ERROR: Poisson initialization error

luca.mi...@gmail.com

unread,

Apr 12, 2019, 4:03:26 AM4/12/19

to FDS and Smokeview Discussions

Hi all,

I have this error :

ERROR: Poisson initialization error, Number= 1, Mesh= 1

in the attached file named height_1cell.fds

This error disappears if the height of the domain is discretized in three cells.

Is there a way to do not have this error having just one cell discretization in the height of the domain??

Thank you

height_1cell.fds

height_3cell.fds

Salah Benkorichi

unread,

Apr 12, 2019, 4:09:32 AM4/12/19

to FDS and Smokeview Discussions

You need minimum 3 cells in any direction for the pressure to function properly.

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+u...@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/a54fb7e5-df82-4e82-876d-ed56f26b1dc4%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

fde

unread,

Apr 12, 2019, 4:13:09 AM4/12/19

to FDS and Smokeview Discussions

Having 1 cell in a direction activates 2D calculation mode. (Example Case: Flowfields/helium_2d_isothermal)

Having problem less than 4 cells in a direction has to do with the poisson calculation as in the error.

When I had this problem before, I found a get around by having a 5 cells in a direction (for example), then filling the 2 cells with OBSTructions, which will leave 3 cells back.

luca.mi...@gmail.com

unread,

Apr 12, 2019, 4:19:55 AM4/12/19

to FDS and Smokeview Discussions

Thank you fde. Placing a an obstruction seems to me a very smart option! I still cannot get why the fact that 1 cell in a direction activates 2D calculation mode can cause an error in the poisson solver.

luca.mi...@gmail.com

unread,

Apr 12, 2019, 4:29:10 AM4/12/19

to FDS and Smokeview Discussions

I cannot use the option of the obstruction, since the mirror surface have to be placed in the boundary of the mesh to work. Is there a way to use fds in 2d, avoiding the problem of poisson initialization error?

fde

unread,

Apr 12, 2019, 4:52:05 AM4/12/19

to FDS and Smokeview Discussions

Can you explain what do you aim to accomplish with your model? Also, do you intend to use 2d model or 3d?

luca.mi...@gmail.com

unread,

Apr 12, 2019, 5:27:19 AM4/12/19

to FDS and Smokeview Discussions

My intent was to use the 2d model with two different velocities imposed at the boundaries

In the chapter 6.3.2 of the user guide, it seems that the only way to obtain the 2d model is to set to 1 the J (in the y direction). (I thought that J was just an example). BTW Using this method it works!!

Thanks

Randy McDermott

unread,

Apr 12, 2019, 6:35:04 AM4/12/19

to fds...@googlegroups.com

J is not just an example. For 2D, you must use the X-Z plane. Gravity is in Z by default.

--

You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+u...@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/76b6c0be-00d5-4632-a2b6-d4000a4e657f%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Sent from my iPhone

Reply all

Reply to author

Forward