Jump to content
GMS, SMS, and WMS User Forum


GMS Experts
  • Posts

  • Joined

  • Last visited

  • Days Won


Everything posted by woodward

  1. Yes the unsaturated curves can cause problems. I also recommend NOT using the spline interpolation, as sometimes this results in a non-monotonic curve. Be careful about the curve spacing and how far away from the water table you specify it (as the last value will be used for all higher tensions as well).
  2. Could be that you have accidentally applied constant head bcs also at the transient boundary? Also with transient simulations, it is often a good idea to start with small time steps, and increase them. You are welcome to send me the project and I can try it out here. Although not sure if my license is up to date.
  3. First thing I would look at is convergence on each time step. Is the model iterating and converging? It sounds like it is not. This could be because your convergence settings are too loose or too tight, or because your discretization is too coarse.
  4. I'm not sure about the effective/total porosity, but the Unsat Curve dialogue in GMS is simply to generate a look up table of values, it is not connected to any of the other model inputs (and the inputs are not stored by GMS). Also there is a bug in the units as discussed here:
  5. Check this old post. Maybe you are right, the problem is the NUMBER of XY time series. It seems GMS doesn't remove old ones so you can exceed the array bound.
  6. By the way, does it make sense to use a 30 day time step when you are simulating movement through the unsaturated zone? Normally you need small grid spacing and time steps for this, although the FEMWATER Langrangian method might help. Check the Peclet and Courant numbers.
  7. Not too many XY series, but maybe some of your XY series are too long. You can edit the unsaturated zone curves in the dialogue box in GMS. There is a wizard for generating them. Or you can generate them outside GMS and paste them in manually i think (that's what i did, I think). It's a while since i used GMS. Or you can edit them directly in the FEMWATER text files produced by GMS. It really important to read the FEMWATER manual and be familiar with how it works, and the text files it reads. There are a few bugs here and there. If you check my older posts here (few years ago) I might have mentioned some of them. By the way DON'T use the spline smoothing option, it can cause errors because it is not guaranteed to be monotonic..
  8. Hi Han It means your model is too big. FEMWATER is unfortunately compiled with fixed array sizes, if your model is too big you will need to recompile FEMWATER with larger arrays or make your model smaller. The source code is available from AQUAVEO. In this case the problem is you have an XY series that is longer than 250 rows. XY series are used for time series and also for unsaturated zone curves. By the way, in my experience FEMWATER has poor mass balance for transport calculations. Simon
  9. Sadly I have not had the opportunity to use FEFLOW, as it is well beyond the budget of my organisation. We tend to stick with the cheaper options. Mostly we use MODFLOW. I'm not sure the best option for fully coupled models. There are other things I have heard of like COMSOL, HydroGeoSphere... PS: FEMWATER can do fully coupled models. There is a limit to the number of nodes and elements, this is hard coded, but if you recompile the exe you can change it. PPS: You could try asking on a general gw modelling forum such as gwmodel, for software suggestions.e.g. https://groups.yahoo.com/neo/groups/gwmodel/conversations/messages https://in.groups.yahoo.com/neo/groups/feflow/conversations/messages
  10. I used to be the FEMWATER support person here, I've done some coding with the FEMWATER source code (developed my own reactive transport version), and I've used it (unsuccessfully) for a small scale problem (I eventually switched to MODFLOW-NWT). My view is that FEMWATER is pretty limited in its usefulness. It does not have many features (e.g. no zone budget) and some of the advertised features do not actually exist (e.g. no river package, some element types don't work). It is an old code that may be useful for some simple problems. The author no longer supports it. Also, I found the transport mass balance to be very poor (this is why I abandoned it). And it is not compatible with modern calibration tools like PEST. Edit: If you decide to go with FEMWATER, I have a bunch of documentation that you might find useful.
  11. I have used PEST with Groundwater Vistas. I assume it's similar in GMS. If you want the realisations to converge, you can change the convergence criterion/increase the number of iterations. The number of iterations is usually set to only 2 for NSMC to make it fast. This might not be enough for your model. Have you checked the model batch file that PEST calls? Can you see where the HK file is created? Maybe the batch file is failing to create the HK file for some reason.
  12. Yeah it doesn't look right. Did you check that the zones are set up correctly?
  13. Hi. There actually is not a River Package in FEMWATER. It is documented and there are code fragments, but it was never completed. For materials properties, the max height about the water table (yes, minimum pressure head, e.g. -40) is used in defining the unsaturated zone curves. The unsaturated curves must be defined between 0 and the maximum height specified (e.g. -40). At distances further above the water table (e.g. -50) the unsaturated zone properties are assumed to be the same as at the maximum height (e.g. -40). So you only need to specify in the unsaturated curves for the region where the properties are changing. By the way, I found that linear interpolation of the unsaturated curves was more stable than the spline interpolation option. Hope this helps. Simon
  14. woodward

