EOS9-infiltrtion simulation with defining capillary pressure as a primary variable

[mitra]

I have a problem for simulating an infiltration test with EOS9. I used van Genuchten model with the Sls=.68. I tried to use 3 consecutive simulations (I used the output file from one of them as an initial condition for the second one) with different capillary pressure at the soil surface. In my case, I used Pcap=-50, -20, -10 cm of water. For the first two simulation (-50,-20) I didn’t have any problem and I received the results at the time I defined, but for the last one (-10) I had some problem with the code convergence. Although I defined bigger time steps, the time step decreased to the value of less than a second (10-4) and after two time steps for which convergence on itr=1 happened, the simulation stopped. The soil parameter that I used and the case I am trying to simulate are real, so I don’t know what is the problem with the last simulation?

Thanks

Maryam

[Charlie Thornton]

Sometimes it can be useful to restrict TOUGH to a smaller time step. I know if you try to run one of the TOUGHREACT example problems (CO2 sequestration) without limiting the max time step, the time step will eventually get too big and the simulation will fail. I think the reasoning is that smaller steps make it easier for the simulator to steer around numerical problems.

However, when the simulation crashes immediately like that its usually because you've gotten into an invalid physical state in one or more of the cells. What does the bottom of your OUT file say after the simulation stops? Maybe there are some error lines that can give us some hints.

- Charlie

[mitra]

Charlie, thank you very much for your answer. In my case, I received this message :

+++++++++ CONVERGENCE FAILURE ON TIME STEP # 5 WITH DT = 0.298023E-05 SECONDS, FOLLOWING TWO STEPS THAT CONVERGED ON ITER = 1

STOP EXECUTION AFTER NEXT TIME STEP

++++++++ REDUCE TIME STEP AT ( 5, 9) ++++++++++++++++ NEW DELT = 0.745058E-06

As far as I know, with decreasing the capillary pressure the flow rate should increase, and I think you are right I may have gotten into an invalid physical state in one or more cells, but the case I defined is physically meaningful, so...?

Thanks

[swenson]

Mitra,

Without more details, it is difficult to answer your question. Would you be willing to send your model and instructions on how to duplicate your problem to support@thunderheadeng.com? We will look at your model and try to identify the problem.

Thanks,

Daniel Swenson

Thunderhead Engineering

[Johny5]

That's pretty good!

simulation

credit

[Johny5]

Thank you for the information you provided. It helped alot.

simulation

credit

[meow44]

thanks for the gelp

simulation assurance vie

[vanesssa99]

Thanks for your idea

simulationcredit

[Guest lindapret88]

This is an interesting post.. thank you for sharing

taux pret auto