swash-users — for questions and discussion related to SWASH

Re: [swash-users] SWASH Tracer Module - Spurious Velocities

[Thomas P. <t.p...@qu...>]

Hi all.

I am currently trying to get the SWASH tracer module working to simulate the advection and diffusion of a neutrally buoyant passive tracer. I have seen very little work in the literature modelling tracers with SWASH, limited to Hilgersom et al. 2018 (GMD - An axisymmetric non-hydrostatic model for double-diffusive water systems<https://gmd.copernicus.org/articles/11/521/2018/>). I have noticed some weird (seemingly stability related) spurious velocities that are generated when the tracer module is turned on. Since this is a passive tracer field (using SALINITY, but DENSITY = OFF), the difference between the two velocity fields should be zero but that is not the case. I've included the .sws files here. The tracer field is a small rectangle with values of 10 while everywhere else the salinity value is zero.

If anyone has any experience with the tracer module in SWASH that may be of assistance, it would be greatly appreciated!


Thanks,


Thomas Pendergast (he/him)
Ph.D. Candidate
BWRC Research Assistant
Queen's University
Department of Civil Engineering

[swash-users] Has anyone worked with waves and discharge in unstructured meshes?

[LEONORA A. P. Z. (Alumno) <LEO...@us...>]

Has anyone worked with waves and discharge in unstructured meshes?
I have some questions and would be very grateful if someone could help me.

Leonora Pereira Z.
Estudiante de Ingeniería Civil
Universidad Técnica Federico Santa María

[swash-users] Can I implement tide and waves on the same boundary?

[LEONORA A. P. Z. (Alumno) <LEO...@us...>]

Hello everyone, can I implement tide and waves on the same boundary?
Greetings,

Leonora Pereira Z.
Estudiante de Ingeniería Civil
Universidad Técnica Federico Santa María

[swash-users] Restrained floating object analysis with SWASH

[Agustin S. <agu...@gm...>]

Hello everyone,

I’m trying to model a floating object in SWASH using the *FLOAT*command. It
seems to work at first,  it reports the water level in the first iteration,
but then it starts showing *NaN* values in every subsequent step.

After I define the  INPGRID DRAFT /READINP DRAFT and run the model, the
PRINT file shows this:

* INPGRID DRAFT REG 323180. 5423030. 0. 818 750 5 5  *
* READINP DRAFT 1 'ShipDraft.float' 3 0*
* FLOAT 0.0, 1.0*
* ** Error            : Data field skipped:0.0
                                                  *
* ** Error            : Data field skipped:1.0  *


I can’t find the right solution. Has anyone experienced the same issue?
I’ve attached the code I’m using below
________________________________________________________________________________
MODE NONST TWOD
$
SET  0.  90.  0.01  200  3  NAUTICAL

CGRID REG 323180. 5423030. 0. 4090 3750 818 750
$
VERT 2
$
INPGRID BOTTOM REG 323180. 5423030. 0. 818 750 5 5 EXCEPTION -99
READINP BOTTOM 1.0
'GrillaSWASH_818x750_5x5m_OrientadaEste_RompeolasCota0.grd' 3 0

INPGRID POROSITY REG 323180. 5423030. 0. 818 750 5 5
READINP POROSITY 1.0 'PorosityLayer.por' 3 0

INPGRID HSTRUCT REG 323180. 5423030. 0. 818 750 5 5
READINP HSTRUCT 1.0 'HeightStructure.hei' 3 0

INPGRID PSIZE REG 323180. 5423030. 0. 818 750 5 5
READINP PSIZE 1.0 'GrainSize.siz' 3 0

INPGRID WLEV REG 323180. 5423030. 0. 1  1 4090 3750
READINP WLEV 1.0 'Nivel.wlv' 3 0

$===================================================================
$  BLOQUE FLOAT –(restrained)
$===================================================================

INPGRID DRAFT REG 323180. 5423030. 0. 818 750 5 5
READINP DRAFT 1 'ShipDraft.float' 3 0
FLOAT 0.0, 1.0
NONHYDROSTATIC BOX PRECond ILUD

$===================================================================
$  END BLOQUE FLOAT
$===================================================================

BOU SHAPE JON
BOU SIDE E CLOCKW SMOO 12. SEC CON SPECT  1.  14.  90. 4. CYCLE 3000 SEC

BOU SIDE N CLOCKW BTYPE RADIATION
BOU SIDE S CLOCKW BTYPE RADIATION

BRE 0.6 0.3
FRIC MANNING 0.019
VISC 0.

