ASME Equation Used for Cylinder

[niravpshah]

I have a cylinder with the flanges welded at the ends. The cyliner will carry 900 degree air with pressure of 115 psi. We want to test this part in the lab and would like to know following.

I have used ASME material property for 2.125 Cr & 1% Moly for the calculation. Also the equations are from ASME section for calculating Circumferantial( t=PR/SE-0.6P) and longitudinal stress (t=PR/2SE+0.4P). I have got good factor of safety for the design. This part is not going to be in fatigue application, its just one time test. My wall of the cylinders are quite thick using the above conservatism.

My question is,

1) Do I have to use stress concetration Kt and multiply my stresses by that number?

2)If yes, what's the place to find the Kt based on my weld geometry?

3) For ASME equation, which has already lot of safety built in, Do we still need to use Kt?

4) If not whats the reason?

I have been hearing two sided opinion for this issue and I wanted to find out from experts the fact.

Thanks in advance for your help.

 

[jte]

niravpshah-

1) No.
2) Although you don't have to, you would find guidance in VIII-2 App. 5.
3) See 1).
4) Because it would be silly to use a fatigue criteria for a single cycle.

It would help if you mentioned more details of your design. Are you sure its 2.125 Cr and not a 2.25 Cr material? I'm presuming you have a cylinder with weld neck flanges on either end, perhaps with blind flanges bolted to the WN's. You could also be using slip on flanges which would be adequate for non-cyclic service.

With WN flanges, you'd have no stress concentration at the weld since it is a butt weld and you can simply grind it smooth or, better yet, grind it flat. With SO flanges you would have two fillet welds which would not be good in cyclic service (see VIII-2 A5-112).

jt

 

[unclesyd]

Come back with a little more information on the physical dimensions and materials.
You might need to do a little with your fabrication.

Dia.
Wall thickness
Cylinder material
Flange material

I dislike one time laboratory cylinders as they never get thrown away.

 

[niravpshah]

Thanks guys for great response.

The material for the cylinder is SA387 Grade 22. It's 2.250 Cr (Jte: Sorry for typo earliar) and 1% Moly.

I wasn't very clear on my geometry information. We have 17 inch OD cylinder (A)of 50 inch length and on the ends we have flange welded. I understand from the above response that for full penetration smooth weld and weld neck flange, no Kt is required For cylinder A flange conncection.

We also have 24 inch OD of cylinder of length 15 inch that goes on top of the smaller cylinder and gets welded by vertical manifold on both sides.



Cyliner B
|--------------|
| | Vertical Manifold
| |
--------------------------------------------
Cylinder A



---------------------------------------------

If I have full penetration of weld between Cylinder A and Cylinder B, Do I need to use Kt?

Again, Thanks for answering my first question for weld neck flange to cylinder A.


Unclesyd, I agree with you for your lab test comment. We may end up running as many as 10 tests (Cycles).
I believe its still low number for going for LCF. Whats your thought for Kt?


For vertical manifold I have used flat wall equation from ASME spec, trc=.707j*SQRT (P/S). If the weld of vertical manifold to cylinder A is full penetration,

Q1) For that joint do i need to use Kt for 10 cycles.

Q2) If no, where can I find the explaination?

Q3) If yes, what's the best resource for finding Kt value without going for FEA?



Jte: I have question for you.

You earlier mentioned that for static loading you don't need to use Kt. Cosider this scenario,

1) You statically load the bar axially (Pure tenstion) to its failure (Rupture).

2) Now, the same bar with a notch in it and load the same way and get the failure point (Rupture).

In both cases you calculate nominal stresses from standard equations. Since you have same nominal stress in the bar, Shoudn't you get the same number for the failure? It may be different if we try to calculate the stress in the notch area. And for that you would want to use Kt.

Awaiting for you response.

Thanks again for great helpl guys.

 

[jte]

niravpshah-

I don't quite understand the "vertical manifold" construction and flat wall equation. Do you have a 17" OD (not exactly a standard pipe!) connected to a smaller (how small?) vertical pipe which connects to the 24" pipe above? In which case you'd design the vertical connecting pipe as a typical nozzle, including req'd reinforcement, per VIII-1 UG-36 and -37. See also App 1-7 if the nozzle is relatively large.

As to defining cyclic service, take a look at VIII-2 AD-160.2. Chances are you qualify under Condition A (at least the steel does!). If you do, then you have a legitimate basis for not getting into cyclic design.

Your question regarding uniform and notched bars in tension: Simple - you just have to read up on primary and secondary stresses in VIII-2 Appendix 4. In particular, AD-130 etc. You'll find the rest of Appendix 4 and Appendix 5 interesting also.

The scenario you proposed puts the bar in primary tension in a manner which allows for no redistribution of the stress - and then takes it beyond yield. This would violate the limit of Pm<k(Sm)even with k set to 1. Since you state that the same nominal stress exists, I'm presuming that the notched area has the same cross sectional area as the unnotched bar for your P/A tensile stress to be equal. If your notched area has a lower A then of course it will fail at a lower applied load.

Generally speaking, for vessels you tend to find stress concentrations at gross structural discontinuities which tend to be able to redistribute stresses which exceed yield. Thus the limit of 3Sm (=ultimate tensile, well beyond yield) for primary and secondary stresses combined. When you get into cyclic analysis, you'll be looking at the peak stresses and combining those with the other loadings to compare against the allowably alternating stress, Sa.

If you don't have Div. 2 available, all of the above will be difficult to follow. I'd suggest getting a copy and getting real familiar with Fig. 4-130.1.

jt

 

[SnTMan]

niravpshah,

Not clear to me what ASME Div. you are designing to, but I hope you will follow rules for welding 2 1/4 Cr matl. It can be kind of troublesome.

I kind of got the impression you applied a FOS on top of Code allowables. If so, not necessary, FS are provided by the tabulated allowables, 3.5 in Div. 1.

Not too familiar with Div.2, but Div. 1 design rules recognize local areas of stress concentration at discontinuities and design rules allow for them.

I agree with jt you will find specific rules in Div. 2 whether or not fatigue design applies.

It might pay do some more studying of Code rules before prodeeding.

Attempting to be helpful!

 

[niravpshah]

Jte and SnTMan,

I really appreciate your responses. I have mangaed to get ASME Div VIII and now the whole world is crystal clear.