Section VIII Division 2 Part 6 Table 6.1

[rickfischer51]

Does anybody have any experience with the first item in Table 6.1? I'm having trouble figuring out the definitions for the diameters and how they relate to the part I'm analyzing.

 

[EnergyMix]

I'm at home and don't have access to the spec and am not sure I'm going to have time at work tomorrow (my days have been pretty crazy busy lately.) Then it's the weekend, and next week is a holiday week, and likely to still be very busy. Can you provide a bit more information. Otherwise, if I have time to look it up, I will, but absolutely no promises.

Want to know the do's and don'ts of Eng-Tips? Read FAQ731-376.
English not your native language? Looking for some help in getting your question across to others or understanding their answers? Go to forum1529.

 

[rickfischer51]

The parts are components of a superconducting radio frequency cavity made from high purity niobium. One piece is half a toroidal shell, sliced on a plane perpendicular to the axis of revoulution, with a pair of ports formed into the end of it. The other is a circular drawn dish. Both parts have surfaces that are curved in two perpendicular planes, so they sound like the "one piece double curved" description in Table 6.1. They are stamped on matched metal dies. I can get the original blank size and the dimensions of any clamping features in the dies. The formula in table 6.1 references Do and Db. Do is the "original outside diameter", which sounds like the outside diameter of the flat blank. Db is "diameter of the blank plate or the diameter of the intermediate product," which also sounds like OD of the flat blank. I would expect the largest strain to be at the smallest radius, so if I take the smallest radius and combine it with any other diameter in the formula in the table, I get crazy high (300%) strains. If I use the formula for a cylinder and apply it to the smallest radius as a check, I get 1.5%, which seems much more reasonable. We've been fabricating stuff like this for decades and they work just fine, so I'm resonably sure the 300% strain number is hosed.

I suspect that this formula is intended for a specific type of geometry and may not apply to my case. A picture or two in the code would go a long way to clearing this up, but I don't see one. As an alternative, I am looking at calculating two strains, one for each radius, with the cylinder formula and treating them as principle strains and calculating the equivilent strain. I need this for an elastic-plastic local failure assessment 5.3.3.1, which is based on equivilent total strain. This approach is not rigorous, but produces convienently believable numbers. How rigorous is finding a largely plastic strain with a simple hand calculation? I assume the cylinder formula in Table 6.1 gives uniaxial strain (hoop strain). So the code is not rigourous either, mixing equivilent and uniaxial strains. I have a little freedom here. The cavity is vacuum loaded, triple protected from accendental over pressure, and will not be code stamped. We use the BPVC for lack of anything better. It's recognized that Niobium is not a code material, and we do a fair amount of conjuring and hand waving to get this far.

Any help or insight you can give would be appreciated.

 

[EnergyMix]

Ok, now I'm really scratching my head. Why are you using the nuclear code if the material isn't found in the code? Do you perhaps mean Section VIII?

Want to know the do's and don'ts of Eng-Tips? Read FAQ731-376.
English not your native language? Looking for some help in getting your question across to others or understanding their answers? Go to forum1529.

 

[rickfischer51]

Oh krapski! I typoed. Should be Division VIII. We use the BPVC because it is dictated by the laboratory safety committee. There is other equipment here that is covered by the code and the safety committee is familiar and comfortable with it, and so for lack of anything better, we use the code here also. We are using Design by Analysis per Part 5 because Design by Rule results in cavity walls that are too thick. Thick walls will result in warpage when the parts are welded together, and we need to maintain our dimensions very closely. The shape determines the resonate electomagnetic frequency, and even very small deviations can cause unacceptable frequency shift and increase power consumption which can get expensive very quickly.

 

[TGS4]

If you really need an accurate number for the forming strain, then you should probably do a simulation of the forming process. For 5.3.3.1, you may have low forming strain exactly where you have your high triaxiality, or maybe vice versa. I do not think that the values in Table 6.1 were intended to bemused in the local failure check - as you said that may he mixing uniaxial with equivalent and/or multiaxial. If it's really important for your complicated part, calculate it yourself.

 

[rickfischer51]

The proceedure for Elastic-Plastic analysis at 5.3.3 says "determine the forming strain .... in accordance with Part 6" at 5.3.3.1.d. I tend to want to stick as close to the code as I can, as the reviewers get nervouse any time they see a deviation. Besides, tooling drawings are not available, and a forming analysis is ouside the scope of this project. Preliminary results show the strains to be low compared to the allowable, so a lot of accuracy is not needed. If I'm calculating this correctly, the uniaxial strain appears to be conservative compared to the equivilent total strain, which would justify using the strains from Table 6-1 in 5.3.3.

 

[TGS4]

Sounds reasonable to me.

 

[EnergyMix]

(By the way rickfischer51, after you did the proverbial hand upside the head, I red flagged your post and asked the system administrators to change the title to Section VIII, which they promptly did. Problem easily resolved and you know how to fix it for the future.)

Want to know the do's and don'ts of Eng-Tips? Read FAQ731-376.
English not your native language? Looking for some help in getting your question across to others or understanding their answers? Go to forum1529.