open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello, could anyone help to clarify in the boundary conditions file how the boundary nodes are ordered, for a domain with multiple holes (i.e. islands)? For example, if my domain includes 5 holes, I understand node id 1 should be the one with minimum x + y, and go increasing order in anti-clockwise direction all the way to the number of nodes along external outline.

However, what about the holes (I understand the ordering is in clockwise direction now)? Which hole should be numbered first, the one again with minimum x + y? It does not seem it is the case after some manual checkup.

Any input is greatly appreciated.

Hailiang

Hello

yes islands are ordered in clockwise order.
The order of island is managed by the mesher and theier order for Telemac is their order or appearing in the boundary condition file.

with my best regards

Riadh

Thanks Riadh. But how nodes are numbered in the boundary conditions file. I understand the file could be created with Fudda or Blue Kenue. But I assume a software has to follow the way telemac is viewing the boundary nodes. Does telemac enforce a certain way for node numbering inside boundary conditions file?

Also I noticed the numbering (all nodes not only the boundary ones) may be changed for model run by telemac. For example, when i check the IPOBO arrays in the input geometry file and in the result file, they are different. Could you confirm whether telemac internally reorder the nodes maybe for the purpose of bandwidth reduction? If so, does it mean telemac does not care whether the input nodes have minimum bandwidth and telemac will handle it internally? Thanks,

Hailiang

My question on boundary node numbering has been answered in another post.

Could anyone comment on the IPOBO array in the geometry file? To my understanding, it is a one-dimensional array. If one node (with id i) is internal, IPOBO = 0, otherwise IPOBO = i. I was wondering how this array is used (maybe updated?) inside telemac. When I use python script (pputils) to check IPOBO array in the result file, the IPOBO values are different from the geometry file.

If I use BlueKenue to view the result file, the IPOBO array is the same as in the geometry file. Seems BlueKenue is not using the array - maybe to be consistent with the geometry one?

Thanks

Hailiang

I did more research on the IPOBO array. I understand it is a 1D array with dimension NPOIN. My observation is that BlueKenue does not define IPOBO array the same way as tel code.

BlueKenue defines IPOBO array like this: IPOBO(1) = 1, IPOBO(2) = 2, IPOBO(3) = 0, ... the array index is node ID and is always the same as value, except value is zero for internal node (not at boundary).

However, tel code defines IPOBO array like this (as parsed with pputils scripts): IPOBO(65) = 1, IPOBO(77) = 2, IPOBO(3) = 0,... the array index is still the node global ID. However, the value 1 corresponds to node ID 65, which has the minimum (x + y). The value 2 corresponds to the ID of the next boundary node in anti-clockwise direction. Node 3 still has value 0 indicating it is a internal node.

Could anyone help to explain how IPOBO array is used in the code? Does the node value really matter for tel code, as long as the internal nodes and boundary nodes are correctly distinguished as BlueKenue currently can? I believe many tel modelers use BlueKenue to create slf files so it would be beneficial to get confirmation from the tel development team on this.

Thank you.

Hailiang