User defined 3D elements

[shenrilin]

Dear all,

I'm now programming my own 3D element, and using single element to test its  ability. Unfortunately, I came across the following problem:

(1) Model description:
single element with one surface is fixed in x1 direction, and applying tension load on the opposite surface. (see attached IUEX input file and solve.UEX )

(2) Problem
The strain are excessively large at the beginning. as shown in figure 1 and figure2.  "time" denotes current time.
 
(3) Part of codes

Code: [Select]

do j = 1,nel
            strn(1) = strn(1) + ul(1,j,1)*shp(1,j)
            strn(2) = strn(2) + ul(2,j,1)*shp(2,j)
            strn(3) = strn(3) + ul(3,j,1)*shp(3,j) 
            strn(4) = strn(4) + 0.5d0*(ul(2,j,1)*shp(1,j)
     &           + ul(1,j,1)*shp(2,j))    ! exy
            strn(5) = strn(5) + 0.5d0*(ul(3,j,1)*shp(2,j)
     &           + ul(2,j,1)*shp(3,j))    ! eyz
            strn(6) = strn(6) + 0.5d0*(ul(3,j,1)*shp(1,j)
     &           + ul(1,j,1)*shp(3,j))    ! ezx
         end do 

Thank all, I appreciate any help.

Best

[Prof. R.L. Taylor]

You should check your mesh using a built in feap element for elastic deformations only so you can be sure you have all the boundary conditions imposed correctly.  Same for your element (how did you get a USER 56?  Standard feap limits to 50).  I do not see any b.c. for the 3-direction???   Be sure to test your element incrementally -- i.e., elastic first, then gradually add your features.

You also need to be sure that using tensor strains for shear you have correct shear behavior -- matrix form usually requires engineering shear (2 eps_xy, etc.).  See the second volume of our FEA book (Zienkiewicz, Taylor & Fox, Elsevier 2013) for details on mapping from tensor (9 components) top engineering (6 components) of stress and strain.

[shenrilin]

Dear Professor,

Thank you for your reply.

I did use a built-in element to test the input file and I am sure that the boundary condition, as well as load, is applied correctly. My senior colleague had increased the limits to 100. I only fixed the surface in y-z plane with x = 0 and applied loads on the surface also on y-z plane with x = 0.1.

In the codes, I choose tensor strain and corresponding formulation to update the stress.

I had double-checked the code but didn't find any bugs. So it's very strange to see strain diverges so quickly. Do you have any suggestions about this error?

Thank you very much


Best,

[Prof. R.L. Taylor]

Please use the solution command CHECk and report the results so we can assess the boundary conditions.