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?

 

[BigInch]

Cold bending is not recommended using cranes. Use a pipeline bending machine.

You can cold bend a pipe in yielding stress range, provided the minimum radius remains in accordance with the minimum cold bend radius permitted by your design code, B31.4 or B31.8. You can hot bend to tighter radaii, provided the resulting "pipe bend" is within what is allowed by your design code.

After cold bending, when the bend is released from the bending machine, there are no applied forces, so applied stresses within the bend are zero. That is the preferred method to install a pipe, since there are no stresses remaining from the installation itself.

If you do not use a bending machine and allow the pipe to bend under its own weight, or if you pull it into some kind of horizontal bend while in the trench, bending stresses are created. If the pipe is not returned to straight before it is buried in its final position, bedded on the bottom of the trench, bending stresses from any remaining curvature will remain in the pipe. Installation stresses (in this case, bending stresses) are created. Any installation stress retained will have to be considered when the pipe is pressured and put into service, meaning that the installation stresses will have to be added to hoop, thermal and all other stresses present under any and all loading conditions and the resulting total combined stress must remain less than the Allowable Combined Stress permited by your design code.

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

 

[KVdA]

Thanks BigInch for your opinion. For clarity, the pipeline is welded above ground (on blocks) and large strings will than be lowered in the trench using cranes. When the pipeline is installed in the trench most of the bending stresses are going to be zero. The remaining welds are performed in the trench, connecting the pipeline strings together.
So the expected stresses (on straight pipeline sections) will only be present during the lifting/lowering and horizontal manipulation of the pipeline.
Looking at your reply, you would consider the stresses as primary limited to the yield value (or yield divided by a safety factor). I agree with you that it would seem logical to do so however is there any reason why the stresses could not be categorized as secondary? Is it not an imposed displacement?

 

[LittleInch]

B 31.3 is not a pipeline code, however as far as I can see para 320 and 302.3.5 c0 apply, i.e. don't exceed the max allowable S given by the tables.

S is in fact the lower of 2/3 SMYS or 1/3 UTS so already has a pretty good safety factor.

Installation there is of course no pressure stress or thermal stress, so in most case so long as you don't have a great big weight or valve in the middle of your lift you will be fine.

I think BI answered a different question, but you confuse the issue by saying "with cranes and the pipeline will be bent.". Now I think BI has assumed this is a permanent strain, but I think what you really mean is "...and the pipeline will be in elastic bending". Hence when in the trench the strain and bending stress will reduce to near zero.
which one?

I don't really understand your primary / secondary point - please explain.

~B31.3 doesn't do buried systems very well - you call it a pipeline, but you're using a piping code - why?

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

 

[BigInch]

You don't know which design code to use? Why?
As LittleInch told you, the B31.3 code should not be used to design a pipeline.
As I said, Design code is B31.4 or B31.8
Your allowed primary stresses and allowed combined stress will be found within those pipeline codes.

Elastic bends stresses remain in the pipe while there is curvature, regardless of what machines do the lifting, lowering and installation. Pipe bending machines bend in the yield range, so when bending force is removed there is no remaining stress (other than some relatively low residual stress).






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

 

[KVdA]

The reason why I did not specifically mentioned a design code is that I have actually the same issue for several design codes.
- design code ASMEB31.3: an aboveground pipingsystem was lifted for modifications on several supports.
- design code EN1594 : lowering a gas pipeline in the trench as explained in my original question.
- design code NEN3650-1+2 : HDD pipeline that needs to be bend to get the pipe in the borehole.

But using different design codes, it is (in my opinion) confusing, unclear, not mentioned, contradictory... Therefore I decided to make the question referring to primary and secondary stresses explained in ASMEVIII div2. The philosophy for defining the allowable stress is for most design codes (I expect) somewhat the same. Primary stresses PL+PB are always bellow the yield (< k*1.5*S with S=allowable stress as defined by LI and k= occasional load factor). Secundary PL+PB+Q stresses are allowed above the yield (<SPS=3*S=2*SMYS).

Also referring to "elastic bending" would be stupid (and I hope my question was not) as it would imply that that the maximum allowable is bellow the SMYS.

The doubt I have is that why should it be limited to the SMYS. Between the lifting points, you would have a sustained or primary stress induced in the pipeline/piping. But the higher stresses are induced due to the displacement at the lifting points. And if I'm correct imposed displacement are categorized as a displacement/secondary stress (limited in ASMEB31.3 §302.3.5 (d)to maximum 1.66*SMYS with SL=0, Sc=Sh=SMYS/1.5).
Or am I wrong to make the link between ASMEVIII div2 and ASMEB31.3, because than why a maximum of 1.66*SMYS and not 2*SMYS?

 

[LittleInch]

I think you're confusing yourself (and me) by looking at all these codes.

