twisted beam

[Irpanni]

Dear All,

I am apply user element for shell element MITC24 in FEAP 7.1 and FEAP 8.3.
In migrating from feap 7.1 to Feap 8.3, I just change the code in definition of material (Inpt2d become Inmate) and loading intensity (d10 become d8).

For case   Parabolic-Hyperbolic, Spherical Dome, Cylindrical Shell Roof, Circular Cylindrical Shell, Both of Feap 7.1 and Feap 8.3 with the same user element (MITC24) definitely give the same result although in ouput file in Feap 8.3 is always show;

 >Element        0 has zero or negative jacobian at nodes:
     Local =  1 Global =        1 Jacobian = 0.00000E+00
     Local =  2 Global =        2 Jacobian = 0.00000E+00
     Local =  3 Global =       15 Jacobian = 0.00000E+00
     Local =  4 Global =       14 Jacobian = 0.00000E+00

    Matrix: IX (CHECk may change order)   
    row/col         1         2         3         4
       1            1         2        15        14

    Matrix: XL (May be local projections)
    row/col     1           2           3           4
       1  0.0000E+00  2.1220-314  2.1220-314  0.0000E+00
       2  5.0928-313  4.9407-324  0.0000E+00  0.0000E+00
 >Element        0 Reverse numbers to fix negative jacobian
            New IX Order:        14        15         2         1

nevertheless Feap 7.1 never show the statement above.

The problem come when I apply for Twisted 2D-Beam, Feap 8.3 give wrong result compare to feap 7.1.is there anyone has solution for problem above?


Thank you very much for your kindly attention.

regards,
Herry

[Prof. R.L. Taylor]

You definitely have a problem. the print of the xl indicates that an error has occurred in computing the coordinates.  This occurs if you have pointer errors or if somehow other precision related quantities are involved.  Until you figure out what is causing this message to appear it is difficult to know what other errors may exist.

Starting in ver8.3 pointers were converted to integer*8 for 64 bit machines.  Ver 7.1 will not have this update.  Generally, if your element does not have history variables the element interface should be adequate.  However, if you use any local pointers you and you have 64bit compile you need to deal with the pointers correctly.

In future, let us know what compiler you use (version too) and if you have any compile warnings.  Compiling with warnings displayed is always best.


 

[Irpanni]

Dear Sir,

I'm using compaq visual fortran under windows xp and microsoft visual studio 2008 with Intel Visual Fortran Compiler Professional Edition 11.1.060 update 5 under windows 8. I did not find any warning during compilation. Please let me know any suitable compiler for ver 8.3.

although the print of the xl indicates the error in computing coordinates, ver83 give correct result for orther shell structures except twisted beam.

I try to debug the program  for case twisted beam. I found that numbering of individual element do not follow counter clock-wise rule (other shell structures follow the rule), although input data/mesh data I have checked correctly.

thank you very much for your kindly attention

best regards,
Herry

[Irpanni]

Dear Sir,

I have already compile with intel fortran composer XE 2013 with VS 2010 under 64 environment, but I couldn't fixed my problem and my user subroutine doesn't have history variables.

During complation there is no compile warnings,

In isw 2, if I bypass call ckisop(ix,x,shp,ndm) it will give correct result.


thank you very much for your kindly attention

[Prof. R.L. Taylor]

The "ckisop" is to check for errors in plane elements with no more than 16 nodes.  Perhaps this is not what you want?  Check the dimensions of your arrays to what are in the subprogram ckisop -- that should be where the problems are.

[Irpanni]

Dear sir,

which parameter should I change or check?

I have tried to plate problem with 9 nodes (4 elements) with my user  shell element, it's still show negative jacobian if I call ckisop

thank you for your kindly attention