Maximum allowable of stresses during pipeline installation 2

[KVdA]

Hi everybody. I have a question that's been bothering me for quit some time. An extra point of view on things would be appreciated. If a pipeline is installed in a trench it is going to be manipulated with cranes and the pipeline will be bent. The bending stress can be estimated but compared to what? Could the stress be categorized as a secundary bending stress? Most codes i have come across seem to limit the allowable to the 90% yield value... What would you chose to be the allowable when the design code is ASMEB31.3?

 

[KVdA]

TGS4, please clarify what other information you would require? The limit state I'm specifically interested in here is stress.
The applied loading is, as stated earlier lifting (imposed displacement which in normal OPE condition is categorized as secondary) during construction (it may/ may not be classified as secondary???) and dead load that's it (which is always primary).
Could you also give me your insight on the questions I had posted earlier?
LI, I agree 0.9*SMYS is in any case safe, and I believe the only correct answer looking alone at ASMEB31.3 . Perhaps I should have not dragged ASMEB31.3 into the question. But by doing so I hoped to play a better advocate of the devil by referring to an imposed displacement.
Secundary stresses are not necessarily discrete nor local, for example thermal loadings can induce stresses which can be local but also general (membrane and bending). Therefore I must agree with TGS4.

Thanks guys for your time and effort,

 

[LittleInch]

My last effort on this is simply to disagree that a crane lift for pipe installation can be classified as an imposed displacement. It's simply a force equal to the weight of the suspended pipe located essentially as a point load.

There is though a significant difference between vessel design and analysis and pipe, though they share common elements it's not the same thing.

good luck

LI

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[BigInch]

Yes installation stresses and operation stresses are different. But yielding in either case is not allowed, or severely limited when it is.

Offshore installation is probably the best example to make.

Evaluating INSTALLATION stresses, for example when lowering the pipeline into the ditch using sidebooms, or cranes, or towing a submerged pipe string out to its offshore route location, stringing off the stinger from the end of a offshore construction barge in a near catenary shape, or maybe dropping a vertical string in a J-lay and evaluating RESIDUAL INSTALLATION stresses that might have been retained from any of those installation methods in combination with OPERATING stresses when in the OPERATING CONDITION are not the same. Yielding is not allowed in any of those installation operations, as all bends are kept within elastic ranges, except when making a cold bend to appropriately limited radaii. Reel lays do use plastic bending to reel and let out the pipe, but pipe is limited to small diameters and radaii are large and yielding well controlled. When the pipe is let out, the process becomes similar to the S-lay technique.

Offshore laying from a stinger in relatively deep water is done in a very long "S" bend. The pipe can easily buckle from excessive bending stress in what can be 1000 ft or more of pipe hanging from the barge down to the mud line as the pipe is strung off the stern of the barge. The pipe can ovalize if bending becomes excessive and yield stress is reached, during which the pipe may collapse and flatten as the top arc and bottom arc are pulled together. Pulling with a very high tension load, usually 100,000 lbs or more, applied by moving the barge forward will prevent the pipe collapse by reducing the compressive bending stress such that local buckling and collapse does not occur. Essentially a hanging string shape is adopted. A string hold in a catenary form is pure tension and it won't buckle. In fact, the compressive bending stress may be reduced so much that the entire cross section is actually put into pure tension. All catenary curvatures are ideally held to elastic radaii that prevent yielding in bending.

B31.4
402.6 Longitudinal Stress
402.6.1 Residual stresses from construction are often present
for spanning, elastic bends, and differential settlement.
Designers should determine if such stresses need to be
evaluated.

403 Criteria for Pipelines
403.1 General
Pipelines within the scope of this Code may be subject to conditions during construction and operation where the external pressure exceeds the internal pressure. The pipe wall selected shall provide adequate strength to prevent collapse, taking into consideration mechanical properties, variations in wall thickness permitted by material specifications, out-of-roundness, bending stresses, and external loads.


