My first recommendation is to review the write-up in the CAESAR II Technical Reference Manual on the Friction Solution. This will give you an understanding of how the solution works, and what the terms mean.

When approaching a friction problem I suggest two models, one with friction, and one without. The model without friction can be used to verify everything falls out the way you think it should, and you have no gross modeling mistakes. The second model is where you add friction. Two things to note here: (1) friction can only be used to hurt you, you can never count on it to help, (2) be careful of your "mu" value - it will likely vary over the life of your piping system.

Now, how do you go about solving a convergence problem?

<font color="#0000ff">First:</font>
Use the [F2] option to single step through the restraints to deterine what isn't converging. I admit the resulting display is somewhat cryptic, but ...

- The three numeric values (such as 0., 1., 0.) are the direction cosines of the restraint. This is an example of a "Y" direction restraint.
- Old state "open", new state "closed", or visa versa indicates the pipe is bouncing, such as on a "+Y" or in a gap.

- "err = xxx" is the percentage change in the normal force on a friction restraint.

<font color="#0000ff">Second:</font>
Identify the nodes with problems. Specifically target gaps and friction. Extremely small gaps can't be built and should be either removed or enlarged. Friction issues take a little more effort...

<font color="#0000ff">Third:</font>
Rather than altering one restraint at a time, it is often better to adjust the friction parameters for the entire job. This is done via the "Configuration Module", on the "Computation Controls" tab.

Friction Stiffness: The default value is 1,000,000 lb/in. Reduce this by half and try the run again. Continue reducing this by half, but I wouldn't go much below 25,000 lb/in, it makes the system too sloppy.

Normal Force Variation: Increase this in steps of 5%. Remember, the larger this value, the more error in your solution you allow.

Angle Variation: Increase this in steps of 5 deg. Remember, the larger this value, the more error in your solution you allow.

<font color="#0000ff">Fourth:</font>
Remember (from the documentation) that there are multiple equilibrium solutions to a friction problem. Review your results and ensure they make sense. Are they in-line with your "no friction" run? What happens if you change "mu"?

Friction problems are not trivial and require a great deal of thought. I hope this information helps.