$ Implicit integration (required for FLOAT)

TIMEI METH IMPL 0.01 0.5

DISCRET UPW UMOM H BDF
DISCRET UPW WMOM H BDF
DISCRET UPW UMOM MOM V

QUANTITY HSIG dur 8 min

$************ OUTPUT REQUESTS *************************
$
FRAME 'GRILLA3' 323180. 5423030. 0. 4090 3750 818 750
TABLE 'GRILLA3' HEAD 'Grilla_sal.txt' XP YP DEPTH WATL HS OUTPUT 000000.000
60 SEC

POINT 'BOYA_GSM' 324870.48 5425365.89
TABLE 'BOYA_GSM' HEAD 'PuntoBoyaGSM.txt' XP YP TSEC DEPTH WATL OUTPUT
000000.000 0.2 SEC

$TEST 1,0
COMPUTE NONSTAT 000000.000 0.05 SEC 001000.000

STOP

[swash-users] Request for help: empty Errfile/PRINT files fixed with GFORTRAN_UNBUFFERED_ALL, but with major slowdown

[Luis S. <sar...@in...>]

Dear SWASH Community,

We would like to share our findings, and ask for your advice, regarding an
issue we have seen mentioned before on this list: after an unsuccessful
run, SWASH was not producing the expected Errfile-* and PRINT-* files.

We were alerted to this problem by one of our users, who is also part of
this community (thank you!!!), who noticed that when his simulations
crashed on Inductiva (our cloud machines are running Ubuntu), the error and
print files were empty. However, when the same simulations crashed on his
macOS laptop, the files were produced correctly.

We confirmed that this issue affects all SWASH versions we have installed
at Inductiva (9.01A, 10.01A, 10.05, and 11.01), and it occurs regardless of
the number of MPI processes.

After some debugging, we discovered that the behavior is controlled by a
Fortran runtime environment variable:

https://gcc.gnu.org/onlinedocs/gcc-4.8.1/gfortran/GFORTRAN_005fUNBUFFERED_005fALL.html

When this variable is set to TRUE, Fortran performs all writes in
unbuffered mode, meaning data is written to disk immediately, which ensures
that the error and print files are complete even when the program crashes.

Setting this variable indeed solved the problem: the Errfile-* and PRINT-*
files are now correctly written after failed runs.

However, we quickly discovered a serious side effect: for simulations that
perform output merging at the end (for example, merging mat files from
multiple ranks), runtime becomes dramatically slower — in our tests, by
roughly a factor of 10x or more.

We assume this slowdown is due to disabling I/O buffering entirely, which
causes frequent disk writes across ranks during the simulation and final
merge. And while this may not be visible for a low number of MPI ranks
(which may be why this is not so noticeable when running SWASH unbuffered
on a laptop) it turns out to have a massive impact when running SWASH on
large cloud machines with 100+ MPI ranks.

Before we commit to a permanent workaround, we’d like to ask the community:

1. Has anyone found a way to ensure error and print files are flushed after
crashes without turning off all Fortran buffering globally?

2. Is there a recommended way within SWASH (or in its MPI configuration) to
flush output safely at runtime or during abnormal termination?

We would be very grateful for any insights, experiences, or suggestions.

Best regards,

*Luís Sarmento*

CEO
+351917372727

*Porto* | Lisboa
*Website* <https://inductiva.ai/>  |  *LinkedIn*
<https://www.linkedin.com/company/inductiva-ai/>  |  *Discord*
<https://discord.gg/XFmXGg6Z>

*Notice: This email and any attachments may contain confidential or
privileged information. If you are not the intended recipient, please
notify the sender and delete the email immediately. Unauthorised use or
disclosure is prohibited.*

[swash-users] R: Merging output files

[Andrea B. <and...@un...>]

Hi Zhiling,
thank you very much for your kind suggestion. At last I myself devised the same solution for postprocessing of results with MatLab.
Nonetheless, in view of data management for graphical representation, it would maybe better enabling the code to write output files for different and contiguous time segments, including a tendblk in addition to the tbegblk, rather than requiring to “mosaic” the output in postprocessing.
To avoid any problem, other contibutors to the forum suggest using different formats instead, e.g. netcdf.
About you last query,

I have been struggling with the issue that my simulation stops with the message
'Terminating error: INSTABLE: water level is too far below the bottom level!’

I suspect this was because of some nearshore points with steep bathymetry but I cannot be sure without further info from SWASH. I wonder does anyone know is it possible to let SWASH output more info on, for example, at which grid point the water level diverges? I assume this could be avoided by setting an extremely small time step but that would be not worth it just because of one or two points in the domain. Thank you all.
***************************************************************************************************************************

you may find indication of the cells with negative depth (in which, or very near to that, instability develops) in the log PRINT files.
However, such occurrences may be very difficult to fix. Beside the drawbacks you mention about reducing the time step, even increasing the depth tolerance is not always a solution, as cells may be cut off abruptely, when the flux across a side is not very small, thus inducing instability in the nearby cells. You can also consider to smoothen your bathymetry – although it is clear that beyond a certain limit you might obtain a too different one from the original 😊 .
Good luck!

