How can I simulate the thermal diffusion (Soret effect) ?

Dear Zaki,

TOUGH is meant for primarily for water with dissolved species and multiphase flow, but maybe T2VOC can handle this mixture. I don't know of any versions of TOUGH in which the Soret effect has been implemented. To do it correctly would require a significant amount of code modification. It may be possible to "trick" the code into crudely simulating diffusion by both a temperature gradient and by a concentration gradient, by using temperature ( x concentration?) as another component with a negative diffusion coefficient. However, I tried to come up with something over a few minutes of rewriting equations with no luck. If you know the Soret equations, with some work possibly you can come up with something. Let us know if you are successful. It would make a very nice TOUGH symposium presentation!

good luck,

Eric

Dear Zaki,

I checked the documentation for TMVOC (the multi-component version of T2VOC), which can handle methane and n-butane.  The Soret effect (mass flow caused by a temperature gradient, when multiple mass components are present) is not implemented.  I agree with Eric that adding it would be a significant effort.  As a new TOUGH user, I recommend that you formulate your problem without the Soret effect first and carefully study your simulation results.  Consider the temperature and concentration gradients and heat and mass fluxes that are present and use the formulation you think is relevant for the Soret equation (I understand that this is an area of current research, and there may be multiple approaches) to estimate how important the Soret effect would be.  If it is significant, then you could consider modifying the code to implement it.  But I recommend that you gain a lot of experience with TMVOC before attempting modifications.

Christine