I made the nonlinear FEM

There are many times that the derivation with respect to degrees of freedom can be trickier that it looks.
I was trying to find the residual and tangent matrix for my isotropic finite element code. I finally did the thing.
I had the problem with derivation since I forgot to take derivation of shape function tensor with respect to degrees of freedom.
I finally did it.

I was working on the Energy Conversion Efficiency of PZT last night. I fixed it later that the due date for Poster. I could not present a poster.

I worked for an hour on PZT material propertis and find out that the best was the one that is offered by Engineering Application of Smart Materials.

I was working for an hour on finite deformation in 2D analysis. I think I will make it work by Monday with Updated Lagrangian and Total Lagrangian. The applicable thing about 2D analysis is that I already have some published result in a book and I can compare my result to them.
I spend large time on fixing my Journal Paper. I will finish it by Monday.

I finally made the LAPACK work on Visual Studio 2008

It was a success. Something that took my time and attention for some time. I had trouble using the LAPACK solver function in FORTRAN 95. I could made that thing work on gfortran gnu. But, for Intel fortran compiler ifort.exe under Visual Studio, I was totally lost.

I made it work first by downloading the given example
http://icl.cs.utk.edu/lapack-for-windows/lapack/LAPACK-VS-Example.zip

Then I downloaded and placed the library prebuild files In the solution library:
http://icl.cs.utk.edu/lapack-for-windows/lapack/#libraries

There was a trick in compiling. I was getting an error in compilation for a file that one library conflict with some others. So I ask the linker in Project property to ignore the file and does the compiling without it.
It was in
Project>XYX Propertis>linker> Ignore Specific Library

It is working now.
I got the example for dgesv function from the
http://faculty.washington.edu/rjl/uwamath583s11/sphinx/notes/html/lapack_examples.html

It was noce to be able to work this package at the end.

LAPACK in windows

I just find a way to use linear solver by LAPACK.
It was not so easy and I am still working on it.

First, I need to install gfortran from

http://gcc.gnu.org/wiki/GFortranBinaries#Windows

Then I downloaded the LAPACK from

http://www.netlib.org/lapack/

Then I just extracted the filed into a library.
after that I download some bath files from

http://gcc.gnu.org/wiki/LAPACK%20on%20Windows

and extracted them into the LAPACK directory.
It worked after I run the batch file:

makelibs.bat

Then I just used a command like this
gfortran -o mypgm mypgm.f90 lapack.a blas.a

to compile my fortran file.
I had my  lapack.a and blas.a copied into the compile directory.

It worked fine.

I have to make this thing work with Eclipse.

I just made the Presiach model to work. Interesting fact about this model is its consistency in definitions. It was working fine for the discrete model. Problem right now will be to make is work for integral definition. I guess the core function of integral in Presiach plate should be something interesting.

This model was not easy to learn. It was something very obscure for more. I just follow the Wikipedia and its links to make it work. I figure that the knowledge is free in these days. But, to do the job it is not only the knowledge that is important. It is the experience. And, no one is going to ask you to do something unless there is problem.

Recursive Modified

I just modified my recursive method for Visco Electro Elasticity. I presented the history variable with respect to the strain and the decay time constant. I saw that the material constants are factor able from the recursive variable. So I factored it. It looks better that before. Not to mention that it took almost 8 hours to modify the equations and implement it into the code.

MAPLE can do it better

It is nice. Now I can effectively use maple to formulate my coefficinet matrix for FEM.
It took me less than few hours to modify a 2D code that I have written for elasticity for Piezoelectric coupling. I used maple and just add another degree of freedom to FEM code and it was done. The Next step will be writing the code for polar coordinate. This will be tricky since obviously in polar coordinate the coordinate variables enter the K matrix. I wish myself luck.
I was thinking to write the code using UEL for ABAQUS too. I guess it will be a very idea since I will learn the UEL too. But, I am planning to finish what I am doing now ASAP and write a report for it.