Problem with statistical integrations crossing OBST

[nicofrapplom]

Dear all,

I am writing since I am experiencing a strange issue, never encountered before or, better, never noticed before, in FDS.

I would like to compute the average velocity through a passage which has a non-rectangular shape. Clearly, the VENT that I use as a control surface is rectangular. In my mind, the VENT that crosses the OBST considers only the gas phase, so that the velocity flux that I compute with

&DEVC ID='VEL_SCALA_MOBILE_LUG', QUANTITY='U-VELOCITY',
XB=13.4, 13.4, -1.60, 1.60, 0.00, 3.60, SPATIAL_STATISTIC='AREA INTEGRAL' /

considers the presence of an OBST in it.

Same for the area that I need then to compute the average velocity (and others cross checks):
&DEVC ID='VEL_SCALA_MOBILE_LUG2', QUANTITY='U-VELOCITY',
XB=13.4, 13.4, -1.60, 1.60, 0.00, 3.60, SPATIAL_STATISTIC='AREA' /

The area that I get is the one of the VENT and not the one of VENT minus the cross section of the OBST cells. The mesh is 0.2 m x 0.2 m starting from [0.0,0.0,0.0].

Is this correct and with DEVC we cannot consider surfaces with obstacles in them, or am I doing/considering/assuming something wrong?

Attached, an example of the cross section, with in black the control surface.

Thank you very much in advance if you can help.

Best regards,
Dr. Nicolò Frapolli

[mcgratta]

The integration of velocity over the entire vent area should result in the volume flow through the tunnel with semi-circular cross section because the velocity should be nearly zero in the blocked regions of the vent. If this is not the case, create an Issue and post a simple input file that demonstrates the problem.

[nicofrapplom]

Dear Dr. McGrattan,
thank you for the answer.
Actually this is the case, it is right. But from the manual there is also the possibility to compute the area which should not include the blocked regions:

Particularly the sentence: "Note that this may differ slightly from what
is specified in the input file due to grid snapping for Cartesian geometries or due to the approximations used
to represent curvilinear geometries." indicates me that the integral of the area should exclude the blockages (OBST). Is it correct?

Thank you in advance for the feedback.

Best regards,
Nicolò Frapolli

[mcgratta]

The paragraph from the User's Guide means that the area of a vent in the model may be different than what the user specifies. If the XB coordinates on the VENT line do not line up exactly with the grid, then the AREA is going to be different than what the user has specified. This is not related to the blockages, but rather the grid size.

[rmcdermo]

I think it is reasonable for FDS to automatically account for OBST when integrating DEVC 'AREA'. Should not be hard to implement for Cartesian.

[rmcdermo]

This should be addressed with PR #16354 and available in the next release.

[nicofrapplom]

Dear all,

thank you for the replies.

For it is already important to know that what I am doing is correct and the output from FDS is also correct.

If in the next release this will be available, even better!

Thank you again.

Best regards,
Nicolò Frapolli