Andrea


Da: Zhiling Liao <tx....@gm...>
Inviato: martedì 21 ottobre 2025 15.36
A: Andrea Balzano <and...@un...>
Cc: swa...@li...
Oggetto: Re: [swash-users] Merging output files

Hi Andrea,

I had the same issue recently and I found mine was because of the file size limit of the merged .mat file to be <8Gb. It was tricky because although my final merged file will definitely not reach 8Gb, somehow the intermediate .mat file during merging can be very large and at some point exceeds that limit, triggering swash to pull the alarm and stop merging.

My solution was to divide my frame for BLOCK output into smaller frames so that each .mat file for each subframe is guaranteed to be small enough.

BTW, although the 8Gb file size limit was documented in user manual Section 5.6, I do hope this and its potential solution (e.g. sub-frame block outputs) can get more highlighted.

I hope this helps your situation.

Best,
Zhiling



On 29 May 2025, at 16:19, Andrea Balzano <and...@un...<mailto:and...@un...>> wrote:

Hi everybody,
I have an issue related to SWASH run in parallel mode not producing the unique MatLab output files after completing the calculations, leaving me with the processor-related output files.
From the forum of this mailing list, it is clear to me that a solution to this issue is producing the output as netcdf or vtk files. Nonetheless, the last simulation that I launched failed a first time due to a power down and, secondly, resulted in not producing the output files, as described. The problem is that the simulation is a long one, taking some two weeks to terminate.
I then ask if someone is informed about the possibility of merging the processor-related output files with some post-processing procedure/script or whatever, so that I don’t have to wait two more weeks to have my results as netcdf or vtk files.
Thank you very much in advance

Andrea Balzano
________________________________________________________________________________
Prof. Ing. Andrea Balzano, PhD
Università di Cagliari
Facoltà di Ingegneria e Architettura
Dipartimento di Ingegneria Civile, Ingegneria Ambientale e Architettura - DICAAR
Via Marengo, 2
09123 Cagliari
Phone.: +39 070 6755304
Fax:    +39 070 6755304
e-mail:  and...@un...<mailto:and...@un...>




_______________________________________________
swash-users mailing list
swa...@li...<mailto:swa...@li...>
https://lists.sourceforge.net/lists/listinfo/swash-users

Re: [swash-users] Merging output files

[Zhiling L. <tx....@gm...>]

Hi Andrea,

Glad to know you came up with the same solution : ) I agree a tendblk feature will be welcomed by the users (the two of us for sure). I tried the .vtk format output but it complains about partitioning the grid during preparation and then stop, I did not dig into that anyway. Netcdf format was also mentioned in the previous email-list as you suggested, but I did not see from the manual how to do that (or do we simply just replace .mat with .nc? I don’t know…) 

Thanks very much for responding to my last query. I wonder how can I let SWASH writes to PRINT file the cells with negative depth because currently all I got is just sth like
	 Time of simulation  ->  001858.625         in sec:       1138.62500
	 ** Warning          : time step is halved! New time step:      0.01250 sec
	 Time of simulation  ->  001858.637         in sec:       1138.63750
	 ** Warning          : new minimal depth for checking drying and flooding: DEPMIN =   0.3215E+00
	 ** Terminating error: INSTABLE: water level is too far below the bottom level!
	 ** Message          :           Please reduce the time step!
I am trying to smoothen my bathymetry but I never liked this solution for obvious reasons. I know you can set the TEST command to output errors at different levels, but it is also not clear how to use that to troubleshoot the INSTABLE water level issue. 

