Jump to content
GMS, SMS, and WMS User Forum
Sean Czarniecki

Solids -> MODFLOW; Grid Overlay with Keq

Recommended Posts

So I'm using the Grid Overlay with Keq method and have 4 materials in my solid. Each one has a different Kh and a Kv set at 1/10 of the Kh. Based on my understanding of the method and the help description ("Rather than simply assigning materials based on which solid encompasses the cell centers, the Keq method attempts to compute a custom Kh and Kv value for each cell. When assigning the material properties to a cell, GMS computes the volume of each solid in the cell and computes an equivalent Kh, Kv, and storage coefficient for the cell that takes each of the solids in the cell into account. Thus, the effect of a thin seam in a cell would be included in the Kh and Kv values for the cell."), I would expect to see my resulting Kv for each cell be 1/10 of the Kh computed. I am finding that the calculated value is often much lower than the expected value.

Does anyone have insight on this matter?

  • Like 1

Share this post


Link to post
Share on other sites

So I'm using the Grid Overlay with Keq method and have 4 materials in my solid. Each one has a different Kh and a Kv set at 1/10 of the Kh. Based on my understanding of the method and the help description ("Rather than simply assigning materials based on which solid encompasses the cell centers, the Keq method attempts to compute a custom Kh and Kv value for each cell. When assigning the material properties to a cell, GMS computes the volume of each solid in the cell and computes an equivalent Kh, Kv, and storage coefficient for the cell that takes each of the solids in the cell into account. Thus, the effect of a thin seam in a cell would be included in the Kh and Kv values for the cell."), I would expect to see my resulting Kv for each cell be 1/10 of the Kh computed. I am finding that the calculated value is often much lower than the expected value.

Does anyone have insight on this matter?

The combination of multiple materials in a newly created cell will result in a representative conductivity value. That is, I think you're seeing a combination of materials instead of a cell containing only a single material.

  • Like 1

Share this post


Link to post
Share on other sites

Right, but it is the combination that I think is incorrect for Kv. Let's say I have 2 materials being combined in a cell. Material 1 has a Kh of 100 ft/day and a Kv of 10 ft/day. Material 2 has a Kh of 10 ft/day and a Kv of 1 ft/day. If the volume of each material in the cell is 50% of the cell, I would expect the resulting Kh to be 55 ft/day and the Kv to be 5.5 ft/day. Kv is 1/10 of Kh. However, in reality, I'm getting results for Kv which are much smaller than 1/10. This is why I'm questioning it.

  • Like 1

Share this post


Link to post
Share on other sites

Okay - I just loaded up GMS7.0 to see if this "issue" I'm having in v6.5 is the same in v7.0.....and the Grid Overlay with Keq isn't even an option any more! Can one of the developers chime in and discuss this at all? This is a great option which should definitely not be removed. I would never have suggested going forward using GMS on my current project if it didn't have this option.

Share this post


Link to post
Share on other sites

Looks like an oversight on our part. I will look into it.

Out of curiosity, why do you use this option instead of using the HUF package? What do you do when you need to calibrate the model as this approach is not very helpful in that regard?

Share this post


Link to post
Share on other sites

Thanks for the response.... An excellent question (regarding the HUF package) - I have never used the HUF package up to this point, but it appears from a quick glance at this that I should to see if it will fit my needs better. The Keq method was used as a good starting point to take around 40 different solid "layers" and boil them down to a 4 layer model. After that, I could do rough calibration by adjusting areas of the layers (the solids had enough flexibility/variability that I could do this without much concern). I agree that the method is not great, but it was fine for my needs. I will look further into the HUF package. Thanks again!

Share this post


Link to post
Share on other sites

Alan,

Do you have any thoughts on my first post in this thread? If I use materials where the Kv is 1/10 of each Kh, I would expect the Keq to come out with Kv being 1/10 of Kh in each cell. It does not do that consistently right now in my complex stratigraphy (I think it only does it when there is one material in the cell). I have not tested it on a simpler stratigraphy.

Regarding the HUF package, I looked into it and feel that, for my current situation, it would actually be harder to calibrate than what I'm doing. It is definitely a good thing to keep in mind as an option for future models though!

Sean

Share this post


Link to post
Share on other sites

I will need to look at the code to see what we are doing to calculate the equivalent K. Can you send us your project and tell me a particular cell in the grid that you have been looking at. That will make it much faster for me to resolve this. Thanks in advance.

Share this post


Link to post
Share on other sites

Here is the equation for the equivalent Kh:

Kh = SUM(Khi*Mi) / SUM(Mi) where Khi is the Kh of the solid and Mi is the length of the solid in the cell.

Equivalent Kv:

Kv = SUM(Mi) / SUM(Mi/Kvi) where Kvi is the Kv of the solid and Mi is the length of the solid in the cell.

You can see how the Kv can be dominated by a small value of Kv. I hope this helps.

I looked at your files and from what I can tell GMS is correctly calculating the Kh and Kv. I did find a different bug unrelated to the calculation of Kh and Kv that should be in the next update of GMS.

Share this post


Link to post
Share on other sites

Thanks for the response - I'm glad to hear that the calculation is doing what it is supposed to do. I guess I disagree with the calculation.

1) I'm assuming that Mi is actually volume rather than length of solid in the cell, right?

2) Why wouldn't the vertical K be calculated the same as the horizontal K? Using my example of 2 materials, if each one has a Kvi which is 1/10 of the Khi, the resulting Kv for the cell should end up being 1/10 of the Kh. Just my opinion. Luckily, for this model, I can go in after the calculation and just use the vertical anisotropy.

Can you tell me if I should be concerned about the bug you found?

Thanks again!

Share this post


Link to post
Share on other sites

Thanks for the response - I'm glad to hear that the calculation is doing what it is supposed to do. I guess I disagree with the calculation.

1) I'm assuming that Mi is actually volume rather than length of solid in the cell, right?

Actually it is just length. The solids are intersected with a vertical line at the cell center to determined the lengths

2) Why wouldn't the vertical K be calculated the same as the horizontal K? Using my example of 2 materials, if each one has a Kvi which is 1/10 of the Khi, the resulting Kv for the cell should end up being 1/10 of the Kh. Just my opinion. Luckily, for this model, I can go in after the calculation and just use the vertical anisotropy.

The calculation is fundamentally different, here is the basic idea. When you calculate vertical K the water must pass through each unit (this is like resistors in series in an electrical engineering problem). In contrast, with horizontal K a layer with a high value of Hk can carry most of the water. If I find a good reference for this I will add it here.

Can you tell me if I should be concerned about the bug you found?

When I was testing GMS would crash.

Thanks again!

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

×
×
  • Create New...