Background:

B31.3-2004, Table D300, Note 13: SIFs for branch connections are based on tests with at least two diameters of straight run pipe on each side of the branch centerline. More closely loaded branches may require special consideration.

B31.1-2001, Table D-1: No note similar to above. (I couldn't access 2004 edition, perhaps it has been added.)

B&PV Code-2004, Fig. ND-3673.2(b)-1, Note 10.(c): Similar note to B31.3 but in terms of arc distances in longitudinal and circumferential direction.

Application:

Multiple pressure reducing stations installed between 1980 and 2002 to B31.1, each with multiple valve legs, carbon steel piping, superheated steam below 700 Deg. F., 400 psig.

All stations typically have full size and reducing tees with other tees, reducers, elbows, WN flanges, and valves directly welded to them, and are to be evaluated under the current B31.3 code.

Questions:

1. Would it be accurate to say to the owner that welding other than two pipe diameters of straight pipe to a tee in steam piping would not have been acceptable under the B31 code unless a shell/brick type finite analysis was performed of the tee and adjacent component?

2. Did earlier codes assume that designers understood the two pipe diameter minimum spacing from other published works? What publications? Was the spacing just not understood to be a potential problem until Rodenbaugh's work in 1987?

3. As Rodenbaugh states in WRC 329, "We would rate the relative complexity of i-factors for pipe, elbows, and branch connections by the ratios 1:5:500." So, at this point in time, are there no other means (by WRC, Section III, or others) to evaluate such tee on fitting or tee on flange/valve fabrications other than shell/brick type finite element analysis?

4. Does the more recent understanding of the limitation of SIFs values for branch connections mean that it is advisable that owners assess older piping systems with congested piping arrangements because stresses in branch connections may be higher than once thought?

5. Or can owners relax since steam systems are typically subject to low cycle fatigue, and will the 20% or so relief for under 3125 cycles and other conservative judgments mean very few systems will begin to leak due to fatigue cracks?

Have I missed the "Recall" notice from the authority having jurisdiction or missed something else altogether?

Q1 Would it be accurate to say to the owner that welding other than two pipe diameters of straight pipe to a tee in steam piping would not have been acceptable under the B31 code unless a shell/brick type finite analysis was performed of the tee and adjacent component?

A1: No the code does not specifically prohibit this but simply states if you do so the SIFs given may not be cuurate.

Q2: Did earlier codes assume that designers understood the two pipe diameter minimum spacing from other published works? What publications? Was the spacing just not understood to be a potential problem until Rodenbaugh's work in 1987?

A2: No, note was added for clarification...

Q3:As Rodenbaugh states in WRC 329, "We would rate the relative complexity of i-factors for pipe, elbows, and branch connections by the ratios 1:5:500." So, at this point in time, are there no other means (by WRC, Section III, or others) to evaluate such tee on fitting or tee on flange/valve fabrications other than shell/brick type finite element analysis?

A3:Yes. Prototype testing or as you mentioned...

Q4: Does the more recent understanding of the limitation of SIFs values for branch connections mean that it is advisable that owners assess older piping systems with congested piping arrangements because stresses in branch connections may be higher than once thought?

A4:Code rules are not retroactive per se. The codes are evergreen and hopefully change as information and understanding grows. See API 579 and take a trip into the realm of fitness for service...


Q5: Or can owners relax since steam systems are typically subject to low cycle fatigue, and will the 20% or so relief for under 3125 cycles and other conservative judgments mean very few systems will begin to leak due to fatigue cracks?

A5: Speculation can only answer this question, but more people than I have often said that low cycles n have saved many a system from failure.

Q6:Have I missed the "Recall" notice from the authority having jurisdiction or missed something else altogether?

A6: Nope, its all nebulous for many reasons some of which have to do with burning trains. Judgement and experience are prerequisites that are essential.

Good questions anybody else have their take on them????

To understand the reference above to the <font color="ff0000">burning train</font>, click this link .

Hello Ken,

Your questions are good. I think that you may be under the impression that some of the theory that Code rules (guidance) are based upon is new. Actually when the "Tube Turns Team" (Markl, et. al.) set out to perform the experiments that resulted in the fatigue rules that were put into the Pressure Piping Code in 1955, they had in-hand, theory that was developed from 1900 through 1942. They were guided by this theory and they essentially proved most of it to be correct. There have been some "adjustments" made to the Codes by the Committees to further refine the rules, but the basic, well-established theories prevail still. The "Markl papers" build upon these theories and adapt them to the "Tube Turns testing" results.