Best,
Zhiling

> On 29 Oct 2025, at 13:06, Andrea Balzano <and...@un...> wrote:
> 
> Hi Zhiling,
> thank you very much for your kind suggestion. At last I myself devised the same solution for postprocessing of results with MatLab.
> Nonetheless, in view of data management for graphical representation, it would maybe better enabling the code to write output files for different and contiguous time segments, including a tendblk in addition to the tbegblk, rather than requiring to “mosaic” the output in postprocessing.
> To avoid any problem, other contibutors to the forum suggest using different formats instead, e.g. netcdf.
> About you last query,  
>  
> I have been struggling with the issue that my simulation stops with the message
> 'Terminating error: INSTABLE: water level is too far below the bottom level!’
>  
> I suspect this was because of some nearshore points with steep bathymetry but I cannot be sure without further info from SWASH. I wonder does anyone know is it possible to let SWASH output more info on, for example, at which grid point the water level diverges? I assume this could be avoided by setting an extremely small time step but that would be not worth it just because of one or two points in the domain. Thank you all.
> ***************************************************************************************************************************
>  
> you may find indication of the cells with negative depth (in which, or very near to that, instability develops) in the log PRINT files.
> However, such occurrences may be very difficult to fix. Beside the drawbacks you mention about reducing the time step, even increasing the depth tolerance is not always a solution, as cells may be cut off abruptely, when the flux across a side is not very small, thus inducing instability in the nearby cells. You can also consider to smoothen your bathymetry – although it is clear that beyond a certain limit you might obtain a too different one from the original 😊 .
> Good luck!
>  
> Andrea
>  
>  
> Da: Zhiling Liao <tx....@gm...>
> Inviato: martedì 21 ottobre 2025 15.36
> A: Andrea Balzano <and...@un...>
> Cc: swa...@li...
> Oggetto: Re: [swash-users] Merging output files
>  
> Hi Andrea,
>  
> I had the same issue recently and I found mine was because of the file size limit of the merged .mat file to be <8Gb. It was tricky because although my final merged file will definitely not reach 8Gb, somehow the intermediate .mat file during merging can be very large and at some point exceeds that limit, triggering swash to pull the alarm and stop merging.
>  
> My solution was to divide my frame for BLOCK output into smaller frames so that each .mat file for each subframe is guaranteed to be small enough.
>  
> BTW, although the 8Gb file size limit was documented in user manual Section 5.6, I do hope this and its potential solution (e.g. sub-frame block outputs) can get more highlighted. 
>  
> I hope this helps your situation.
>  
> Best,
> Zhiling
>  
> 
> 
> On 29 May 2025, at 16:19, Andrea Balzano <and...@un... <mailto:and...@un...>> wrote:
>  
> Hi everybody,
> I have an issue related to SWASH run in parallel mode not producing the unique MatLab output files after completing the calculations, leaving me with the processor-related output files.
> From the forum of this mailing list, it is clear to me that a solution to this issue is producing the output as netcdf or vtk files. Nonetheless, the last simulation that I launched failed a first time due to a power down and, secondly, resulted in not producing the output files, as described. The problem is that the simulation is a long one, taking some two weeks to terminate.
> I then ask if someone is informed about the possibility of merging the processor-related output files with some post-processing procedure/script or whatever, so that I don’t have to wait two more weeks to have my results as netcdf or vtk files.
> Thank you very much in advance
>  
> Andrea Balzano
> ________________________________________________________________________________
> Prof. Ing. Andrea Balzano, PhD
> Università di Cagliari
> Facoltà di Ingegneria e Architettura
> Dipartimento di Ingegneria Civile, Ingegneria Ambientale e Architettura - DICAAR
> Via Marengo, 2
> 09123 Cagliari
> Phone.: +39 070 6755304
> Fax:    +39 070 6755304
> e-mail:  and...@un... <mailto:and...@un...>
>  
>  
>  
>  
> _______________________________________________
> swash-users mailing list
> swa...@li... <mailto:swa...@li...>
> https://lists.sourceforge.net/lists/listinfo/swash-users

[swash-users] Troubleshooting Water level too far below bottom level

[Zhiling L. <tx....@gm...>]

Dear all,

I have been struggling with the issue that my simulation stops with the message
'Terminating error: INSTABLE: water level is too far below the bottom level!’

I suspect this was because of some nearshore points with steep bathymetry but I cannot be sure without further info from SWASH. I wonder does anyone know is it possible to let SWASH output more info on, for example, at which grid point the water level diverges? I assume this could be avoided by setting an extremely small time step but that would be not worth it just because of one or two points in the domain. Thank you all.

Best,
Zhiling

Re: [swash-users] Merging output files

[Zhiling L. <tx....@gm...>]

Hi Andrea,

I had the same issue recently and I found mine was because of the file size limit of the merged .mat file to be <8Gb. It was tricky because although my final merged file will definitely not reach 8Gb, somehow the intermediate .mat file during merging can be very large and at some point exceeds that limit, triggering swash to pull the alarm and stop merging.

My solution was to divide my frame for BLOCK output into smaller frames so that each .mat file for each subframe is guaranteed to be small enough.

BTW, although the 8Gb file size limit was documented in user manual Section 5.6, I do hope this and its potential solution (e.g. sub-frame block outputs) can get more highlighted. 

I hope this helps your situation.

Best,
Zhiling


> On 29 May 2025, at 16:19, Andrea Balzano <and...@un...> wrote:
> 
> Hi everybody,
> I have an issue related to SWASH run in parallel mode not producing the unique MatLab output files after completing the calculations, leaving me with the processor-related output files.
> From the forum of this mailing list, it is clear to me that a solution to this issue is producing the output as netcdf or vtk files. Nonetheless, the last simulation that I launched failed a first time due to a power down and, secondly, resulted in not producing the output files, as described. The problem is that the simulation is a long one, taking some two weeks to terminate.
> I then ask if someone is informed about the possibility of merging the processor-related output files with some post-processing procedure/script or whatever, so that I don’t have to wait two more weeks to have my results as netcdf or vtk files.
> Thank you very much in advance
>  
> Andrea Balzano
> ________________________________________________________________________________
> Prof. Ing. Andrea Balzano, PhD
> Università di Cagliari
> Facoltà di Ingegneria e Architettura
> Dipartimento di Ingegneria Civile, Ingegneria Ambientale e Architettura - DICAAR
> Via Marengo, 2
> 09123 Cagliari
> Phone.: +39 070 6755304
> Fax:    +39 070 6755304
> e-mail:  and...@un... <mailto:and...@un...>
>  
>  
>  
>  
> _______________________________________________
> swash-users mailing list
> swa...@li... <mailto:swa...@li...>
> https://lists.sourceforge.net/lists/listinfo/swash-users