Your OP states " What would you chose to be the allowable when the design code is ASMEB31.3? "

My response was use Sa as listed in table A-1
ASME VIII is a pressure vessel code and will just confuse you.

When you lift a pipeline or section of piping for installation purposes, whether off supports or into a pipeline trench, you should not create stresses which yield the pipe. How much below that figure you go is your safety factor and usually 0.9 is accepted, if a little high. Once you've completed your operation the resultant loads and stresses are either created by the span between support (yr first bullet) or hopefully near zero (your second bullet). If you create stresses > SMYS you risk bending and buckling the pipe leading to potential failure of your pipe (most times you will then be removed from the project....).

The stress induced in an HDD is different because that bending stress is permanent (in some sections at least) and hence needs to be added into the equivalent stress calculation in whichever pipeline code you are using. The sections above ground during installation will normally need support, but this is the same as the installation stresses noted above.

I looked at 302.3.5 d) and can't see where your figures come from, nor how Sl is zero noting that d) is really looking at fatigue and stress cycles.

LI


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

 

[BigInch]

You need to check stresses in accordance with the pipe design code. NOT ASME VIII div2

The bending stress induced in an HDD and during offshore pipeline installations as well, are elastic, so they remain after installation, becoming part of the combined stress calculations.

Allowable stresses are always less than SMYS. PLEASE STOP YOUR WORK AND FIND SOMEBODY THAT KNOWS WHAT THEY ARE DOING BEFORE YOU GET SOMEONE HURT.

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

 

[KVdA]

Little Inch,
You state use Sa as listed in table A-1, I suppose you mean Sh as per §302.3.5(c) or just plane old S per §302.3.1 . Sa is a combination of Sc, Sh and Sl as per §302.3.5(d).
I know ASMEVIII is a pressure vessel code but it is the design by analysis section (=§5) I'm referring to, so when choosing FEM you will end up using it's rules.

In some cases the biggest bending stresses are due to the lifting operation other bending stresses are from dead load. The case I'm keen to receive your reply is when the dead load stresses are minimal compared to the lifting stresses.

I'm also well aware of the fact that buckling etc requires verification if above yield.

Big Inch,
Sa=f*(1.25*(Sc+Sh)-SL))
- Suppose no meaningful temperature difference => Sc=Sh
- Suppose SL is minimal (near zero).
- Suppose f=1 but can also be 1,2 according fig. 302.3.5.
As already stated S=min(Sy/1.5;UTS/3), so lets suppose the Sy value determines the allowable stress.
Sa=1,2*(1,25*((Sy/1,5) +(Sy/1,5))-0)=2*Sy.
So thank you very much for your refreshing input which helped me to solve at least one of my questions.

Before being verbally attacked, I agree it just doesn't feel right, but is there any real argument why it should never go above the SMYS.
Consider a piping system where due to thermal loadings (the impossible happens) and the stress value goes above the yield. Local yielding occur and stresses drop. A little later the product is redirected causing the piping temperature to drop and so does the thermal stresses. Is it not somewhat similar. Thermal stresses are secondary but so is imposed displacement...

I just keep on trying

 

[LittleInch]

KVdA,

You're looking at the wrong section in B 31.3

302.3.5 (d) is all about fatigue. I don't agree Sl is zero and Sc and Sh are limited to 20 ksi anyway.
Sl is Stresses Due to Sustained Loads, SL. The sum of the
longitudinal stresses due to sustained loads, SL, e.g.,
the pressure and weight in any component in a piping
system (see para. 320), shall not exceed Sh,


You want to use 304.7.2 d)

(d) detailed stress analysis (e.g., finite element
method) with results evaluated as described in
Section VIII, Division 2, Part 5. The basic allowable stress
from Table A-1 shall be used in place of the allowable
stress, S, in Division 2 where applicable.

It shouldn't go beyond yield because you bend the pipe....

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

 

[BigInch]

Just keep your stresses below allowable stresses unless you're doing a controlled cold bend with a pipe bending machine. I have nothing else to say about this.

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

 

[KVdA]

Well that was the hole point of the thread, to discuss what the allowable stress should/could be (as it is not mentioned in the ASMEB31.3).

Please notice that §302.3.5(d) is also valid with 1 load cycle, ASMEB31.3 surely does not exclude it.
Sl can be zero, if no pressure, no dead load (or piping fully supported) and the only load (causing stress) is thermal temperature difference.
But the example was just to show that there is a link between ASMEB31.3 and ASMEVIII div2 part5. And of coarse that under certain loading conditions it is allowable to exceed the yield value of the used material.

§304.7.2 d) is for unlisted material but actually that doesn't matter. The only thing I whish to discus is why ASMEB31.3 §302.3.5(d) is not applicable (or why can the stresses not be considered as secondary).

