Simulation error

[SebastianM]

Hello,

I´ve encountered a problem with a simulation. I´ve simulated a concrete block, which has 3 adiabatic borders and on the remaining surface, a measured climate is attached. The simulation shows, that the concrete block loses water (in the first few years). The more water is lost, the more clearly can yearly moisture cycles be seen. On one point, a turning starts to occur. The concrete block starts adding up water again while the amplitudes of the cycles are increasing. Under Control Parameters I´ve increased the Maxit [to 1.5e6] and Convergence Crit. [to 5e-8] by about the power of 3. This seems the delay the turning point when the concrete is starting to add water. Physically once the water content has dropped to a certain limit, the yearly moisture cycles should fluctuate around it in more or less a steady state behaviour.

Is there any solution for this problem? Thank you for your tips.
I´ve attached a picture that describes the problem better.

Best regards,

Sebastian

[Christian Bludau]

Dear Sebastian,
can you post a picture of your construction? Is there air included?
Christian

[SebastianM]

Hello Christian,

thank you for your response.

The construction is a layered concrete wall with thin "monitoring" layers in between. The material characteristics are the same, except the moisture storage function. There are no air layers in the model.

Sebastian

[Christian Bludau]

What grid did you use? And what boundary conditions?
Is the picture of the water content posted the total water content or did you evaluate a special layer?

Christian

[SebastianM]

I´ve used a measured climate which is very similar to alpine region weather data from the TRY (2011) data. The water content in the picture is the water content of the whole construction. The grid size I´ve used = min = max = (x=75,y=75). Could using a too small grid size such erros?

Sebastian

[Christian Bludau]

Hello Sebastian,

I´m not really sure if this is an error. It may be, but I have seen similar results due to effects of coupled redistribution and drying.
If it si an error, it can be caused by to less grid elements or, what is more possible an unfavorable course of the moisture storage function. Maybe try to approximate them based on the measured values.

Christian

[Manfred Kehrer]

I agree with you that this should not look like that. I guess you have areas in your model having max. saturated elements. We know there is an unsolved numerical problem for this case, Fraunhofer IBP specialists may help you better to solve this.

[SebastianM]

Thank you for your response and help!