FREEZING MODEL

[mrcddnn]

Dear CODE-BRIGHT users, CODE-BRIGHT team,
I am using CODE_BRIGHT to model THM triaxial test performed on frozen soil for my Master thesis.
Actually, I am focusing only on the thermo-hydraulic problem.
------------------------------------------------------------------------------------------------
The main phase of the model are:
1. Initial step: Surface temperature T=5°C. Initial pore water pressure = 0 and all boundaries are undrained.
2. A ramp of decreasing temperature from T=5°C to T=-5°C is applied at the top boundary of the sample.
3. A constant temperature T=-5°C is maintained on the top boundary until stationary condition is achieved.

For simplicity I am considering a sample constituted only by water, so I am using a porosity value of 0.998.
------------------------------------------------------------------------------------------------
For the freezing model I am adopting:
Retention Curve: ITYCL=18 (Van Genuchten)
Liquid Phase Relative Permeability: ITYCL=12
Conductive Flux of Heat (2): ITYCL=5
------------------------------------------------------------------------------------------------
As output I have observed the following:
- when Temperature is positive -> LiqSatDegree = 1, and Ice Pressure is negative.
(1) Shouldn't ice pressure be equal to 0 since for T>0 water is present only as liquid phase?

- when T approaches to 0 (i.e. T=0.5°C), LiqSatDegree value is already lower than 1 (i.e. Sl=0.8) even if Suction is negative. I performed a parametric analyses changing retention curve parameters but I still observe this behavior.
(2) How is it possible?
(3) Van Genutchen model should provide liquid saturation degree only for positive value of suction? Should it be a bug of the code?


Thanks

Marco