ASME VIII Division 2 - Fatigue Analysis Article 5.5.2.3 1

[Morts81]

Forum Members

With regards to ASME VIII Division 2 Article 5.5.2.3 Step 4 (a) & (b). This relates to thermal cycles in vessels and determining whether a fatigue analysis is required. (For the purpose of this question, I'm ignoring pressure cycles)

I'm trying to understand whether this temperature differential is the final steady-state differential or at any point in the transient period.

Our main concern right now is that we have a 12mm thick S/S vessel (insulated) that could theoretically have LNG introduced at -160degC while the vessel has been sitting at ambient for a period of time. This is likely to occur a few thousand times over the vessel's life cycle.

Intuitively, if the vessel shell is at 20degC and the LNG is introduced, there will be a very short period of time before the through-thickness of the shell reaches steady state. Is it the intent of ASME VIII Division 2 to account for this short transient thermal cycle or should I just be analysing for the steady state case, in which case I don't envisage any differential above 28degC and therefore according to table 5.8 my effective cycles would be zero.

Thanks in advance for any help. Very much appreciated.

Morts81

 

[XL83NL]

I dont believe one needs to account for transient cycles, however there is a means in ASME VIII-2 (if I recall correctly) where temperature differential has an impact, i.e. a delta T of say 180 C (from 20 to -160 deg C) has to be accounted more than a delta T of say 140 deg C (from -20 to -160 C).

Other (committee-)members like TGS4 can and probably will provide you with a better answer, so dont put your bets on me .

 

[davefitz]

I am not qualified as a sect VIII div 2 fatigue analyst, but fatigue damage is independent of time ; it is instead a function of how many times the severe temperature differential occurs. Even if the differential only occurs for 0.1 second it still counts as a fatigue cycle.

"Nobody expects the Spanish Inquisition!"

 

[pdiculous963]

I agree with davefitz, transient effects must be considered. The firm I work at performs transient thermal mechanical stress analysis when a fatigue analysis is required due to thermal cycles.

 

[XL83NL]

How can you account for transient effects? Either a cycle is there, or it isnt. Does it matter of It lasts 10 or 12 seconds?

 

[TGS4]

It is intended account for short transient thermal cycles. It is the instantaneous temperature difference that creates thermal stresses, regardless of the duration of the temperature difference. Your situation is indeed the case that this was intended to cover.

Without performing a transient heat transfer analysis, I would use a delta-T of 180°C in your Method A calculation - which would result in a factor of 8. Your calculation will definitely fail the Method A screening. And, such a conservative assessment may fail a Method B screening, as well. Based on my experience, you will need to perform a transient heat transfer analysis to justify any non-conservative metal temperature differential.

Good question - thanks for asking.

 

[pdiculous963]

No it does not matter how long a loads is applied a fatigue analysis, as davefitz mention. I was referring to the transient temperature distribution (i.e. not steady state), which would cause stresses in excess of the steady state stresses in your situation.

 

[Morts81]

Thanks a lot for the responses guys, they are all very much appreciated.

One further question on this issue.

We have a project specification for cyclic loading and it states (verbatim):

A "cyclic service" is defined as one in which the equipment will be subjected to periodic changes in stress (due to variations in temperature, variations in pressure, or both) which would require integral construction by the rules of ASME VIII, Division 2, Paragraph 5.5.2.3.

Now, my understanding is that the author of this statement has misinterpreted Paragraph 5.5.2.3. The screening process in 5.5.2.3 determines whether a fatigue assessment is required. An input into this screening is whether the equipment is of integral or non-integral construction. In fact, the fatigue assessment can be carried out on a non-integral vessel and show that the equipment is suitable for the design cycles.

I just want to ensure that my understanding is correct and there there is no 'cut-off' in Div 2 where one must use integral construction in lieu of non-integral?

For example, if I look at Clause 4.2.5.5 for Category D Welds, I believe that as long as my fatigue assessment is successful then re-pads are acceptable?

Note that I'm not designing a new vessel here, I'm trying to determine suitability of an existing piece of kit and we will engage an experienced fatigue engineer in due course but I'm trying to do some gap analysis for management in the short term.

Regards

 

[Morts81]

Sorry guys

Just to confirm, the vessel I am looking at is an ASME VIII Division 1 vessel.

My undertanding is that Div 1 does not expressly cover cyclic loading and fabrication/design requirements and that the cyclic fatigue analysis in Div 2 is generally used in this case. Based on this, are the fabrication/design requirements stipulated in Div 2 mandatory?

Thanks again

 

[TGS4]

Morts81 - you are correct on all accounts. The screening criteria in 5.5.2.3 is suitable for VIII-1 construction as well. It is extremely conservative - if it passes, then you can be assured that there is unlikely to be fatigue damage in your vessel provided that you have evaluated all of the cyclic stresses.

Since this is an in-service vessel, however, in the event that the screening criteria is not achieved, I would NOT recommend doing further fatigue analyses. Rather, you should be performing a fracture-mechanics assessment to determine inspection locations, techniques, and frequency as described in this blog post (not my post, but a co-worker's):

http://becht.com/blog/use-of-fracture-mechanics-for-fatigue-design

 

[Morts81]

Thanks TGS4. This is a weird circumstance where we haven't actually started up the refinery yet so this equipment is installed but not yet commissioned. It's not expected to start up until end of year, maybe early next year (we all know how schedules go).

The plan going forward was to carry out the fatigue analysis and determine whether it was suitable for the 30 year design life. If yes, then we are good to go. If no, then we would look at sourcing a replacement vessel. Considering it is 15m tall by 3000m diameter we aren't going to get a replacement in time so the intent would be to determine an operating philosophy in the mean time. The 400 cycle limit in ASME Div 2 would take us to 2-3 years so we have time to put a plan in place.

One step further from this could be the fracture mechanics analysis you recommend and determine an inspection regime to go along with this temporary operating philosophy for extra peace of mind.