Effect of the diffusion coefficient(D0) on the TOUGHREACT
Dear Sir or Madam,
Sorry to disturb you. I found a problem when I run the example model in the 8.1 Problem 1 – Aqueous Transport with Adsorption (Linear Kd) and Decay (EOS9) in TOUGHREACT User Guide Version 2.0.
Firstly, I got the perfect results when I followed the model instructions given in that User’s Guide. However, I faced a problem when I changed the diffusion coefficients. In the example model, the numerical dispersion(Dnum) is 0.01 m2/day and diffusion coefficient is 0, and I built 4 models and just varied the tortuosity(τ) and diffusion coefficient(D0) so that the τ*D0 = 0 (model 1); τ*D0 = 0.001m2/day = 0.1*Dnum (model 2); τ*D0 = 0.01m2/day = 1*Dnum (model 3); τ*D0 = 0.1m2/day = 10*Dnum (model 4). In the first three models, the numerical solution and the analytical solution matches well, but in the last model (the diffusion coefficient is large), the numerical solution deviates from the analytical solution.
Is there anybody can explain to me? Thank you!
Best regards,
Lang
2 replies
-
Lang,
This is most likely a time stepping issue, since the diffusion coefficient is much larger when the results don't agree. If the diffusion coefficient is increased by a factor of 100, then decrease the maximum time step by a factor of 100, and check the results.
regards,
Eric
-
Eric Sonnenthal
Dear Dr. Sonnenthal,
Thank you for your prompt reply.
I decreased the maximum time step by a factor of 100 in two approaches: 1) reduced the DELTMX (upper limit for time step size) in the flow.inp from 8.64E+03 to 8.64E+01. I found the approach could not converge, and could not run to the stop time. 2) set up the REDLT =100(factor by which time step is reduced in case of convergence failure). I found the second approach could not make any change compared with my previous model. So could you help to see my input files? Thank you!