Is it realistic simulation? (Heat Transfer Simulation)
We are trying to do a heat transfer simulations with the below settings.
We think that the temperature distribution is not so correct(particularly in the area which has been circled in the pictures in the attachment)
you can see the pictures of meshing of our simulation and the residual as well.
In summary, We have a machine which has been subtracted by a block(as a room).Now we have a machine inside the room with 6 surfaces.
The Machine has the constant temperature(26'C) and the 6 surfaces has a different constant temperature(21'C) and without any Velocity or mass inlet.
We used just one region(air) because we have the constant temperature of the machine and the block's wall.
The pictures are for our simulations with these conditions:
-Segregated Flow-Fluid Temperature
-Ideal Gas(we used also Constant Density with Boussinesq Model with thermal expansion coefficient(0.003) and nearly same result was achieved)
*The coupled solver also used but the residual was not convergent.
Meshing: BaseSize : 11mm
-Prism Layer(3) (and also 12-no change)
Is there anybody know why the temperature distribution is not correct?
Thank you in advance.
Based on just looking at the images (not a great way to check to validity of a simulation) I would say the results seem believable.
The points of relatively low temperature that you point out may be areas where convection is causing a significant flow over the machine (You could check this by looking at velocity vectors or scalars).
What is the turbulence intensity? Is it by default 0.01? It is a good try if you put TI=0.04.
Can you change the place of plane section and see the result?
make sure that you determined the temperature of all surfaces to be 26 degree.
I checked the velocity streamline.but it seems there isnt any significant streamline in the mentioned areas.(the pictures is attached)
Also, I tried mostly the laminar model,,Do you know how can we find which model should be used for my simulation? laminar or turbulence?
I,ve read somewhere that if the area is not so big,then laminar case can be used,but im not sure.
Thank you both again in advance.
I think because the velocity is less than 0.3 m/s, you can use laminar model.
I hope you could show us another section at the middle of your geometry.
I tried to decrease again the mesh size of my shape.I increased the number of mesh from 9 million to 15 million cells,(with 12 prism layers on the machine) but again there are some problem particularly in the Heat Transfer Coefficient of the machine.
I attach some pictures of the new simulation.
Thank you in advance again.
|All times are GMT -4. The time now is 09:48.|