Dws and Dww

manexi · Post by **manexi** » Tue Sep 06, 2011 11:02 pm -1100

Hello,

I'm sinthetizing my work about WUFI and I have some doubts about Dww and Dws.

According to the help, liquid transfer depends on the RH by:

g,liq=-Dphi*grad(phi)

And Dphi = Dw * dw/dphi

Then, if I got it right, Dw can either be Dws or Dww depending on the boundary conditions of the element.

So if Dws is used when there is liquid water to absorb (e.g. rain), and Dww is used to distribute the water once there is no more rain (which should be most of the time for most of the grid), how does WUFI make the choice between Dws and Dww?

I mean is Dws only used in the external surface grid or does the choice depend on the water content of the adjacent element? (i.e. what happens when we put sources?) and if it does, on which basis does it choose?

Thank you =).
Manexi

Christian Bludau · Tue Sep 06, 2011 11:47 pm -1100

Hello manexi,
In general you are right, the DWS should only be used for the boundary elements of a construction during the rain time. But according to measurements we did and compared them to the calculation, for highly porous materials the water front is not sucked deep enough into the material. So this assumption is not working. For this reason we assume, that the whole construction is sucking during rain events (DWS is used for every material/ grid element) and in the rest of the time the whole construction is redistributing, which is as you mentioned the case during most of the time (DWW is used).
Compared to measurements, this works quite well.
The DW do not depend on the sources. So the source is distributed by the actual state. If there is rain, DWS is used, else DWW.
Hope that helps,
Christian

manexi · Post by **manexi** » Wed Sep 07, 2011 2:21 am -1100

This helps a lot! I was thinking about local Dws/Dww depending on each element of the grid, but now it's easier to explain.

Thank you!
Manexi.

manexi · Post by **manexi** » Wed Sep 07, 2011 3:45 am -1100

Hello Christian.

Concerning you answer I have another question:

regarding the free saturation of the materials, shouldn't Dw be cancelled above free saturation? (since no more capillary action is possible)

Dw(w>wf) = 0

Ok, if Dw is cancelled then water might stay in the structure, but then shouldn't the direction be taken into account? Like Dw = 0 for incoming water, but Dw>0 for outgoing water.

for example: If gl is the liquid flux (-Dw*dw/dphi*grad(phi) ), dSext the surface vector outwards the element then,

Dw (w>wf) = 0 if gl.dSext < 0, Dw if not.
-----------------------------------------------------

I used a similar method ( Dws(w>wf)=0 for Dws only, whatever the direction) for some models where I was having trouble with the amount of water absorbed by rain.

Best regards,
Manexi

Christian Bludau · Wed Sep 07, 2011 8:19 pm -1100

Hello manexi,

if you cancel the Dw at the free saturation there will be no further transport. The Dw only gives the speed of the transport and uptake of water. If the material is saturated respectively supersaturated, it is sucking/redistributing quite fast.
The amount of rain is only indirect influenced by the Dw. It is calculated by the pore volume and the existing water in the pore volume. So if the pores are filled, Wufi still is sucking/ redistributing, but no water can be added by rain.

Further information you can find in the PhD thesis of H.M.Künzel and M.Krus:
http://www.ibp.fraunhofer.de/Images/mk_ ... -30730.pdf
http://www.ibp.fraunhofer.de/Images/hk_ ... -30731.pdf

Have fun,
Christian

manexi · Post by **manexi** » Wed Sep 07, 2011 10:22 pm -1100

Hello Christian,

if you cancel the Dw at the free saturation there will be no further transport.

Totally right, I just didn't know how to stabilize the element and it seemed to work since it gives time to the unsaturated elements to redistribute water, thus increasing the humidity gradient between saturated elements and unsaturated elements (accelerating the transfer afterwards). I have to admit it only "worked" because only Dws was cancelled thus allowing the transfer in "sunny" days, and limiting it when it rains.

If the material is saturated respectively supersaturated, it is sucking/redistributing quite fast

Ok, I get your point: in super saturation, the material might give as much as it takes... but I know by experience that the simulation might crash under these circumstances.

The amount of rain is only indirect influenced by the Dw. It is calculated by the pore volume and the existing water in the pore volume. So if the pores are filled, Wufi still is sucking/ redistributing, but no water can be added by rain.

That's where it becomes tricky for me. Shouln't this condition be applied to free saturation instead of maximum saturation? I mean if no more water can be added by capillary action after free saturation, that means that the material cannot stock up any more extra water by capillary transfer (under normal circumstances), so it can suck up as much as it gives away, but shouldn't exceed wf. Am I right?

Thanks
Manexi

P.S. Thanks for the dissertations, I'm going through them.

Christian Bludau · Wed Sep 07, 2011 10:52 pm -1100

Hello Manexi

Ok, I get your point: in super saturation, the material might give as much as it takes... but I know by experience that the simulation might crash under these circumstances.

Yes, we had the same problem, but as I´m mainly a user , I have no idea how this was solved in Wufi. Maybe Manfred can tell you this.

Shouldn't this condition be applied to free saturation instead of maximum saturation? I mean if no more water can be added by capillary action after free saturation, that means that the material cannot stock up any more extra water by capillary transfer (under normal circumstances), so it can suck up as much as it gives away, but shouldn't exceed wf. Am I right?

Yes, you are right, for the amount of water uptake it has to be the free saturation, not maximum saturation. But even so you can not stop the transport at values higher than the free saturation...

Regards,
Christian