After all, code is a law, which changes with time!
If present is considered a new normal, we are finding shut down every 2.5 to 3 years and considering sustained load with pressure and weight which happens once in blue moon in startup or shutdown. But, the operating condition load is not checked for piping and we are finding welds getting damaged in piping due to high operating load, although
within expansion range allowable.
If today may day is called a day remembered by labors only with socialist view, not of freedom, why can't we change operating load as sustained and cold to hot load as expansion?

We are indebted to the great nation of US to give us May Day, although we are in changed times Worldwide!
Reg,
Sam

Dear sam,

I think you can re-write and propose a more accurate version of the Appendix P, which would release your doubts about operating conditions.

It may be useful also to provide an example in which the system is qualified by B31.3 and failed in actual operating condition with parameters less than those considered in analysis.

Alternate sustained cases with lift-off are easy to swallow... at some temperature, a given support is going to either be on its support or it's not. It's a binary condition.

However, system temperature is not binary - it is analog. You will have a difficult time proving that the stresses you see were consistent through the length and life of the system, and that they were the cause.

It is true that some support get lift off all the time in operation giving rise to flow induced vibration and we are blaming the operating load on recurrent weld failure, too, as to prevent vibration, we made piping excessively rigid in axial direction, without removing root cause.

Our operating temperatures are not always high enough to have full 100% relaxation in small duration.

Many times we face difficulty in explaining situations with code and doubt appears in mind! I will think over the problem afresh.

It is very difficult to assess as-built support stiffness and structural steel support elevations in successive supports at site which can give rise to lift off which can not be caught in alt sust cond
Loads.
Reg,
Sam

As we learnt from PRG Webinar 044(Part 2), large thermal expansion load can be treated as primary load on equipment nozzle, similarly piping made too stiff by rigid strut to raise natural frequency to min 16 Hz, it is possible for thermal expansion load to cause crack in weld due to primary stress from expansion load.

For eg, for SS pipe, if we have temp drop of 100C between two anchors in straight length configuration, stress will be E*alpha*DT i.e. 2E6*0.018/1000*100 = 3600 kg/cm^2 .. which can cause primary failures in weld.

Maybe in our case, we have caused failure at welds to avoid imaginary two phase vibration by making the piping too stiff!

reg,
sam

Do you mean your system is qualified by B31.3 with SE< =SA and still has troubles due to thermal strain?

If yes (and the answer is based- more or less- on the example you've given), this would be really surprising vs. the fact B31.3 includes now the range stress due to axial loads in the calculated displacement stress range, SE. In brackets, it includes also axial load stress intensification factors.

About your remarks: I really don't know a criterion to understand when a thermal expansion of a piping system has the potential of behaving like a sustained load for system components. Usually is not more than what is called "an especially weak location in the system" and here is quoted the vessel nozzle connection. That's why the piping loads are often considered as sustained or primary load for vessel nozzle connection and I think is also a very prudent approach (but I can understand that nobody wants to have a junction nozzle-shell in a state "ready to become" a plastic hinge!).

However I'm afraid that considering "especially weak locations" everywhere in the piping system would end in the failure of the philosophy that has been successfully applied for almost a century. That's why I cannot confirm the similarity you've mentioned.

Hello...

Just wanted to share my idea regarding this topic. I think this is somewhat related to elastic follow-up phenomena of piping operating at creep range where there are some unrelieved part of secondary thermal stress and converted to primary type stress. If I remember right, EN code considers 1/3 of thermal expansion stress to be added to primary sustained stress for piping at creep state. But I think the allowable given in ASME Code at high temperatures already cover this so there is already sufficient safety factor so as not to consider secondary as primary stress.

The August 1992 COADE ME News discuss about this elastic follow-up phenomenon.

Just my thoughts and subject for correction and more opinion from others.

Cheers!!!

Quoting:

<< For eg, for SS pipe, if we have temp drop of 100C between two anchors in straight length configuration, stress will be E*alpha*DT i.e. 2E6*0.018/1000*100 = 3600 kg/cm^2 .. which can cause primary failures in weld.>>

I don't want to offense anybody, but...To provide two subsequent anchors on the same straight run pipe is an obvious proof of incompetence and misunderstanding of elementary statics and piping flexibility!! The potential failure is not necessarily related to creep phenomenon, it is due to general material yielding (collapse) overall pipe section.

In addition, to accommodate large thermal expansion and requirement for vibration mitigation, the solution is definitely not the the rigid struts, but the snubbers employment instead.

