Hi,
In Caesar-II, we have option to define elbow type : 1. Single flange 2. Double flange. Question is do we must have to activate this option for elbow when it is with single flange or double flange? I mean do this added feature in C-II is due to B31.3 requirement?

As we have done so many projects earlier without using this option.

We have not used this option in our current project, but we got comment from Client to define elbow as single flange type or double flange type where it is applicable. Activating this option will affects flexibility and if it is connected to nozzle then nozzle load increases. At this stage of project it is difficult to ask equipment vendor to change allowable loads or change pipe routing to qualify nozzle loads

Please let me know your experience about this.
Thanks for your time..!!

Hello,

Yes, it is a requirement in ASME B31.3 Appendix D. However, you can try ASME B31J if it is there or FEA using FEA Tools and see if this will help. Maybe the flex factor using the alternative methods might be different.

If you have a high temperature system, using hot elastic modulus will help to reduce nozzle loads.

Just correct any wrong statement I've made.

Warm Regards,

Thanks Borzki..!!

Good, good. So now that appendix D is dead what is the resolution. If appendix D is no longer valid, we can do whatever we want?

I think that we tend to be superficial. We are concentrating on one thing and we are ruined something else.
In my humble opinion only the 19/early 20 century engineers were impressive. We can not compete. These gentlemen were indeed engineers with a complete education. No offence to nobody, me included.
What I an stunned is that I am not a great engineer, however I did not expect the failure of the greatest.
We are humans after all.

going further.
So appendix D empirical formulas are replaced by B31J empirical formulas (more accurate but still empirical). what's the point. Anyone benchmarked the benefit before implementing? now everybody is confused and have to rely on the software because can not really check anything because too complex.
I think someone told that this is the trend and everybody goes. Like for the appendix P. Eventually we will all go with the FEA but this was from the beginning the aim. If there is a soft that can do, we will adopt. Some companies are going this way which is very good but there is a thing. Now stress calculations inside a project are mass production. Ever consider this? You will have hundreds of calculations on a small project. In the same time nobody want to pay this amount. So the push is to reduce the time requested for a calculation, but with what cost?
Can tell much more.. but.....

The empirical formulas in B31J are based on physical testing of (pipe fitting) specimens in a lab AND Finite Element Analysis. The empirical formulas are provided in B31J so that implementation by Engineers and software vendors is possible. These B31J equations f(which are dependent on D, T, d, and t) are more accurate and more realistic that the old Appendix D equations (which were only a function of the D/T of the header).

If you doubt the new B31J equations, use FEATools and the FEA option. We have been recommending this approach since 2013.

Thank you. It is not that I do not trust, but look how much time it takes to implement and how difficult is to manage. Also not all the companies have the FEA software available.

Regards,

Mr. Ay,

For piping under B31.3, for most applications, the number of cycles expected is much less that 7,000, typically few hundred in 20 years of operation and maybe dozen of them between minimum ambient temperature and maximum in operation.

Which is the benefit calculating more accurate and more realistic SIFs under B31J or FEA (an EPRI report makes a good comment saying "the SIFs actually are fatigue correlation factors") but maintaining an engineering approach contemporaneous with old works of Appendix D, conservative enough to be considered lack of realism?

I know, the best answer would be "this is what the Code asks for"... but why we change the Code? Is just our pride to be more accurate in a particular zone of engineering, or we can see accidents due to this kind of fatigue by poor calculated SIFs?

Tx.
M

Better is the enemy of good

Yes, Dan, I think the problem is when necessity is no longer the mother of "better" as innovative response...unfortunately this aspect has become the rule.

The issue is not just the SIFs, but equally (or perhaps more) important are the flexibilities. Back in the 1950s Markl recognized that the flexibility factor of 1.0 for tees was inaccurate - it was Markl who published the phrase "more applicable data". B31J gives empirical equations to provide realistic flexibility values, and the FEA option of FEATools gives the Engineer an even more accurate estimate of the true flexibility of tees.

(Note that some people assume B31J will lower system stresses. This is an incorrect conclusion. B31J, or its FEA alternative in FEATools, will provide more accurate results for system behavior. Stresses may increase in some parts of the system and decrease in other parts. The goal is accuracy. Everyone with CAESAR II has access to FEATools.)

CAESAR II Version 13 will incorporate the empirical equations of B31J for both SIFs and Flexibilities.

Just my opinion: "the goal is accuracy" may be amended as " the goal is accuracy qualifying the system for at least 7,000 cycles of displacement/ strain- 1 cycle every day, 20 years life time".

Amazing is industry is not able to really understand this goal, we'll be great to make it more clear in B31.3 where "7,000" is -for the majority of Clients- just a figure somewhere "in details". Their perception is the goal of improving SIFs is to improve the accuracy of the sentence "is collapsing under temperature/ is not collapsing under temperature case".

I would like to add also that it was Markl who published the following paragraph "On the other hand, for average applications, the question whether and to what extent fatigue should be considered is not readily answered. This is particularly true in view of the large amount of pipe which is giving satisfactory service in proportion to the relatively few fatigue failures which have been encountered".

