B31.3 App D Note (1)states "Stress intensification and flexibility factor data in Table D300 are for use in the absence of more directly applicable data (see para. 319.3.6). Their validity has been demonstrated for D/ T ≤ 100."

Now, even with the avaiability of FEATOOLS at their disposal, if any engineering A/E firm considers App-D SIF & k values in a non-conservative manner as FESIF gives higher SIF than App D values, is the intent of note 1 above is not violated ? If so, is the responsibility of the safety of the underdesigned hardware falling on the A/E firm ?

I wish to get the opinion of our elders here as in B31.3 jurisdiction we are working with thin cryogenic piping of large diameter of appreciably high pressure.

regards,
sam

As no one reacted on this issue so far, I am adding a rejoinder here.

In 1981, at the start of my engineering service, 500 MWe steam generator units were introduced in India and we were used to analyse steam hammer loading in main steam, hot & cold reheat steam piping connected with 500 MWe steam turbine by equivalent static method only. When force spectrum method of piping stress was first introduced for that project to economise on number of snubbers, most engineers were enthusiast enough!

Similarly, people in industry are quite happy to use FESIF or NozzlePro to qualify a equipment nozzle load rejected by equipment vendor using WRC-107/297 method to avoid a change in layout.

But, the same people justify FESIF & NozzlePro as unreliable FEA, when d/D near to 0.5 & D/T > 50 combination in some flexible locations ask for SIFs much higher than predicted by experimentally derived old & reliable B31.3 App-D values! From a simulator, we need to hear the good news only, not the truth!

regards,
sam

I believe this is an excellent discussion because it draws attention to consideration of what is the right answer, and to who is responsible for the right answer.

B31.3-2012 Appendix D Note 11 states that out-of-plane SIFS for certain reducing branch connections may be non-conservative and that the selection of the appropriate SIF is the designers responsibility. It was an awareness of this condition that prompted the writing of WRC 329 which uncovered a number of other both conservative and non-conservative issues with the B31.1, B31.3 and Section III piping codes. The Section III Codes have long since implemented the recommendations of WRC 329 but the B31.3 Code to this date (4/16/2014) has not. (One reason for this is that the Section III implementation was clearly spelled out in WRC 329 while B31.1 and B31.3 changes were not.)

One of the largest errors in the then current (1987) version of B31.3 cited by WRC 329 was that torsional SIFs of 1.0 were used in the B31.3 stress calculations for branch connections. This omission exists in the 2012 version of B31.3 with the exception that now a place holder for the torsional SIF in B31.3 (it) is given in para 319.4.4, which states that the torsional SIF should be equal to 1.0 in the absence of more applicable data. Many users of B31.3 in all pipe stress programs ignore the directive to use more applicable data and apply the default of 1.0, which is known to be off by 5 times or more as a function of D/T, d/D and the branch connection type.

The claim is that, “nothing has failed.”

It can easily be shown that failures could have occurred in practical situations and that even if failures have not occurred, the intended safety factor is not provided.

So then the question becomes: “Because we don’t think failures have happened to us yet, can we continue a behavior that a prudent man would likely not approve upon review?” I think we all know that answer. Since SIFs affect the safety and integrity of piping systems, and errors in some SIFs can be quite large, if we are going to continue practicing pipe stress, should we ignore those errors today, just like we ignored them yesterday? Most of us would consider letting a surgeon using 1970’s technology to operate on us if the technology hasn’t changed. If the technology was radically different however, we would likely have second thoughts. The same should certainly be true with pipe stress, especially since process, welding, fabrication and manufacturing technologies have changed so significantly since the 1970’s.

Fatigue failures (leaks or cracks) due to analysis errors committed yesterday may not occur tomorrow, or even next year, but are more likely to occur later in life after some number of fatigue cycling has occurred. Stating that a system hasn’t failed yet is no guarantee that it will not fail tomorrow.

The B31.3 Code in Note 11 states that there can be non-conservatisms in common geometries using the present Appendix D, and that the correction to non-conservatism is the designers responsibility. It seems to be fairly clear that there are answers provided by the application of the existing B31.3 Appendix D that are not “right” (if a non-conservative answer can be considered a not “right” answer in piping), and that it is the designer’s responsibility to select conservative (or “right”) answers.

