Hi..the grade of meshing depends on degree of freedom and the desired accuracy..it is better to obtain the correct grade of meshing by performing trail analysis and based on the trail analysis arrive the suitable mesh. The trail analysis can be done by varying mesh dimensions and mesh density (coarse, medium, fine and very fine mesh). it is better to go with fine mesh for getting better accuracy of results..ok all the best..

Hi

You must do the grid independence. For this, you must choose the smaller grid until the answers remain almost constant.

The criteria are based on the distortion measures described above together with the element length. These are set by the engineer and are used in the quality check of the element mesh in the pre-processing program before the boundary conditions are applied. Elements that violate these criteria must be regenerated by the engineer and it can be regulated for example Mapped meshing Free meshing 16 5 Finite Element Analysis by an increase or decrease of the element size. The criteria are based on the distortion measurements: aspect ratio, skewness, mid point alignment and mid point deviation.

The most basic way is to refine the mesh successively until the results no longer change or so that the changes are judged small enough. Sometimes studying an engineering quantity of interest resulting from your calculation can be more telling than looking at the residuals.

Otherwise, there are automatic mesh adaptation methods that are VERY effective, but this software may not have any.

In general, the mesh should follow the variations of your unknow fields. Depending on your physic, it can be complicated.

Thanks everyone for your answers with best regards.

Refined mesh sizes (i.e., mesh with small sizes) provide more accurate approximate solutions than those of coarse sizes. But refined mesh comes with a computational cost at it takes more computing time to get a solution depending on the processing speed of your computer. However, surgical combination of refined and coarsed mesh can be selectively applied in your numerical model where needed in order to circumvent the computational cost and at same time optimise your approximate solution.

Hi,

The type of model element is also essential in determining the finite element mesh size. In terms of energy, the finite element model is generally more rigid than the actual structure. This problem, called shear locking, causes the structure model to be less deformed than the actual structure. Therefore, sometimes reducing the mesh size may not significantly alter the response. Surprisingly, the answer is also wrong! For example, the weakness of the full integration elements in bending can be noted. In this case, a second or reduced order element may be used to solve this problem.

Thank you for your answer.

There is another criterion called wavelength introduced by Semblat and Brioist [1] in the seismic analysis. Thus, “classically, the element size is chosen around a tenth or a twentieth of the wavelength.”

[1] J.-F. Semblat, J.-J. Brioist, Efficiency of higher order finite elements for the analysis of seismic wave propagation, J. Sound Vib 231 (2000) 460-467.