You seem to be interested in "getting to the bottom" of the Code rules - good for you. If you were willing to "do some homework" I would very strongly recommend that you obtain a copy of:

TID25553 - SURVEY REPORT ON STRUCTURAL DESIGN OF PIPING SYSTEMS AND COMPONENTS. Circa 1970. Creators/Authors: Rodabaugh, E.C. ; Pickett, A.G.

See description at:

http://www.osti.gov/energycitations/product.biblio.jsp?osti_id=4097595

This is itself now an old (36 years old!!!) out-of-print publication but it can be had from the National Technical Information Service - they can reproduce it from microfilm. The "survey" is a "final report" of a literature search commissioned by Oak Ridge NL, and it gives references to most of the important published work that piping structural design is based upon. But it is more than that. It summarized all important piping design theory up to 1970 in one document. In it, Mr. Rodabaugh wrote a narrative, which guides us through the various published works that is a real gem for understanding how the theory comes together (for example the Section 7. on "curved pipe and miters" comprises nearly 50 pages). Mr. Rodabaugh essentially distills the referenced works and not only gives us the important equations but he also comment upon them as only he can. For ANY serious student of structural design and analysis of piping systems, this report is the "Holy Grail". Having said that, the referenced reports/papers may not be easy to obtain. A grand parade of Rodabaugh et. al. papers and reports (e.g., WRC reports) since then have built a monumental basis for piping design.

The problem that the Code Committees have is to find a way to incorporate the theories that we have known for a long time. It is a problem because the Committees are charged with the task of keeping it as simple as is consistent with a safe (i.e., conservative) design. Even the earliest theories showed that the most accurate calculation of pipe stresses is usually outside the realm of linear elastic beam theory so what we are left with (if the "keep it simple" method is to be followed) is linear elastic approximation with the inclusion of some "design margin" (factor of safety) to address the many vagaries. What IS new however is the growing capability of piping engineers to do more sophisticated structural analyses (computers have come of age). That is good because there are some “gray areas” in theory that could profit from being revisited. For example, you seem to be quite interested in "end effects". All we can say is that the SIF's and FF's developed for Code rules were based upon testing that included certain specific conditions within a certain specific proximity to the component (e.g. elbow or branch connection) being tested and the "end effects" that are implicit from these tests (with these conditions) cannot currently be predicted by equations based upon linear elastic beam theory. There is some "recent" 3D FEA work that can be enlightening. The Codes included rules for "pressure stiffening" and "flange stiffening" for bends/elbows but so far nothing for other deviations from the conditions prevailing in the "Tube Turns (Markl) testing". However, by reporting the "end conditions" that were included in the testing, the Code is not precluding designs that do not reproduce these end conditions - so as Mr. Luf has pointed out, that is the answer to your question 1.

Q2. The Code assumes nothing as there really is no "two pipe diameter minimum spacing limitation". The "spacing" is one of the possible conditions of proximity generally termed "end effects". The "end conditions" beget "end effects". As the FF for branch connections is unity, "end effects" do not affect the flexibility analysis although "end effects" doubtlessly DO affect the flexibility of the branch connection. For branch connections the SIF accuracy may (as the Code notes) be affected if the minimum spacing is less than two pipe diameters but since the flexibility would be diminished the Code SIF may still be conservative. The changing bend/elbow flexibility that results when there are geometries other than straight pipe close to the bend/elbow curvature cannot be accurately described simply as a function of D/r in an equation. The prudent piping engineer will familiarize him/herself with the theory on which the Code is based (see "Markl papers"). If the engineer wishes to take leave of the Code rules it should be by employing design/analyses techniques having rather more rigor than the beam theory that the B31 Codes are based upon (and I would add, documenting it). If it can be shown (via Code beam theory analyses employing current B31 Code rules) that the design complies with the Code (with an understanding of the limitations of beam theory), the design will result in a reasonable service life.

Q3. Branch connections are complicated. As we all know, we would have to buy the fittings and have them in-hand as we developed the 3D FEA model - no two manufacturers make them to the same dimensions (thickness, crotch radii, etc.) so “who knows” what the purchasing department will procure for installation into the piping system you design. In the B31 Codes we use a FF of 1.0 at branch connections although we know they are more flexible that this. We then impose the SIF's on the branch connections that resulted from (limited) testing. There is obviously some additional conservatism (factor of safety, or "index of ignorance") in this that has been shown by experience to result in reasonable service life. Using existing B31 Code beam theory (with appropriate SIF's and FF's) for these analyses has been shown to result in acceptable results. If you would vary from the prescribed B31 methodology, you must substitute more rigorous analytical methods (e.g., 3D FEA using ASME B&PV Codes, Section VIII, Division 2 rules and allowable stresses). I would suggest that the need for this would not comprise more than a few percent of the systems we ever design.

