surface thickness/obst

[Pai]

I have a couple of dubts regarding the obstruction and the surface
thickness;

1) I have a wall which is 300mm thick but is made up of 2 different
layers. So in that case will I have a single obstruction with two
different surfaces (thickness equal to that of each of my layers)?

2) I have a wall with one material (brick), so is my surface thickness
going to be equal that of my wall thickness? I will have one
obstruction

I went through the manual and couple of issues posted earlier, but I
was a bit confused.

[drjfloyd]

1) At the beginning of Section 8.4 is an example of a SURF containing
two MATL layers. As noted in the caution statement in the box below
the example when a SURF is applied to an OBST, FDS always assumes that
the 1st node for the wall heat transfer is the first MATL in the
SURF. So if your OBST is exposed on both sides to the computational
domain, then you will need two SURF definitions with the MATL layers
reversed for one of them.

What specifically about this is confusing you in the manual? Is there
a particular point you would like phrased differently or see more
discussion of?

2) See the box at the end of section 8.1 of the User's guide. The
OBST thickness (i.e. the XB) defines how the hydrodynamic computation
sees the surface. The SURF THICKNESS defines how the heat transfer
computation sees the surface. The two parameters are independent.

What specifically about this is confusing you in the manual? Is there
a particular point you would like phrased differently or see more
discussion of?

[Pai]

The manual says ''The thickness of the wall indicated by the OBST line
need not match that indicated by the surface line''. Tbis is a bit
confusing for me if I have a single obstruction that is made up of a
single material.

For example if I have a timber floor (an obstruction) of thickness
100mm made up of oak, then in the material command, I specify the
various properties. But in the surface command, what should be the
thickness of this? Is it 100mm or something else (I assume that the
heat transfer computation needs the total thickness of the obstruction
too).

Would it not be better if we have an OBST command which is assigned
with a material and a thickness?


Thanks.

[drjfloyd]

SURF always has the true thickness. OBST may or may not have the true
thickness depending upon the needs of the hydrodynamic computation.
For example, I am modeling an large room that has smaller a enclosed
space inside of it. The walls of the enclosed space are typical gyp
board on wood or metal stud construction with a thickness of ~ 11 cm.
I can define the enclosed space as a series of OBST lines representing
each wall and the ceiling. In this case, yes I would probably give
each OBST the true thickness of 11 cm. But what if I don't care about
computing the flow inside that space? It would be far easier in this
case to specify just one OBST whose dimensions are that of the
enclosed space and assign to the sides of the OBST a SURF with the
true thickness of the walls.

[Pai]

I am a structural engineer and I am just learning the basic aspects
and concepts of CFD. If i have a case where an entire wall burns
(timber) due to a very large fire, in that case specifying the exact
thickness of the obstruction is important (thats my guess). Now when
we specify the thickness of the surface (say 100mm) of that
obstruction should this value be equal to that of the obstruction? I
am aware that in terms of smoke movement, this may not matter but when
it comes to burning and heat released due to the burning, should the
obst thickness not match the surface thickness?

I apologise if I am not understanding some of your explainations.

Many thanks.

[Kevin]

The gas and solid phase models are geometrically uncoupled because it
is not always possible to resolve the solid obstruction with the given
gas phase grid. This is a fundamental assumption of FDS and most CFD
models.

[drjfloyd]

There are two length scales at work here. With LES, for typical fires
and compartments, the grid size needed to resolve the flow field is
Order(cm). If you have a material with a low thermal conductivity,
the grid size for the heat transfer calculation is Order(mm). To do
this combined calc you have two approaches:

1) define everything using the finest mesh required (i.e. mm
everywhere), this is not feasible.
2) define those regions that are gas phase with a gas phase mesh and
those regions that are solid phase with a solid phase mesh. Now you
have a number of grid cells that is feasible.

The OBST inputs define how surfaces appear to the gas phase.
The SURF inputs define how surfaces appear to the solid phase (i.e.
heat transfer)

What this means is that with OBST you define your surface to give the
impact to the flow field that is needed. You apply a SURF to OBST so
that the surface removes heat according to its actual thickness. Yes,
the surface area given by the OBST may not be the true surface area
and this will have some impact on the solution; however, keep in mind
that no computer model is reality. They are all reality abstracted in
some manner to allow for a model to be done. In the case of OBST and
SURF, it may be that for a given gas phase mesh size that you may or
may not get as good a representation of the surface area available for
heat transfer as you would like. At that point as a user you need to
decide whether or not you can live with that error or if you need to
expend more resources to refine the gas phase solution further.

On Jul 28, 12:12 pm, Pai <arunpai...@gmail.com> wrote:

[Pai]

Thank you drjfloyd and Kevin.

[Pai]

Can I have the source code of FDS or may be a link that takes me to it?

[Gregor]

On 29 Jul., 11:00, Pai <arunpai...@gmail.com> wrote:
> Can I have the source code of FDS or may be a link that takes me to it?

see: http://code.google.com/p/fds-smv/

[Franck Didieux @ LNE]

I use to consider OBST as obstructions to the gas flow dynamic. Their
thickness (in cells) has only a meaning in terms of where the gas flow
is influenced by the obstruction in the computational domain.
These obstructions have SURF applied to, just as they were wearing a
piece of clothing, that give them properties.
A SURF should be considered as "what the environement see of the item
seen from that side", on a properties point of view (including of
course thermal properties).
Consequently, I agree with drjfloyd when he writes that the SURFs on
opposed faces of an OBST have to be defined with materials listed in
reversed order. Each layer defined in the SURF block should have its
real thickness coded here: this leads to a correct calculation on a
thermal point of view. These "thermal" thicknesses (unit 'mm') of the
SURF have no link with the computational thickness (unit 'cell') of
the OBST the SURF are applied to.
Be careful to also use the correct BACKING option : in case of a wall
exposed from both sides, you should use an obstruction one cell thick,
and apply the SURFs with BACKING = 'EXPOSED'.
One last thing : even if on opposed sides and using the exposed
backing option, SURF are independant.