I'm just looking how people try "savant" explanations for obvious elementary mistakes and misunderstandings of fundamental background theory!

Please note that B31 Code theory is based on the NORMAL and ELEMENTARY assumption that piping layout and support arrangement design are carried out by competent and responsible people. Sound engineering practice means to design piping layout that is primarily working under BENDING loadings & deformations. If huge axial loads are developed, then..."there's something rotten in Denmark"...

I am sorry for attracting such harse comments on an innocuous post!

"Do you mean your system is qualified by B31.3 with SE< =SA and still has troubles due to thermal strain?"

Theremal cycles undergone are too small in number, at temperature much below creep temperature! So, doubting the expansion load to be primary - like considered for pipe support design.

Had both piping and support structural details considered in same input, this load could have been considered as secondary, too; isn't it right!

"Provide two subsequent anchors on the same straight run pipe" - one does not provide in normal situation.

In abnormal situation of high frequency multiphase flow, the requirement of raising natural frequency above 16 Hz is a requirement which can be attained by putting rigid strut. This give rise to a situation of restricting thermal expansion significantly. In size reduction - reducer small diameter end - weld leakage is appearing regularly.

It is a highly constrained and axially stiff piping with many SIFs located very near to each other - a special case I referred to.

reg,
sam

PS; We never doubt in code!

Hello Stressers!!!

My opinion regarding this topic is the Code sets minimum requirements only. A competent stress engineer who can identify those special situations such as high strain concentrations I think is a must to avoid pre mature cracking of welds. But as we all know welds have many uncertainties involved including the quality of weld.

Just my thoughts and subject to further opinion from other practicing stress engineers.

Anyway, so far all the calculations I have performed uses the normal ASME B31.3 primary sustained stress and secondary expansion stress even the Radiant Coil Piping which is operating at 600 deg C for 3 years now. But that calculation is taken from a licensor who has lots of experience in Radiant Coil Piping. I even tried using 1/3 of thermal expansion stress added to the sustained stress but it's failing for one existing furnace line that is currently operating.

I remember one of my senior stress engineer who doesn't like to weld pad on piping. I was confused then why he wants to avoid it because during that time I want some calculations to prove to me that it's not working. But during the course of my experience I have read that pad at high temperature creates high thermal stress locally due to heat transfer rates (large mass, small mass difference). So, Why perform a very detailed calculation if you can avoid it by just omitting it from design.

My statement is open for correction as I have seen operating plants with high temperatures (550 deg. C) with pad on the support. Maybe the insulation helps to reduce the thermal difference.

Cheers!!!

Sam,

I just repeat for the third time my question
"Do you mean your system is qualified by B31.3 with SE< =SA and still has troubles due to thermal strain?" The only thing to add is I would like to know SE has been calculated considering stress due to axial loads.

My additional question would be you intend to share your case and listen to other opinions or you prefer to say in different forms - I have a crack in weld and I have also the solution which is to consider everything as primary load and stress, hence the Code must be changed because, BTW, "We never doubt in code!" but we'd like to change it?

Dear Mariog,

Code is an alive document of our ever learning efforts of mechanical engineer!

I do not think we have taken all the inputs correctly in pipe stress analysis, still!

But, question remains ...If we consider piping inline components designed for secondary load criteria even equipment box,Les similarly by fea like nozzlepro, are we not too conservative regarding pipe support detailiing!

With the case in hand with recurrent weld failure, I have doubt of scc of duplex steel even at not so high temperature due to presence of contaminants.
Reg,
Sam

About supports, ASME III NF (far more specialized than B31.3 and B31.1 on the supports subject) says in NF-3121.2- "For evaluation purposes, stresses induced in the support by restraint of free end displacement [NF-3111(e)] and anchor motion [NF-3111 (f)] of piping are considered primary stresses.".

NF changed the classification "for evaluation purposes" in 1982.

B31.3 and B31.1 may be silent on classification of stress in supports, but the practice "for evaluation purposes" is the same by the allowable considered.

Which is the connection with your case? You are not talking about stress induced in supports but induced in piping- if I understand well.

In performing a detailed stress analysis of piping components or pipe support attachments (that is practically the relevant case of pipe support analysis), not only the induced stress origin (e.g. Sustained/Primary, Secondary/Expansion etc.) is to be considered, but also the stress types should be taken into account - meaning Membrane Type / General (Pm) or Local Membrane (Pl, Pb), Secondary Type (Q = meaning LOCAL BENDING due to discontinuity effects), Peak Type (F = due to local discontinuities with relevance for Fatigue Failure).

