Wall Assembly with multiple metal layers - Initial RH question

[DevlinNick]

Hi All,

I'm trying to study a wall assembly that includes a metal faced SIP panel but would appreciate some guidance on the initial conditions. The wall is for an offsite fabricated system, constructed in a factory. The wall assembly is (from inside to outside):

15mm Plasterboard
12mm Plywood
VCL (Sd=50)
90mm mineral wool (with metal stud in the actual condition) - (various options breaking this down into 3 layers of 1/88/1mm for example)
1mm steel sheet (inner face of SIP panel - actual thickness is 0.5mm)
78mm PIR @ 0.023 W/m.K (various options include breaking this down into 3 layers of 1/76/1mm for example) [SIP insulation]
1mm steel sheet. [outer face of SIP panel actual thickness is 0.5mm]
Vented rainscreen cavity
Timber / other rainscreen outer layer

With 80% RH Initial conditions
I have tried various configurations of the geometry, including omitting the outer timber and cavity and replacing the outermost steel layer with an sd-value (and turning off the rain). I've also tried the same approach to omit the VCL layer internally. However, For a 5 year simulation, there are always convergence errors and B1 and B2 values that are not sufficiently similar. Eg -0.32 and -0.419 kg/m2 respectively.

With 70% RH initial conditions:
There are zero convergence failures and B1 / B2 values of -0.118 / -0.119 kg/m2 respectively. This would appear to be a satisfactory numerical result.

The question therefore, is how do I evaluate whether it is acceptable to start with the 70% RH initial conditions throughout. In both cases, the total water content in the assembly is lower at the end of the first five years than at the start, suggesting that the initial RH is not too low:

70% RH starts at 1.68 kg/m2, ends at 1.58kg/m2 - reduced by 0.1kg/m2
80% RH starts at 1.99kg/m2, ends at 1.68 kg/m2 - reduced by 0.32kg/m2.

All comments and observations are gratefully received. Thank you.

[Christian Bludau]

Hi Nick,

from my perspective, it would be more appropriate to set up the model in such a way that it can also be evaluated reliably with an initial relative humidity of 80 %, rather than reducing the initial moisture level solely for numerical reasons.
The observed B1/B2 balance discrepancies at 80 % RH are most likely caused by condensation forming within the mineral fibre insulation or other layers with very low moisture storage function. There are two practical approaches that may help to stabilise the simulation:

Test an alternative insulation material
As a first step, it may be helpful to run a comparative simulation using a different insulation material to check whether the numerical instability is specifically related to the hygrothermal properties of the mineral wool.

Splitting edge regions
Similar to the “Split Air Layer” approach discussed in the WUFI forum, the boundary regions of the insulation layer can be separated into thin sub‑layers and modelled individually. These edge layers can then be assigned the material “Mineral Wool (heat. cond. 0.04 W/mK)”.
See the following forum post for details:
viewtopic.php?f=32&t=1777&p=5226#p5226

In addition, the adaptive time step control in WUFI should be enabled. This setting is particularly important when condensation occurs or when strongly non‑linear moisture processes are present, as it significantly improves numerical stability.

Best regards
Christian