Design of High Pressure Steam Line on pipe rack

[Vikoll]

Dear Colleagues, in a course of review of stress analysis in oil treatment plant I came across of NPS 24” Sch.100 high pressure steam line at 320 deg.C and 11000 kPag. Line is on a pipe rack about 220m long and has 7 expansion loops. The issue is that there are only 4 line stops. Two pairs of loops do not have line stops in between. I was explained that these 4 small loops were designed to act as 2 big loops, so thermal growth will be absorbed by both loops at the same time in proportion to loop size. CII shows 93% expansion stress in one of such loops. Stress is not liberal.
I doubt that relying on friction and not designing proper line stops between loops is reliable theory. In my view result of such stress analysis is just computer numbers based on ideal model with perfectly distributed friction forces that are in balance. In real life there will be start-ups and shut-downs and line will not be evenly heated. Also there is construction tolerance that may result in supporting structural steel being on slightly different elevation that will affect friction and consequently expansion stress.
Can you please let me know your opinion on whether or not such design can be dangerous and are lines stops between expansions loops absolutely required?
Thanks for your expert advice!

Vikoll

 

[DSB123]

Poor design approach. As you say you cannot rely on the friction effects. Positive control of thermal displacements is required. Mickey Mouse Design House approach!!!

 

[Vikoll]

Thank you, DSB123, for your opinion. The question remains: does this absolutely need to be fixed considering the fact that line is installed and modification will cause start-up delay not to mention huge cost? Additional info to consider: straight runs without line stops between loops in one case is about 30 m long and in a second case about 20 m long. Spacing between supports is 6 m. Installing line stops on available supports may also requires steel reinforcement.
Thanks!

 

[TGS4]

Your original post mentioned that you were reviewing a design, not that it was already installed and you were performing some sort of post-facto review. That changes the situation significantly.

What it doesn't change is that I agree with DSB123's assessment that such a design is poor engineering practice (although I take umbrage with the Mickey Mouse slam - the Disney imagineers have a remarkable reputation as producing very good quality engineering, but that's beside the point...).

Nevertheless, the choice as to whether it's going to work or not is your call to make. Someone is paying you to make that decision. So, make it. Consider all of the failure modes that any good piping engineer would consider (you do understand the failure mode rationale behind the different allowable stress bases, right?), and make a conclusion. The only one here with all of the applicable information is you, and it is going to take a frighteningly long time period to tease all of that information out of you if you continue to persist in outsourcing this engineering review to the internet. [/rant]

 

[DSB123]

TGS4,
I did not intend to make disparaging remarks about Disney Engineers - I was relating to the fact that whoever came up with that design approach was a "clown" or maybe you call them a cowboy (or have I now made an insult to the "home on the range" brigade??)

 

[TGS4]

DSB123 - no offense taken, and no apology required anyway. Just adding some levity to the discussion. However, I completely agree with your sentiment that whomever engineered this originally was most certainly not following good engineering practice. I would likely use terms that are not repeatable in polite company, myself

 

[DSB123]

TGS4,
I have never been a PC type of guy - I just say it as it is whether people can accept the truth is their problem!!!

 

[Vikoll]

Thank you all for such great exchange of opinions. This line was designed and DWG singed and sealed by PE who took personal professional responsibility for it. Second PE who was involved agree that even design is not perfect, it still does not require modification. The dilemma now is that in my opinion it is beyond of any doubt that this design is wrong from "good engineering practice" point of view, even if it passes computer stress analysis. I made it clear to company management that there is a risk associated with this line. The question now: how high the risk is? Will line likely fail if left as is? This line is not in a cyclic service. It will be shut down maybe once a year for maintenance. Friction forces will increase over the time due to corrosion and contamination so it will become more difficult to overcome them, but at the same time stress will reduce due to creep. Could it happen that two small loops with no line stop in between will not act as one big loop and line will fail? If answer is yes, that means I should take responsibility for the delay of mechanical completion and huge cost associated with modification. So my question was not really about engineering review but rather about the right balance between the theory (Caesar analysis) and practicality under given circumstances. This practicality maybe only based of experience and good engineering judgement, that is why I posted this thread so engineering community can share the valuable experience that will definitely help me to proceed with further proper actions.
Thanks!