For Connected Equipment Nozzles and welded Pipe Support Attachments, it appears indeed that the relatively recent editions (e.g. post-2007) of ASME Codes (both VIII-2 and III-1 NB/NC) brought supplementary provisions that mainly require to include piping restrained thermal expansion loadings (and the assimilated ones with externally imposed movements) within PRIMARY Stress category.
This means an additional conservative requirement to include the external piping loads induced by restrained thermal expansion in the Primary Loads when Pm (General Membrane) and Pm+Pl/ Pm+Pb (Local Membrane) stresses are to be qualified.

However, to go further and to qualify the resultant Operating (P+W+T) stresses as B31 Sustained ones (e.g. against Material Basic Allowable Stress), either for a piping component (spool, elbow, intersection) or for a welded attachment (trunnion), definitely would mean an unrealistic approach.

B31 Codes do not include General Membrane (Pm), Local Membrane (Pm+Pl, Pm+Pb) and Secondary (Q) stress categories, as are defined by ASME VIII-2 and III-1 NB/NC Codes.
B31 Sustained Stresses are evaluated using Stress Indexes, which are established by limit load analysis and are not identical with Primary Stress Concentration Factors. Besides Sustained Stresses, the Displacement Stress Ranges (EXP) may be considered as Peak (e.g. Pm+Pl+Q+F) Stresses directly. B31 Codes do not include Secondary Stress (Pm+Pl+Q) concept.
The resultant stress limitation Pm+Pl+Q+F <= Sa, as provided by ASME VIII-2 Part 5, might be considered similar to B31 Displacement Stress Limitation SE <= SA = f*{1.25*(Sh+Sc) - SL}, which might be written as SE+SL <= f*1.25 * (Sh+Sc). In such case, ASME VIII-2 "Sa" would be similar to f*1.25*(Sh+Sc) stress limit and "SE+SL" overall effective stress might be seen as "Pm+Pl+Q+F" overall Peak stress. Of course, there are some inadvertances regarding B31 SIFs employment, which are 50% from the overall ASME VIII-2 Stress Concentration Factors, but this is another discussion...

The point is that when Stress Concentration Effects in Piping Systems are to be analyzed accurately (meaning pipe fittings - e.g. intersections, equipment nozzles or pipe support attachments), B31 Codes' equations and stress categories are not sufficient and therefore more refined concepts and methods are required - such those provided by ASME VIII-2 Part 5 or ASME III-1 NB-3200.

But such detailed assessments are required only for non-typical circumstances - such as Fitness for Service (FFS) assessments, for instance.
In my opinion, there is no need to consider B31 (31.3, 31.1) Codes as inappropriate or insufficiently safe/accurate any more. When piping layout is correctly designed and the system is well-balanced, B31 Codes equations ensures a safe and proper design.
Regarding Pipe Support design (e.g. welded attachments analysis), there are several specific design methods validated by common engineering practice: Kellogg Line Method, ASME III-1 Non-Mandatory Appendix Y-1000/2000/3000 (that are former ASME Nuclear Code Cases N-318,391,392), WRC-107 etc.

As final argument, I would bring into discussion the initial problem I found unacceptable in my previous comment:
<< For eg, for SS pipe, if we have temp drop of 100C between two anchors in straight length configuration, stress will be E*alpha*DT i.e. 2E6*0.018/1000*100 = 3600 kg/cm^2 .. which can cause primary failures in weld.>>

Sam still might disagree or dislike it, but...such design is WRONG, does not match the most elementary requirements of flexibility and safe design and is definitely AGAINST B31 Code Design Philosophy.
Please remember B31.3 Para. 319.2.3 statement:
<system....>>.

LOCAL YIELDING is definitely out of discussion when we deal with an uniform compression load that induces axial compression stress exceeding significantly material Yield Limit. In such case, piping fails by general material yielding (collapse) overall pipe section. Under such circumstances, we cannot check SE against SA any more, there is no basis for that any more!

When B31 Code mentions LOCAL YIELDING, the BENDING loadings are specifically regarded. This means to design the layout by including lateral offsets for thermal growth "compensation".