Depending on friction between pipe and mud on the bottom of the ocean and the lay stresses developed during installation, some axial tension might not completely relax and be held within the pipe as the barge moves well ahead. Axial tension can remain in the line after construction has been completed. Of course the pipe will retain any elastic horizontal, or vertical bend stresses into the operation phase as well. Those must be combined with stresses from pressure (internal and external), currents, waves, thermal excursions, mud subsidence, seismic loads and/or any other stresses that might become present during its operating lifetime. To evaluate the effects of those with operating stresses, you use the combined stress limits in the pipe design codes.

403.3 Criteria to Prevent Yield Failure
403.3.1 Strength Criteria. The maximum longitudinal stress due to axial and bending loads during installation and operation shall be limited to a value that prevents pipe buckling or otherwise impairs the serviceability of the installed pipeline. Other stresses resulting from pipeline installation activities such as spans, shall be limited to the same criteria. Instead of a stress criterion, an allowable installation strain limit may be used. Stress values for steel pipe during operation shall not exceed the allowable values in Table 403.3.1-1 as calculated by the equations in this Chapter.

Maximum Tension stress is often limited to 90% of SMYS to prevent yielding in tension. Where yielding has occurred or is used in a particular installation method, strains are limited to 2%

403.3.3 Strain Criteria for Pipelines. When a pipeline may experience a noncyclic displacement of its support (such as fault movement along the pipeline route or differential support settlement or subsidence along the pipeline), the longitudinal and combined stress limits may be replaced with an allowable strain limit, so long as the consequences of yielding do not impair the serviceability of the installed pipeline. The permissible maximum longitudinal strain depends upon the ductility of the material, any previously experienced plastic strain, and the buckling behavior of the pipe. Where plastic strains are anticipated, the pipe eccentricity, pipe out-of-roundness, and the ability of the weld to undergo such strains without detrimental effect should be considered. Maximum strain shall be limited to 2%.

The problem is not limited to pipelines.
Effect of bending on Wind turbines.

http://www.sciencedirect.com/science/article/pii/S0263823113002085

Richard Feynman's Problem Solving Algorithm
1. Write down the problem.
2. Think very hard.
3. Write down the answer.

 

[dhengr]

KVdA:
If you find something in a code someplace, I suspect it will be... ‘during lifting, thou shalt not cause yielding or buckling of the pipe, elst you will kink it..., and that’s a big big sin.’ You guys know the ASME and API codes (or whichever other codes) you use every day much better than I do. But, it seems to me that we have become so dependant upon codes and stds. to do all our thinking for us that we’ve lost the use of all common sense and engineering judgement. And, I’m not particularly throwing darts at anyone. If I were lifting and installing your pipe spools, et.al. into a trench, I would not want to leave any residual stresses after my lifting operation. That is my obligation to you. You just don’t know how to design your pipe line around these, because I could do so many different, damaging things, in almost infinite ways and locations, if I were not making every effort to avoid same. The lifting loads (primarily pipe dead loads plus some impact, maybe wind) and stresses are primary loads and stresses to me and for my part of the project. And, while some of your codes for pressure vessels or pipe lines may allow primary design stresses (pressure, temp. changes, pipe spanning, etc.) of .9Fy, I would probably limit my design considerations (allowable stresses during lifting) to .75-.8Fy, and no buckling or concentrated wall kinking. This is because there are a fair number of uncertainties in my part of the project, not the least of which, we have a bunch of crane and tractor jockeys and line foremen who are of the get-er-done mentality to gain a few minutes at lunch time.

My guess would be that anyone who’s done this a few times and has a bit of experience, will literally have a flow chart, spreadsheets, etc. for this design phase of their work. It will involve pipe material, mechanical props., section prop., pipe dia., wall thick. and thus weight/ft. Now, I could calc. my max. span length btwn. lift points for a plain straight run of pipe (spool?). I would look at the end pipe section, a cantilever, and the first couple interior spans; with span length as the primary variable to arrive at a max. allowable bending moment and lifting reaction for the given pipe section props. This leads to a tabulation of max. distance btwn. pick points. Then, I would look at pipe curvature which might add unusual loading/stresses (torsion, etc?) to the string, and also added heavy loads, like heavy valves, etc. Normally, these would be applied by superposition to my basic tabulated stresses, span lengths, etc. when they occur.

