Issue with generating waves from JONSWAP 2D wavemaker

185 views
Skip to first unread message

raphael Krier Mariani

unread,
May 3, 2022, 2:23:13 AM5/3/22
to funwa...@googlegroups.com
Dear FUNWAVE community, 
I have been facing some issues generating waves over an idealized 2D reef bathymetry using the JON_2D wavemaker. Thus, I have a few questions that require your expertise and I would like to pick your brains. 

Goal: Generate incident random waves with Hs=2m to ensure wave breaking at the reef edge (breaking index used: 0.82 from Tonelli and Marko (2009)).

Issues: 
1) With the present setup, I obtain large significant wave height values on the seaward side of the wavemaker, which exceed the desired Hs (>1.5 times greater than the H0 value in the input file). The largest value is obtained on the northern side of the domain (Fig.1). It is strange as it does not seem to be any reflection on the eastern and western boundaries that could generate resonance.  

2) The presence of undulations in Hs over the ramp also seems odd. These oscillations gradually increase toward the reef edge (Fig. 2), which leaves me wondering if these oscillations are generated by a partial reflection of waves at the reef edge or if they are just noise generated by "truncation error" due to the 2m grid interval on x. 

3) The distribution of Hs seems to be non-uniform alongshore (larger near the northern and southern periodic boundaries, smaller near the centreline of the domain)

4) For some reason there are large isolated peaks in free surface elevation within the sponge layer at the beginning of the model runs (Fig. 4) and I wonder why. It also seems that the free surface elevation jumps to reach maximums almost twice larger than H0 specified in the input file (Fig. 5) (which could happen if waves are partially reflected at the cliff edge).

Questions:
1) Concerning Hs extreme near the seaward boundary
- Are significant wave height values on the seaward side of the wavemaker related to the setup of the wavemaker? Or by the sponge layer on the eastern wall? Or both...? How can I fix this? 
- How is the width of the wavemaker specified?

2) Concerning Hs oscillations in the domain
- Are the undulations over the ramp and near the platform edge noise related to a poor spatial resolution on the X axis? For instance, there are only 4 gridpoints covering the forereef slope, is that an issue? 

3) Concerning longshore variations in Hs over the ramp.
Why Hs is un-evenly distributed alongshore? Is that because of the directional spread specified in the input file (also apparent in Fig.4)?

4) Concerning jumps in free surface elevation
Why do I obtain large values for eta near the wavemaker and within the sponge layer?

Previous suggestions on these topics on FUNWAVE forum.
- grid spacing needs to be smaller than wavelength: 50-80 points per wavelength.
- Lower CFL to reduce noise


The details of the domain characteristics are annotated on Fig 1 and 2. I have also attached to this email the input file I used for the simulation. I have the feeling that something needs to be changed in the wavemaker parameters or the sponge layers. In the meantime, I will try to drop Dx to 1m and see if this makes a difference. 

Thank you in advance for your help and for your time. 
Regards, 
Raphael









 
input.txt
Fig.4 isolated eta anomaly in the sponge layer 12sec.jpg
Fig.5 Large eta along wave crest 1012 sec.jpg
Fig.3 Hs top view.jpg
Fig.2 2D significant wave height (Profile view).jpg
Fig.1 2D significant wave height (oblique view).jpg

fyshi...@gmail.com

unread,
May 4, 2022, 11:37:42 AM5/4/22
to funwave-tvd
>1) With the present setup, I obtain large significant wave height values on the seaward side of the wavemaker, which exceed the desired Hs (>1.5 times greater than the H0 value in the input file). The largest value is obtained on the northern side of the domain (Fig.1). It is strange as it does not seem to be any reflection on the eastern and western boundaries that could generate resonance.  

This may be caused by the sponge layer where the parameter R_sponge was set too large. Try to reduce it to 0.85. In addition, Csp is also large.   

>2) The presence of undulations in Hs over the ramp also seems odd. These oscillations gradually increase toward the reef edge (Fig. 2), which leaves me wondering if these oscillations are generated by a partial reflection of waves at the reef edge or if they are just noise generated by "truncation error" due to the 2m grid interval on x. 

 It may be associated with wave refraction/reflection.  If numerical noise is a concern, try to enlarge your computational domain (in lateral) to see if you still get this pattern.  

>3) The distribution of Hs seems to be non-uniform alongshore (larger near the northern and southern periodic boundaries, smaller near the centreline of the domain)

It may be related to wave coherence. See the recent publication Salatin et al. 2021 JGR.  

>4) For some reason there are large isolated peaks in free surface elevation within the sponge layer at the beginning of the model runs (Fig. 4) and I wonder why. It also seems that the free surface elevation jumps to reach maximums almost twice larger than H0 specified in the input file (Fig. 5) (which could happen if waves are partially reflected at the cliff edge).

