I wrote a Fortran programme use WUFI equation to solve an 1D question (with initial condition are temperature 10C and rh 0.6 and constant indoor and outdoor temperature 10 C and relative humidity 0.6, 0.8 ). I use different liquid transfer coefficient(suction & redistribution) to do calculation and compare those two results with WUFI results, it showed that the WUFI software results match better with the “suction” results. But there is no rain was set in the simulation. As I know, “suction” happened only when there is driving rain. I think it is not reasonale.
I was wondering can I decide which coefficient is used in WUFI simulation? I cannot unlock material data, is it possible in WUFI light or only can happened in full version WUFI Pro?
Here is the compare result of my programme and WUFI:
thank you so much for your kind help.
Dear Lu,
WUFI uses two liquid transfer coefficients (suction & redistribution) and as you wrote it is switched using the rain in the boundary condition. To use a different one you would have to change the coefficients in the material data.
In WUFI light it is not possible to edit material data or enter new ones. Therefore you would need the full version.
Christian
moisture transport in WUFI always is the sum of liquid transport and vapor diffusion. For a reference solution in which only liquid transport occurs, you must suppress vapor transport in the material by setting the mu value of the material to a very high value (9e9, say). Also set the sd-values of both surfaces to very high values (9e9, say) to avoid vapor exchange between the component and the surroundings.
Christian Bludau wrote: ↑Thu Mar 14, 2019 8:21 pm -1100
Dear Lu,
WUFI uses two liquid transfer coefficients (suction & redistribution) and as you wrote it is switched using the rain in the boundary condition. To use a different one you would have to change the coefficients in the material data.
In WUFI light it is not possible to edit material data or enter new ones. Therefore you would need the full version.
Christian
Thank you so much for your help. I have another question.
If I use user-defined constant climate condition(only temperature and relative humidity were input) instead of use any climate data, which mode (redistribution or suction) was used? As the result the moisture transport so fast. I think it might be a suction mode but no rain was set in the climate page. I was very confused. Did I use it properly?
Thank you!
Dear Lu,
with constant boundary conditions only the redistribution is used for calculation.
Speed will depend on the transfer coefficients of the boundary, diffusion/transport coefficient of you material, and also on the water content.
Without more information I can not tell, if you did it properly.
What initial water content did you use?
Christian
Christian Bludau wrote: ↑Tue Mar 26, 2019 9:51 pm -1100
Dear Lu,
with constant boundary conditions only the redistribution is used for calculation.
Speed will depend on the transfer coefficients of the boundary, diffusion/transport coefficient of you material, and also on the water content.
Without more information I can not tell, if you did it properly.
What initial water content did you use?
Christian
Thank you so much for your kind help!
The initial condition: relative humidity=60% temperature=10C
climate is constant: ex-rh=60%; in-rh=80%; ex-temperature=10C; in-temperature=10C
I think it is sorption(hygroscopic) region case.
In my understanding: in sorption region, Dϕ is calculated by wet cup-test(Mr. Kunzel's thesis). In capillay region, Dϕ equal to Dww or Dws. I think Dϕ values showed in WUFI material Data is hygroscopic region it is very small. But the calculation results suit better with it is in the capillary region(Dww), it turns out moisture transport very fast.
Did I use the software properly? What is Dϕ should equal to in hygroscopic region? Did surface diffusion means the liquid flow(=Dϕ·▽ϕ) in hygroscopic region?
Attachments
equation.jpg (21.96 KiB) Viewed 14000 times
material data.jpg (62.75 KiB) Viewed 14000 times
initial condition.jpg (20.25 KiB) Viewed 14000 times
the difference between Dϕ and Dw is not that the one applies to the hygroscopic region and the other one to the capillary moisture region, they don't.
The difference is: They are used for different formulations of the liquid transport equation.
If you write the liquid transport equation using the gradient of the water content dw/dx as the driving force, then the transport coefficient to be used in this equation is Dw. The liquid flux density g is
g = -Dw dw/dx
If you write the liquid transport equation using the gradient of the relative humidity dϕ/dx as the driving force, then the transport coefficient to be used in this equation is Dϕ. With this choice, the liquid flux density g is
g = -Dϕ dϕ/dx
Both formulations are possible and which one you use depends on your preference or the data you have. But Dw and Dϕ are not identical. If you describe the same flux density g by both formulations, you have
Dw dw/dx = Dϕ dϕ/dx
from which follows
Dϕ = Dw dw/dϕ
(Künzel's equation 19). So you can convert Dϕ into Dw, or Dw into Dϕ, but to do that you have to know the derivative dw/dϕ of the moisture storage function w(ϕ). WUFI expects Dw as input, so if you have Dϕ you have to convert it into Dw for use with WUFI.
Dr. Künzel's equation (23) is derived in this way: If you have a porous capillary-active material exposed to a gradient of relative humidity, you will have both a vapor flux density gv and a liquid flux density gw in the material.
The vapor flux density gv is driven by the gradient of the water vapor partial pressure. Assuming no temperature gradient, we have
where δ is the diffusion coefficient of air, μ is the diffusion resistance factor of the material, δ/μ is the diffusion coefficient of the material, and psat is the saturation vapor pressure at the prevailing temperature.
The liquid flux density gw is driven by the gradient of the relative humidity:
gw = -Dϕ dϕ/dx
We choose this formulation rather than -Dw dw/dx, because the relative humidities are given as boundary conditions.
If the μ-value of the material is moisture-dependent (because a noticeable liquid transport occurs simultaneously with the vapor transport), then the moist material will have a higher moisture flux gv* which corresponds to the μ-value μ*. In reality, however, gv* is not a pure vapor flux density, it is the original (dry) vapor flux gv with some liquid flux gw added:
gv* = gv + gw,
so that we can find gw:
gw = gv* - gv
Inserting the expressions from above, we have
Dϕ dϕ/dx = δ/μ* psat dϕ/dx - δ/μ psat dϕ/dx
which gives
Dϕ = δ psat (1/μ* - 1/μ)
However, WUFI expects the liquid transport coefficients to be given as Dw, so you have to convert the Dϕ into Dw, using the moisture storage function of the material.
Also note that during the measurement there is a continuum of humidities in the material specimen, ranging from ϕ on the left surface to ϕ on the right surface. If the intensity of the liquid transport is moisture-dependent (as is to be expected), you therefore have a continuum of different liquid transport coefficients Dϕ in different parts of the specimen, and the result of your measurement is some kind of average Dϕ over the humidity range present in the specimen under the chosen measurement conditions. It will be difficult to determine for which water content the found average should be tabulated in WUFI.
Also, there may be interactions between vapor and liquid transport in the material which are not taken into account by the above simplified analysis (That is, suppose liquid transport is stronger than vapor transport. Then some water may be transported deep into the specimen by liquid transport, evaporate there and travel the shorter remaining distance as vapor. Liquid transport would then act as a partial 'short-circuit' for the vapor transport).
We do not actually use equation (23) for determining liquid transport coefficients. I think Dr. Künzel only included this equation because it allows to show what the order of magnitude of the liquid transport in the hygroscopic region is in the different materials. If we measure the liquid transport coefficients, we determine moisture profiles during a suction experiment. This method only has low accuracy for moisture levels in the hygroscopic region, but the effect of liquid transport at the higher moisture levels usually dominates the simulation results anyway.
