Jump to content
GMS, SMS, and WMS User Forum
amy

Instabilities with Bouss2D

Recommended Posts

Hello.

I am experiencing some instabilities while running BOUSS-2d on 3500*2300 (m) grid with 10*10 (m) cells.

I've managed to make a successful run with a test: small duration (400s).

When I set the model to run for longer periods (1700s - still below the recommended duration), the model generates the aforementioned instabilities.

I have managed to select a time window for the animation that allowed me to see what happened: the model was working fine, and then it appears a huge burst near one of the extremities of the wavemaker, stopping the model.

Trying to solve this, I lowered the time step in order to decrease the Courant number, set the grid in order to keep the wavemaker at a constant elevation, then performing a grid smooth to avoid abrupt elevation changes.

None of these actions managed to overcome the instabilities.

Do you have any suggestions to what is causing this and how can I solve it?

Thank you in advance.

Best regards,

Amelia

Share this post


Link to post
Share on other sites

Amelia,

There are a few things I have seen that can cause instabilities as waves reflect into the wave maker. This particularly occures when the wave maker does not cover the entire row or column in which it is located (end effects).

If the wave maker is interior, you may need to put a thicker damping layer behind it. You may also want to move it a little further away from the structures that are reflecting back into it.

If these suggestions don't help, I would probably need to look at the specific characteristics of the simulation to give further suggestions. It sounds like you have made the standard corrections to address this.

Alan

Share this post


Link to post
Share on other sites

Alan,

Thank you for your quick answer.

I have an internal wavemaker that covers the entire column. Behind it, I already had a 100m damping layer with a coefficient of 0.9.

I performed another test increasing the layer to 300m and changing the coefficent to 1.0 and still no good results were obtained for 1700s.

The wave maker is 1900m far from shore.

Do you have any further suggestions, or is it better I send you the project to give a look at?

Thank you, once again.

Amelia

Share this post


Link to post
Share on other sites

Amelia,

As I said, it sounds like you have applied all the normal approaches for increasing stability. It also sounds like you have enough space behind the wave maker and enough distance to shore.

One other thing that can impact is waves bouncing off the side of the grid. What is the angle between your principal wave direction and the wave maker? If it is oblique, do you have damping on the sides of your grid?

You indicated that you modified the depths so that the wave maker was at a constant depth and smoothed the bathymetry already. What method did you use to smooth, and did you smooth the entire grid, or just limit the slope?

Alan

Share this post


Link to post
Share on other sites

Alan,

The principal wave direction is 0ยบ (shore normal projection), and I have damping layers on the sides too (40m with a coefficient of 1.0).

I have smoothed the entire grid with the the respective model tool, with several iterations. After that, I had to modify the grid manually in the harbour structures, in order to keep them above water.

Next, I've performed a single smooth grid iteration near the coastline to avoid the abrupt changes (I have also tried to run the model without these last steps - with submerse harbour structures - but still no good results for larger durations)

Thank you.

Amelia

Share this post


Link to post
Share on other sites

Amelia,

As I said, it sounds like you are doing things right. There are a few approaches at smoothing that are a little more complex that can be truer to the original geometry. Have you looked at where the instabilities are occurring? Are they in a shoreward area or closer to the wave maker?

I am afraid I don't have any more general suggestions. If you want make your project available to me, I could give it a quick review.

Alan

Share this post


Link to post
Share on other sites

Thank you Alan for all your suggestions. Regarding your last question, during the run, the model warns about some instabilities in a shoreward cell, but the burst I've managed to see in the animation occurs in one of the extremities of the wave maker.

I just uploaded the project files (Amelia_Boussinesq.zip) to your FTP site. I will be grateful if you can have a look at the project. Thanks.

Amelia

Share this post


Link to post
Share on other sites

Amelia,

BOUSS2D often reports initial instabilities like what you report on the shoreward cell. Those can often be ignored. The "burst" is the more important thing. SMS tech support got your message and notified me that they have the files. I have not had a chance to look at it yet, but I have them. What kind of a deadline are you working with? I will try to get to your files in the next day or so.

I assume that since you are able to see the burst, you have modified your output to include high temporal resolution around the time period of the instability. Have you noticed any waves reflecting to the wave maker that trigger the instability?

Alan

Share this post


Link to post
Share on other sites

Amelia,

We were able to get a version of your grid to run by removing the damping on the sides of the grid and setting the maximum depth of the grid to 25 meters.

The impacts of these two changes are:

1- The damping on the sides was causing distortion at the wave maker. I am discussing this with the developer of BOUSS2D to get clarification as to why the damping was causing problems at the wave maker and how he would address this issue in general.

2- A depth of 25 m should be sufficient to represent an 8 sec wave. If you want to run longer waves, you could use the original grid. The grid should work with the wave you have, but setting the cells that were deeper than 25 m to 25 reduced the slope in the area of the wave maker. This change should not be needed, but I need to test that more.

When I get more information, I will post it.

Alan

Share this post


Link to post
Share on other sites

Amelia,

The issue of damping on the lateral boundary is a little sensitive. There are planned features in the model to allow the lateral boundary to simulate an open boundary, so that no damping will be needed at all. In the mean time, you should use a weak damping on those boundaries (something like 0.1) if needed for oblique waves. If the wave is mostly parallel to the grid, you should try without damping.

I hope this has been helpful. Let me know if you run into more complications.

Alan

Share this post


Link to post
Share on other sites

Good morning Alan,

I am glad to inform you that with your last suggestions the problem was solved; I was able to get the model to run without instabilities.

Thank you very much for all your suggestions and your support.

Amelia

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

×