My max. lifting reaction would be based on the types of things TGS4 brings up in his 13JUN17, 05:51, para. 2. And, I think all wounds could be healed if he changed his para. 1, from “Unfortunately, it appears that our resident pipeline experts don't fully understand the issue(s),” to ‘unfortunately.... resident pipeline experts (due respect intended).... misunderstand the OP; we are not talking about cold bending, except to prevent it during pipe spool handling.' I suspect he will agree with this rephrasing. My max. reaction will be based on the type of lifting equip. we are using, and the newer roller/cradle lifting devices which spread the reaction over 4,5-6' of pipe length, on several sets of rollers, and allow some longitudinal pipe or cradle movement are far superior to the older simple sling. Once you get a pipe dia., or so in length, away from this lifting point you should have no pipe stress/strain or permanent deformation problems, under normal conditions. Any permanent deformation will induce residual stresses, my fault, infinite possibilities, which again, you don’t know how to account for in your pipeline stress analysis. But, if you haven’t preped. your trench properly, to grade, and with a nice clean base, you can be introducing pipe stresses every bit as significant as some of the things I could do in lifting. I also agree that this kind of problem is not something which should be done, the first time, without some good, experienced local mentoring.

 

[BigInch]

Yes that phrasing helps a lot.

I think my just prior to post identifies B31.4 code sections that detail the exact requirements you mention... but in "code language".

Richard Feynman's Problem Solving Algorithm
1. Write down the problem.
2. Think very hard.
3. Write down the answer.

 

[TGS4]

Due to the challenges in categorizing the stresses in such a situation, I would generally not perform this type of analysis using a linear-elastic approach. I would simulate the actual true stress-true strain curve. Understand that some plasticity is inevitable, since the proportional limit of many pipeline steels is quite low, even compared to the pseudo-elastic "allowable stress" limit.

I would evaluate the buckling failure mode, ensuring that a margin of at least 2 is achieved (using an LRFD-like elastic-plastic buckling approach, while imposing the maximum of the out-of-roundness tolerances).

Otherwise, the residual stresses would be examined and compared to a reasonable limit based on conditions such as SCC, or Cl-SCC. If the pipeline is in sweet service, I probably wouldn't really care about the residual stresses - the weld residual stresses would be high anyways. Unless you PWHT your welds, in which case I would leave the residual stresses equivalent to what you have after PWHT (generally 0.2*Sy).

 

[BigInch]

Sure. The nonlinear property of the steel above proportional limits, (70% +/- of pipeline steel's SMYS). The allowable stress methods used in pipeline design in the USA are not used everywhere. Norway, Holland and U.K. behavior. I think ASME B31.4 and 8 use linear properties for the usual reason, simplicity for the no spreadsheet era. The Offshore Part of B31.8 allows design to be based on strain. Using the nonlin method, I would guess that the required wall thickness decreases a bit.

Richard Feynman's Problem Solving Algorithm
1. Write down the problem.
2. Think very hard.
3. Write down the answer.

 

[LittleInch]

So for pipeline steels API 5L type, what is the limit of proportionality?

I had a search and it looked like ~80-85%, but would be good to see some reference or typical graphs. Given that actual stress / strain and "yield" is often higher than SMYS, indicates a good reason to limit such stresses to a certain percent of SMYS??

I realize some situations require this level of investigation and analysis but the vast majority you want to not get so close to the limit that some transient event (wind, jerking of the lift etc) causes the pipe to deform or buckle. IMO.

Been an interesting discussion though

LI




Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[KVdA]

Thanks again everybody. We all have different backgrounds, and possibly therefore somewhat surprised of each others reaction/point of view on things. I'm starting to realize that my question was formulated wrong, I'm now also aware that this is quite a dangerous subject as no clear guidance is provided in most codes (so left to the interpretation of the engineer) hence the first reactions I received (which I now beginning to understand). However if one is in the situation where it is obvious that certain damage is done, it is less obvious to simply go to the owner and say " you will need to buy 70 new pipes because your contractor has gone above 0.75 the SMYS". The owner is going to ask "on what basis are you refusing, is there any code requirement not fulfilled....", in any case he is going to do the extra mile to ensure that the piping/pipeline is going in service (as they generally are more interested in time schedules and being as cost effective as possible).
So BI I certainly hope you realize my intention was not to send out a message letting contractors do whatever they want during installation. At the contrary, in my opinion to little people are aware of how a pipe flex's (in my case) and are unaware of how easily they are trespassing the great unknown. I also find it very surprising that B31.4 allows 2% strain, that seems a lot... (or am I missing something).

