How to define heterogeneity in TOUGH2-ECO2N

Hi

I am a new TOUGH user. I am interested in learning about defining fluid injection models with TOUGH2. I would like to learn about building heterogeneous models - which Kx and Ky differ from one cell to another. What algorithm is used in these models?

 

Thank you

Xu

4replies Oldest first
  • Oldest first
  • Newest first
  • Active threads
  • Popular
  • Hi Xu,

    You can specify permeability modifiers for each element (see variable PMX in block ELEME.1). You can also generate random fields internally (see Section 7.5 of the manual). These fields internally generated fields do not have any spatial correlation. In case you have iTOUGH2, it has GSLIB integrated, so you could specify spatially correlated random property fields.

    Note that the anisotropy ratios are maintained as given by the ratios of PER(1), PER(2), and PER(3) are maintained, i.e., the assumption is made that the randomness in Kx, Ky, and Kz is 100% correlated. In my opinion, this is a defensible assumption, as it is applied on an individual gridblock. Larger-scale anisotropy would have to be implemented by specifying spatially correlated random fields with anisotropy.

    Stefan

    Reply Like
      • Ramin
      • Ramin.1
      • 4 mths ago
      • Reported - view

      Stefan Finsterle 

      Stefan,

       

      Regarding GSLIB, is there any typical value of ^number simulated nodes to use^ ? I am a bit confused with this! 

       

      Thanks,

      Ramin. 

      Reply Like
  • Ramin,

    It's the number of simulated (rather than measured) nodes used to evaluate the semivariogram when generating the value for next point. I don't know what a typical value is (a few tens?). As always, I suggest you simply change it and see what happens!

    Stefan

    Reply Like
      • Ramin
      • Ramin.1
      • 4 mths ago
      • Reported - view

      Stefan Finsterle 

      Stefan,

      Thanks for your reply. It seems it does not have a great effect in my case, as I change it.

       

      Ramin. 

      Reply Like
Like1 Follow
  • 1 Likes
  • 4 mths agoLast active
  • 4Replies
  • 784Views
  • 5 Following