From what I understand, the use of B31J will show not only more accurate sif values, but also a more accurate behaviour of the system. Having different values of the flexibilities, the system will distribute differently the loads. So we will have different results for the same system. Now the next question would be how much different. Not only how much different in one specific tee or in a specific elbow, because this is more or less obvious looking directly at the numbers in pre and post B31J. But how much different overall? Looks that quite significant since App. D was completely cancelled. Does it mean that it was wrong all along? I do not minimise the efforts of the people that worked on B31J, I think is a great work. I only try to understand why App D was wipe out completely. Until now, we were told that the stress programs accurately predicted the systems behaviour. Appearently not looking at this decision.
Maybe it was considered that it was a period of transition, when app D and B31J were both considered valid and that now it is time to move forward with only B31J. I think that the period was insufficient because I do not think that it was tested at a large scale this change, actually I think that almost nobody applied this method in their projects globally. Maybe on some specific points when it was really important. It is a big difference when things are done in some special cases or globally. Maybe with the new release, these things being incorporated in the program will be fine, people will apply without objections. But now is quite a hit in terms of time required to issue a calculation. I doubt that clients are willing to give more money, to relax the schedule, managers to accepts delays for this. The burden will be on the people that acctually work directly.
Just my opinion.

I agree that calculating flexibility with more accurate data is a progress because we have a better understanding on how a system works. When calculate more precisely fatigue to an extent that never occur in the piping life (applying also a conservative range of temperature per cycle) we just experiment a nice theoretical exercise and I cannot see any progress. It is also worth to understand the consequences: is time consuming for calculation and, more important today and for future, the system may be (and in practice often is) unnecessary flexible in field.

I would expect that people that can not agree on the code (eg. B31.1 vs B31.3) would have difficulties to agree also on this subject. Seems that they can agree. So why not an unified code?

There are a number of presentations on the Hexagon PPM website discussing the benefits of B31J and when it should be used. PRG suggests B31J is applicable (producing more accurate results for overall system behavior) when:

- A high operating cycle system (N>3000)
- A system with rotating equipment
- A tight system (flexibilities can have very large affects on stress)
- Systems undergoing dynamic analysis (natural frequencies will be lower)
- Glass or refractory lined systems

In the real word, more than 90% of the systems do not satisfy criteria 2,3, 4 and 5; almost all of them (exception are systems working in temperature cycles) will never be subject of 3000 cycles (or more) due to a temperature range between minimum design metal temperature and maximum design metal temperature, as they are defined in Projects.
For the rest of (let's say) 10% systems, accurate flexibility is important (and an improvement in calculation was/is necessary) but fatigue under cycles of temperature MDMT-MDT should be confirmed case by case as a realistic scenario.

In the end, how many systems are really subject to such fatigue MDMT-MDT, case for which the benefits of B31J SIFs are obvious? And why we apply them to all systems?

Not the accuracy of B31J SIFs is questionable but the extent of such fatigue for all systems- this is my opinion (and as I've written previously Markl had the same remark before I was born...).

and just to finish my dissertation....
We have today better tools FEA and B31J to evaluate flexibility of components- that is really great.
We have today better tools (again FEA and B31J) to evaluate SIFs, fine!
But we have a lot of cases in which the number of cycles is somehow arbitrary highly set for services where there is no evidence such cycles will ever appear. For such cases a rational engineering approach would be to set (conservatively anyway!) the number of cycles as coincidentally with the new accurate values for SIFs to maintain the same level of safety (against fatigue) that was with the old values of Appendix D. I think this is a rational engineering approach because it appears there is no real need to increase the level of safety against fatigue for such regular applications.
Of course applications where such fatigue is a real concern must be evaluated with a higher number of cycles and accurate SIFs.
In the end, the evaluation of number of cycles would be the Owner responsibility, B31.3 has such tools in 302.3.5 (d) but not giving the freedom to set the number of cycles below a high limit (3000 or 7000 cycles). That limit can be lower for regular applications- that's the point.

How about the safety factor already considered? how much in percent is this gained accuracy?

Also if there are only some cases when B31J should be used, why to apply to all? Where is the practical side?
One more thing, if B31J is good, we do not need FEA, right? I'm asking for a friend

Getting back to the original query for Elbows, please note that B31J did not publish any NEW data for Elbows. Thus FEAtools or FEBend would need to be used for "more applicable data". I recall that PCLgold (FEPipe suite stress program) will also modify the Elbow SIFs and Flex based on Close-Coupled ELLS; FEAtools may do this also??? For example an elbow pair will be more flexible with a 6D vs 3D (FMU) length between elbow work points.

Perhaps someday in the future, we'll have a complete operating audit of every line in and out of service.

Considering the cost of hard drive capacity, it's somewhat surprising that we don't see more in the way of temperature and pressure tracking.

However, I don't expect we'll see wide-scale deviation from the standard working practice of 7000 cycles until such an appreciable amount of information can be obtained.

But I also note that no facility in the world will be completely demolished and replaced after 20 years on the premise of 7,000 permissible cycles, either.

So I don't think I'd want to readily stray from 7,000 cycles being the standard in the absence of better information. I've worked on lines that are 50 years old, and while I have a good feeling that it still is nowhere near 7,000 cycles, I do not have any desire to be seated before a panel of jurists or a judge to explain why I elected to deviate from standard practice without that specific criterion in hand.

Well said and great advice again Michael.

Possibly for the recent wave of LNG plants, FERC (USA Government agency) rules MAY require a FFS evaluation for operation past the so called design life of 20 to 25 years. But from my observations also, the typical chemical plants and the LNG plants will still be running at 75+ years including revamps on top of revamps.