 

[rhino27]

If you have the theoretical friction coefficient in your piping analysis along with the appropriate design temperature and pressure and your only coming out to be at 93% of the allowable, I think the system will be fine. You are right, the construction tolerances, actual friction factors, etc. will not be exactly as you have placed into your stress run. If you are using material allowables from the B31.3 code, these values have a safety factor of 3. This means at 93% you are only approaching the safety factor of 3 limit. "Let'er rip", (the design; hopefully, not the actual pipe.)

Hope this helps.

The other guys are correct in that you would want to control the design as much as possible with guides, line stops, etc.

 

[TGS4]

rhino27 - your assessment of the design margins is completely incorrect, especially when it comes to the expansion stress range. The design margin of 3 that you referred to is the margin on the allowable stress (typically used for pressure and other sustained loads) against the engineering ultimate strength. The limit on the expansion stresses is much higher, but that's because the failure mode isn't catastrophic plastic collapse (or burst), but ratcheting/elastic shakedown.

Your cavalier approach belies a serious misunderstanding of the failure modes of piping systems and the associated design margins against the specific failure modes.

If it were me, I would include in my assessment a sensitivity study into the effect of different values of the coefficient of friction. Friction only obeys one law: Murphy's! If you need it, you can guarantee its not there, and if it "hurts"you, you can guarantee that it's value will be MUCH higher than you expected. Flow-induced vibration will also tend to mess up whatever you expected, in the long term. Proved with extreme caution.

 

[rhino27]

TGS4... I disagree with your thoughts. The allowable stress is used in the expansion stress range. Thermal ratcheting/shakedown occurs, but so does stress relaxation and the possible fact that appendix V could be valid in this situation in which the pipe network only is running at the maximum design conditions for a short amount of time (2000 hours per life), therefore this further increase the safety margin.

Vikoll has already mentioned that delays with be very costly and delay start up. His analysis passes... is it the best design practice? No. Is your partner the hottest person in the world? No, but he/she is good enough as is Vikoll's design. Is it "cavalier"? Yes, but I believe the "Cavaliers" won the NBA finals last year... short point, they were good enough. I have had to design similar systems that have two loops with no support in between the loops. This is on reformer piping operating up and down consistently of various ranges extended up to 1700 F. They have been in operation for 15 years with no failures...

Will increasing cost, upsetting the client, etc. be worth adding two supports for a piping system that is already passing? This does not make financial sense at all.

A lot of young piping engineers spend thousands of extra dollars for supports that are not needed and they never think about additional pipe supports reducing the flexibility of a piping system which cause higher stress areas and decrease stress relaxation of the system. Not to mention additional pipe supports could mean additional structure steel, engineering time on structure steel etc. (more money) You have to consider your Engineering Economics courses as well as Engineering materials, etc...

Last point.. the pipe is going to look disfigured after years of operation regardless of how you support the system unless you cold spring every location to make it appear uniform in the hot position, but again no one does this and it doesn't make economical sense either. B31.3 design is for 15-20 years. Most plants only have a 10-15 year contract before they recertify/replace piping and equipment.

Adequate design is and adequate design... It is not a perfect design.

 

[saplanti]

Vikoll,

I agree with TGS4, especially in the sensitivity study on the last paragraph of his latest post. If the sensitivity analysis gives you no problem you should try it. If the analyses show you there are some problem areas available you should report them. At least the PE or the owner of the job will know where the higher expected stresses (including failure if available), displacements, and support/nozzle forces are. I suppose PE and the owner will be very happy if they saw the problems earlier in case there are some. This way the operation/maintenance/failure of the system will be more cost effective.

rhino27,

You would be careful with the code issues in your posts with TGS4. First check his background on the code involvement, especially B31.3.
You remind me my supervisor, that I worked with 25 years ago for about 8 years, with your wording and several subjects in your latest post that I do not want to get into discussion here.

I am also sure that TGS4 was very honest with that word "cavalier" as well if you read it fully and carefully. I believe you misunderstood him (cultural clash on the use of word), and you should not be offended.

Kind regards.

 

[TGS4]

rhino27 - what is the design margin against ratcheting/shakedown? Most certainly not 3.0 - much closer to 1.0. That was my primary beef with your assertion. The concern is not with adding additional vertical supports, but with directional anchors at existing supports, and the control that they offer.

