jimkirby / nhwave Goto Github PK

Hi,

I've been having a problem with the output probe filenames - probe file names are writing as hundreds of files with names of random strings, for example, probe_0001E1??e34??1?x, and many of the probe files are empty. I think I've narrowed the bug down to the line
STAT_FILE = TRIM(FDIR)//'probe_'//TRIM(FILE_NUM)
in the subroutine probes. I've made a bodge fix for it by changing it to
STAT_FILE = TRIM(FDIR)//'probe_'//TRIM(FILE_NUM(1:4)),
which solves the problem, and implies that the TRIM isn't working. No idea why.

On a slightly related note, I'm not sure why the probes subroutine isn't crashing when running in parallel - as far as I'm reading the code it looks like the file should be being opened at the same time by multiple cores, which should crash it...

Cheers,
Murray

We need to add Dimitry's model, or adjust the two-layer model to do either rheology

I'm looking to implement NHWAVE, in the future for the purpose of considering landslide generated tsunamis. We've had a first pass at setting up HYPRE and NHWAVE, both have compiled, and I've subsequently gone to run some of the supplied example cases - periodic wave over submerged bar and the Enet-Grilli lab slide. In both cases, the model ran to completion. However, the water level time series for both appear to be indicating the input surface deformations are not propagating rapidly through the grid, see the attached plot which indicates the surface water level near the left linear wave boundary condition, sampled at 1s intervals (period = 2s)

The appearance is similar to an overdamped vibrating string for example. I'm wondering if you've seen this issue pop up before, and if you might have any suggestions for its resolution? I was also wondering whether there might be an issue with the performance of the solver as it's been set up, as it is difficult to see how the problem could be attached to the nhwave source code.

Thanks and regards

Zac Couper

Cheng, create your own branch from the master, add the changes, and then request a pull.

I meant to tag the commit on Dec. 2 2016 as version 3.0, but it seems to be located at the most recent commit. I haven't figured out how to move the tag, or remoce it and place the tag where I want it.

Add the additions that were implemented for Kitimat study

I don't know how to do what I want to do here without causing trouble. If you look at the master-kirby branch, you will note that I have a lot more separated parts that were extracted from nhwave.F. Gangfeng has edits to nhwave.F and possibly two_layer-slide.F that will conflict with this, so that we can't write the two versions back over top each other. I'm not sure how to proceed. If I commit my changes now, the file structure in Gangfeng's and Cheng's versions will not agree with this. You'd either both have to split your files up the way mine are, or I should just drop this and wait for you to to get your changes made using the file structure that you both copied from the master. I guess I'll do that.

Dear developing group, is there any example for the set up of a floating obstacle in the NHWAVE input file? We are considering applying the model to simulate wave energy converters. Thanks.

The berkhoff example does not run. HOTSTART, RNG, DepConst, CUR, etc. are all missing from input.txt

Dr. Ma and Dr Derakhti, I found in the following subroutines some parts of code are supposed to be defined in parallel.

1. subroutine focused_wave_left_boundary: myid is defined in series code.
2. subroutine RandomU: myid is used in getting U,V W.
3. subroutine les_dyn(ISTEP): phi_3D_exch is called in series code.
   I can fix the first and third ones, but I dont know how to modify the second one.
   Could you please fix it?

Cheng

Hi NHWAVE Groups,

When trying the current benchmark example (Hilo Harbor), I got output named random code. (Shown in the following figure) Could you give me any suggestion or hint to fix it? I'll appreciate it a lot!

(Can I ask those basic questions like this one here when I'm learning use this model? It seems the forum on Sourceforge (https://sourceforge.net/p/nhwave/discussion/) is not used.)

Sincerely,
Chen

Someone needs to add these modifications.

The dimension for the variable xstat should be made dynamic (for writing nesting data)