Simulating smoke

[cumulo_nimbus]

Hi all,

I have been trying to simulate smoke and fire in a confined space. Attached is the case I have been trying to run, but I just couldnt get the scalar and temperature to convect out into the air volume. Can anyone kindly help me out on this one? Thanks!

[srhusain]

Hi

looking at a the model briefly here are some points to consider:

1. Film coefficients are applied to internal surfaces which is not a valid setup- what is the intent of this condition when the fluid can naturally remove the heat from the heat source
2. Velocities are being applied to internal surfaces which is not a valid setup. If you want to model an internal fan i would specify that volume as such.
3.  Looking at the size of the fluid volume it appears to be about 100 meters tall- is this the size you intended?

[cumulo_nimbus]

Regarding the 100m size, I suspect it might be a matter of units conversion when CFD read in my .sat file. Back in my inventor, I actually moved the fire block above the ground by 50cm. Will get that fixed.

But are there any critical mistakes that I've made in the configuration?

- I know the run has to be transient but are the boundary conditions set to steady state? like the scalar, the velocity inlet, etc?

- is the time step of 0.05 appropriate? there isnt any units so I'm assuming it's in seconds.

- is incompressible auto forced convection the correct choice to use?

- I've set the air to variable but are there any internal changes required?

- ultimately I cant seem to convect the smoke and heat up into the room after a few hundred iterations.

[srhusain]

• I think applying proper boundary conditions and resizing the enclosure with proper dimensions will be key to get your analysis running correctly.
• The boundary conditions do not have to be transient.
• Running incompressible with Air-variable is Ok, it just means that the density is only a function of temperature and not pressure.
• You are correct in assuming that the time step size is always in seconds. For the time step size, you can experiment by using larger values if you want to get to steady state faster (assuming that the physics of your setup permits a steady solution). Having a very small time step just delays the system response (flow fields, temperature, etc) and will necessitate a correspondingly larger amount of time steps to yield a well developed solution.

[cumulo_nimbus]

thanks. managed to get the convection to occur 🙂