[Libmesh-users] system matrix

[li p. <li...@ya...>]

Hello all,
I have been worked with elastic problem for a long
time. My main work is to implement a non linear
elastic problem for an organ, for which I need a
energy function. Well, after several months fighting
with papers and dissertations of other researcher I
found one. I put this in the attachment because of the
expression in email. As I tried to test it with some
data, I can't get reasonable results. I implemented
this problem as quasi-static with incremental
solution. At each timestep system matrix and rhs will
be calculated with current displacement. Since it's a
non linear problem, I applied a Newton-method. But the
convergence was too bad. 
Roy told me last time. There could be some thing wrong
with system matrix calculation. I checked it. Didn't
found anything. My question now is,
if the algorithm was implemented well, are there any
other reasons? For example, geometrical form, force,
element, anything.

best

pan


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com