Could the staff at COADE list out the Load Cases wherein Corroded thickness is used for analysis by the software for B 31.3, B 31.4 & B 31.8 as default.

Is it as per their respective codes?

I believe that by default the software considers uncorroded thickness for sustained and Occasional cases as the load case will be conservative due to weight etc.

Also by default corroded thickness is considered by the software for expansion load case only as this is conservative.

Do confirm on the above.

Another query is whether +12.5% mill tolerence is actually used in CAESAR-II stress analysis Calculations (Operating, Sustained, Expansion and Occasional checks)?

12.5% Mill tolerence is used in min. wall thickness computation only. It is not used in flexibilty analysis.

Corrosion in CAESAR II is handled on a "code by code basis". The software uses a corroded section modulus for the stress calculations only - in whatever load cases the active piping code says to do so. If the code doesn't say to corrode a particular load case, CAESAR II won't assume that case is corroded.

If you want to change the default behavior of the software with respect to corrosion, you can do so by checking the All Cases Corroded checkbox in the Configuration Module. This switch is on the SIF and Tees tab.

Dear Richard,

we run a B31.4 line with the option "All Case Corroded" as "False". The system passed.

Then we run the same system enabling the "All Case Corroded" as "True" and the system failed miserably in operating, expansion and sustained load cases.

Considering your words, my conclusion is that B31.4 does not consider a reduced section modulus due to less thickness in normal condition ("All Case Corroded" as False). Is my understanding correct?

If I'm right, the option must be always enabled if we are designing the pipeline for 30 years.

How does it work for B31.8?

Please advise.

Thanks in advance

For flexibility calculations:
B31.4 paragraph 402.1 says "Nominal dimensions of pipe and fittings shall be used in flexibility calculations."
B31.8 paragraph 832.3(e) says "Properties of pipe and fittings for these calculations shall be based on nominal dimensions, and the joint factor E shall be taken as 1.00."

Both B31.4 and B31.8 state that flexibility calculations are to be performed using nominal dimensions. The stress equations use the term Z, defined as the section modulus. Corroded cross sections are not mentioned.

Therefore if you want to consider corrosion in your stress calculation, you must set "All Cases Corroded" to "True" in the configuration.

Richard, Dave, thanks for your replies.

I did a little test to understand how the software works.

1)First Run

- "All Case Corroded" enabled (true)
- 6 mm corrosion allowance defined in Caesar as per Project Design Basis

2) Second Run

- "All Case Corroded" disabled (false)
- 0 mm corrosion allowance defined in Caesar.
- Deducted 6 mm to all WT in Caesar.

I was expecting to get exactly the same results but in the end the first run turned out to be much worse from a stress point of view.

Anybody can explain?

In CAESAR II - corrosion does not affect system stiffness (flexibility is based on nominal dimensions) and corrosion does not affect load (e.g., weight based on volume times density).
In CAESAR II , when corrosion is included - the denominator in the stress calculation (i.e., the A in F/A and the Z in M/Z) is reduced due to corrosion.
So your first run is not identical to your second run as your second run has reduced stiffness and reduced load.

When the piping codes address corrosion, they state "corrosion is removed from the section modulus to make the stress calculation".

Deducting corrosion from the wall thickness in non-conservative because:
a) with a smaller wall thickness the system is more flexible, [K] is reduced.
b) with a smaller wall thickness the system weight is reduced, {F} is reduced.

Both alterations are non-conservative, i.e. "wrong".

OK, it's now clear to me that when enabling "All Case Corroded" for B31.4 & for B31.8 the ONLY parameters that are affected are the denominators in the stress calculation (A and Z).

Few more considerations:

- I suppose that for hoop stress calculation Caesar considers the thickness without the allowances as outlined in both codes (Clause 402.3 in B31.4 & Clause 805.2.3 in B31.8).

- I wonder why the ASME committee decided to not take into account corrosion for the flexibility calculation. Corrosion happens, and it's the most likely reason of pipeline failure, then why neglect it?