But ok, thanks anyways for your insights.

 

[TGS4]

I have some experience in designing lifts using cranes for pipe (both pipe to-be-buried and pipe-rack situations). Unfortunately, it appears that our resident pipeline experts don't fully understand the issue(s).

First, in my experience, localized buckling (kinking) of the lifted pipe at the lift location(s) is the biggest threat. You're inducing not only longitudinal compressive stresses, but also radially-inward stresses. I'm not aware of a hand-calc method that can save you. And you would need to examine a non-linear buckling analysis from Part 5 in VIII-2.

Secondly, you need to understand the purpose of the 3S limit - it is a ratcheting limit to ensure shakedown to elastic action. It is not applicable to a one-time-only load. Nevertheless, some stresses may exceed yield on a single load application. However, classifying and categorizing the stresses from an elastic analysis may be quite challenging, because most pressure vessel engineers with experience, have that experience for pressure vessels and not piping or pipelines. I would recommend the services of an expert in this regard.

Finally, your service may play a large part here. If you are in sour service and require a low level of residual stresses, then your answer will change. Otherwise, yield-level residual stresses will be the likely result.

 

[KVdA]

TGS4,

What do you use for allowable stress for the design of the lifts? Looking at your reply you would also stay bellow the SMYS/SF correct?

Do you always perform a buckling analysis per ASMEVIII-div2 in those cases?

I thought that once above the yield value the system shakeddown (or is it not a verb)to a stress strain function perpendicular to the original. But I think I start to realize why introducing additional compressive stress and strain to a system is perhaps not that healthy for piping, because it may limit the use of the system after installation. Do you agree?

Thanks

 

[BigInch]

Thank you TGS for the barb. Atually this time it's simply that I would rather not encourage the OP to do work that IMO he isn't qualifed to do. My choice. Your statment, "I would recommend the services of an expert in this regard." indicates you apparently agree. You don't actually offer any real help anyway, so did you come here today just to toss a few barbs or what?

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

 

[KVdA]

BigInch,

If your replies on questions are only going result in insulting people than I would rather have no reply of you at all in this thread.
Also understand that stating incorrect things in combination with questioning peoples qualification damages your credibility.
I can understand that you might have misunderstood the question, or perhaps I didn't do a good job in expressing myself. In any case please let us focus on the question or discuss (without being offended if people don't agree on certain topics). I'm still more than eager to hear a well argumented point of view of an expert.

 

[BigInch]

I don't see the insult. It's more like a fact that several people can apparently agree on.
Better to do as TGS says and hire an expert.

Anyway I'm more than happy to leave this alone. Just wondered what TGS' problem was. Usually he has pretty good advice. Oh yes, there it is. "hire an expert".


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

 

[LittleInch]

Ok, this thread has bounced around a bit. Let's reset here.

primary question as far as I could see was about lifting pipes for installation. Pipeline design codes restrict equivalent stresses at 0.9 SMYS, what was the limit in B 31.3. was the original question.

I still maintain its the allowable stress value in table A1 and provided a few references.

Now my deep understanding of buckling and pressure vessels isn't massive, but one thing I like about ET is that it forces you to go and look things up. So my somewhat simplistic view of primary and secondary stresses is that primary is the overall stress inputs (pressure, bending, axial etc), whilst secondary is more for discrete local areas where the stress is self relieving or limiting if the material yields.

Lifting pipes for installation should not result in reaching yield or deformation / buckling of the pipe.

When you calculate that using FEA or similar I don't believe lifting loads are secondary stresses as if it starts to yield at the lifting point it will just keep on yielding, plus the stress in the pipe are not local. Hence my point that Sl cannot be considered as zero until there is no load or lift on the pipe. That's not the point here.

Now when the pipe is in it's final position, you can analyses the stresses based on the initial loads and stresses, if any. To me piping and buried flat pipelines should have no or very low residual stress levels.

HDD pipes in the curved zone will have residual stresses and these need to form part of the start point for any analysis.

comments?

LI

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

 

[TGS4]

BigInch - no barb intended. It's just that looking at lifting a piece of steel is much more complicated than evaluating an operating pipe/pipeline. For starters, the concept of design margin for the specified failure modes is something that needs to be considered in the context of a risk assessment. Since the pipe/pipeline is not actually operating, I would have a difficult time justifying using only the Code-mandated operating design margins without any other information. That, an LI's understanding of stress categorization is, by their own assessment, lacking.

 

[LittleInch]

TGS4 - Ok, but what's your opinion on what the design margin should be for a pipe installation evaluation?

why would you have a difficult time with 0.9 SMYS? Like I said, the aim is surely to install the pipe without damaging it?

sincerely - in simplistic terms for stress categorization was I right or not? I like to learn as well...

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