Error with inverse method

A discussion forum for users of the new HYDRUS 2D/3D. HYDRUS is a software package for simulating water, heat and solute movement in two- and three-dimensional variably saturated media. Happy Posting!
Post Reply
David Radcliffe
Posts: 25
Joined: Wed Oct 07, 2009 3:32 pm

Error with inverse method

Post by David Radcliffe » Mon Apr 29, 2019 12:37 pm

I have made good progress in running my hillslope model using Version 3. However, when I try to run the inverse method to start a calibration process, I get an error right away about writing to an output file. I have attached some screen shots. Thank you for your attention.
Attachments
Forum April 29 2019.docx
(1.13 MiB) Downloaded 8 times

mortsde
Posts: 3
Joined: Mon Sep 08, 2014 10:38 pm
Location: USA

Re: Error with inverse method

Post by mortsde » Tue Apr 30, 2019 11:49 pm

I'm experiencing a similar issue on a unique problem.

I can successfully simulation with "0" iteration steps. My initial hydraulic parameters are nearly equivalent to my existing parameters to ensure that my calculations are reasonable for a first run.

Jirka
Posts: 4671
Joined: Sat Mar 16, 2002 3:47 pm
Location: USA
Location: Riverside, CA

Re: Error with inverse method

Post by Jirka » Wed May 01, 2019 2:57 am

I do not see such problem with Version 3.01.1380 that I have installed on my PC (or 3.01.1470, which I've just installed).

There are still some problems with the project:
1. There is only one data point (in the objective function). Is that all that you want to fit?
2. This data point is for time=250d. There is no reason to have the final time of the simulation to be 365 d if measured data points end before that. The simulation will stop at the time of the last measured point.
3. Longitudinal dispersivity seems rather small for such a large domain.
4. There is certainly no need to have print time every day in the inverse problem.
Untitled.png
Untitled.png (266.15 KiB) Viewed 116 times
J.

David Radcliffe
Posts: 25
Joined: Wed Oct 07, 2009 3:32 pm

Re: Error with inverse method

Post by David Radcliffe » Wed May 01, 2019 12:32 pm

Thank you for the suggestions. When I check for updates using the Help menu, it says that I have version 3.01.1460 which is the latest version. Do I need to go to ELIS to get version 3.10.1470?

Jirka
Posts: 4671
Joined: Sat Mar 16, 2002 3:47 pm
Location: USA
Location: Riverside, CA

Re: Error with inverse method

Post by Jirka » Wed May 01, 2019 9:57 pm

You can always get upgrades in ELIS. If I recall, the only update in the computational modules was the option to allow the Walker equation for both negative and positive exponents. Otherwise, the computational modules should be the same and changes were done mainly in the GUI. J.

David Radcliffe
Posts: 25
Joined: Wed Oct 07, 2009 3:32 pm

Re: Error with inverse method

Post by David Radcliffe » Fri May 03, 2019 4:12 pm

We have tried running the model on two computers, both with Version 3.01.1460, and the inverse will not run. Is there some file we could look at that might give us a better idea of why it is failing on our systems?

Jirka
Posts: 4671
Joined: Sat Mar 16, 2002 3:47 pm
Location: USA
Location: Riverside, CA

Re: Error with inverse method

Post by Jirka » Sat May 04, 2019 12:44 am

I ran your project both at my desktop (with Windows 7) and notebook (with Windows 10), and I did not encounter any problems. I do not know what else I can advise.

The message 'Error when writing to an output file !' can appear after reading Sink Information when writing to output files Sink.out, h.out, th.out, V.out, ObsNode.out, and Balance.out. I do not write the file name as this should never happen. Perhaps, you can see, which of these files exist and in what states. Note that only ObsNode and Balance are text files, the other files are binary.

J.

David Radcliffe
Posts: 25
Joined: Wed Oct 07, 2009 3:32 pm

Re: Error with inverse method

Post by David Radcliffe » Wed May 08, 2019 1:00 pm

The solution to our problem was to open HYDRUS as Administrator.

Post Reply