abaqus中六面体单元对比四面体

        四面体网格生成算法在自动化、效率、可靠性以及几何通用性上要强于六面体网格生成算法，可是大部分工程分析人员更愿意花费更多的时间和精力使用六面体网格开展数值分析和计算，究其原因是六面体单元在数值计算中有四面体单元无法比拟的优势，主要表现在以下方面：

 图1：规则立方体区域由一个六面体（顶部）和五个四面体（底部）[1]

        六面体单元的计算精度明显要高于四面体单元。
        在实体离散过程中，如果采用相同大小的单元尺寸，六面体网格划分的单元数量要比四面体网格划分的单元数量少得多。较少的单元数量意味着较少的有限元计算时间.

        六面体网格更加接近被模拟的物理场。

尤其是在复杂多场耦合模型中，四面体单元计算的结果会严重失真。这种情况下比较好的解决办法有两个：

1、加密网格

2、将一阶精度单元修改为二阶精度

3、将四面体单元修改为六面体单元

分享论坛上的一些讨论：https://www.eng-tips.com/viewthread.cfm?qid=314206

pereiraICL (Bioengineer)

(OP)

17 Jan 12 11:13

Dear all,

I am having problems with my low-permeability model in ABAQUS using tetrahedrons C3D10MP.

A resulting pore pressure with a checker-board-ish pattern appears.

Is there a way of tackling this permeability locking?

Best,

Andre

JayTung (Computer)17 Jan 12 22:45

Hi Andre,

I encounter the same problem before. That's some kind of numerical oscillation problem. The problem could be your mesh size irregularity. So do finer the mesh and test.

Jay

stungj@gmail.com

pereiraICL (Bioengineer)

(OP)

18 Jan 12 07:55

Jay,

Thanks for your reply.

A considerably finer mesh will still exhibit those problems and I can't afford to refine it further. Using a hexahedral mesh seems to solve it, but it is not the desirable solution either.

Andre  

JayTung (Computer)19 Jan 12 03:35

What element type do you use ?

I used to use CPE4. Which did you choose ?

pereiraICL (Bioengineer)

(OP)

20 Jan 12 13:03

I used C3D10MP in the pictured model. I made a hex mesh with C3D8P elements, that apparently solve the problem. Though hex meshes are very hard to develop in organic geometries, so I would like to have a solution that involves tet elements.

Thanks

Andre

JayTung (Computer)31 Jan 12 22:55

I am going to construct a tet element for my model. Let you know if it works well.

Jay

参考：

[1]Benzley S E, Perry E, Merkley K, et al. A comparison of all hexagonal and all tetrahedral finite      element meshes for elastic and elasto-plastic analysis[C]//Proceedings, 4th international          meshing roundtable. Sandia National Laboratories Albuquerque, NM, 1995, 17: 179-191.

[2]https://www.eng-tips.com/viewthread.cfm?qid=314206