open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

I want to represent a piled jetty in my simulation. The dimension of the region of study is in UTM-zone 42,is in the range of 696464 E - 698092 E, 2289302 N - 2293593 N. The diameter of each circular pile of a jetty is 1 meter. In Bluekenue I tried creating a circle with 1 m diameter and then redistributing it at 0.2 meter interval, however bluekenue is not able to resolve 0.2 m dimension, that is I cannot get interval along the line of 0.2 m. Is there any way to get this dimensions in Bluekenue, and generate a grid of 0.2 m to represent the jetty piles.

One of my friend who happens to be an expert in opengl graphic has mentioned to me that before drawing any object in bluekenue drawing area, the graphic window on which we are drawing should be set with desired pixel format. If it is true then is there any way to change it.

Also does TELEMAC2D support grids of dimensions as small as 0.1 to 0.2 m?

With Best Regards,
Gourish

Hello,

The one trick I sometimes find helpful is to translate the x and y coordinates of the mesh so that the lower left most node is close to 0,0. In your case, this will avoid having 7 digit coordinate values, and may work in representing a 2 m structure.

Pat

Hello,

On Telemac side, the coordinates are read in the geometry selafin format file that is generally with single precision and are then stored and treated as double precision, so very small elements are possible. Double precision is also possible in the selafin format (this is supported by Telemac with keyword GEOMETRY FILE FORMAT), but maybe not in all pre and post-processors.

The problem with large numbers in coordinates is when areas are computed (e.g. integral of test functions). They are of the form X1*Y-X2*Y1, so if the Xs and Ys are very large and the element very small, we may loose digits in the truncation process, when the difference of 2 very large numbers gives the area. This is why shifting the coordinates to have the smallest numbers, like suggested by pprodano is a good idea (when there is no georeference needed).

With best regards,

Jean-Michel Hervouet

Hi Pat and Jean-Michel Hervouet,

Thanks for the reply. I tried the method as suggested by Pat, and yes it does work.

However if I am give forcing at the liquid boundaries using tidal constituents, that is when
OPTION FOR TIDAL BOUNDARY CONDITIONS = 1 and
TIDAL DATA BASE = 2,
I also need to specify the GEOGRAPHIC SYSTEM and ZONE NUMBER IN GEOGRAPHIC SYSTEM, implying that I have to specify the option ORIGIN COORDINATES which will be the coordinates X and Y referring to 0,0 in the mesh.

Thanks and with best regards,
Gourish

Hi Gourish,

My suggestion is to run a course mesh model (no refinement) and use the tidal forcing as you have initially defined it. You can then use the coarse mesh run to extract boundary conditions at the locations of interest. Then, refine the mesh, translate the coordinates, and impose the extracted boundary conditions at the liquid boundaries.

I recognize that this is really inefficient, and I defer to others for a more elegant solution.

Pat

Hi Pat,

Thanks for the suggestion. I have done the same technique earlier for some other simulation where there were no piles present, and I have also got some good results with calibration. However I have not tried with the case of pile structures, I shall try it and make a comparison with both the model runs to check the efficiency.

Presently I am trying the earlier suggested method.

With Best Regards,
Gourish

hi gourish
can you please explain me the steps for creating a circular profile in blue kenue
thanks

Hi,
two options come in my mind:

1. use a CAD programm, draw a circle and import the xy coordinates into BlueKenue (xyz ascii file)
2. more elegant way: in Excel or somewhere calculate the xy coordinates of the circle by means of the equation for the circle, cartesian or polar. save the coordinates into a xyz ascii file

Best regards,
Clemens

Thank you Gourish

I generated using matlab