open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,
I have just tried to post this message but it has not appeared on the forum, so please don't consider it in case the first one finally gets published.

I am trying to model a bed with two layers (surface and subsurface) having different grain-size distributions. I am using init_compo.f to specify the initial bed composition with:
NCOUCHES=2
ES(J,1)=1
AVAIL(J,1,I)= ... and AVAIL(J,2,I)= ...

My first question is about the layers thickness: should I also specify the second layer thickness ES(J,2) or is this directly computed by Sisyphe according to the total bed thickness?

Secondly, do I need to use the keyword NUMBER OF BED LOAD MODEL LAYERS in the cas file? According to the manual, the default value for this is 2, which I don't really understand...

Thirdly, I get an error message 'ACTIVE LAYER TOO BIG/STRATIFICATION' when I launch this simulation. According to init_avail.f, this message occurs qhen the active layer is thicker than the first layer. But I get the message even when I set the first layer thickness to several meters, which is much larger than what the active layer should be. Is the active layer indeed computed as three times the mean sediment diameter, as explained in the v6p0 manual?

If anyone has some insights about this, it would be very helpful and appreciated!
Thanks

Caroline

Hi Caroline,
1) the thickness of the last layer is computed by Sisyphe in order to get the height of sediment (sum of ES) equal to ZF - ZR
2) Effectivement, by default the number of bed load layers is 2
3) This is a tricky point...I'd suggest you to perform some sensitivity analysis on the value of this parameter. My student is also taking a look to this sensitive parameter...so your feedback will be very useful. By the way, in the literature there are not too much information about this issue.

all the best,

Pablo

Bonjour Caroline!
If I understand : the first layer in your model (the top layer) should be the active layer. Its layer is specified by keywords 'ACTIVE LAYER THICKNESS'. So make sure in init_compo, that the top layer is equal to the top layer ...
The second layer does not need to be specified : the subroutine init_avai will fill it up in order to insure the total bed layer thickness is equal to the difference between bed layer and rigid bed : (ZF-ZR).
Good luck!
Catherine
PS: the default value being 2 you don't need to specify the number of layers in your steering file

Hello,

Thank you Catherine and Pablo for your suggestions.
It works indeed if I specify the active layer thickness in the .cas file equal to the top layer thickness described in init_compo.f.

However, I think this does not corresponds to what I want to model: I want to model am armored bed, where the surface material (say, first 10 cm for instance) is coarser than the subsurface material (everything below 10cm). So I would represent this as a layered bed with top layer of 10cm and below, another layer.

It seems to me that this top layer is not the same thing as the active layer: Initially, the active layer (thickness ha) should be part of the top layer (I don't see any reason by the way why the active layer has to be of the same thickness of the top layer), but then, as scour occurs, the top layer (surface material) is removed, so the active layer should be composed of material from the subsurface.

In other words, I see that the top and below layers described in init_compo.f have fixed coordinates (10 cm below initial bed surface) whereas the active layer has moving coordinates (ha cm below current bed surface). So the active layer can be part of the top layer or of the lower layer depending on how the bed evolves - scour or deposition.

Perhaps I have misunderstood completely the purpose of the layers description?
Or, if indeed that what the init_compo.f means, I should be able to define independantly the top/bottom layers of the bed material, and the active layer properties. How could I do that?

Thank you for your advice,
Caroline

Hi Caroline,
I also wish to do something similar to what yiu describe & am having problems understanding how to do this.

I wondered if you had solved youir problem, as I noticed theer did not seem to be any more replies to yiuir post.

Best wishes
John

Hi John,

No I have not really solved it, I still don't understand why the description of the active layer (sediment transport property) has to be coupled with the description of the bed composition.

Anyway, if this can be helpful, a configuration that works for me is the following:
Describe two (or N) layers in your init_compo.f file (top and bottom layer, say). Then in the .cas file, used an active layer thickness which is smaller than your top layer, and indicate that the number of bedload layers is 3 (or N+1).

I have heard recently that the new release of Sisyphe v6p2 should incorporate a new algorithm to compute active layer independently of the bed composition. Perhaps this will help, but I have not found it until now in the current version available on SVN.

Good luck
Caroline

Hello Caroline,
Thanks for your reply. I will try your suggestion. However, I am now using version 6.2 so the implementation may have changed.

I am attending the 2012 telemac workshop/conference & will try and find out more there and post any progress.

John