[swash-users] Problems with unstructured mesh: "(corner)point outside comp. grid"

[LEONORA A. P. Z. (Alumno) <LEO...@us...>]

Hello everyone,
I’m trying to work with unstructured meshes, but when I try to apply a boundary condition I get the following warning:
 ** Warning          : (corner)point outside comp. grid
       Set of output locations: BOUND_01  coordinates:       38.75       43.71
 ** Warning          : (corner)point outside comp. grid
       Set of output locations: BOUND_01  coordinates:        0.35       43.67
 ** Warning          : (corner)point outside comp. grid
       Set of output locations: BOUND_02  coordinates:       38.94        3.74
 ** Warning          : (corner)point outside comp. grid
       Set of output locations: BOUND_02  coordinates:       38.98       40.94
 ** Warning          : (corner)point outside comp. grid
       Set of output locations: BOUND_04  coordinates:        0.35       42.09

I want to emphasize that this error only appears when I try to implement boundary conditions. Could anyone help me understand this warning? I’d really appreciate it. I’m attaching the .node, .ele, and .sws files.

Leonora Pereira Z.
Universidad Técnica Federico Santa María


Leonora Pereira Z.
Estudiante de Ingeniería Civil
Universidad Técnica Federico Santa María

[swash-users] unstructured mesh: "(corner)point outside comp. grid"

[LEONORA A. P. Z. (Alumno) <LEO...@us...>]

Hello everyone,
I’m trying to work with unstructured meshes, but when I try to apply a boundary condition I get the following warning:
 ** Warning          : (corner)point outside comp. grid
       Set of output locations: BOUND_01  coordinates:       38.75       43.71
 ** Warning          : (corner)point outside comp. grid
       Set of output locations: BOUND_01  coordinates:        0.35       43.67
 ** Warning          : (corner)point outside comp. grid
       Set of output locations: BOUND_02  coordinates:       38.94        3.74
 ** Warning          : (corner)point outside comp. grid
       Set of output locations: BOUND_02  coordinates:       38.98       40.94
 ** Warning          : (corner)point outside comp. grid
       Set of output locations: BOUND_04  coordinates:        0.35       42.09

I want to emphasize that this error only appears when I try to implement boundary conditions. Could anyone help me understand this warning? I’d really appreciate it. I’m attaching the .node, .ele, and .sws files.

Leonora Pereira Z.
Universidad Técnica Federico Santa María

[swash-users] make ser do MPI compilation

[Mikaël G. <mgu...@un...>]

Hello,

I try to compile with gfortran 13.3 :

"make ser" end with this error :

    swanparll.f:43:10:

        43 |       USE MPI
           |          1
    Fatal Error: Cannot open module file ‘mpi.mod’ for reading at (1):
    No such file or directory
    compilation terminated.


Mikaël

[swash-users] AOCC flang compiler compatibility

[Mikaël G. <mgu...@un...>]

Hello,

Would it be possible to consider a version compatible with the flang 
(aocc) compiler with openMPI?
Currently, I get the following error during configuration:

# make config Current Fortran compiler 'flang' not supported.... make: 
*** [Makefile:205: config] Error 25

Thank you in advance.
Mikaël

[swash-users] Porosity layers after v4.01 implementation

[<hil...@im...>]

Dear community,

I recently noticed that the implementation of porosity layers in SWASH changed with v4.01 (some time ago, in 2017) from a two-term formulation (laminar and turbulent frictional forces following Engelund, 1953) to a three-term formulation (laminar and turbulent frictional forces plus an added-mass term, following Van Gent, 1995).

It seems that most of the validation and testing reported in the literature are from before these changes. Is anyone aware of any recent successful implementation of porosity layers in SWASH using the new formulation that could be shared? I am particularly interested in wave agitation studies involving both rock revetments and quay walls.

Any feedback would be greatly appreciated.

