Thermal Cycling

[Sail_Fish]

Hello,

I have a vessel (drier) designed to ASME VIII Div 1 that is undergoing daily temperature swings from 100F (38'C) to 380F (193'C) within half an hour. The design conditions are 260 PSIG @ 500F.

Is there any guidance on allowable temperature swings before you need to assess for thermal fatigue?

 

[davefitz]

A simple approximation of the magnitude of the thermal stresses that occur due to cyclic temperature swings can be found in the formulas of EN 12952-3 annex C, and similar formulas in the prior german boiler code trd 301 . knowledge of the material's thermal diffusivity, wall thickness, UTS, yield stress, and stress concentration factor of the highest stressed local part of the cylindrical vessel is needed. If the ramp rate of temperature caused a significant loss of fatigue life over the design life of the vessel, then a more sophisticated fatigue analysis as per sect VIII div 2 , usually an elastic plastice finite element analysis , would be recommended. It may be less costly to alter the design details of the highest stressed part than to contract for a fatigue specialist to provide the analysis.

"...when logic, and proportion, have fallen, sloppy dead..." Grace Slick

 

[TGS4]

I disagree with davefitz that an elastic-plastic fatigue analysis would be needed. For drivers like this, the cycle count is so high that the stresses need to remain well in that elastic range.

In VIII-2 there is fatigue screening Method B available. Try that to see if you pass. What matters is the temperature gradients - if the temperature rise and fall are gradual, you may not have any issues.

Still, a thermal fatigue analysis is not particularly expensive...

 

[Sail_Fish]

Thank you for the replies.

I took a look at the screening method B (5.5.2.4) and it looks as though I am interested in Steps 5 - 8. correct?

In steps 5 - 7 it asks for a maximum temperature difference between to adjacent points, with the adjacent points calculated using L=2.5SQRT(Rt).

Would this not assume I have a circumferential mal-distribution between two locations on the drum? I am trying to assess the drum heating and cooling equally over the entire drum. Am I missing something here?

 

[Sail_Fish]

Thank you for the replies.

I took a look at the screening method B (5.5.2.4) and it looks as though I am interested in Steps 5 - 8. correct?

In steps 5 - 7 it asks for a maximum temperature difference between to adjacent points, with the adjacent points calculated using L=2.5SQRT(Rt).

Would this not assume I have a circumferential mal-distribution between two locations on the drum? I am trying to assess the drum heating and cooling equally over the entire drum. Am I missing something here?

 

[TGS4]

Not necessarily. Do you understand the temperature profiles? You may need to take temperature measurements or perform some thermal FEA. Most likely the issue will be at your inlet nozzle.

 

[Sail_Fish]

Yes I have temperature probes on the drum with an accurate temperature profile. Just to stress there is no issue with the drums, such as cracking etc. The question was asked because the drums operate in 'cyclic' service, albeit low frequency (they heat up, steady state and cool down over a 48 hour period), do we need to assess for thermal fatigue.

Appreciate the help by the way.

 

[TGS4]

If you have good temperature profile data, then you should be able to complete a Method B Fatigue Screening assessment.

 

[racookpe1978]

Not answered (yet!) is the thickness and length of the metal being cycled: Implicit in each comment of the "thermal profile". A long narrow PV may be "ok" across the middle of the PV wall, but the two ends or a thickened flange at the center may be overstressed.

Does the whole vessel cool down to the 39 degrees, or does most stay hot when the vessel is opened and refilled? Is the interior rapidly chilled when it is refilled with cold ? product, or does the inside wall slowly cool?

 

[Sail_Fish]

OK I have attached the temperature profile of the drum. Hopefully it is easy enough to understand. Green Line is the top head temperature and the blue line is the bottom head temperature.

Thickness of the drum is 32mm with a total height between Tan's of 6100mm. ID is 1675mm. Heads are 2:1 Semi 33mm thick.

 

[TGS4]

You could try the screening with the maximum delta T in your plot. If it passes, then you're done. If not, you will likely need a fatigue analysis. That's a very steep gradient on the cool down, which is, in my experience, likely to cause large thermal gradients and hence large stresses.