High Convergence & Balance Error in Wall Assembly with imervious layer

[Nick Devlin]

Hi there,

I have a historic brick cavity wall constructed in the 1950's. The two leaves of brick are separated by a 10-15mm unventilated air cavity. The outer face of the inner brick layer (facing the cavity) was originally painted with a a bitumen paint layer. It is assumed that this is a water proofing layer. I have modelled this bitumen layer as 1mm Bituminous Paper from the database. I am evaluating internal wall insulation options with diathonite and wood fibre insulations with a Proclima VCL etc.

When the calculation is run, there is a huge difference in the Balances (more that 100x difference 9kg vs 1111kg) and c. 2000 convergence errors as the total moisture content increases but does level out over 50 years. There is no significant improvement in the convergence errors or balances when using Adaptive Time Step control. For the north elevation, the calculation time period needs to be 50 years to see a levelling out of the total moisture content in the wall because at 25 years it is still increasing (all other elevations follow a general drying out trend). At the end of the 50 year modelling period, for the north elevation, the RH levels of the woodfibre insulations are shown to have annual averages of c. 85%. See the graph. This is at the interface of the woodfibre and diathonite layer.

By contrast, if the bitumen layer is omitted from the component assembly, the calculation reaches equilibrium in 5-10 years, the balances match almost perfectly and there are zero convergence errors.

I would be interested to receive advice on how best to interpret these results. The model with the bitumen layer is unstable whilst the model without the bitumen is less representative of the actual case under investigation. One option would be to omit the outer brick and cavity layers, and replace the bitumen layer with an Sd-value.

All thoughts are greatly welcome.

Kind regards

Nick