Hello. Maybe there is already excessive deformation in the pipe (plastic state) and behaving non-linearly and at this state it's have to capture using a beam element software. Just assuming.

If it's caused by primary based load (then there is a tendency not to redistribute stress). But elastic-plastic analysis using FEA may capture the effect after yield, and since the piping is still intact, you would expect the analysis to converge, but low cycle fatigue will complicate if present in the system (if the system will experience it over and over again). This can be a good candidate for fitness for service analysis which I haven't tried before (determining how many cycles the system will endure).

Warm Regards,