Counterquestion: Why are you developing it, if you don't know what it can be used for?

Of course it won't be useful!! Many are doing this, like 4sided, 5sided, even 20sided!! but all at the end can be reuced to a 3sided triad model... instead of making 1division 4sided u can make 2division 3 sided! not even a single change it ll cause!

@ Vedran and Ram : Thank you for your comments.

At one point even i thought the same. But, When i really got interested,I felt there is something more I can explore.

1. I am developing a five sided coarse mesh. It will not be reduced to triangular.

2. Using Triangular(3 node) and quadrilateral (4 node), 5 sided (5 node) must give higher accuracy.

More over I would like to say, i have considered 2 DOF per node.

I feel its a compromise between 6 node triangular and other lower order elements in terms of accuracy and computational time. In many literature they mentioned about flexibility meshing using polygonal elements. So five sided may work better..!!

5 sided means once again it can be formed to 3 triad meshes... penta means 5element and 5node but it can be divided to 3triads with addition of 2more elements.. . new information I gthered was that for sheetmetals and shell parts they will analyse with quad and not triad.. The reason is the elements in quad will be equally distributed and will be simpler for analysis.. but in case of triad the nodes will be like that but the distribution look little clumsy... Traids are generally used for plastic parts

Counterquestion: Why are you developing it, if you don't know what it can be used for?

@Ram : thank you for the information.

If the 5 sided is again divided in to triads it will not be a coarse penta/ 5 sided mesh. you can not call it as 5 sided mesh. and I found it one of the papers about the same. In my case its a pentagon shaped element. But not a regular pentagon.

@Martin : Its is very general concept in FEA that as the no of nodes increases- Number of degrees of freedom also increases, it also increases the accuracy. Its is assumed that it is a compromise in computational time too. that's it..!!

The question i posted here because, if somebody has some good experience in this particular field can explore more in this regard, which can help me to develop the element. Hope i have answered your question.

New elements can be developed in FEM to address many issues.

Like addressing the Modelling of a new material with different properties, coupling of structures, of even to increase the accuracy of FEA results

Finally for what application Penta meshes can be used and how your penta mesh uniques from the existing? Coz sheetmetals components uses quads for mesh. Plastic parts and few engine parts uses triads!! thats why.. and pentas are existing so how urs differs from others??

5 sided is not useful. Polygonal is becoming very useful. Voronoi polygonons are the dual to triangular meshes and are far more isotropic. In addition, elements which capture physics (rather than just math) are useful. Google exterior calculus and FE or Discrete Calculus. 5 -sided has been done 25 years ago.

I am confused...When you say 5 sided element, do you mean in 3D or 2D? What are the DOFs you are considering... You can have any number of DOFs per node if you are developing the element yourself. Also, what field is this for? Solid mechanics? Fluid Mechanics? Coupled Problems?

@ram

What exactly do you mean by no change happens when you use different types of elements? That statement does not make sense to me at all.

@blair :

Can u plz give me the link regarding 5 sided element.

@Prajwal

Its 2DOF per node. 5 side, pentagon shape(not regular pentagon)

solid mechanics will be my concern.

@ram

literature study says 3 things

1.polygonal meshing is flexible. I should check with pentagonal shaped element

2.Accuracy will be higher compared to 3node tri, and 4 node quadri element

3.computational time will be less from 6node triangle, 9 node quadri

Do you know meshless methods? I am working with Radial point interpolation method and i had great results. It's simple to implement and appears that this method is robust and eficient.

I completely agree with Danesh, there is still room. For example today there are few tools to simulate thin structures with solid elements. Higher order FEM is a good approach, but the standard tools does not have them. If you work in field of process simulation (effects of temperature) or adhesive bonds in a GFEM you cannot solve all aspects with shell elements.

Another topic is the geometrical optimization. The exchange between FEM and CAD is important. The isogeometric elements are maybe good way to solve this gap. Furthermore it gives researches about the finite cell method. Here you are able to make topological optimizations.