Clarissa Posted June 17, 2019 Report Share Posted June 17, 2019 We are about to start a 3D dam seepage model. We have done several similar models over the last few years in a finite element, unstructured mesh model code similar to FEMWATER. This time, we've decided to start with a MODFLOW model due to funding and scheduling issues. I've been wondering if a MODFLOW-USG model will meet our needs, and eliminate the need to switch to an FEM model sometime in the future. I took a MODFLOW-USG class several years ago, and have been looking for an opportunity to use it, but have not stayed up to date with the current condition of the interface in GMS. I've done a little testing over the last few days, and I see two possible hindrances. I think I've got workarounds for both, but wanted to see if anyone else had additional ideas. 1. We definitely need to be able to look at a cross-section of the dam and show head contours on the cross-section. With a regular grid, this is straight forward as long as your cross-section is along a row or column. I don't see any cross-section cutter in the USG model. With an unstructured mesh, we are used to using the control key to select a line of elements to isolate. I do not see that that hotkey is available with an unstructured grid, but it looks like we can select and isolate a group of cells and get a similar effect - either draw a box or select with a polygon. Is that our best bet? 2. Once we have the model solution, we compute a factor of safety against uplift at all the surface nodes in the FEM model - basically we need the gradient for upward flow. We have an external code that will automatically do this for an unstructured mesh by comparing heads and elevations at nodes located at the same x,y location. If we use a regular structured gridded MODFLOW, the computation is pretty trivial using the MODFLOW arrays for the centroid elevations and computed heads and the data calculator. I am not finding the same options in the unstructured grid. I realize that the grid is not required to have layers and that the computational points are not necessarily at exactly the same x,y locations, but it seems that the only way to build the unstructured grid in GMS is with layers, so if you're using the GMS interface, it's likely a layered USG. At this point, I'm thinking writing an external code or maybe an excel macro may be my only option here and I can probably steal code from the routine we use on unstructured meshes. Is there anything I'm missing in the GMS interface that could help me here? Thanks for your advice. Quote Link to comment Share on other sites More sharing options...

## Recommended Posts

## Join the conversation

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