    2D mesh number

    Hi Alan. I was checking back through my ancient notes, and the elements issues might be only for transport, it seems the Lagrangian particle tracking approach only works for hexahedral elements. It's a long time ago. I think I might have used octahedral elements at some point for testing, but never the tetrahedral ones.
  15. woodward

    2D mesh number

    The FEMWATER/GMS mesh has some bugs. Not all element types work. You should use hexahedral elements. Also there is an issue with the mesh node numbering. Here are my notes from some time ago; "*.3dm: Geometric data. For regular hexahedron elements, this file can be generated by program “argusmesh1.f” or the Matlab script MakeMesh1.m. Otherwise, it can be generated using GMS. However, there is a bug in FEMWATER with regards the element orientation specified from GMS, and the elements must be rotated 90 degrees. I will write a Matlab script to do this."
  16. So if you only run MODFLOW transient you still get the message? Then it is not caused by PEST. Which version of MODFLOW? Have you got the parameter estimation routine of MODFLOW2000 turned on by mistake?
  17. I haven't used PEST with GMS, only with Groundwater Vistas (GV), but it looks like the problem is that the model is not returning a valid answer. Could be because of (1) invalid calibration data (2) parameter values that result in invalid model output (3) MODFLOW crash. Could also be a problem if a parameter value that is log scaled is 0 or less. In GV there's a switch that will cause MODFLOW always to terminate even if there's an error, so you should set that. Also put bounds on the parameters to make sure you'll always get a valid simulation.
  18. It's a long time since I've done this, but the menu command might only appear when you have the mesh selected.
  19. Hi Alexxander This error is simply because your model is too big for the array bounds that are hard coded into the compiled version of FEMWATER that is supplied with GMS. MAXNPK is the maximum number of mesh nodes. You will need to contact Aquaveo to get a source code for FEMWATER, then recompile with a larger value of MAXNPK (and possibly some other array bounds as they tend to be interrelated). I have the source, and am happy to share with you if Aquaveo agrees. My version has a couple of small bug fixes too. Simon
  20. MODFLOW-CFP is a version of MODFLOW-2005 that includes the CFP package which simulates flow in lateral macropores ("conduits") in one of two ways (1) by explicit specification of pore geometry or (2) by implicit specification of macropore effects via a non-linear flux-head-gradient function. I am doing some transient modelling to simulate 3D flow in a small hillslope that has a large number of water table piezometers. There are 3 material layers (ignimbrite sand) but adjusting the Ksat with PEST does not give a good fit to the water level data. I would like to try simulating CFP (there are certainly macropores at the site) but have not been successful in getting mf2005cfp running. Has anyone had any luck with this? I set up a MF2005 model, which ran correctly, then added the CFP package (the input file is quite simple) and attempted to run with mf2005cfp.exe, but it crashed on the first time step with a convergence failure. Any suggestions? I put conduits in every layer with mean diameter 1.0, and critical Reynold's Numbers 1.0 and 2.0. I used T = 20 degrees. Thanks!
  21. Are you saying your mesh only has a single layer? Then Femwater is not the right model to use. Femwater model unsaturated-saturated flows, and unsaturated flows required relatively fine vertical mesh spacing. It is quite different to MODFLOW. I would say Femwater is not a good model for modelling saturated flows only. I'm not sure why you are having trouble compiling the exe. One thing is that you need to allocate enough memory in the compilation - I had to increase the stack size in the compiler options at one stage to get this to work.
  • Create New...