Dears,
I'm reviewing a piping system on a coke unit with a leak at welds at tee and support shoe locations. The line is designed to operate at 433C temperature but the contractor expansion stress is 98% using liberal and when I remove liberal it becomes 150% expansion stress. Are we allowed to use liberal for such cyclic high temperature systems?

Is it transfer line? From heater to coke drum? If yes number of cycles in these system are normally high. It will be having crosses instead of tee.

They are the lines from the drums to fractionator "representing coke drum overhead vapor". What do you mean by crosses instead of tee?

Mostafa,
I assumed that this is transfer line system where slurry is expected. Normally cross fittings are used at change in direction to allow removal of slurry. In overhead vapor system slurry is not expected but this process has high thermal cycles. It will easier to understand your problem if add some sketch for failure location.

IMHO for such system with very high number of thermal cycles, it is better to avoid use of "liberal" stress equation.
I hope effects of coke deposition inside pipe, banana effect of the coke drum and various temperature scenarios are also considered for overhead lines.

Should you not perform fatigue evaluation on this line?

Is temperature in the creep range for the material. CAESAR II 2017 (9.0) added creep evaluation per the EN code.

The banana deflection itself could be considered a one-time deflection and could be construed to have a higher allowable.

In such cases, I have treated them all as regular allowables with new piping to be on the safe side, but my settlement values were in the 4 inch range, which I'm given to understand that coke drums experiencing banana effects are much higher.

I would also give special considerations at adjacent shoe locations. If the pipe is kinking at that location, you might want to have a shoe design that's less likely to puncture the pipe.

Michael,
Banana effect is not one time. Banana effect is experienced in coke drum because of quenching which is batch process. Please correct me if my understanding is not correct.

Sorry,banana effect is permanent deformation that means it's one time deformation.

Dears,

I believe banana effect is cyclic, and I like the proposal to have a shoe that doesn't penetrate the pipe since one of the leaks is at a pipe shoe while the other is at a tee-fitting.

We tried adding a small loop between 36" header (collecting from 6 drums) and the 24" header since failure is on the 24" header. Stresses are lower and the line is much better but the problem is the client wants a fast solution without shutdown since the planned turn-around is after two years. I tried increasing the wall thickness at failure points and also changing supports configurations and locations but the best solution is the loop. Anybody have any ideas we can try to solve without a shutdown , cutting pipes and adding loop.

I would agree that the banana effect isn't completely one-time. There are likely smaller cycles that it experiences on a day to day basis that are difficult to quantify. However, the threat of permanent deflection exists as the vessel may have already yielded.

A photograph of the vessel when it is on versus off will indicate the permanency of the banana effect. However, assuming that it's cyclical isn't a bad idea, but this also means the vessel should be looked at so it itself doesn't fail.

You may want to consider what happens if it is now permanently in the shape that it's in, and ensure that there's sufficient make-up no matter what.

With that said, try replacing the shoe with a spring support. It can't puncture the pipe if it readily deflects. Your travel may be high, though.

I can't speak for the tee.

Hi Mostafa,

Is the leak at welds at tee and support shoe locations are already happening at site or as per stress calculation? If as per stress calculation I think you can use the creep evaluation in Caesar II since the temperature is somewhat high already, although is not a mandatory requirement in ASME Code but in EN Code is an option and if you fail you can do an elastic plastic analysis. I would suggest also not to use the liberal allowable since this is an indication that you have a constrained layout configuration which in creep regime can cause elastic follow up where elastic strains is converted to creep strains although the effect is long term weld embrittlement. Not sure nowadays if weld technology can address this weld embrittlement issue. Just take note that my suggestion of not using liberal stress is somewhat subjective and I think not mandated by Code. As we know "elastic follow up" phenomena is addressed through design by rules (what configuration will cause it such as small bore to large bore connection, etc.) as it's hard to predict it in design by analysis. Some design by rules based from experience of some clients is not use welded support pads and pad reinforced branch connection at creep regime (some say it starts at 427 deg. C, some say at 538 deg. C, some say 30% of melting point of metal) (high temperature design). You can check also client standards regarding high temperature design.

Also try reading "Coade ME News" Aug. 1992 where it has a very good explanation of the creep phenomena and what causes elastic follow up.

Please correct me fellow stressers in case there is a wrong statement I've made.

Cheers!!!

Based from the discussion, I noticed that it's already operating. How many years the plant is operating? In this case I think using loop to relax the stress at failed tee is a good solution. And maybe recommending to use a clamp shoe instead of welded shoe.

If it's already operating at more than 11 years, then maybe just replacing the failed tee and failed shoe with a new one will reset it's creep life to say another 11 years?...

Just my wild idea fellow stressers!!

Cheers!!

Anyway the work will still involve a hot work, by patching up the leaking tee and shoe. Might as well replace it with new ones, I think the work involve maybe almost the same and can be permanent solution resetting the life of tee and shoe to say another 11 years?

Thanks Borzki, the plant has only been operating for three years. I believe the main issue is that the line is overstressed at leak locations "shoe support and tee" so we will propose changing pipe material from A105 Gr B to A691 9CR since this is the material of the main 36" header. Also the additional loop is relaxing the line a lot since the relative movement between both headers is quite high. A good idea is the clamped shoe support but the pipe support standard does not have this at this temperature.

Hello Mostafa

There are a lot of standard and code prohibit welded shoes on high temperature like EN 13480-3 which talking about wleded pads for supports use until 300 c only so I think it's better to replace welded shoe with clamped shoe

You can also consult a metallurgical engineer if you have one in your company. He can do a forensic study for the root cause of failed tee and shoe. As we know, there are still other factors which can cause that failure such as weld quality, fabrication, etc. But this field can be performed by an expert in metallurgy who can detect what might cause the failure on a microscopic level.

As a stress engineer we can recommend (by taking a look at macroscopic level) to add loop and clamp shoe. Maybe this is a case of elastic follow up, but this is only one of the possible cause and I think it's hard to prove it, as elastic follow up is not a straightforward to do in stress analysis. Even performing an elastic plastic analysis will not be so accurate as it's hard to have a proper boundary condition.

Any other opinion is greatly appreciated.

Cheers!!