ERROR(819): VENT null is .....; with multiple meshes model

[Vanillama]

Good morning everyone,

I am currently running a quite big model I am trying to reduce using PERIODIC boundary conditions.

I already successively launch the case using only OpenMP parameters.

I would like to try to run it using multiple meshes in order to accelerate the calculation using MPI processes with the cmd line BUT I am encountering some issues I don't succeed to solve
"fds_local -p 5 -o 2 ModelName.fds"

It is related with the ERROR(819)

Below the screenshot of the cmd console

I enclose to the message the case .fds file in case of need (.txt format).

BuildingVentReduced_Meshes.txt

Thank you by advance for your help.

Regards :)

[Vanillama]

For information,
The problem seems to come from boundary condition lines, here

[rmcdermo]

Try

&VENT PBY = -1, SURF_ID = 'PERIODIC' /
&VENT PBY = 6, SURF_ID = 'PERIODIC' /

[mcgratta]

I made the correction that was suggested and the case is running for me. The -1 and 6 are the lower and upper bounds of the y direction.

[drjfloyd]

This case is 21 million grid cells. It is going to take a decent amount of free RAM to run this. If you only have 32 or 64 Gb it may not be enough.

[mcgratta]

The case is still running on a linux cluster after 400 time steps. I cannot run this on my Windows computer because it requires too much memory.

[Vanillama]

Thanks for your reply all

The model is a 5.8 millions cells (50016670). It's big but I thought the computer could have made it.

I will try to increase the size of the cells to 20 cm. It 's a bit big but I will see how it's going. Unfortunately, I don't see how I can reduce the mesh size. 10 cm / cells seems already quite big for me.

I am closing this topic

[mcgratta]

The case is still running on our cluster. Each of the five processes is using about 2.5 GB RAM. You might need a bigger boat.