ASME Section VIII - Cyclic Loads

[UKCats]

How does ASME Section VIII handle cyclic loads. Does it use lower allowable stresses??? or in other words, apply a penalty.

I have a vessel that is steam sparged and it shakes. It has failed..cracked. To replace it, I want to apply a lower allowable stress based on fatigue which would lead to a thicker shell, but I cannot find S-N data (fatigue data) for Alloy 825 material to determine my new allowables...so I was wondering if Section VIII just adds a reduction factor....like ASME B31.3 Process Piping code.

ASME B31.3 uses a deduction factor (or penalty) based on # of cycles...I don't see the same methodology in Section VIII.

 

[prex]

In ASME VIII Div.2 (no rules in Div.1, Div.2 used when necessary) the fatigue analysis is to be conducted with a quite detailed method.
The important point to note is that fatigue strength is normally governed by design details (especially nozzle connections, flanges and supports) rather than by the thickness of shell walls. I suppose you didn't get the failure in plain wall (BTW: what kind of rupture did you get?).


prex

http://www.xcalcs.com

Online tools for structural design

 

[UKCats]

The failure was at the toe of a filet weld. The vessel shell is Alloy 825 and the support lugs are carbon steel. The leg that is in contact with the vessel has developed a crack at the toe. On the new vessel, I want to build up the weld, then feather it down by grinding to eliminate any stress riser at the toe. I also want to remove the CS support lugs and go with matching base metals 825/825.

Are you saying I don't need to go with a thicker shell???

Should I still put a 825 pad between the vessel wall and support lug???

One last thing, I don't understand how there are no rules in Div. 1 for cyclic loads... UG-22 says they need to be considered.... is it that Div. 1 gives no other guidance than that???

 

[jte]

UKCats-

I'll go backwards with your questions: Div. 1 does not give guidance for cyclic loads. From the Foreword: "This Code contains mandatory requirements,... The Code does not address all aspects of these activities... The Code is not a handbook..." Division 2 is most frequently used for guidance on fatigue evaluation (though not design).

As prex stated, the key to fatigue design is close attention to details, not a brute force "make it thicker" approach. I will not allow fillet welds in any vessel I'm involved with which is in fatigue. They crack. Depending on your vessel size and type of fatigue (thermal, pressure, and the rates of change), you might want to consider going to a ring support if the vessel is supported on sides or a skirt if the vessel is supported from the bottom. If you still stick with legs, put full penetration welds in and, as you noted, grind them smooth.

jt

 

[prex]

Agreeing with jte, I will add some more thoughts:
- how to avoid the same problem in future constructions critically depends on type and reason of failure. To make a judgement I would need to know the exact geometry and loads.
- from what you say I understand the ruptured weld is the weld of the support doubler plate to the shell, so the failure doesn't seem related to a differential expansion fatigue, more likely to an excessive mechanical load. In the first case the use of 825 for lugs would be beneficial, in the second one you need to make the support more robust and perhaps less subject to bending moments.
- yes I definitely suggest to have a doubler plate between the support lugs and the shell. Increasing the shell thickness could be beneficial (though this is not granted), but don't think it is a good idea, unless bending stresses due to support loads are high in the shell.
- as jte states, fillet welds are generally not good for fatigue, though they still need be used even under high fatigue. Of course however they must be continuous and subject to moderate loads. Rounding and flattening the weld may also be helpful.
- for the lug to shell attachment, if you are really in a critical situation with high loads, you could consider a high quality (and costly) solution: make an insert into the shell plate with a plate of higher thickness (double?), full penetration welded to shell.

prex

http://www.xcalcs.com

Online tools for structural design

 

[davefitz]

The above replies are correct. The fatigue analysis , using any credible method, will involve your determination of the maximum alternating stress or strain , which in the case of a vibrating sparger will involve a vibration analysis and probably detailed finit element analysis. The vibration analysis can have its accuracy improved if you have field test data from a prior sparger prior to its failure.

The method of modifying the design to preclude failure from high cycle vibration fatigue will likley involve both reducing the vibration potential ( either structurally by reducing teh spacing between supports or modifying the process of sparging) as well as improving the weld detail, liklely replacing the fillet weld with a full pen weld or some otehr detail.It is NOT solved by making the vessel thicker.