Thank you!
Best regards,

Hilario Castro Lara
Engineer-Advisor
[cid:image001.png@01DC088F.5C9E3550]<http://www.imdc.be/>
Van Immerseelstraat 66
B-2018 Antwerp, Belgium
Mob. +32 467 808 634
Tel.  +32 3 270 92 95
IMDC is looking for new colleagues! View the vacancies here<https://imdc.be/en/careers>
[signature_860041298]

ENGIE Mail Disclaimer: https://www.engie.com/disclaimer

[swash-users] Error during compilation of latest version

[Nicolas G. <nic...@ou...>]

Dear devs,

I tried to compile latest version using main branch from gitlab repo but I get this error:

$ cmake --build .
[3/28] Building Fortran object lib/CMakeFiles/swash11.01.dir/SwashBCspectrum.f90.o
FAILED: lib/CMakeFiles/swash11.01.dir/SwashBCspectrum.f90.o
/usr/bin/f95  -I../src -O -w -fno-second-underscore -ffree-line-length-none -Jmod   -fpreprocessed -c lib/CMakeFiles/swash11.01.dir/SwashBCspectrum.f90-pp.f90 -o lib/CMakeFiles/swash11.01.dir/SwashBCspectrum.f90.o
../src/SwashBCspectrum.f90:337:28:

  337 |        call random_seed(put=tseed)
      |                            1
Error: Size of ‘put’ argument of ‘random_seed’ intrinsic at (1) too small (1/33)
[24/28] Building Fortran object lib/CMakeFiles/swash11.01.dir/SwashQuanOutp.f90.o
ninja: build stopped: subcommand failed.

It seems related to this change: https://gitlab.tudelft.nl/citg/wavemodels/swash/-/commit/c4257fc693139cae42d7d95c94a03e633ac32837#2ce7c825b5b808abffd4ec32380f3ce8b24735be

Kind regards,
Nicolas

[swash-users] SWASH usage

[Martin P. <mar...@as...>]

Hello,

I am using SWASH for the first time trying to model propagating of waves in an open water body with some land masses around. I am using a bottom grid file with depths in each point which also includes land points given with nan. In the end I would like to be able to plot a three-hour developed wave state on a map for this particular area in a .vtk/.vts file. I have been struggling to set up the input file with the different commands, especially selecting the timesteps and in the output I do not get any values for Hsig for example over the whole domain. I only get a proper plot for the sea depth and land points. Could someone look into the input file in the attachment and give some directions as to what should I do further?

Best regards,

Martin

[swash-users] Merging output files

[Andrea B. <and...@un...>]

Hi everybody,
I have an issue related to SWASH run in parallel mode not producing the unique MatLab output files after completing the calculations, leaving me with the processor-related output files.
>From the forum of this mailing list, it is clear to me that a solution to this issue is producing the output as netcdf or vtk files. Nonetheless, the last simulation that I launched failed a first time due to a power down and, secondly, resulted in not producing the output files, as described. The problem is that the simulation is a long one, taking some two weeks to terminate.
I then ask if someone is informed about the possibility of merging the processor-related output files with some post-processing procedure/script or whatever, so that I don't have to wait two more weeks to have my results as netcdf or vtk files.
Thank you very much in advance

Andrea Balzano
________________________________________________________________________________
Prof. Ing. Andrea Balzano, PhD
Università di Cagliari
Facoltà di Ingegneria e Architettura
Dipartimento di Ingegneria Civile, Ingegneria Ambientale e Architettura - DICAAR
Via Marengo, 2
09123 Cagliari
Phone.: +39 070 6755304
Fax:    +39 070 6755304
e-mail:  and...@un...<mailto:and...@un...>

Re: [swash-users] SWASH error "the area of the centroid dual is negative..."

[Clayton H. <cl...@ca...>]

It sounds like you have either a zero area element (three colocated
points), or an element in which the points are oriented in the clockwise
direction (yielding a negative area).  Some griding software like Blue
Kenue have tools to find and correct orientation problems.

Cheers,

Clayton

On Wed, Apr 16, 2025, 9:55 a.m. Ana Catarina Zózimo <acz...@ln...>
wrote:

> Dear all,
> A student of mine is currently running the unstructured grid SWASH model
> and encountered an error message: *"The area of the centroid dual is
> negative or zero at the vertex" *after enabling the BREAKING option. This
> issue causes the program to terminate. However, the model operates
> correctly when BREAKING is not activated.
>
> Could you please assist us with this issue?
>
> Thanks and best regards,
>
> --
> Ana Catarina Zózimo
> Doutorada
> Núcleo de Portos e Estruturas Marítimas/Departamento de Hidráulica e Ambiente
> LNEC - Laboratório Nacional de Engenharia CivilAv. do Brasil 101, 1700-066 Lisboa, Portugal <https://www.google.com/maps/search/Av.+do+Brasil+101,+1700-066+Lisboa,+Portugal?entry=gmail&source=g>
> Tel. (+351) 21 844 3756
> ---
> Ana Catarina Zózimo
> Postdoctorate
> Ports and Maritime Structures Unit/Hydraulics and Environment Department
> LNEC - National Laboratory for Civil EngineeringAv. do Brasil 101, 1700-066 Lisboa, Portugal <https://www.google.com/maps/search/Av.+do+Brasil+101,+1700-066+Lisboa,+Portugal?entry=gmail&source=g>
> Tel. (+351) 21 844 3756
>
> _______________________________________________
> swash-users mailing list
> swa...@li...
> https://lists.sourceforge.net/lists/listinfo/swash-users
>