There is no excuse to restrain piping axial thermal growth by subsequent anchors on same straight pipe run (buried pipelines are excluded in this discussion). In case of axial vibration susceptibility, the layout should be changed by providing offets/loops that would allow intermediate stops, or to provide SNUBBERS (NOT Rigid Struts!) that would allow thermal expansion growth but would lock sudden chocks or vibratory movements.

Hi again,

Just a short errata - A B31.3 Code quoting that disappeared:

Para. 319.2.3 statement:
<< In contrast with stresses from sustained loads, such as internal pressure or weight, displacement stresses may be permitted to attain sufficient magnitude to cause LOCAL YIELDING in various portions of a piping system. >>

Thanks Dorin for that detailed explanation. Nowadays, the Elastic Plastic analysis is becoming common to overcome this requirement in ASME VIII Div.2 to consider restrain free end displacements (thermal piping loads) as primary membrane stress in case the shell/nozzle fails with this criteria, if I remember right there is a 2.4 factor on load for this EP Analysis. But I think from piping, the factor of safety is still sufficient provided good engineering design practice is followed. As stated above if potential vibration is an issue and an axial restraint is really required while allowing for thermal expansion then snubber might be a good solution.

Any other opinion is greatly appreciated and just correct me in case I have made wrong statement.

Cheers!!!

Borzki,

I'm confused about the possibility to overcome ASME VIII Div 2 classification of Nozzles' stress attributable to restrained free end displacements of attached piping - within the limits of reinforcement or outside the limits of reinforcement.

I think there is little in the way of interpretation for classification of stresses under the rules in table 5.6 which are clear about that.

Hi Mariog,

Sorry to cause a confusion. Maybe "overcome" is not the right word in my statement. Anyway I have got this idea from one of the webinar by Mr. Paulin & Mr. Hinnant from Paulin Research Group (entitled "External Loads on Nozzles and Pipe Intersections"). Refer to the link below:

https://www.youtube.com/watch?v=WTQMgGeUgwQ

In this webinar, the Piping Code and Pressure Vessel Code approach was compared. At the end of the webinar it introduces the elastic plastic analysis using the operating combined load (Thermal, Weight, Pressure) multiplied by a factor and see if the solution will converged (structure will still be in equilibrium in other words).

Please correct me in case I misinterpreted the webinar. Anyway, EP analysis is not a straightforward analysis than elastic analysis and maybe it's better to use this as a last resort (say fitness for service evaluation?).

Any other opinion is highly appreciated.

Cheers!!!

What I tried to say is that- for piping point of view, and only as my opinion- using elastic-plastic analysis of pressure vessel together with B31 approach means a mix of concepts. Better to remain on the safe part, considering stress classification method, as an engineering approximation, simple and mature, even in some cases brings certain difficulties in correct classification. But the last is not the case with Table 5.6.

Maybe the future will be to consider elast-plastic approach in all system- piping and pressure-vessels, reflecting structural damage limit state, which will release also Sam's concerns....

I see. Got your point Mariog. I think better not to use the EP analysis from piping point of view. Actually, it's a quite complicated requirement also. And I think if you perform EP analysis, all other requirement such as ratcheting and fatigue must be EP analysis also to be consistent as there is also an approach to this in the code and when I read it's quite complicated and requires tedious task to perform. From design point of view a simple elastic analysis together with Code requirements is sufficient and proven approach based on years of experience.

Just correct me if I have got your point. But I agree that a consistent approach must be maintained so as not to cause confusion and we might miss something that we don't know if we mix the approach.

Cheers!!

I have read this article and I think maybe this is one of the reason why a very detailed stress analysis is usually not undertaken from piping point of view due to some practical reasons. The use of the Code plus a competent stress engineer is I think is sufficient enough based on years of experience in the piping industry.

"First, pressure vessels for typically represent a large investment and an extensive stress analysis thereof represents a relatively small portion of the total cost of the vessel. In contrast, a typical piping system will include many different piping components, each of which individually represents a relatively small investment. The detailed stress analysis of these piping components, however, can be quite costly."

Any other opinion is highly appreciated.

Cheers!!!

"And I think if you perform EP analysis, all other requirement such as ratcheting and fatigue must be EP analysis also to be consistent as there is also an approach to this in the code".

To clarify this doubt I have captured a statement in ASME VIII Div. 2 Part 5 section 5.1.1.2 stating:

"If multiple assessment procedures are provided
for a failure mode, only one of these procedures must be satisfied to qualify the design of a component.

Any other opinion is highly appreciated. Corrections are welcome.

Cheers!!!