Hi,
I'm attempting to solve a pile on Winkler foundation problem which involves applying ground motion in the form of a number of displacement time histories at each spring support. This leads to 41 constraint equations.
Initially I used the plain constraint handler and got the wrong solution. On realizing my error, I changed it to Transformation Method. However, it gave me the same solution as earlier. Does the transformation method not work if there are a large number of single point constraints?
I tried Lagrange multipliers, but do not get convergence inspite of trying different algorithms and system solvers. I believe this is because of the large number of constraints in the problem. Does anybody have any idea how I can improve convergence of this system?
Finally, the penalty function method was giving the expected solution only in certain cases.
Any help on any of these issues will be greatly appreciated.
Thank you
Jayram
Nonhomogeneous single point constraints
Moderators: silvia, selimgunay, Moderators
transformation should work if no dof is constrained more than once, e.g no equalDOF and sp. we don't yet follow the path.
for the lagrange you probably have to change the convergence test, disp increments for penalty and unbalance for lagrange.
as for the penalty it should give you the correct response, if it is not it is probably because your convergence test is not tight enough. try relative tests instead of absolute norms.
for the lagrange you probably have to change the convergence test, disp increments for penalty and unbalance for lagrange.
as for the penalty it should give you the correct response, if it is not it is probably because your convergence test is not tight enough. try relative tests instead of absolute norms.