Q4. Reanalyses? No I do not think so (unless the system comprises A-335, P-11 alloy steel). The 1955 B31 Code Committee "knew what they did no know", ergo they built-in conservatism. The Committee knew that the tests were on NPS 4, standard wall, carbon pipe and fittings and they knew that the extrapolation of that data down to 1/2 inch and up to 72 inches included some inaccuracies. Subsequent B31 Code Committees have continued to look at the data evolving from recent testing and FEA work (MANY ORNL papers are available and recent (and continuing) cyclic component testing work by Rodabaugh/Woods has been reported to the Committees). Some Committees have seen fit to make a few changes based upon a better understanding of the physics and mechanics of piping components. I believe that the rules as they exist continue to support designs having good service life. These Codes are “living documents” and they will continue to develop to reflect “lessons learned”.

Q5. The B31 Pressure Piping Code (post 1955) has proven to be reliable in supporting piping designs that result in reasonable service life. At this point, the most prevalent mode of failure in pressure piping is corrosion. Adjustments have been made to the Codes in a timely fashion when it has been shown that adjustments are needed (e.g. reduction of allowable stress in A-335, P-11 material due to creep/fatigue interaction). New rules have been put into place when needed (Category M fluid service, High Pressure Piping, etc.) and new Pressure Piping Sections have been developed to provide a "better fit" for specific systems (e.g., B31.9, B31.12). These Committees do not have their collective "heads buried in the sand". Owners need only to insist upon systems that have been shown by EXPERT ANALYSIS to comply with ALL THE DESIGN REQUIREMENTS of the Codes and Standards (and to be fabricated, erected, examined and tested in compliance with the Codes).

I can hardly wait to see your questions on flange design and flange rating. You can go into the B16.5 Standard and find an acceptable temperature/pressure rating for flanges to be used in your design. Then if you use the ASME B&PV, Section VIII, Division 1, Appendix 2 rules to evaluate the flanges some of the flanges will not "pass" i.e., be shown to be "not acceptable" (sometimes even with zero forces and moments at the flange pair). But that is grist for another mill (thread).

As context for what I write below, I’d like to say that my bias is that I have an ever greater appreciation for the breath and depth of the B31 code and what it has and continues to accomplish.

However, my experience has been that most owners and engineering managers do not have the same perspective.

They ask:

1. Why does it take such extensive and time-consuming analysis to complete a stress analysis now, when 20-30 years ago it was done quickly with simple methods?

2. I can show you dozens of steam piping systems where the tees have other fittings, flanges, or valves welded directly to them. Why were those systems successfully accomplished then, and I can assure you no shell/brick finite element analysis or prototype was built to prove the tees were okay, but you need these methods to accomplish the same now?

3. Such systems are still working today, decades later. Do you hear of tees failing in carbon steel steam piping? The code is overly conservative and highly specified with no benefit over the designs of decades ago.

4. “The code is not precluding designs that do not reproduce these end conditions…” (John B.) No, it is not precluding such designs, but doesn’t provide an answer within the code as to whether a very common piping fit-up is okay or not. Why was it okay by code to weld other fittings/flanges/valves to tees decades ago (agreed the code does not, per se, state it can not be done so now), but now says “may require special consideration”, read “the SIFs may be higher, may be lower”, so the code can not tell me whether that tee that I’m looking at right here in this 20 year old station is okay or not? Tees are often fitted to other fittings/flanges/valves, so would not share the perspective that such configurations are a small percentage of systems that are evaluated.

5. “Owners need only to insist upon systems that have been shown by EXPERT ANALYSIS to comply with ALL THE DESIGN REQUIREMENTS of the Codes and Standards (and to be fabricated, erected, examined and tested in compliance with the Codes.” (John B.) The same could be said for a code that says: Build it so it doesn’t fail. What good is a code that doesn’t address common piping arrangements other to say you need to go to another code or conduct special analysis to know if the system is will not fail.

The Luf in me would like to respond: “Well, piping designs can be produced quickly and appear to work on paper. Just give a recent graduate the software and tell them you need the system designed in 2 weeks and you’ll get your quick design. Do you feel lucky? Just because you can weld hunks of metal together, doesn’t mean you have a structurally sound system. Go ahead and make my day.”

But I’d like to bridge the gap of understanding a chunk better than that.

