Creating user defined elements in ABAQUS Explicit

[fred1980]

Hi there,
I am trying to develop a user defined element subroutine in abaqus explicit, VUEL. I'm starting with a simple 4 node linear solid element. However still there are some questions regarding VUEL structure. I'm wondering if anyone can introduce me a good source that explains how to write a VUEL or perhaps an example.

thanks

[StephenKiernan]

Hi Fred,

I had to write a 3-D hexahedral VUEL as part of my Phd and apart from the ABAQUS documentation the only book I looked at was 'Finite Element Method: A Practical Course' by Liu G.R. It doesn't give VUEL examples but it explains the formulation of each type of element (1-D, 2-D, 3-D) in detail and helped me a lot.

[fred1980]

Thanks Stephen,
Well, I'm relatively familiar with finite element formulation although I definitely need to take a look at the book you introduced to finish the job,
Right now my biggest concern is that ABAQUS explicit uses logarithmic strain rate. In my previous experiences I used to use the relation e = B*U where e is strain, B is the derivatives of shape function and U is the deformations of the node to derive the strain at any point. This formula in face represent the fact that in my previous experiences strain, let’s say in x direction, is defined as e = dU/dx. However, I’m not sure what the approach is when the strain rate is logarithmic. Like I said right now my problems are more related to VUEL structure itself. Also if I want to write a nonlinear VUEL I need to know how to derive geometrically nonlinear logarithmic strain. Therefore, at this stage I really need to find some detailed explanation of VUEL.

Thanks again,

[Issac44]

StephenKiernan if u come across any other book that give VUEL examples in detail plz do tell......thnx

[StephenKiernan]

@fred1980
Maybe I'm wrong but isn't the strain that is calculated from B*u non-linear already? This is the equation I use and the strain that comes out is the true logarithmic strain, not engineering.

@Issac44
If I see one I'll let you know

[FEguru]

I believe with VUEL you must handle the stress and strain definitions yourself in the coding. ABAQUS provides displacements and displacement increments, and expects the right-hand-side vector. You don't have to use logarithmic strains by the way just because ABAQUS uses them for its own formulations - that's part of the freedom of using your own elements! However, if you do use logarithmic or any other "advanced" strain measure then you have to derive the required B matrix relating those strains to displacements. In that case, I recommend reading books/papers on the topic of large deformations, such as the one recommended by StephenKiernan. If the element you're developing is 1D that process will be much easier than if it is 2D or 3D.

[fred1980]

StephenKiernan:
well, I know the strain that ABAQUS provides for VUMAT is logarithmic, therefore I thought we should use the same logarithmic strain in VUEL but considering what FEguru said I think he is right and one do not neccessarily have to use logarithmic scale. However, in VUEL example provided by abaqus (the cases of trusses) logarithmic strain are used. Of course for a 3d brick element I think it should be much more difficult than truss elements to use logarithmic strain.
I think if you just simply calculate the strain from B*U it would be linear unless you consider the rotational terms in B matrix.

FEguru :
thanks for your reply, still I hope I can somehow implement logarithmic strain in VUEL. I'll look to see if I can find required formulations.

[yufei]

Hi there,
I am trying to develop a user defined element subroutine in abaqus explicit, VUEL. I'm starting with a simple 4 node linear solid element. However still there are some questions regarding VUEL structure. I'm wondering if anyone can introduce me a good source that explains how to write a VUEL or perhaps an example.

thanks

Hi fred,
I am also want to develop a VUEL with plane 4 node solid element. However I met some problems, and the followings are my ideas. Could you please give me some advice or point out mistakes? thank you.
1. The material is linear elastic, so I think {rhs}=-[K]{u}, [K] is the element stiffness matrix, {u} is the vector of displacement at the end of the current increment provided by vuel.
2. When calculating the energy, I use energy(kblock,iElIe)_n = 1/2 * {u}^t *[K]*{u}, {u}^t is the transpose of {u}. By the way, I wonder if it is necessary to be defined.
3. I set dtimeStable(kblock) = 1e-7, a very small value to guarantee convergence.