Hi,

I realize that ASME B31.3 directs engineers to use the nominal thickness when calculating SIF values per the code. My question is, when calculating SIFs using FEA would it be appropriate to use the corroded thickness in the calculation or should the nominal thickness still be used.

For reference, I can most commonly looking to determine SIFs for dummy legs and stanchions (things that are not directly covered by ASME B31.3 formulas).

Thank you

Either calculated as per ASME B31.3 App. D / B31J, or evaluated by FEA, the SIFs are used for Displacement Stress calculation purpose.

Since B31 codes require the displacement stresses to be calculated using Pipe Section Modulus established with NOMINAL (not-corroded and not affected by mill under-thickness tolerance) wall thickness, it makes sense the SIFs to be quantified in same manner - e.g. using NOMINAL wal l thickness.

It might be discussion about SSIs, which are used in conjunction with B31 Code Primary (e.g. Sustained and Occasional) stresses that are calculated with corroded wall thickness. However, as per my knowledge, common engineering practice does not explicitly require to use corroded wall thickness/sizes for SSIs evaluation, and typically the nominal section characteristics are used.

Other opinions are welcome.

Regards,

I agree with Dorin. Additionally, how would you propose to use a corroded thickness? Simply reducing the wall thickness (by increasing the inner diameter) is way over conservative - real corrosion is NOT increasing the bore of the pipe.

Hi Richard, I understand that corrosion is not increasing the bore of the pipe; but what happen with an abrassive slurry. In my experience, when you transport mineral ore or tails, the inside diamter change in time.

thanks

I agree, depending on the process corrosion can be very different. However, as far as I know, there have been no (Markl type) tests on corroded fittings. Even the testing for the current work on B31J uses uncorroded fittings.

Let's turn this around and look at this idea from a different viewpoint. Assume you did compute SIFs using a corroded (thinner) wall. Now assume that your system fails in the field (at or near the fitting) and there is a subsequent law suit. How can you defend your analysis and assumptions (since there are no published tests or justifications) for the corroded SIFs?

Take a step back. Our task is to ensure the pipe is installed to code. The code says to use WT = WT(new).

Let's say we have a pipe A. It has OD = OD. It has WT = tmin + CA + round-up to nominal pipe thickness. We use WT = tmin + CA + round-up to calculate SIFs.

Let's say we have a pipe B. It also has OD = OD. But because we're sensitive to weight on this pipe, and corrosion isn't an issue, we purchase custom piping where WT = tmin. We use WT = tmin to calculate SIFs.

I would argue that using the more conservative of the two would still be in compliance with the code, as the code is seen as "minimum requirements." However, you wouldn't be incorrect with adhering to the piping code, either.

Hi

In my process, the corrosion is less imporant than the abrassion (copper ore, copper tails) so I use the corrosion thickness as abrasion thickness. We calculate the abrasion rate according to our experience and calculate the minium thickness accoding the the usefull life of the pipe.

I uderstand that the stress calculus use the thinner thickness (corroed) because is the more defavorable case, and for calculus of the load on restraints use the total thickness because have more weight. Please let me know if I in a mistake.

By other hand, how can I get the stress in the begining and at the end of the usefull life?

Separating abrasion and corrosion is moot, in my opinion, so when we say "corroded," I would take it to mean "abraded/abrased" as applicable.

If you want initial stresses, set the corrosion allowance to 0 for the model.

For end of life, use initial thickness with corrosion allowance.

Yes, it uses the strength of the thinner material and the weight of the thicker material. We're not doing this because it's "unfavorable" and because it's "conservative."

We're doing this because we cannot predict with certainty that any given point in the system will be the first to corrode and everywhere else is as thick as it was installed.

If you're measuring abrasion through the life of the pipe, there's nothing stopping you from creating a "current day stresses" model with present day thicknesses and 0 corrosion allowance.

There's also nothing stopping you from creating "next year's projected piping at present rates."

And the year after.

And the year after.

It'll be up to regular inspection to verify your predicted rates.

It'll be up to you... or your successor... to adjust the analysis accordingly as new information makes itself known.

With respect to SIFs on fittings, if you want to get the most accurate handle on stresses, put CNODED anchors on the fitting welds, and model it in FEA, again, attempting to predict the future based on the info you have now.

Thank you Michael for your answer!!