Here’s what I am finding on these steam stations with tees that don’t meet Table D300, Note 13: The cold startup expansion stresses are at, or slightly above the code allowable, 75 to 150%. But the switchover between one leg of the station to another have stresses of 500 to 700% of the allowable stress. The switchovers will occur two to three dozen times in the life time of the system. I speculate that the systems were designed for the cold startup stresses, but without regard to the switchover stresses.

I understand that I will have to gut the system I am to revise and provide the spacing around the tees, or gut the system and perform a shell/brick finite element analysis on each tee that does not meet the Table D300, Note 13 requirement, or perform NDE to check the current status of the piping and perform a API 579 fitness for service analysis to see what can be left in place and added to.

I’d like to yet ask:

1. “…since the flexibility would be diminished the Code SIF may still be conservative.” (John B.) But may not be conservative. So, how can I learn more about how the Flexibility Factor, k, factors into all of this? I don’t see any code equations using k, so is it used in the linear elastic beam theory equations? I didn’t see Rodabaugh use it in WRC 329, though he discusses it. Why does the code give the values but not explicitly use it … did I miss where it is used?

The responses were very helpful. Thank you John L. and John B. for taking the time to respond and for all that you wrote. Richard – I usually need my spouse to translate many of Luf’s references, but I did get his reference to burning trains on my own.

Phew!!!! Well I try to be clear but I guess I should continue my life’s work on that front....

I feel your frustration!!!! That is one of the reasons I got talked into assisting the committee... and I continue to try to push for more clarity when it is desirable and somehow you can get enough votes for your proposal vs against it.

I will cite you something from my own career... I was asked to check a Dynaflex run on some thin walled high D/t ratio piping when I was in my 20's ... The analyst had use a SIF value of 1.0 for all the tees because the code values were "too high". I refused to sign off on it saying that maybe the SIFs were excessive but the real numbers had to be farther away than that of 1.0.

The system was built and for 15 years all I ever got from the clever piping designer was a lot of yapping! You’re too conservative… the code is baloney etc. Then one day one tee and then another, and another, all fatigue cracked! Was I vindicated.... not really by that time the designer and his big mouth had permanently damaged my reputation and nobody was bright enough to relate the failures to his ****py design and my correct call.

The lesson.... being right is never a move that will win friends and influence people favorably, and when pipe fails its never an over load its always bad tees bad welds etc.

Have I changed from this stupidity?? No I still try to do what I feel is appropriate knowing it will never win me fame or fortune, but also knowing that I have done my work in an ethical manner.

What to do in your specific instance???? Report the overload and the possibility that the geometry of the stations may not allow for a reliable answer as far as a simple B31 analysis is concerned. State the consequences of the overload, which is, reduced cycle life, which may or may not be a problem.

I know this is slightly off subject - but when you said "Such systems are still working today, decades later. Do you hear of tees failing in carbon steel steam piping? The code is overly conservative and highly specified with no benefit over the designs of decades ago" - please have a look at the pictures on the links below.

There was a major incident at BP Grangemouth in Scotland a few years ago [unfortunately one of a sequence of major incidents in a very short time it has to be said..] An 18" steam main, running in a trench beside a public road suffered
"a catastrophic failure .........(in) a section of the MP steam pipeline. The failure, which occurred suddenly and without warning, involved the rupture of a tee-piece on a 450mm (18” diameter) section of line operating at a pressure of 14 barg (200 psi). Failure occurred in a horizontal section of pipework..... immediately before the isolation valves........... The failure resulted in an “open end” to the MP steam supply..."

Have a look at these pictures from the HSE report into it

http://www.hse.gov.uk/comah/bpgrange/incident/figure11.htm [rupture photo]
http://www.hse.gov.uk/comah/bpgrange/incident/figure12.htm [steam main photo]
http://www.hse.gov.uk/comah/bpgrange/incident/figure13.htm [rupture photo]

It's scary how the tee has been ripped open. The cause was put down to condensation induced water hammer. Failures occur for many reasons, not all of them accountable in the design analysis.

My ususal response "just because it hasn't broken doesn't mean it meets code".

John L. - Your personal example is encouragement to persevere in learning the science and reporting the conclusions openly. My interest is not fear in delivering unwelcomed results as much as wanting to give the client the best answer the science and tested judgement can provide.

Cpt. Kenny - "Failures occur for many reasons, not all of them accountable in the design analysis." Yes, CIWH events are off topic, and even though keenly studied in the nuclear industry and by Wanye Kirchner in industrial/district heating realm, represent the burning train issues.

