dimensions by opening Model->Summary. The length units in Elmer are m by default.

HTH,

Matthias

[annier]

Hi Chatpall,
1.) as per my understanding, your total timesize = timestep interval *timestep size = 20 *1 = 20 s.
The solver will iterate 20 times and each timestep will be 1 seconds.
(A time scale size being very small may also give a simulation result like above for a given geometry during transient simulation.)

2.) In Salome, the dimensions of a geometry can be measured in geometry module by selecting the required geometry and using the following:
Measures>>>>Dimensions>>>>Bounding Box
(A domain size or length dimensions very larger than the defined heating power can give a result with a very narrow heated portion at the bottom for a given total time in transient simulation)

Yours
Anil Kunwar

[chatpall]

Matthias, anil.

Thank you very much for your answers!

I have made mistake considering my model from salome in mm, but it was in m!!! There was the discrepancy. When I changed the model dimensions I am obtaining "real" results! Thank you very much for your time!

P.

[annier]

Hi Chatpall,
1. You can change the model dimensions
or,
2. Provided the geometry dimensions are in m, you can use coordinate scaling in SIF:

Code: Select all

Simulation
...
Coordinate Scaling = 0.001
...
End

Yours
Anil Kunwar