For any given Finite Element Analysis (simulation) it's essential to define each part of the geometry with one or more elements depending on the types of structures and analysis. As far as I know, definition of element is necessary to ensure perfect Mesh Refinement. For few geometrical models it's convenient to decide the element types because any kind of element selection generates almost similar kind of results. But most of the case it's confusing to select right element type from various elements (Beam, Shell, Plate Bending, Plane stress, etc.).
Is there any quick strategy (from experience) to choose the exact element type? Any tactic or tricks to share? Thanks in advance!
I do not think there is no a strategy to select the element type. The element type selection depends on the problem. But you can rely on the studies in which similar cases was analysed for example to have a quick decision.
You are right Dr. Mazid, element selection depends on the problem type that’s what I mentioned even in my query. Despite knowing this, I want to hear some experience. For instance, it’s better to (most of the case) select Plain Stress Element for Brick except for few specific cases.
Nevertheless, thanks for your answer.
Thanks Mr. Claudio Pedrazzi for your comprehensible answer and join in the discussion. Unequivocally, dimensionality (D) of the problem is one of the key parameters to select the element type. Most of the 'Element Library' manual of the commercial softwares describes about the element types but I noticed that, the majority of the students (Undergrad, Postgrad) are reluctant to go through those manual 'deeply' as those manual loaded with tons of complex mathematical equations. Ha ha ha. :)
Nevertheless, it is paramount to understand those equations to clarify the suitability of the 'specific' element for the 'specific' purpose of FEA. It would be great, if there is a 'general' guideline for the beginners to make a 'simple' start as you stated in your crawl-walk-run rule.
Thanks once again for your answer.
Best Regards!
Following Claudio's comment, I propose a phylosophical answer too.
Luckily, the definition of the "model" still remains a prerogative of the engineer (i.e. there is no "authomatic modelling button"). So, in my opinion no general rule can be stated a priori, as each engineering problem needs to be critically analysed independently. Maybe a general attitude can be identified: use the simplest model geometry (2D instead of 3D), exploit symmetry whenever possible (e.g. axisymmetric model), start with simple material model (linear elastic vs plasticity) to help your physical understanding of the problem. Last, but not least, always rely on the theory of 'strength of materials' to get a physical understanding of your FE results (never trust numerical results per se).
In some circumstances, you may even find that a simple mechanical/analytical model works very well, in place of a time-consuming FE simuation.
- Badges
- Science topic