SLH - And I'll add: Just because it hasn't broken doesn't mean it meets code AND WILL REMAIN UNBROKEN.

Would there be any interest left in answering my question regarding Flexibility Factors:

So, how can I learn more about how the Flexibility Factor, k, factors into all of this? I don’t see any code equations using k, so is it used in the linear elastic beam theory equations? I didn’t see Rodabaugh use it in WRC 329, though he discusses it. Why does the code give the values but not explicitly use it … did I miss where it is used?

Check out the October 2002 Newsletter article on the flexibility factor for more info.

http://www.coade.com/newsletters/oct02.pdf

Author and colleauge Glynn Woods quotes Markl to us occasionally one of his quotes is my favorite.... " The code represents a compromise between scientific truth and day to day practicality" or something like that....

I have a youngster learning the trade right now who is also constantly frustrated, and bewildered.... maybe because he is working with me???

Dave - Thank you, your article in the newsletter is an enlightening distillation of what goes on out of view within the flexibility matrix analysis with regard to flexibility factors for elbows. I can extrapolate from there for tees if other than k=1 were used.

1. I find it interesting that the 2PD of straight pipe diameters for intersection runs does not also apply to the intersection branch. Would not the flexibility of the intersection be affected by a stiffer element close on the branch as well as close on the run?

2. Do any piping codes (international) use an approach other than von Karman/Hovgaard/Beskin/Markl basis for developing the fatigue stresses in fittings? (That is: for metal piping, not GRP piping.)

I believe that BS 3278? specifically states that elbows may not be welded directly to each other and must be given a straight spool of pipe between each weld line....

Hello again Ken,

First of all, I think your questions are very good. It is never my intention to show disrespect to you; on the contrary, your postings on this board indicate that you have given and are continuing to give uncommonly good in-depth thought to piping design. I applaud you. Some of what I will write here is actually aimed at people who lurk and do not pose their own questions. Please accept in advance my apologies for any tone of harshness that might creep into my opinions.

You surely did provide some very important references: "von Karman/Hovgaard/Beskin/Markl basis for developing the fatigue stresses in fittings". The publications by these fellows also give you the chain of logic that leads to the methods used in the B31 Codes for calculating flexibility factors. Mr. Rodabaugh describes the von Karman flexibility and stress index equations on pp 7-22 - 7-25 of TID25553.

Q1. "I find it interesting that the 2PD of straight pipe diameters for intersection runs does not also apply to the intersection branch. Would not the flexibility of the intersection be affected by a stiffer element close on the branch as well as close on the run?"

Yes it will. Again, a suite of equations would be required to calculate the flexibilities inherent in all the possible branch connection geometries. If they are to use linear elastic theory they would have to be approximate. In absence of better data (there still does not appear to be a theory developed that will support simple parametric linear approximation for calculating FF's for generic branch connections) the Codes use a FF of 1.0 (in concert with the Markl developed SIF's). This will be conservative for static analyses and perhaps not so conservative for dynamic analyses. However, the use of these Code rules has proven to be adequate for assuring acceptable service life of branch connections (if ductile materials are used and very significant creep (and creep-fatigue interaction) is not included in the service life).

"I don’t see any code equations using k, so is it used in the linear elastic beam theory equations?"

Yes. The FF is used in the development of the system stiffness matrix (the Code prescribed FF for branch connections is 1.0). The effect of the FF's on the stiffness matrices will affect the calculation of the forces and moments. The calculated forces and moments are used in the Code equations for calculating the various stresses (stress ranges). The Code stress equations apply the SIF's that result from the increased flexibility.

Q1. Why does it take such extensive and time-consuming analysis to complete a stress analysis now, when 20-30 years ago it was done quickly with simple methods?

Why do we do this at all? Because the building codes of the States and Provinces require it (and therefore it is the law). It is the law, because it has proven to be a good set of minimum requirements for piping design.

20 - 30 years ago (mid 70's) we took more time to do the work and the results were not as accurate as they are now (and the systems that we designed, for the most part, proved to be satisfactory). If someone in the business thinks we are doing more now than was called for (required) in the mid-'70's, their memory is faulty. Before then, before computers were generally available, some so-called design firms (drafting shops actually) were taking chances with the health and safety of a lot of people ("just do it like we did on the previous contract"). It took me (and I presume everybody else who did it correctly) much more time to evaluate the stresses in piping systems using "square corner technique" and "Flex-Anal" charts than is now required to do the same systems using CAESAR II.

