Hello,

Which of below is a proper approach to resolve iteration problem of calculation

a>change friction stiffness value at configuration set up and re-run analysis

Or

b>Remove support friction at lifting support and re-run analysis

I prefer the latter as it is specific to the single restraint. You can review results and evaluate the significance of the change. That is, if the restraint is active, your "error" is mu*N - if that's small, then don't worry about it.
Of course, changing just one restraint may be insufficient to cause convergence.

Thanks Dave.

But if the analysis involves high temperature and large diameter piping then result without considering friction at few support may be not acceptable.
How to deal with such situation ?

Could you please explain what is exactly friction stiffnes ? and its's significance in analysis.

Thanks in advance

---Durgesh

If I have eliminated friction at just a few supports in order to converge, and if I think this friction is significant, I would reanalyze specific load cases where I model the friction force by hand or lock the pipe from sliding by my own (entered) perpendicular restraints.

Friction stiffness is the stiffness applied to those (artificial) restraints that the program uses to model the "stickiness" of a support.

If the internal system load perpendicular to the restraint does not exceed the friction load (mu*N), the pipe should stick rather than slip. The program adds two restraints mutually perpendicular to that normal load to prevent this sliding and these restraints have a stiffness as set in the configuration file - this friction stiffness.

Why would this friction stiffness be different than the default stiffness assigned to rigid restraints?

Let's assume you have a series of Y supports with friction along a straight run and this run has some load applied at one end. Assume, too, that the pipe does not slip so friction stiffness is applied in the X & Z directions at all friction supports. In the analysis, that large load will be carried by the closest friction restraint. If mu*N is less than the applied load, the pipe should slip and CAESAR II will replace those friction restraints at that node with mu*N against the slide. The next iteration will do the same at the next friction point down the line. High friction stiffness will isolate the changes to only the next friction support in the line. A lower friction stiffness will allow some of this applied load to "leak" down the line an reduce the number of iterations to a consistent (final) solution. A lower friction stiffness reduces iteration.

However...

Several years ago, when PCs were much slower and iteration became tedious, we shipped a version of CAESAR II with a lower value for this stiffness - 50,000 lbf/in, I believe. Iterations were reduced but we also saw other response that wasn't reasonable. We traced this unusual response to the lower friction stiffness. We returned to the higher default value for friction stiffness in the following release.

I wouldn't change friction stiffness. But that's my choice.

Thank you for explanation.

Dave / Durgesh,

Could you give a an explanation as to why the friction causes the non convergence? I am having this same issue and I am rmoving the friction from every third restraint. It is a shot in the dark.

Could you give a an explanation as to why the friction causes the non convergence?

Thanks,
- Jonathan

Two or more nonlinear conditions on restraint interact with one another in such a way that a they, together, do not provide a correct solution.

Thanks Dave. Sounds like a black box issue. I have your replies from other post on ways to troubleshoot the problem. I appreciate your answer.