When CAESAR II is calculating branch stresses at an intersection point in a model as per equation B31.3 319.4 (20), how is the corrosion allowance incorporated into the calculation of the effective section modulus?

I also note that as I change the corrosion allowance the effective branch wall thickness value as reported in the misc report does not change.

Generally in the commercial piping stress software, the sectional properties(moment inertia and sSectional modulus) for calculating flexibility and moment do not include corrosion(C.A)or mechnical allowance(M.T). However, C.A and M.T are considered to calculate stress level only, that is, these allowances are deducted from nominal pipe wall thickness in stress equations.

Sun Wee is correct. Also be aware that the various codes treat corrosion differently depending on the load case. B31.3 for example, considers corrosion in the SUS and OCC load cases, but not the EXP load case. If you want the EXP case corroded, you have to change the configuration and set the "ALL_STRESS_CASES_CORRODED" flag (on the "SIF & Stress" tab).

The effective section modulus, Ze, is used in calculating branch stress for reducing outlet branch connections. Equation (21) shows Ze = pi*r2^2*TS where pi is pi, r2 is the mean radius of the branch, and TS is a function of the nominal thickness of the branch or header.

Corrosion et. al. (c) will not affect flexibility or load in CAESAR II. Corrosion will adjust stress as stated previously. In keeping with that concept, in CAESAR II, this TS will be reduced by the entered corrosion allowance. Or, Ze = pi*r2^2*(TS-c).

Thanks Dave, exactly what I wanted to know.

I have been looking at some calcs where the corrosion allowance is incorporated into the effective Z calculation in the following manner:

Ts = lesser of (nominal thickness of header - C, or i*nominal thickness of branch - C)

and r2 = mean branch C.S.A - corrosion allowance

which gives a much more conservative stress value than the method CAESAR II uses.

But looking at appendix J in B31.3, it is clear the nominal thickness should be used for the terms in calculating the effective section modulus, and hence I don't believe the corrosion allowance should be subtracted in the above manner.

However I also find it hard to get my head around taking the corrosion allowance off the Ts term (ie Ts - C) as the i*tb term is not a 'real' thickness, hence it does not seem correct to take the corrosion allowance off this value either.

I’m thinking about approaching the B31.3 committee for clarification on this one...

A simple question, have you read through the code thoroughly? If you send us a question which is clearly apparent in the code you may not like the non answer you get from the comittee.

Read appdx z in preparing your RFI....

I am unsure what is unclear here, read 319.5.3 for flexibilty calcs and para 302.3.5 for sustained loads. Seems clear enough to me.

Thanks John - any approach to the committee for clarification would obviously be after extensive internal brainstorming sessions within the company I work for.

I am actually comparing two pipe stressing packages, and the other package gives us higher code stresses that CAESAR II, for an identical model to B31.3.

I have established that the difference in results is attributed to the way the corrosion allowance is considered when calculating the effective section modulus, hence my query.

Yes B31.3 302.3.5(c) states that the thickness of the pipe used to calculate Sl shall be the nominal thickness minus mechanical, corrosion and errosion allowance, but when using formula 319.4.4 (21) and hence calculating the value of Ts, according to appendix J the wall thickness to be used for both the branch and the header are also nominal values, hence it is not clear to me how the corrosion allowance is to be considered in this instance. I can not see anything in B31.3 that tells me that Ts-c is the way to apply the corrosion allowance in this instance.

CAESAR II determines what value "t" (either Th or iiTb) should be used to compute Ze, using nominal values during the Error Check phase of the analysis.

Then, for each load case, Ze is computed using "Ts = t - corr", as necessary. The "as necessary" phrase goes back to what I said above, SUS and OCC will be corroded, while EXP will not be.

I believe your confusion is software and not code related. You get 2 different answers from 2 different programs!


Try a simple model in both programs which you have done a hand calculation for and see which program gets it correct.

This should be easy enough to do. So code or software deal or no deal?

See para. 302.3.5c which gives an explanation on how allowances are normally treated in calculating stresses due to sustained loads. Also, as you stated, the equation for effective section modulus requires use of nominal pipe wall thickness, so that is what is required to be used in that instance.

Thanks every one.

Yes John, I had already built a basic model (with SUS only load) in both programmes and then carried out the manual calcs as well. CAESAR II and the other programme give quite different results when the model is run to B31.3 with the corrosion allowance, (although if I change the wall thickness in my CAESAR II model to the corroded thickness and then have 0 corrosion allowance then the results do agree.)

I know that the corrosion allowance is to be considered for B31.3 SUS stresses and I understand how both programmes calculate the Zeff value, so in the end it looks like it is coming down to interpretation of the code.

Well if code case 178 becomes part of the book maybe things will clear up....

Sounds like other program applies corrosion to r2 as well as TS in Ze. CAESAR II leaves r2 alone as it is defined as nominal.

You're right Dave, they do apply corrosion to r2, but they also apply it to the Th or Tb terms before calculating Ts as well, which again in the code are defined as nominal (resulting in a very conservsative value of stress).

Well after all this some questions come to my mind Ms Itchy...

1)Chuck Bechts post above seems clear as to what B31.3 requires, seeing as how Chuck is an author of a reference book on the code, and is currently the comittee chair do you agree that B31.3 is clear on this matter?

2)If your answer to number 1 is yes then have you asked the other software vendor why they use the corroded properties?

3)Do they intend to modify their software to use the nominal wall properties as discussed here?

Just wondering what the rest of the story is????

Hi John,

Yes it is clear to me that B31.3 requires the use of the corroded thickness for a sustained load case (and yes I have Mr Becht's book on B31.3 on my desk).

But in my personal opinion (womans logic maybe?) it is still not clear to me if I follow the code religiously where the corrosion allowance should be applied - I do not disagree with how CAESAR II applies the factor, to me it is the logical way to do so. But the code has been vague on other issues in the past, ie code case 178.

Yes I will be approaching the other software vendor for clarification on why they use corroded properties. I will update this post once I have their comments and finish the story.

Ms Itchy any update on this???