WRC 329 states that this non-conservatism is accompanied by a number of other issues in the then, and still current, B31.3 that should also be addressed.

B31.3 Chapter 1 Paragraph 300 (b) defines the responsibilities of the owner, the designer, and the manufacturer. The designer is responsible to the owner for assurance that the engineering design compiles with the requirements of the Code. The Code in note 11 says that the designer is responsible for the selection of an appropriate SIF, and defines conditions where some SIFs are non-conservative. Is the designer using Appendix D in non-conservative areas stating that he or she is satisfying the requirements of the Code? Do we require that in some cases the Code requires us to be “non-conservative,” or to select something that is more applicable?

We find that in practice these issues are far more complicated than appear in these few sentences, and are a function of the operation of the piping system, the material of construction, the dimensionless parameters of interest, (D/T, d/D and t/T), the type of branch connection studied, and the location of manufacturer.

In these cases it seems that one answer can never apply in all situations, and that when a particular situation is considered, the details of that situation should be spelled out clearly so that a specific, response can be formulated. There are clearly cases where SIFs that are two times to low have not resulted in thru-wall cracks. Fortunately, we can usually identify these cases, and know that when we confront them, the designer has selected an appropriate SIF even when it is non-conservative in accordance with Note 11 of B31.3 Appendix D.

When safety factors are less than the error, a SIF that is appropriate, and non-conservative for one problem though, will still be non-conservative, but may not be appropriate for the other. We feel that the prudent designer must recognize this latter condition, and we have attempted to write software to make that recognition and its safe resolution easier to achieve.

Regards,
Tony Paulin
Paulin Research Group

Using more accurate SIFs should be consistent with a proper calculation procedure recognizing that, for a refinery piping system, it is virtually impossible to have 7000 full displacement cycles during the expected service life of the piping system (i.e one full displacement cycle, every day, more than 19 years).
I would consider the actual service of the piping system is the reason for which "nothing has failed."

Best Regards,
M

On a slightly different but related subject, one area that can help you specially in thermal expansion load case, is the flexibilities of Tees. Current Appendix D does not provide flexibility of the tee, so if you can compute that and include that in your analysis you can get more accurate and in some case more economical solution.

Using more accurate Sifs and more accurate flexibilities for Tees goes hand in hand.

It can help you build a more accurate model of actual piping system (more accurate hanger selection..) and give you much needed flexibility in tight systems or ones governed by thermal cases.

Thank You Mandeep for reminding us the use of FESIF/NozzlePro calculated flexibility at branches. It will give us an accurate solution. Many times we introduce error by ignoring these flexibilities, thinking that we are remaining conservative.

Just WRC-107 & 297 are inaccurate in curve's extrapolated regions, beyond tested ranges of piping App-D SIF & flexibility can be inaccurate.

As Tony Paulin reminded, we do not shy away from using most advanced medical tools like video camera inside our ear to check when paper pieces go inside or watching fast recorded voice box parts in slow motion to detect minor throat problems in ENT clinic in not so developed parts of India, we should use your FEATOOLS just like Caesar-II in this age, when thin pipe sizes like 50" -100" are commonplace in Indian & Chinese Process & Power Plants for which App-D SIF & k can be inaccurate enough!

regards,
sam

PS: In India,presently it is election summer when we could VIPs reaching out to the masses by rare visits earlier; now, they are reaching us through video conferencing, even to the remotest places! In piping stress analysis, too, we should not lag behind, when Intergraph & PRG have such good tools ready for us all!

Together we can make the Earth a peaceful & safer place! Let us pray!

God heard our prayers! We are having Caesar-II & NozzlePro. Now, we will get FEATOOLS, too!

regards,
sam

About flexibilities on Tees, someone knows what's the procedure for include in CAESAR II these flexilities?

Regards,
M

Watch the Webinars on FEATools, there are several, but this one will get you started.

Application of FEATools