[swash-users] [PLEASE HELP!] Salinity transport in unstructured grids

[LEONORA A. P. Z. (Alumno) <LEO...@us...>]

Hello everyone,
I'm working with the new version 11.01, which includes constituent transport for unstructured grids.
I started with a very simple simulation. First, I ran it without any boundary conditions—everything worked fine. Then I added a small discharge on the left side of the domain—still fine. But when I added salinity to this same setup (with the discharge), the model crashes. According to the .prt file, the time step keeps halving until the water depths at some nodes become negative, which makes no sense since my model is just a box with a flat bathymetry of 10 meters across the entire domain.
Honestly, this is driving me crazy. Could someone please help?
Best regards,
Leonora Pereira Z.
Estudiante de Ingeniería Civil
leo...@us...
Universidad Técnica Federico Santa María

[swash-users] Request for Support Regarding NaN Results in Floating Object Analysis with SWASH

[Semih C. <sem...@do...>]

Dear all,

 

Recently, I have been working on analyzing floating object behavior using
SWASH. I aim to compare the model results with physical experiments. To this
end, I set up a wave flume domain and successfully simulated both regular
and irregular waves without any issues. However, after introducing a fixed
floating object into the model, the "WLEV" outputs across the entire
computational domain turned into "NaN" values. To resolve this issue, I have
tried the following approaches:

 

.              Tested various combinations of computational and input domain
sizes (10x40m, 20x40m, 10x100m, 20x100m, 40x100m, 100x100m, etc.).

.              Used different spatial resolutions (1x1m, 0.1x1m, 0.01x1m,
0.1x0.1m, 0.01x0.01m, etc.).

.              Conducted analyses with various time steps (0.1s, 0.01s,
0.005s, 0.001s, 0.0001s).

.              Defined regular and irregular waves approaching perpendicular
to the channel with wave heights between 0.1-1m and periods between 3-6s.

.              Configured the bottom file with both three sides closed and
all sides open boundary conditions.

.              Applied porosity to closed boundaries and compared results
with and without porosity.

.              Added sponge layers at boundaries without incoming waves and
observed the effects with and without sponge layers.

.              Tested both INIT WATLEV and INIT ZERO as initial conditions.

.              Defined water level across the entire domain.

.              Used the SET command to define SWL combinations between 0-5m.

.              Adjusted bathymetry between 0.5m-10m to investigate
bathymetry effects.

.              Defined the "draft" file both across the entire domain and
only at the structure location.

.              Ran analyses using BODY DIM and BODY SOLV commands.

 

Throughout these tests, although I occasionally encountered stability
issues, I was able to resolve them with appropriate adjustments. However,
the issue with NaN "WLEV" output persists specifically when using the DRAFT
and FLOAT commands. Upon reviewing the Draft output file, I can confirm that
the draft file is correctly defined and recognized by the model but despite
all efforts, I continue to receive NaN results.

 

I would greatly appreciate any suggestions or insights you might have
regarding this issue.

 

Thank you for your time and support.

 

Best regards,




Semih CAN
Civil/Coastal Engineer, MSc.

 

Dolfen Engineering&Consultancy

Üsküp Caddesi (Çevre Sokak) 20/3,

Kavaklıdere, 06680 Ankara

T: (312) 468 5930 pbx

F: (312) 468 5941

E:  <mailto:sem...@do...> sem...@do...

W: <http://www.dolfen.com/> www.dolfen.com

 

[swash-users] SWASH error "the area of the centroid dual is negative..."

[Ana C. Z. <acz...@ln...>]

Dear all,

A student of mine is currently running the unstructured grid SWASH model 
and encountered an error message: *"The area of the centroid dual is 
negative or zero at the vertex" *after enabling the BREAKING option. 
This issue causes the program to terminate. However, the model operates 
correctly when BREAKING is not activated.

Could you please assist us with this issue?

Thanks and best regards,

-- 
Ana Catarina Zózimo
Doutorada
Núcleo de Portos e Estruturas Marítimas/Departamento de Hidráulica e Ambiente
LNEC - Laboratório Nacional de Engenharia Civil
Av. do Brasil 101, 1700-066 Lisboa, Portugal
Tel. (+351) 21 844 3756
---
Ana Catarina Zózimo
Postdoctorate
Ports and Maritime Structures Unit/Hydraulics and Environment Department
LNEC - National Laboratory for Civil Engineering
Av. do Brasil 101, 1700-066 Lisboa, Portugal
Tel. (+351) 21 844 3756

[swash-users] Submerged breakwater representation

[DEKNEUDT Q. <que...@ar...>]

Hello everyone,
I would like to represent a submerged breakwater in SWASH 1D to estimate its effectiveness and wave attenuation behind it. What do you think is the best way to represent this structure? Is it directly in the bathymetry .bot or via a porous structure with porosity, height, and grain size?

