Dear CalculiX users,
the present implementation of the C3D8I-Element comes from Prof. Otto Bernhardi of the University of Applied Sciences of Karlsruhe, Germany. Now, he sent me an improved version, leading to better results especially for coarse meshes. After checking the implementation I plan to replace the present version of C3D8I by the new one, however, I would like to know whether people actually use this element, and whether there is serious objection to the replacement.
Dear Dr Dhondt,
this type of element can be very useful, I often give it a try in various tests and comparisons. Here’s one such example: Elasto-plastic bending - elements, accuracy
I’ve also seen several other threads on this forum where people used C3D8I elements in their studies.
It would be great to have an improved version of these elements implemented in CalculiX.
I haven’t been using linear elements. Even in your user manual it states that better results are obtained with quadratic elements. That being said, it never hurts to improve your element formulations. I can’t imagine anyone saying oh my god, please don’t make ‘something’ better. I like it worse.
dear dr. Guido,
many researcher in my areas still prefer linear incompatible hex element (C3D8I) to be used to perform an analysis with contact and plasticity of multipart objects. so, it could be a benefit if an improved version of these type element being updates in CalculiX.
**edited : an example application and document references
Hello Mr. Dohndt,
do you have and doubts or what is your concern about serious objection to the replacement?
Well, sometimes people want to be able to reproduce exactly the same results they got, let’s say, a year ago. If I change the formulation, also the results may (and will) change and I do not want to keep the old formulation. It just overloads the code. So people will have to keep the old version in order to reproduce their old results.
A change in results is for the industry usually not pleasant: the new formulation has to be benchmarked etc… and this means a lot of extra work.
hope, an improved versions are still considered to be include in next future release of CalculiX.
is this possible to use another element type or user element names? e.g C3D8S and UHE8S where S is denote to strain as many reference notes (link *1 & *2 ), or something else appropriates for initialed.