Newton-Raphson with initial stiffness

Forum for OpenSees users to post questions, comments, etc. on the use of the OpenSees interpreter, OpenSees.exe

Moderators: silvia, selimgunay, Moderators

Post Reply
vpapanik
Posts: 36
Joined: Tue Jun 15, 2010 12:53 am
Location: Thessaloniki
Contact:

Newton-Raphson with initial stiffness

Post by vpapanik » Thu Nov 16, 2017 1:44 pm

Is there any serious drawback when using a NR algorithm with initial stiffness for a dynamic analysis, compared to using NR and revert to NR w/ initial stiffness only upon non-convergence ? It seems that NR w/ initial stiffness always converges in my model, while NR does not.

Is the tangent stiffness matrix updated at the end of each iteration when using NR w/ initial stiffness ? (otherwise it would be exactly the same with modified NR w/ initial stiffness I assume)

Thanks a lot in advance !

fmk
Site Admin
Posts: 5883
Joined: Fri Jun 11, 2004 2:33 pm
Location: UC Berkeley
Contact:

Re: Newton-Raphson with initial stiffness

Post by fmk » Mon Nov 27, 2017 6:11 am

the main drawback is time. from a purely numerical point of view there are situations where it would results in a different answer being found for a given step (you could draw some curves on a sheet of paper and find cases where the two methods will give a different result), you will probably find these do not happen for the typical model.

Post Reply