- My system, which was passing with a stress ratio around 90%, is now failing miserably when enabling the "All Case Corroded" option (I'm talking about 240%!). True, I have 6 mm CA; true, I have high pressure and temperature (92 barg and 110 deg); true, I have long straight legs of pipe and failure is occurring at bends. It's all true but...from 90% to 240%? Needless to say that if this is a correct figure, my design will completely change. How much this result can be seen as realistic?

- When enabling "All Case Corroded", also Hydrostatic Load Case is failing. Should I neglect this result?

Thanks again for your support

I can't speak to the decisions made by the various Code Committees. If the hydrotest case fails, and you know this will happen ONLY on a new system, set it up as a duplicate model without corrosion.

OK, maybe I'm wrong but looks like the two codes have different definition for "nominal wall thickness".

B31.4 define the nominal thickness as the calculated thickness plus the allowances (para. 403.2.1).

B31.8 define the nominal thickness as "the wall thickness computed by or used in the design equation in para. 841.1.1 [..]". Going to para. 841.1.1 I cannot find any single reference to allowances. This make me think that nominal thickness, as per B31.8, does not consider corrosion allowance.

if my understanding is correct, Caesar should consider by default (i.e., "All Case Corroded" set as False) the corroded thickness when running the analysis with B31.8.

One more thought about one of the replies of this discussion.



I agree it is less conservative, but isn't the "All Case Corroded" scenario too much conservative? I explain better: if corrosion will reduce my section (thus reducing A and Z), also weight should decrease, hence also the numerator F and M. I think my second run is less conservative to the first (and this is certain) but closer to the real situation (and this is more a personal opinion, debatable indeed).

Thanks for this discussion.

I think it's fair realistic to consider that the corrosion is more aggressive in some sections or elements than in others. That's why it's conservative to remove CA from the section modulus to make the stress calculation but not in other parts of calculation.

Resurrecting this thread from the grave.

I would expect that "all cases corroded" would have CAESAR assume the pipe has the strength of corroded pipe, but the stiffness and weight of uncorroded pipe.

Running a basic test with the option enabled/disabled, I see SIFs, flexibility characteristics, and pipe weight remain the same (CoG report).

I know that if we included corroded SIFs, we'd have a lot harder of a time getting things to pass stress. In my experience, we'd see a lot of 50% increases in stresses.

Thus, all cases corroded doesn't flatly increase all attributes, but should conventionally increase stresses across the board.

Is this correct?

Michael,

Correct. The configuration directive "ALL_CASES_CORRODED" simply means that the section modulus and cross-sectional area used to compute the bending and axial stresses are corroded.

Why do the piping codes recommend that corrosion is handled in this way?

First, corrosion physically is a helter-skelter affair where local areas and spots of the pipe thin. Corrosion is not a uniform boring of the interior of the pipe. The problem is you don't know where these corroded spots are. To reduce the pipe weight by the corrosion allowance would be non-conservative because you would be reducing the load.

Second, to my knowledge, there have been no SIF tests on corroded pipe. Even if there were such tests, your pipe would corrode differently than the specimens used in the tests. So SIFs and K-factors are based on uncorroded pipe.

Third, due to the unknowns of the corroded areas (location and depth), the flexibility analysis is performed with the uncorroded cross section. This means the solution of [K]{x} = {f}, which is {x} is the same whether you include corrosion or not. The only difference is in the final stresses. All other results (displacements, forces, moments, and restraint loads) will be the same. This is because [K] and {f} are (assumed to be) not affected by corrosion

Thanks, Richard. This re-confirms what I've been soapboxing.

Mill Tolerance isn't treated in the same way, though. It affects permissible pressure, but largely doesn't impact CAESAR's performance.

The logic behind that being manufacturing is controlled and losing MT in large regions of piping is unheard of, whereas corrosion is generally uncontrolled?

Yes, I agree.