 

[BigInch]

2% strain limit can most likely be safely applied as a tensile strain limit.
The onset of wrinkling or buckling typically occurs for compressive strains ranging from 0.3% to 0.6%

Page 3

https://www.researchgate.net/public...derations_in_the_Assessment_of_Pipeline_Spans

C 8.2

https://www.researchgate.net/publication/232405640_Strain-based_design_criteria_of_pipelines

Ask for the Tensile Test Diagram to see the Proportional limit.
It is the highest stress at which the curve in a stress-strain diagram is a straight line

Richard Feynman's Problem Solving Algorithm
1. Write down the problem.
2. Think very hard.
3. Write down the answer.

 

[KVdA]

Apparently 2% strain is not that uncommon see §833.5 of ASMEB31.8(I'm not that familiar with using strain as limit state).
also :
A842.1.4 Allowable Strains. Instead of the stress
criteria of para. A842.1.3, an allowable installation strain
limit may be used. The maximum longitudinal strain
due to axial
and bending loads during installation shall
be limited to a value that prevents pipe buckling and will
not impair the serviceability of the installed pipeline.
BI, Could the strain calculation you are referring to be found in API RP 1111?

 

[KVdA]

Something just doesn't feel right here. Allowing 2% of strain during installation would seems like a whole lot of allowable stress. And also residual stress.
I'm used to 0.2->0.5% strain (hence my original reaction). I know there's also buckling which will define the upper strain limit but still? Am I missing something, perhaps everything becomes clear after reading the references.

 

[BigInch]

page 41

https://www.researchgate.net/publication/232405640_Strain-based_design_criteria_of_pipelines

Richard Feynman's Problem Solving Algorithm
1. Write down the problem.
2. Think very hard.
3. Write down the answer.

 

[LittleInch]

KVdA Note that the section you refer to is for offshore pipelines and not cranes etc onshore.

Offshore installation analysis is normally much more rigourous and in depth than land based installation due to the cost of getting it wrong and also the much higher stress and potentially strain levels involved in laying pipe, especially in deeper waters. If someone buckles a pipe lifting it in by a crane or sideboom, it's much easier to see by all the supervisoirs and inspectors.

The 2% limit referred to in the codes refers to loss of support or e.g. ground movement or other single cycle events, but doesn't talk about installation.

2% also covers cold pulled bends of 40D.

Usually the owner is telling the construction contractor that he needs to replace things and its the CC who is trying to say it's all OK...

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[TGS4]

All of these strain numbers being tossed around are starting to make my head spin.

First off - buckling is a phenomenon that does not require any plasticity - elastic (Euler) buckling can occur in the elastic regime. So, it is foolhardy to think that some pseudo-elastic (or plastic) strain limit somehow protects against buckling is just wrong.

Secondly, the proportional limit (which I typically define as 1e-6 plastic strain, but there isn't much difference if you use 1e-4 plastic strain) depends on the ratio of yield to ultimate. So, for API 5L pipe, it really depends on the Grade.

The magnitude of the strain (unless you approach the ductility limit) is generally irrelevant. It's the residual stress that is important.

 

[KVdA]

Buckling might be an important parameter (or even the most) to check. But I believe that we will never correctly predict it, as it is to depended on so many different parameters which are for most (actually all) constructions unknown. You can spend so much time at it, that it would be far easier and more reliable to just send an intelligent pig down to actually measure the ovalisation, dents and all other information you want.
So in my opinion buckling doubt = inspection.
Stresses or strains are more difficult. The slightest change during OPE might cause failure, and for underground piping there are a great deal of parameters that are unknown.
TGS4- Did I just noticed a contradiction in your replies? "It's the residual stress that is the important". (Which I completely agree with). Than earlier post:" If the pipeline is in sweet service, I probably wouldn't really care about the residual stresses - the weld residual stresses would be high anyways. Unless you PWHT your welds, in which case I would leave the residual stresses equivalent to what you have after PWHT (generally 0.2*Sy)." Or am I missing something here?