I also take serious umbrage with your statement

B31.3 design is for 15-20 years. Most plants only have a 10-15 year contract before they recertify/replace piping and equipment.

I have more experience than I would like to have with have with B31.3 (and B31.1) piping that is older than 30 years, and some approaching 50 years. A design for 15-20 years, based typically on design corrosion rates, is only at the EPC-stage, and it is for economic purposes only. Actual installed life will be much much longer.

The OP's design may indeed be "good enough", but that is their call to make, after sufficient due diligence. Neither you nor I can make that assessment without all of the details. But to make that decision needs a complete and correct understanding of the different failure modes applicable.

 

[rhino27]

Again... b31.3 piping design is mostly for 175,200 hour design basis in the EPC-stage, take chafe or umbrage if you would like, but this is a fact. I am sure that you will agree the EPC company is not going to design for 50 years design life unless the client specifies to do this because it will be more costly. Therefore, Vikoll is probably only using the standard range. I concur that you nor anyone else could make an assessment without all of the details. As I completely, disagree with your choice to assume that I was taking a cavalier approach with my first reply. If you are designing to b31.3 the design stress are based on a 20 year rupture life some can extend up to 30 years depending on the material selected and some can be as low as 15. B31.1 which you mention will have longer design life... most pushing that of 40 years...That is if the lines are running at the design case (highest temp & highest pressure) limits non stop for 15-20 years. Most engineers know that this will not be the case as it will normal be at the normal operating range; therefore, the pipe life will extend further past the 15-20 years. However, this should be monitored. Furthermore, the designer can calculate the exact design life if desirable if the surge temperatures and pressures are known using appendix V. This appendix allows the designer to pick the service hours of design which you can also see in this example the standard design life of 175,200 hours mentioned. You can manipulate this number to 50 years if you like, but that would be a client specified requirement. If the agreement is to design per b31.3 in the design basis, the b31.3 regular allowables in table A will be used and the overall design life at design conditions will only be 15-20 years as I have previously mentioned. Before extending operating licenses, most plants will have to get the plants/piping inspected and the remaining life will be calculated before proceeding to 50+ years most of the time per an API inspector. I have attached a pdf that has b31.1 and b31.3 comparison to further assist in your understandings.

 

[DSB123]

TGS4 - With you on this one. You mention Cavalier approach - I prefer "wearing of chaps and a stetson" !!!

 

[rhino27]

I am taking another economical approach and not replying to this post further. For you can lead a horse to water , but you can't make the horse drink.

 

[TGS4]

rhino27 - you attachment is (as I have typically found with such explanatory manifestos) unconvincing, and still littered with the inaccurate language of someone who is completely illiterate in the concepts of designing for multiple failure modes. Do you actually think that design margin against buckling is 3.0? or fatigue is 3.0? or ratcheting is 3.0? Think again! The 3.0 factor is merely one of the many factors that go into determining the safety of a piping system - it is only the design margin against minimum-specified tensile-strength for pressure (only) design. Many system will fail because of yielding, so the margin of 1.5 is more accurate in describing the "safety". And what, pray tell, do you think is the design margin for bolted flange joints?

Except for materials designed for (and operating in) the creep range, there is no inherent design life in any of the piping design Codes. This imagined limit of 15-20 years (or some magical 175,200hr limit) may be typical of your experience with EPCs, but it is not found anywhere in any published Code or Standard. And it is most certainly not a fact - perhaps an alternative fact in your world-view.

The only thing that limits the operating life below the creep threshold is corrosion (and erosion), which is based on a corrosion rate that is highly dependent on the process fluid - and (with a few notable exceptions) the ASME Codes are agnostic with respect to process fluid; and the ASME Codes don't even provide any guidance on how to select said corrosion or erosion allowances.

DSB123 - thanks for the support. That is a good description, too. It's ok - from a business perspective I don't mind designers doing the wrong thing with the wrong mentality - the owners typically end up calling me in after 5-7 years of operation to fix the original deficiencies anyway. I'm maybe halfway through my carer and I've already seen a career's worth of screw-ups caused by this type of mentality.