Ordinary Stress Intensity Plot to Qualify Vessel 1

[LS_SMS]

A colleague shared an interesting twist with me today and I'm not sure if it's sufficient to qualify a vessel. They were doing Sec. III work. They plotted stress intensity on their ANSYS model. It was not a "linearized" stress intensity plot; instead, it was just an ordinary stress intensity plot. They said that because the stress intensity everywhere on the model showed it was below the S Code limit, the vessel therefore qualifies. In other words, everything in their plot was below the Primary Membrane limit, S, which is the most limiting qualification, which means the vessel therefore passes all qualifications (i.e., general membrane, local membrane, primary membrane + bending, primary + secondary).

They never inserted any stress cut lines to do linearization and touted this method as a shortcut to avoid having to create many linearization paths throughout the vessel. This sounds like it could be a viable shortcut, but I have never heard of this approach before. I wonder if regulators would buy into the story if they saw this in a report. Can anyone shine some light?

 

[centondollar]

What do you mean by "ordinary stress intensity plot"? Is this a steel pressure vessel or something similar? If so, the von Mises stress would be the interesting parameter (it considers all stress components) and it can be directly plotted in many softwares for e.g., each element integration point or element mid-point.

 

[r6155]

1) Nuclear is a serious issue. Experience is required in design, fabrication and inspection.
2) Contact a pressure vessel engineer.

Regards

 

[DriveMeNuts]

Did you actually linearise the stresses to verify your colleague's claims?

 

[LS_SMS]

What do you mean by "ordinary stress intensity plot"?

Hidden inside my question was the ANSYS nuance. My apologies. I kind of assumed/hoped that an ANSYS user would have read my question and understood that there is a difference in ANSYS post-processing, in terms of plotting ordinary stress intensity versus linearized stress intensity. To clarify, ordinary stress intensity is merely a plot on the exterior skin of the model. In other words, the stress intensity is calculated from the outermost nodes/elements and plotted on the surface. In essence, you are not given any glimpse of the through-wall stress response. Conversely, linearized stress intensity is along a user-defined path that cuts through the vessel wall. You get to see the stress response through the wall along the stress cut line.

Is this a steel pressure vessel or something similar? If so, the von Mises stress would be the interesting parameter (it considers all stress components) and it can be directly plotted in many softwares for e.g., each element integration point or element mid-point.

Yes, it's a steel heat exchanger. von Mises doesn't apply for most Sec. III work (von Mises is Sec. VIII work). Rather, stress intensity (derived from Tresca) is the governing metric in Sec. III.

At the heart of my question was...do we linearize along a stress cut line or not? My colleague proposed a situation where he believes stress cut lines are unnecessary, therefore saving time and schedule.

Contact a pressure vessel engineer.

I am one. However, I don't claim to be omniscient. Therefore, when faced with a new pressure vessel situation that I've not seen before, I sought to ask other folks in this forum.

Did you actually linearise the stresses to verify your colleague's claims?

No. It's his model and he hasn't shared it with me. We have only had a few web meetings to glance at a couple things here and there. It's a commercial project with typical time and budget constraints. I told him I had never heard of this and urged him to test his theory before betting the farm on it. Whether or not he does that, is on him. I personally am uncomfortable with his hypothesis and would not likely proceed with it, if in his shoes.

 

[r6155]

Do you have ASME III?
Making calculations without knowing the manufacturing and inspection details is not a good idea for ASME III.
No forum replaces a training course.
No offense.
Regards

 

[LS_SMS]

Do you have ASME III?
Making calculations without knowing the manufacturing and inspection details is not a good idea for ASME III.
No forum replaces a training course.
No offense.
Regards

Geez. You replied twice and had nothing meaningful to share. Yes, I realize the Code document is the governing authority. I can crack it open and it's going to tell me to linearize the stress. So if I follow the exact letter-of-the-law, then I shall linearize, just like always.

On the flipside, I was wondering/inquiring if my colleague revealed something I was overlooking, since I confess I'm not omniscient. In other words, was my colleague saying "there's no need to do all that because you already proved it is below the allowable."

Yes, I've had formal Sec. III training.
Yes, I've worked in nuclear a number of years.

 

[TGS4]

Here's something to consider - you may have been shown a surface stress intensity plot on the outside, but were you also shown a similar plot showing the inside?

Remember that a surface plot of stresses (stress intensity in your case) is a proxy for the membrane-plus-bending stress intensity. Unless you are simultaneously shown the inside surface, you really don't know what the value of the membrane stress intensity is (as you noted - without doing a formal linearization).

That said, if both the OD and ID surfaces had all stresses less than S, than it is much more likely that the membrane stress would also be less than S, and applying some engineering judgement could avoid extensive stress linearization.

 

[LS_SMS]

Here's something to consider - you may have been shown a surface stress intensity plot on the outside, but were you also shown a similar plot showing the inside?

Remember that a surface plot of stresses (stress intensity in your case) is a proxy for the membrane-plus-bending stress intensity. Unless you are simultaneously shown the inside surface, you really don't know what the value of the membrane stress intensity is (as you noted - without doing a formal linearization).

That said, if both the OD and ID surfaces had all stresses less than S, than it is much more likely that the membrane stress would also be less than S, and applying some engineering judgement could avoid extensive stress linearization.

Great feedback! Yah, that's kind of the angle my colleague was taking. The inner and outer plots are only proxies. But he hypothesized that if it showed everything below S, there is no need to linearize.

Hmm, at a minimum, it's an interesting twist on vessel design that I've not heard before.