PhreeqcUsers Discussion Forum

Reactive Transport => Reactive Transport Modelling => Topic started by: bessie000 on March 10, 2015, 09:45:36 PM

Title: How to model explicit time and distance
Post by: bessie000 on March 10, 2015, 09:45:36 PM
Dear all,

I am a beginner of Phreeqc, and I am trying to model a 100 m field (composed of lots of minerals) reacting with acid mine drainage. I would like to see the changes across the whole 100-m distance after different times, say 1 yr, 10 yr and 100 yr. But I am so confused about between shifts, time_step, and their relationship to distance. Also, how to calculate flow rate based on these parameters? Is it true that 1 cell needs 1 time_step?
This is what I list as an example:

-cells 100
-length 1.0
-shifts 20000
-flow_direction forward
-time_step 1440
-boundary_conditions flux flux
-punch cells 1-100


Thanks.

Bessie
Title: Re: How to model explicit time and distance
Post by: dlparkhurst on March 11, 2015, 02:16:43 AM
The TRANSPORT data block is described in the manual.

PHREEQC's transport algorithm could be called a method of characteristics because the characteristic velocity is used to simulate advection, and dispersion is simulated by mixing. So yes, in one shift the water in one cell is advected to the next cell. So the velocity is the cell length (1 m) divided by the time step (1440 s).
Title: Re: How to model explicit time and distance
Post by: bessie000 on March 25, 2015, 03:39:44 PM
One point I still don't quite understand is the option Phreeqc has about modeling implicit or explicit time. What are their differences? and they seem to give quite different results even when the rest of the files are the same, especially in terms of running time, it is dramatically slower when I specify a time_step, even though it is a very small one.
Title: Re: How to model explicit time and distance
Post by: dlparkhurst on March 25, 2015, 07:13:03 PM
I don't know what you are referring to. The only explicit and implicit that I can think of is related to -cvode in KINETICS, which uses a different solver. Time steps would be the same with or without specifying cvode.