Convergence problem is due to system non linearity and need not be due to friction alone.In your question, you have highlighted about the convergence problem , but is it solely due to friction ? When the iterative processs goes on to solve {P} = {K}{X}, you can see by using F2 why there is non convergence. You may get message like "old state open, new state closed ", this obviously indicates that one directional support like +Y or support with gap can be the cause. If only you get messages like "old state sliding, new state non sliding ", the problem is with friction.However for different node points you may get different message.
Now coming to your question on Friction stiffness, we have to admit that using numerical techniques, we have to resort to approximation.When a pipe starts sliding ( a change from static friction to kinetic friction ) the laws of static equilibrium do not hold good , so how are we solving a staic analysis problem in that situation? The answer is engineering approximation. I fully agree that a change in friction tolerance can make a system " pass or fail".But we have realise that Friction itself is indeterministic, say at the same location you need not have the same friction coefficient all along the thermal history, it will vary depending on surface condition ( but not on surface area) and the kinetic friction coefficient is function of velocity also.So using 0.3 does not mean it is 0.3 all along or along all the points.
So the entire process is not which will simulate the exact friction history that the system undergoes. CAESAR or any pipe stress software attempts to address this stocastic problem using the method originally suggested in Sobieczanski's paper.Besides CAESAR II allows the user to try with not just one but several options of friction stiffness.The objective of friction analysis should be to get a feel of the frictional effects on the sytem, but it need not be taken as a 100 % representation of the actual friction history.
What I will recommend you to do is run a file with friction and one without friction and compare the results. You may find that your system was having advantage due to friction !
Anindya Bhattacharya
Stress Analyst
Bechtel Corporation
_________________________
Anindya Bhattacharya