Q2. "I can show you dozens of steam piping systems where the tees have other fittings, flanges, or valves welded directly to them. Why were those systems successfully accomplished then, and I can assure you no shell/brick finite element analysis or prototype was built to prove the tees were okay, but you need these methods to accomplish the same now?"

Because for the very large majority of piping systems it is not necessary (albeit in a few designs it will be necessary - not very many) to do shell/brick finite element analysis or to build prototypes to assure that the systems will provide adequate service life. The (fatigue) rules in the Code books have been in use in one form or another since 1955 and they have served us well. If the piping engineer understands the design parameters and gives due attention to the system being designed and performs professionally in developing a representative analysis model and diligently applying the Code rules, there is a good likelihood that the piping system will provide adequate service life. We have all heard horror stories of piping system failures and presumably we have all applied "lessons learned" (from the root cause analyses) to our designs. I can assure that the Code Committees have done so.

Q3. "Such systems are still working today, decades later. Do you hear of tees failing in carbon steel steam piping? The code is overly conservative and highly specified with no benefit over the designs of decades ago."

Actually, I have seen more branch connections fail due to pressure pulsation (typically in feed water and condensate piping) than I have seen in steam and process lines. Today, some generators are trying to INCREASE the plant design pressure and run closer to theoretical stress limits in an effort to reduce the kilowatt-hour-to-cost ratio. In the last 10 years I have seen many plants being run near or above the Code (stress) pressure limit in their feed water and condensate piping. When I think about that in context with what I know about erosion-corrosion, I am glad that I have been so lucky that my hearing loss is my only enduring personal plant injury.

The B31 Code is what it is - in many jurisdictions it is the law. If it is someone’s stated judgment that the B31 Code is OVERLY conservative and highly specified with no benefit over the designs of decades ago, that person is speaking out of ignorance. We have learned from doing and we have used hard-won wisdom to DECREASE conservatism where it can appropriately be decreased (e.g., in some cases higher allowable stresses and "f" factors greater than one).

Q4. “The code is not precluding designs that do not reproduce these end conditions…” (John B.) No, it is not precluding such designs, but doesn’t provide an answer within the code as to whether a very common piping fit-up is okay or not. Why was it okay by code to weld other fittings/flanges/valves to tees decades ago (agreed the code does not, per se, state it can not be done so now), but now says “may require special consideration”, read “the SIFs may be higher, may be lower”, so the code can not tell me whether that tee that I’m looking at right here in this 20 year old station is okay or not? Tees are often fitted to other fittings/flanges/valves, so would not share the perspective that such configurations are a small percentage of systems that are evaluated."

"doesn’t provide an answer within the code as to whether a very common piping fit-up is okay or not."

It is not the Code's charter to list every construct that would represent adequate design/construction. The Code is not a design manual. The Code provides a finite set of design rules based upon the experiences and knowledge of thousands of past and present Code Committee members. Some Codes provide warnings regarding using certain constructs in specific types of systems and services ( e.g., B31.1 Chapter II, Part 6 - I guess these are examples of "don't do it").

"Why was it okay by code to weld other fittings/flanges/valves to tees decades ago .... but now says “may require special consideration”, read “the SIFs may be higher, may be lower”.

There has been no "rule change" here; the advisory note was added for information. The Codes tell the designer that if the designer has better data the designer should use the better data. In the case stated, the note at issue intends to make the designer aware that when components are used in close proximity such that the stiffness of a branch connection might be affected (and consequently the SIF may be affected), the designer should give additional design consideration to the construct. One of the possible "additional considerations" might comprise the application of knowledge of many years of successful service of similar constructs in other piping systems (i.e., engineering). The Code does not pretend to provide specific design assistance or rules for every design challenge. It is the responsibility of the piping engineer to satisfy him/herself, that the design is adequate. Perhaps (my opinion) that the assumption is that the piping engineer is educated in piping engineering and is willing to apply that education (albeit, here in this forum we are frequently provided with evidence to the contrary).

".....so the code can not tell me whether that tee that I’m looking at right here in this 20 year old station is okay or not"?

Correct, the Code cannot do this (and it does not attempt to). And, it is not in the Codes charter to do so. To get to the "whether that tee that I’m looking at right here in this 20 year old station is okay or not" point, one would evaluate the remaining life fraction of the component while applying fresh structural analysis data. Also, a prudent amount of new nondestructive evaluation would be required to make an assessment - e.g., what is/are the remaining wall thickness(es), are there any surface indications of fatigue cracking and surface replication to establish the degree of creep damage.

Q5 "....What good is a code that doesn’t address common piping arrangements other to say you need to go to another code or conduct special analysis to know if the system is will not fail..."