Thank you for your responses.

I have already conducted a comparison test and the results are presented below.
[cid:image003.png@01DBA8BA.CBB19D40]

Best regards,
Quentin Dekneudt

[swash-users] [HELP!] Issue with Freshwater Appearing in Deactivated Cells in River-Ocean Interaction Simulation

[LEONORA A. P. Z. (Alumno) <LEO...@us...>]

Hello everyone,
I am trying to simulate the interaction between a river and the ocean, specifically to observe the behavior of the plume. For now, I only have a discharge (DISCH) as a boundary condition. My initial conditions are, of course, the bathymetry and a salinity file. My bathymetry is quite simple: the ocean has a steeper slope than the beach slope, and the river is channeled by a 3.5-meter wall. In my initial salinity condition, I set a salinity of 35 throughout the entire ocean domain and the river, and for everything that is "land" (negative value), I set a value of -999 because in my .sws file, I specify that I want the cells containing the value -999 to be deactivated (to avoid unnecessary salinity calculations and reduce simulation time). What happens when I do this is that, at 1 second, freshwater starts appearing at the mouth of my river, specifically around the cells I deactivated, which channel the river (gradual expansion). In my opinion, this is incorrect because freshwater should only appear where the river starts, and therefore where the discharge is. I want to clarify that freshwater does appear at the start of the river, but it also appears in my deactivated cells.
I will attach a photo to make it clearer. Does anyone know why this is happening? Is it inappropriate for me to deactivate the cells that represent my wall? Is it better to leave the entire domain with an initial salinity of 35? Is there any way to speed up the simulation without deactivating the cells?
I am also attaching my .sws, .bot, and .sal files.
The simulation did not crash; it runs fine, but I am disturbed by seeing this issue that shouldn’t be there.
Thanks.

[cid:9bb2b137-fa4e-47b5-8c81-3102a4f6d93f] [cid:157ec24b-b2db-4010-b854-c5d0e43b6530] [cid:36880485-ccec-43ed-a795-c6eda7ffa96f]
I would like to clarify that the blue color represents freshwater and the white color represents saltwater or 35

Leonora Pereira Z.
Estudiante de Ingeniería Civil y MSc. en Ingeniería Civil
Universidad Técnica Federico Santa María
Chile

[swash-users] Problems to register in swash users mailing list

[Ana C. Z. <acz...@ln...>]

Dear Sir/Madam,

A colleague of mine, in CC to this e-mail, is having problems in 
registering in the swash users mailing list. He already filled the form 
to register and updated his profile, but he still can't manage to access 
the group.

Can you please help us on that?

Thank you and best regards,

-- 
Ana Catarina Zózimo
Doutorada
Núcleo de Portos e Estruturas Marítimas/Departamento de Hidráulica e Ambiente
LNEC - Laboratório Nacional de Engenharia Civil
Av. do Brasil 101, 1700-066 Lisboa, Portugal
Tel. (+351) 21 844 3756
---
Ana Catarina Zózimo
Postdoctorate
Ports and Maritime Structures Unit/Hydraulics and Environment Department
LNEC - National Laboratory for Civil Engineering
Av. do Brasil 101, 1700-066 Lisboa, Portugal
Tel. (+351) 21 844 3756

Re: [swash-users] [PLEASE HELP!] Issues Running SWASH in Parallel with OpenMPI on Ubuntu 24

[Bunnik, T. <t.b...@ma...>]

Hi Leonora,

Could it be that you accidentally build a serial version of swash instead of the mpi version?

Best regards,

Tim


dr. ir. Tim Bunnik
 | Senior Researcher | Research & Development|Monday, 9 am – 5.30 pm​
Tuesday, 9 am – 5.30 pm​
Wednesday, 9 am – 5.30 pm​
Thursday, 9 am – 5.30 pm​
Friday, 9 am – 5.30 am​
T +31 317 49 34 29 | M +31 654 23 42 81 | t.b...@ma... | www.marin.nl
From: LEONORA ANTONIA PEREIRA ZAPATA (Alumno) <LEO...@us...>
Sent: Tuesday, March 18, 2025 3:02 AM
To: swa...@li...
Subject: [swash-users] [PLEASE HELP!] Issues Running SWASH in Parallel with OpenMPI on Ubuntu 24

Algunos contactos que recibieron este mensaje no suelen recibir correos electrónicos de leo...@us...<mailto:leo...@us...>. Por qué es esto importante<https://aka.ms/LearnAboutSenderIdentification>

Hello everyone,



I have been trying to run SWASH in parallel, so I downloaded OpenMPI on Ubuntu version 24.



I ran a simulation with 15 cores, however, I get the impression that it is running the same simulation 15 times, meaning not in parts. In the terminal, the typical percentage that appears when running a simulation shows up 15 times. I also notice that the simulation takes MUCH longer than if I were running it serially.



Has anyone figured out how to resolve this?