I'm trying to model an experiment where I had a cylinder filled with sand and some headspace on the top. The sand had some solvent in it. The whole cylinder was initially at about 100 degC but then let cool off to room temperature. from the headspace frequent gas sample were taken for analysis during the cooling period. A very long and narrow diffusion tube was also connected to the cylinder to ensure the system is always close to atmospheric pressure (but not completely open and not losing all the solvent vapor accumulating in the headspace). Anyway, In TOUGH TMVOC, a cylindrical grid was generated and initialized with the material. Then, to represent the diffusion tube, I added a grid block using the "extra cell" option in petrasim and attached it to one of the blocks in the headspace section, with very large volume and atmospheric pressure. Also, for the lab temperature, I defined a time-dependent Neuman thermal boundary condition and attached it to the outer blocks using again "extra cell".
Just to test, I ran the model without the diffusion tube and the outside temperature imposed on the outer blocks and it goes all the way to the end, 7 hrs, smoothly (fully closed boundary condition). However, as soon as I attached the two boundary conditions, the simulation crashes. I tried different things, without any success yet. the model is attached herewith. Any advice would be greatly appreciated.