I would suggest that Codes that were and are being used around the world to provide guidance in the design of many, many successful piping systems (and were used as the basis of the development of other county’s piping Codes) have demonstrated that they are of SOME value (?). Upon reflection, I think that your interpretation - "...other (than) to say you need to go to another code or conduct special analysis to know if the system is will not fail..." - is an inaccurate overstatement. The Code does not mandate that you go to any other Code (except for external pressure design) or that you conduct any other special analyses (except for "unlisted components" pressure design). Ken, would you rather that the phrase "...may require special consideration" NOT appear in the Code? The Committee that included that phrase did so simply to inform the Code user that under the stated circumstances the SIF's provided in Appendix "D" might not be accurate. It just waves a flag and says "think about this while persuing your engineering responsibilities". Reference is made to paragraph 319.3.6.

".........Here’s what I am finding on these steam stations with tees that don’t meet Table D300, Note 13..."

Ken, some of what you have said seems to imply that you are interpreting the Code such that not having at least two diameters of straight run pipe on each side of the branch centerline is a violation of the Code rules. Appendix D, Table D300, Note 13, does not imply that branch connection not having at least two diameters of straight run pipe on each side of the branch centerline are a violation of Code rules. Rather, it says (in my opinion) that the SIF's at the branch connection, calculated by the equations in Appendix "D" might not be accurate in such a construct. There is no "...requirement..." stated on note 13.

Appendix "D" provides guidance for piping system analysis based upon the limited testing done by Markl's team. It is "for what it is worth"; anyone having better data should use it. Ken, would you suggest that it would be a better Code if Appendix "D" were deleted? With all due respect, is there an attitude of "well if the collective piping Codes cannot be totally comprehensive and perfect in all areas of piping design then they are of no value at all"? The body of research work (theoretical and experimental) is out there. TID25553 gives all the right references (later documents have supplemented those references). It cannot be expected that the Codes reprint all the underlying theory or reprint even the references to these works. Taken as a whole, the entire body of research work is not sufficient to cover all the design issues that we will encounter. Code Committee members would be the first to tell the world that the Codes are a distillation of the best knowledge that we have (albeit incomplete and imperfect) for piping design. In my view, that does not invalidate the worth of the B31 Codes for Pressure Piping.

"...But the switchover between one leg of the station to another have stresses of 500 to 700% of the allowable stress..."

Yes, "tie-in piping" between two or more units in the same power plant often seem to exhibit indications of "less than good" analysis. Sometimes I wonder if there has been ANY analysis of these "tie-in line". The design team responsible for each of the units seem to abdicate responsibility for the "tie-in lines" ("its not in OUR scope"). As you point out, the problem is often found in the analyses of the various modes of operation - some units "coming up" and other units "going down" or into "hot stand-by". Some full cycles and many partial cycles especially when base-loaded units are placed into "peaking" service.

Most plant owners are now looking at the remaining useful life of their units so that they may know which "critical components" will have to be replaced and in what order (this will affect the projected cost per kilowatt-hour of various stations and will dictate what plants are put into "cold stand-by"). Some plant owners have placed the "major pieces of equipment" in their "critical component" programs but ignored the high energy piping. They may find out the hard way that the piping systems' most likely mode of failure is low cycle high stress (the data shown over on the extreme left of the S-N "curve") due to inadequately evaluated "modes of operation". Some will see the value of having an engineer tell them that their piping systems are in trouble.

"....how can I learn more about how the Flexibility Factor, k, factors..."

At the risk of being repetitive, begin with the "Markl Papers". Sift through the TID25553 references (some of these will be tough to find). Then there are selected WRC bulletins and a wealth of Oak Ridge National Laboratories (ORNL) reports. Read the John Brock written section on flexibility analysis in the Fifth Edition (it was dropped in subsequent editions) of Reno King's Piping Handbook.

Again, my apologies for any unintended Luf-isms found in the above. Also, all of this is just my opinion and not the opinion of ASME international or any ASME Code Committee.

Best regards, John.

Luf isms indeed well John Breen all I can say to you is @#%W$%^&^%$*^^%**&)^&&%#$)_)^!%$!%

John –

Your responses to the questions of my alter ego/owner/manager were very helpful to me and I hope to others too. I’d like to respond to just a few issues that linger dispite such careful and thoughtful input.

1. Regarding the fact that the Code does not provide specific guidance for every possible configuration… the Code helps one evaluate tees, but not necessary every fitting-to-fitting group that doesn’t comply with D300, Note 13. You offer: This refers to 319.4.1(a) “duplicates, or replaces without significant change, a system operating with a sucessful service record.”

