请推荐一个shell element algorithm
大变形，4节点等参单元，5/6变量（u,v,w,theta1,theta2,theta3）最好是基于位移法的，非H-R, H-W变分的。
是 strain-space非线性，stress-strain线性的。
厚度可以是定常，但必须有厚度方向积分。
其实就是和simo/bathe的薄壳模型类似，但是厚度方向存在积分。
不知道各位朋友有没有了解类似这种比较成熟的壳单元。
多谢！ 有文献最好。 厚度方向有积分的话，难道是三维蜕化壳元吗？
只玩过线性平板型壳元，翻翻ABAQUS、Adina的理论手册，里面应该有单元参考文献的 大变形？要不要求大应变？
最好是基于位移法的？纯位移法的似乎难度较大吧，现在一个个的都是混合的吧？ MITC4 is an element with not so bad properties andalso easy to implemented. Just google MITC4 MITC4 is an element with not so bad properties andalso easy to implemented. Just google MITC4
hillyuan 发表于 2010-11-22 07:40 http://forum.simwe.com/images/common/back.gif
I have checked out the MITC4, which doesn't have guass integration through thickness.
Thanks anyway.
I have checked out the MITC4, which doesn't have guass integration through thickness.
Thanks anyway.
caoer 发表于 2010-11-22 23:07 http://forum.simwe.com/images/common/back.gif
Are you sure? If I am not mistaken, MITC4 shell element allows 2 Gauss integration points through the thickness for linear elastic analysis. For nonlinear analysis, it allows up to 6 points in the case of Gauss integration rule and 7 in the case of Newton-Cotes rule.
Are you sure? If I am not mistaken, MITC4 shell element allows 2 Gauss integration points through the thickness for linear elastic analysis. For nonlinear analysis, it allows up to 6 points in the ...
tonnyw 发表于 2010-11-22 10:44 http://forum.simwe.com/images/common/back.gif
2 Gauss integration points is ture but it is for the plane stress, shear stress and bending, not the thickness. I am saying the thickness not the transverse deflection w.
If you saw any paper relevant to this, would you mind upload a copy of paper? Thanks! 7# caoer
I am confused. What exactly do you mean by 必须有厚度方向积分?
I don't see why there is transverse deflection w in shell element formulation either.
In general, the shell element has three translation dofs which are global and two rotations around the vector normal to the mid-surface of the shell element which are local.
The number of integration points is only the way to get the stiffness matrix or load vector by numerical approximation.
You may think it as solid element + 2/3 rotation variables. Thanks! You may think it as solid element + 2/3 rotation variables. Thanks!
caoer 发表于 2010-11-23 10:35 http://forum.simwe.com/images/common/back.gif
Well, the most popular shell element is always considered as the degeneration of three-dimensional brick model. You still didn't answer my questions. 同样疑惑，为什么厚度方向非要有积分呢？
模拟汽车碰撞吗？MITC4单元肯定是够用了，性能要比ABAQUS中的S4R5单元要稍微好一点 本帖最后由 caoer 于 2010-11-23 11:14 编辑
To: tonnyw, yes, degenerated solid approach is typically carried out numerically, my current algorithm is based on the classical shell theory, which it the totally same to the SIMO's method around 1990, check out the paper ' on a tress resultatn geometrically exact shell model, I-VII', those are brilliant papers.
TO: pasuka. 虽然是薄壳，但是想引入核辐射作用，厚度方向对电磁波的吸收想考虑进去，
厚度方向对辐射的吸收就不是那么均匀了，
那么常规的取中面计算的薄壳单元就不好用了。
to both:
如图所示：
S/R Hughes Liu
S/R co-rotational Hughes Liu
2种似乎是我想要的，厚度有7个积分点。只是这么多年过去了，想知道有没有更好更新的算法。 anyone can download these two papers? Thanks!
1. D.J. Allman: “Evaluation of the constant strain triangle with drilling rotations,” Int. J. Numer. Meth. Eng., vol. 26, pp. 2645–2655, 1988.
2. C. Pacoste: A flat facet three node element for shell analysis—some theoretical and numerical aspects, Royal Institute of Technology, Department of Structural Engineering, Technical report 1999:20, Structural Mechanics, 1999. To: tonnyw, yes, degenerated solid approach is typically carried out numerically, my current algorithm is based on the classical shell theory, which it the totally same to the SIMO's method around 199 ...
caoer 发表于 2010-11-23 23:19 http://forum.simwe.com/images/common/back.gif
If this is the case, hillyuan pointed out the right direction. You worry about this is no integration through the thickness. Now I tell you there is integration through the thickness. MITC shell element takes care of the variation of the shell thickness. 本帖最后由 caoer 于 2010-11-23 15:07 编辑
Thanks for the reply.
The MITC is a mixed interpolation method to cancel the locking issue during bending of plate/shell. It has another advantage over the contiummn method that the interpolation function could be low order instead of 5th/7th/9th order. See the page 445 of Finite Element Procedure by Bathe and the paper by Dvorkin and Bathe in 1983.
Obviously, this algorithm could be extended to the thick shell from the thin shell, but it does not focus on the integration through-the-thickness of shell. I was looking for the integration through thickness method not the basic MITC idea.
However I have gotten a good candidate paper for this problem.
Thanks you guys show me the tips. Evaluation of the constant strain triangle with drilling rotations 271981Evaluation of the constant strain triangle with drilling rotations
nongda 发表于 2010-11-27 21:18 http://forum.simwe.com/images/common/back.gif
thanks dude, I have transfers 100 Fangzhen dollors to your account.
thanks dude, I have transfers 100 Fangzhen dollors to your account.
caoer 发表于 2010-11-28 12:34 http://forum.simwe.com/images/common/back.gif
谢谢，另一个没有找到，不知道收藏在哪个库 膜拜大牛啊，看不懂你们说的什么，看来路还很长啊 caoer 发表于 2010-11-28 12:34
thanks dude, I have transfers 100 Fangzhen dollors to your account.
你好版主，我最打算开发一个VUEL的壳单元，也看了Bathe的书和其他的一些资料，但是一直有以下几个疑问不解，希望能得到版主的指点：
1. 一般的壳单元都是3平动加2转动的5自由度单元，两个转动根据由节点法向定义的局部坐标系定义，但是ABAQUS中VUEL必须要定义6个自由度，那么这样的单元应该怎样构造呢？是否有相关的文献？
2. 这个问题其实跟第一个有关系。壳单元在节点处的法向一般是不连续的，这样的话应该如何定义不同单元的法向，因为法向不同的话转角也不同；另一方面，如果根据退化壳的原理根据壳厚度上下的节点定义法向，那么就又回到第一个问题，如何定义第三个转动自由度？还是说第一个问题中的三个转动自由度都是相对于整体坐标系来说的？
3. 如果要对壳单元中的材料定义局部材料局部坐标系应该怎么操作？能否使用定义转动用到的节点坐标系作为材料坐标系。或者另有其他的办法？
多谢指点！
页:
[1]
2