hi all,
i want to know about the reason of non convergence specially in case of spring hangers. actually i m facing a problem with one main steam line. actually one spring hanger is located just after 3 meter vertical rise. when i done static analysis its showing nonconvergence error, but when i shifted it 250mm from its earlier position then its got converged. i just want to know reason behind this. actually i know all convergence and non convergence concepts for friction modelling ans analysis, but in case of spring hanger please tell me the reason of non convergence

Hello Raju,

PLEASE! You may use the search function provided by this board to search for previous discussions of your question. Look at the top of the page and find the "Post a Poll" button. Just below it you will see the "search" hypertext. If you "click" on the "search" hypertext, it will take you to the page that will allow you to search previous postings by "keyword".

When I searched using the keyword "convergence", nine previous threads were found. One of those threads was this one:

http://www.coade.com/cgi-local/ultimatebb.cgi?ubb=get_topic;f=1;t=000299#000000

Also look at this:

http://www.coade.com/cgi-local/ultimatebb.cgi?ubb=get_topic;f=1;t=000119#000000

There are several things related to restraints that can introduce non-linearities to the solution. These are addressed by CAESAR II iteratively. Note that the inclusion of friction in your restraint description is "one of the usual suspects" (if you describe a "Mu" friction for a restraint that is lifting off (not supporting) you will introduce a possible non-convergence issue). Also, if you describe restraints as "single acting" types (using for example +Y or -Y instead of (double acting) "Y" you will compel CAESAR II to iterate. This might also be the source of a non-convergence issue under some circumstances. As you can imagine, you are not the first person to have a convergence problem with an analysis model. Always try the “search” option first and experiment with the advice provided to/by the community. THEN, if you still have questions post them here.

Regards, John.

Raju,

The spring hanger in and of itself is not the cause of your non-convergence. Non-convergence is ALWAYS caused by nonlinear boundary conditions such as uni-directional restraints, friction, or large-rotation rods. Having said that, changing a hanger location will change the load distribution in the piping system such that a nonlinear boundary condition elsewhere in the model is affected.

In a nutshell, nonlinear boundary conditions require Caesar to alter the boundary condition between iterations. Sometimes it gets stuck in an infinite loop because it changes the boundary condition back and forth on successive iterations. Of course this situation is exacerbated by large numbers of nonlinear boundary conditions. Sometimes as innocuous a change as a 5 degree temperature increase will make a previously converging model become non-convergent. This is simply the nature of the numerical methods employed and does not reflect in any way on the design of the piping system.