It's normal. Wave height inside the wavemaker should be larger than the target wave height. refer to Wei et al. 1999, CENG. 

raphael Krier Mariani

unread,
May 13, 2022, 4:46:55 PM5/13/22
to fyshi...@gmail.com, funwave-tvd
Hi Shi and all FUNWAVE users, thanks for the recommendations. 
I reduced the dissipation coefficient in the sponges on either side of the domain and it helped a lot. However, I still obtain oscillations in Hs. I did a series of sensitivity analyses on several parameters, which I am attaching to this email as I think it could be helpful to other funwave users. 

As you mentioned, the oscillations in Hs could be due to reflection. So I  generated directional spectra (attached) from gauges located at the shoreline, at the seaward extremity of the reef platform, at the bottom of the reef edge and near the wavemaker. However, there is no sign of reflection at these four locations.

These spectra were obtained from a 2D model using the same input (except for the wavemaker and domain dimensions) as the input file attached. 
I modified slightly the domain by adding a wider sponge on the western boundary corresponding to the wavelength of the IG waves at the shoreline. I also change the location of the wavemaker so it is now one wavelength away from the bottom of the ramp and one wavelength away from the eastern sponge.

I still obtain oscillations in Hs. Do you have any idea why these are generated? 

Thanks a lot for your time and help.

Regards,
Raphael

--
You received this message because you are subscribed to the Google Groups "funwave-tvd" group.
To unsubscribe from this group and stop receiving emails from it, send an email to funwave-tvd...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/funwave-tvd/82436209-4fb1-4ea6-a7ec-419a44f66e49n%40googlegroups.com.
input.txt
Shoreline.jpg
SpongeWidth.jpg
Top reef edge.jpg
Wave Maker.jpg
Bottom reef edge.jpg
Csp.jpg
CFL.jpg
cd.jpg
CDsponge.jpg
deltaWK.jpg
dx.jpg
Rsponge.jpg

Matt Malej

unread,
May 17, 2022, 5:13:46 PM5/17/22
to raphael Krier Mariani, funwave-tvd
See here about the coherent structures and try this new wavemaker for your Hs variability: https://fengyanshi.github.io/build/html/wavemaker_coherence.html?highlight=coherent



From: funwa...@googlegroups.com <funwa...@googlegroups.com> on behalf of raphael Krier Mariani <raphael...@gmail.com>
Sent: Friday, May 13, 2022 4:46 PM
To: fyshi...@gmail.com <fyshi...@gmail.com>
Cc: funwave-tvd <funwa...@googlegroups.com>
Subject: Re: Issue with generating waves from JONSWAP 2D wavemaker
 

fengyan shi

unread,
May 18, 2022, 6:32:43 AM5/18/22
to raphael Krier Mariani, funwave-tvd
Hi Raphael
I think your test results are helpful to the community. Regarding the oscillations in Hs, they may be caused by wave coherence as I mentioned last time. In the FUNWAVE  package (GITHUB) updated recently, there’re several wave coherence tests done by Salatin (see Salatin et al., 2021 JGR paper). You can modify your input files following those cases. 
/simple_cases/wave_coherence/

Hope it helps.
-Fengyan

To view this discussion on the web visit https://groups.google.com/d/msgid/funwave-tvd/CAExzas%3D-2D%2Bk2riBCsuCO%3Dxx29O1_X8JOwxt3SqZvArA7X_PqA%40mail.gmail.com.
<input.txt><Bottom reef edge.jpg><Csp.jpg><CFL.jpg><cd.jpg><CDsponge.jpg><deltaWK.jpg><dx.jpg><Rsponge.jpg><Shoreline.jpg><SpongeWidth.jpg><Top reef edge.jpg><Wave Maker.jpg>

raphael Krier Mariani

unread,
May 19, 2022, 11:30:19 PM5/19/22
to Matt Malej, funwave-tvd
Thanks, Fengyan and Matt, 
I tried FUNWAVE3.6 and there are some improvements in the simulations. Just a last question concerning the output I obtain from funwave3.6.
I observed differences in the gauge outputs generated by the two versions of the model and I wonder why. Before trying the new wavemaker I used the same input file that I previously used in V3.4 to run funwaveV3.6.

I run simulations of 2348sec to generate spectra over 2048sec observation period (excluding the first 300 seconds of the simulation, STEADY_TIME). The stations' sampling frequency is 2Hz (PLOT_INTV_STATION=0.5)

In the station output of funwave 3.4 I obtain 4696 samples, which corresponds to 2348 sec (no problem here, that makes sense).
However, in the station output of funwave 3.6 I obtain 5000 samples, which is 152sec longer than the duration of the simulation. After sample 4993, the time in the left column reaches 2348sec and becomes constant for the rest of the simulation. I suspect that from that point forward, the rest of the samples in the station file can be neglected (this might be due to the new station buffer introduced after V3.4. If not where does this difference come from?). However, the station timestamp reaches 2348sec at sample number 4993 instead of sample number 4996? Is the 3 sample difference due to the adaptative time-stepping or is it a bug? 
I might be a bit picky here but this could be important for higher-order spectral analysis, especially for cross-spectra correlation.

Thank you for your time and great guidance,
Raphael
input.txt
sta_1072 funwave3.6
sta_1072 funwave3.4
Reply all
Reply to author
Forward
0 new messages