 

[BigInch]

I'm not much interested in predicting anything correctly accurately (at least to more than 3 significant figures). Practicality usually means knowing how to stay well away from the number to the conservative side anyway, whatever it might be and whichever side conservatism might actually be on.

That's right. Residual stresses flip. I noticed that too. But for the most part, they are ignored in pipeline design because they are low. Those flip-floops might be one of the reasons why. I think they are typically of more concern when considering H2S corrosion.

If you're interested in preventing buckling during construction, obviously an intelligent pig isn't going to help you much with that problem.

Buckling is best prevented during construction of pipelines by keeping the bending moments small and when they are not small, keep cross section in tension. Bring enough cranes and sidebooms to the party to keep unsupported lengths of installation strings short. In operation you restrain the pipeline by burial at a depth sufficient to prohibit (upward) movemet should compressive loads near buckling loads. Predicted buckling stresses are limited to 60% of Euler bucking load. Full lateral support from soil weight above the pipe keeps the potential columns very short and Euler buckling loads above those that might develop when a fully restrained pipe reaches design temperature.

OK, now that we have the stresses relatively boxed in, I've lost sight of the reason you came here. would you mind trying to re-explain, if you can, what it is that you are actually trying to accomplish? A diagram representative of the general problem could be quite helpful.

Richard Feynman's Problem Solving Algorithm
1. Write down the problem.
2. Think very hard.
3. Write down the answer.

 

[KVdA]

BI, Well I totally agree, that with buckling it's better to be on the conservative side in order to prevent it during construction. Sometimes however the possible damage is already done. Hence the solution that can be provided using an intelligent pig after the construction phase.
Therefore I'm not that interested in predicting buckling (as I also have some criteria that can be used for that).
As stated earlier, I'm more interested in the stresses (after installation). I don't understand that you can say that generally they are ignored. Perhaps you can, if you are staying bellow the yield, but actually (and I'm almost scared in saying this)I'm interested in the case where during construction you are above. I'm still convinced that they (the imposed displacement)are secondary. But they should be taken in account as you bring an additional compressive (residual) stress in the pipe which should be taken in account for all other following loading conditions for which you are designing the pipeline/piping for.
Perhaps simulating the true stress true strain curve is a better option (as brought to my attention by TGS4), but for the moment I need to do some research on that topic.
I also did not yet check how the strains come in the equation (in detail), but I would find it strange that if you are respecting/correctly categorizing all involved loads you would have a problem with that. But please do correct me if I'm wrong.

 

[LittleInch]

"As stated earlier, I'm more interested in the stresses (after installation). "

Ah ok, that got lost in the noise of the debate we've been having. In that case (your pipe has yielded), as opposed to the actual lifting operation, then that is an imposed displacement.

The key to me is then whether that displacement is accommodated within either vertical or horizontal support from the earth or whether it is under permanent load / stress to try and force the pipe back to a straight condition. Otherwise you would have trouble at every cold formed 40D bend (< 2% strain) in every pipeline ever built.

As a very junior engineer I'll never forget seeing an excavator on full weight ( tracks off the ground) forcing a pipe down so that it would tie-in to a replacement section. I did think at the time that there would be a certain residual stress level there, but no one else seemed to be concerned...

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[BigInch]

If the displacement is within the elastic range, the stresses caused by that load remain present as do the forces applied to the pipe that caused them and these stresses cannot be ignored. They are carried into the operating phase. They might still be nearly at elastic limit stresses.

If the displacements were in the plastic range and the displacements remain after the loads causing them have been removed , the stress caused by active forces have been reduced to ZERO and any remaining stresses are the residual stresses resulting from the previous yield. They seldom exceed 20% of SMYS and are usually ignored.

This is why pipe bending machines that produce plastic range bends are used to shape pipeline bends in the field during construction and they are preferred to methods that may only result in elastic range displacements, which can leave high stresses trapped in the pipe's installed configuration. These stresses can be significant.

CLEAR?

Richard Feynman's Problem Solving Algorithm
1. Write down the problem.
2. Think very hard.
3. Write down the answer.