I have no doubt that such “duplicate” applications exist, but I started analyzing piping 25 years ago and I can’t say I’ve seen any duplicate systems, especially when it comes to branch connections. I’m also not sure that if I did know of a duplicate system configuration in operation for 20 years, how I would know it wasn’t just a few cycles away from fatigue failure, without performing NDE and as much analysis as it would take to just analyze the duplicate system in the first place.

So, I don’t find the use of “duplicates…with a sucessful service record” of much practical value in the design of branch connections in plant piping.

2. On the same topic that “the code can’t address every configuration” you comment: No, I certainly see the importance of the statement. And I do understand the Code is not saying that it is a violation of the Code to weld other than straight pipe within two diameters.

But I don’t understand why in the same Note 13, the code doesn’t include a straight run distance for the branch of the tee also, and only addresses the distances from the branch centerline. Since the code is specific regarding the header, it seems to lead one to think that such an inaccuracy does not exist with the branch. Is that true? I don’t think so.

3. No, that is not my attitude, but I do find that attitude more often than I’d like. It is in comments like “We have analyzed systems like this for years sucessfully without stress analysis programs, now with analysis programs like CAESAR II its not enough, we have to perform FEA to know whether a design meets code.” Of course, that statement holds certain assumptions: that the designers would have been informed of problems had their design not been sucessful; that the designs performed prior to when analysis programs were used, actually met the code.

4. Though I “know” this, it was helpful to read it. I was beginning to think that the Code could reference some of the fundemental literature that is the basis of the code. Not as “required” reading, but as recommended reading in an Appendix. What better way to quickly impress upon the first time reader, and returning readers, that there is a significant knowledge base and history that forms the basis of the code.

In a broader sense, I guess what this whole thread demonstrates is that it is hard for me to accept that after doing this for enough years that I shuld have known better, that I didn’t know better. I didn’t know that stiffer components welded too near to tees can result in higher stresses than indicated by the standard SIF values.

To the moderaters,

Would it be possible to 'sticky' this post? Personally I found it enlightening and sound in thinking.

Hello Friends,

Getting through (in-depth) various technical papers, references, experimental lab tests, etc. will lead you to an endless chain with no conclusion or judgement.
I believe, as far as analysis stisfies piping design code requirements, there's no need to go further, except those cases which are instructed to you by clients or other authorities.
You should trust codes and standards and follow the rules. That's why codes are changed frequently to incorporate new findings.
On the other hand, engineer must also trust software he/she uses by knowing all capabilities and getting trained properly.
For above mentioned problem, you may use FEPIPE for shell/brick finite element analysis. You can export CAESAR II models right away to FEPIPE.
Thanks to all friends that send valuable posts on this topic.

Farhad,

You are saying "....you may use FEPIPE for shell/brick finite element analysis. You can export CAESAR II models right away to FEPIPE.".

Can you please explain how to export CAESAR II model to FEPIPE, I am really interested?

Thanks in advance and kind regards,

Ibrahim Demir

"On the other hand, engineer must also trust software he/she uses by knowing all capabilities and getting trained properly."


My friend Farhad, a sword is a powerful weapon, but only if the swordsman is skilled. Otherwise is useless.

Best regards,

Mr. Ibrahim, FEPIPE can read CAESAR II neutral file.

Mr. Danb, at last, you should use your own skills. It's up to you to learn good piping engineering knowledge and CAESAR II software. No choice, get skilled.

Dear All,

I went through the paper that John Breen was referring (TID25553 - SURVEY REPORT ON STRUCTURAL DESIGN OF PIPING SYSTEMS AND COMPONENTS. Circa 1970. Creators/Authors: Rodabaugh, E.C. ; Pickett, A.G.), and unfortunately could not find the answer we are looking for,(Sorry John, I have trusted and payed 45 USD for the paper, but I am glad to have it for its references).

Dave's reference is not answering it either.

My guess is there are variaty of flexibilities depending on the geometric combination of fitting-to-fitting applications. However, we need to consult to FEA to see those properly. I wish FE-PIPE had these option, but Mr. Paulin is doing better, and providing 3D model of the systems wherever necessary.

I look forward to working on the newest version of FE-Pipe.

Kind regards,

Ibrahim Demir

a guy who did not buckle up did not be killed in highway accident,
it is not meaning all people did not need to buckle seat belt up, it is just because this guy is lucky in this accident, and none can garanttee he does not be killed in next accident if he still does not wear seat belt in next time,