I have a problem regarding the behavior of a UDS in a transient model.

I am modelling air flowing through a potato pile. The pile is modeled as a porous zone (fluid: air, solid: potato). In order to model the moisture content of the potatoes, I defined a user defined scalar (units: kg water/kg potato). Water in the air is modeled via species transport. I implemented source terms for the solid (potato) part in this way that the uds (potato moisture content) will drop in time due to respiration and evaporation. Water diffusivity in the potato is defined as a UDF and implemented in the material-properties for potato as UDS-diffusivity.

After convergence of the flow field (in a steady state run), I did a transient run for the energy and mass equations with adaptive time steps, after patching the initial UDS-value.

 After only a couple iterations, the UDS-value drops immensely and unrealistically. I checked the values of the sources, but none of them explains the immense drop in the UDS-value. I tried varying the UDS-diffusivity, but this does not help either.

 I would like to ask advice on this problem, since I really don’t know what to try next. Does anyone have an idea of what could cause this drop in UDS-value?

Thanks in advance!
Kind regards,