 

[UKCats]

First of all, it sounds like when I re-build the tank, I need to have a fabricator familiar with Div. 2 rules if I want fatigue reliable analysis performed...right??? Also, a Div. 2 fabricator usually is stronger in tools like FEA and can design an better support lug...right??? Each support lug has a base plate and 3 perpendicular gussets...looks very typical to my untrained eye.

Second, the vibration is due to the collapsing steam on the inside of the vessel. IF the vessel steam nozzle suffered from fatigue cracking, I would agree to stiffen the steam line... but since that's not where the cracks are, I don't think pipe supports will help. I'm thinking of putting an internal stray nozzle at the end of the steam pipe to due finer sparging, and therefore, perhaps, less dynamic loads due to collapsing steam. Any thoughts???

Third...I'm not familiar with vibration analysis...or how to approach this problem. Is there a software program or standard to help me. I know CaesaerII can do dynamic modelling of piping...is there something similar for vessels??? I'm ignorant in this area.

Fourth... Currently, I do not have a pad between the 4 carbon steel support lugs and the Alloy 825 shell. I assume the doubler plate you're referring to is what I call a pad??? How do I get rid of fillet welds??? I understand I could cut a hole in the vessel wall and insert a plate and full-pen weld it in place...however, if I just back bevel the pad, and weld it on, is this enough penetration??? or should I go thru the vessel wall.

Fifth/Lastly... I contacted, Special Metals (INCO), a major supplier of Alloy 825. They said nickel welds don't wet very well like stainless steel, in that, at the toe, they don't feather out well. They also don't recommend leaving a concaved weld profile, in that nickel will tend to crack at the throat. So I was going to oversize my convex weld, then grind it down in a smooth concaved profile. According to them, this will greatly reduce stress risers in the weld. thoughts???

You have been great.

 

[prex]

I don't think that your key problem is having a Div.2 fabricator, as in fact you don't know (and probably will never know) the amplitude of the vibrating load: FEM is fine, but if you can't determine the loads...
If you add a pad (or doubler plate) you'll immediately get a beneficial effect: if that weld brakes again, you'll be able to repair it with no worries about the vessel.
Also: if the broken weld was due to high bending stress in the shell at the lug junction, increasing the shell thickness with a pad will correspondingly decrease the bending at both the lug attachment and in the shell (at pad periphery).
I'm not pushing you towards the solution of the shell insert: I think that a pad continously welded with a fillet all around will be OK; it will be at least a forward leap in the strength of the vessel with respect to present situation.
By the way I consider a mistake not having provided a pad from the start: this is current practice in vessel design, especially with dissimilar materials, and also for the most current vessels, where fatigue is not a worry.

prex

http://www.xcalcs.com

Online tools for structural design

 

[jte]

UKCats-

I mostly agree with prex's above reply. FEA/FEM is fine - but there's a tremendous amount of interpretations as to what needs to be analyzed. If you don't know what you are doing with the FEA tool, you might be getting very little value from it. A Div. 2 fabricator is not your only option. All you need is a consulting engineer or firm who has good familiarity with designing fatigue vessels. Consultants and good engineering are expensive - until you consider the cost of unplanned shutdowns due to fatigue failures.

Unlike prex, I would strongly push you towards using insert plates instead of re-pads. As I mentioned before, I don't allow fillet welds on vessels in fatigue.

As to the dissimilar metal concerns: Have you considered making the first foot or two of the columns (or the entire column) out of 825? Of course, I'd still push for losing the legs and going with a skirt (again, at least for the first foot or so). What you wind up doing is placing the dissimilar metal weld out of the pressure retaining envelope. This has three significant advantages: First, a crack is less likely to develop and propagate since the weld is removed from the energy causing the failure. Second, if a crack does occur, it doesn't threaten you with a loss of containment (provided, of course, that the vessel doesn't fall over!). Third, repairing the support structure for a vessel should not require jurisdictional or AI authorization and will thus be significantly easier.

jt