Jump to content
GMS, SMS, and WMS User Forum
Sign in to follow this  
Hydrology

Unstable Model, But Still Converges

Recommended Posts

Can you have an unstable MODFLOW model that still converges every time? I have a steady state model that runs to completion each time, but it gives me a different answer each time I run it. If I run the model over and over I get the same answer. But, if I close out GMS, then reopen GMS and run the model, I get a different answer. Or, if I change a k and remap, then change it back to the original and remap, I can get a different answer. This is driving me nuts and its put this project way behind schedule. Any help would be appreciated.

Edited by Hydrology

Share this post


Link to post
Share on other sites

What version of MODFLOW are you using? In our testing we have noticed that MODFLOW-NWT and MODFLOW-USG can sometimes give slightly different numbers with exactly the same model files. If so, you could try changing the LINMETH in the solver. Also, if you could send the model files to tech support (http://www.aquaveo.com/technical-support) we'll see if we can track down the problem.

Share this post


Link to post
Share on other sites

I have tried ver. 9.2.6 and ver. 10.0.4, and I get the same varying results. The differences are not small. I am running a pump test and the latest 2 runs show a drawdown of 5.7' 25 ft. away from the pump well and the next run showed 6.6 ft' drawdown 25 ft. away from the pump well. Both measurements were taken at the same cell. I am using Modflow 2005, Parallel, with the PCG2 solver. I will email the files to you.

Edited by Hydrology

Share this post


Link to post
Share on other sites

I tried to check the 64bit box in the Global settings for MODFLOW thinking that it might give me greater accuracy, but the program immediately crashes when I try to assign a starting head to the model if the 64bit box is checked. Update, I don't think that it is the 64bit check box, because it sometimes happens without it being checked. It seems that if you run it one time without setting the starting gw (ie let the gw default to the top of the layer), then set the gw to the correct elevation prior to the second run, then GMS wont crash. By the way, you get a different answer if you set the gw vs if you let it default to the top of the grid.

Edited by Hydrology

Share this post


Link to post
Share on other sites

I have also noticed that on some runs, areas within the upper layer of the model dry out. They are discontinuous random blobs of dried out cells. These are in areas of the model where absolutely nothing should be happening. They are a long way from where my well is, plus, the well does not go to the bottom of the layer, so nothing should ever dry out. The well is not drying out, so things further away from the well should definitely not dry out. Sometimes when I rerun the same model, the whole grid dries out.

Share this post


Link to post
Share on other sites

I was thinking that maybe the grid had gotten corrupted, so I deleted the grid frame and grid. Then created a new grid frame and grid. I created a new modflow model and mapped my model to modflow. On the first run, the whole of layer 1 was brought down to elev. -3.5. No slope, just a flat gw elev across the whole model, it seemed to ignore all of the boundary conditions. I then remapped the model without changing anything and ran it again. This time it picked up the boundary conditions and wells correctly, the the groundwater drawdown is different from any of the other runs that we have had.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×