Poll question:

When calculating SIFs for bends on elbows utilizing FE software, do you consider corrosion allowance? Please explain your rationale.

Thanks.

I don't. There could be a better reason for this, but personally, I think Markl tests were not performed based on corroded condition. And no adjustment of the result were made to include corrosion. The SIF calculation methods of B31 codes also do not take into consideration corroded pipe dimensions. It's only considered in the section modulus calculation.

But, when it comes to calculating stresses using FEM, I do consider corrosion allowance. In that case SIF is immaterial of course.

Well, B31.3 "asks" me to consider a number of cycles that may be, for the most cases in practice, equivalent to one thermal cycle (between the metal temperatures I defined) per day, during a system life of 20 years. I don't think make sense to consider accurate SIFs (and accurate peak stress) calculated considering CA consumed in the first day and having no corrosion in the rest of 6999 cycles. In addition (not related to your question) I don't think it's worth to improve SIF values by FEA and use them in calculation together with conservative values for temperature, defining a range that would appear maybe three times in 20 years.

In addition, most Piping Codes state that in a flexibility analysis nominal dimensions are to be used. Corrosion is to be removed from the section modulus before making the stress calculations.