ASME SECT VIII DIV 2 Fatigue assessment & adjacent points 1

[andyross333]

I am trying to get clarification on considering the thermal gradient between adjacent points using the fatigue analysis screen Method A of ASME SECT VIII DIV 2 Section 5.5.2.3. I understand in general how to calculate the temperature difference but I am unclear in a certain number of situations. The vessel is a horizontal drum with several step changes in ID. The ends are capped with flat plates.

1) For the drum in the meridional direction, using L=2.5sqrt(Rt), how do I account for a step change in the ID? In thread794-300509 there is mention that the average value of Rt should be used when Rt is not constant. Is this calculated once across the entire drum? We have significant changes in ID but in the area we are most concerned about, the thickness remains constant. Do we use Rt just for that local area or across the entire drum?

2) For the flat plate end cap, I intend to use L=3.5a with 'a' being the ID of the drum at the end cap. Is this correct?

3) Given a point on the flat plate near the drum, do I use 3.5a to consider the second location on the drum? In my case, 3.5a is significantly larger than 2.5sqrt(Rt).

Sorry if my explanation is not clear enough, I cannot upload model images as we are under NDA. I will attempt to clarify if need be.

Thanks in advance,
Andrew

 

[TGS4]

1) The local area of interest.

2) No. "a" is defined as the radius of hot spot or heated area within a plate. Do you have that?

3) Not sure what you're asking here. Are you wrapping around from the flat plate onto the shell?

 

[andyross333]

1) Thanks, this makes more sense to me. Over a length that crosses a step in ID, do we take the larger or smaller of the two radii or average?

2) The flat plate is a bolt flange at the end of the drum. It is heated from the inside of the drum. So as I understand that definition, the heated area is the ID x flange thickness. Correct?

3) Yes, I'm wondering if you need to consider wrapping around from the plate to the shell. You said it better than I did. So far, I've been taking adjacent points in the meridional, circumferential and radial (through thickness) directions but was unclear whether it is required to take points in other directions. See the attached sample image for clarification.

Thanks

 

[TGS4]

1) Average

2) Not quite. The heated area is the heated area. Is the entire ID of the vessel heated?

3) I tend to NOT wrap around corners of different geometries (nozzle/shell).

 

[andyross333]

Thanks for your guidance, great responses. That cleared everything up.

With respect to the flange, the maximum temperature gradient across it in all directions is so minimal anyways (~20C) that it is no longer a concern.