Reinforced Flathead 1

[jhollin1138]

The company I work for primary product line is the manufacturing of butterfly valves. I need to design a pressure test chamber a hi-performance, zero leakage valve we are currently building. The chamber will consist of a rolled shell and a flathead. The chamber will then be connected to the butterfly valve and valve’s disc will basically act as the other head. Due to the size, about Ø82”, the test pressure and since we are a Code shop I am designing the chamber per Section VIII, Div 1.

I reviewed UG-34 and I am connecting the flathead to the shell as shown in Fig UG-34(f). When I calculated the required thickness for the flat head, using C=0.2, it worked out to be about 1-1/2" thick. Not very particle for a test chamber I am going to use 1 time. So I decided to use a waffle/honeycomb grid to minimize the thickness required. The problem is I cannot figure out how to calculate the reinforced flathead per Code. I reviewed UG-47, but that really doesn’t apply. Can anyone point me to the correct section or offer and advice?

 

[codeeng]

Go out and buy a thin dished head.

 

[jhollin1138]

So are you saying it can't be done? I realize I can use a dished head, but why do that when I have plenty of scrape in different thicknesses (no 1-1/2") laying around that I can use. Plus, when I am done with my testing, I can reuse some of flat sections if needed. I have no need for a scraped Ø82" dished head.

 

[waskillywabbit]

Sounds like it would be more appropriate and economical to design a pressure test chamber that could accomodate the product line of butterfly valves and not just one.

OR

just get the 1.5" plate and add it to the scrap pile when you are done.

Seems like there are several threads on this topic in the forums here.

Good luck.

Brian

 

[waskillywabbit]

Update:
It would seem that rib stiffened flat heads are not allowed by Div 1.

Try this post:

http://www.eng-tips.com/viewthread.cfm?qid=93914

Brian

 

[codeeng]

You can calculate each stiffener as a beam including a portion of the cover. By the time you cut all these stiffeners and weld them all together you'd be better off with a dish.

 

[jhollin1138]

Ok, thanks for everyones time. I'll look into going with the dished head vs. the price of a 1-1/2" thk. plate.

@waskillywabbit: The thread really doesn't say ribs aren't allowed. It has just as many people saying "yes" as it does saying "no."

 

[waskillywabbit]

"It would seem..." I generally tend to stay away from things the code does as well...the fact that they don't discuss it in detail or detail how to do it speaks volumes to me.

Good luck.

Brian

 

[codeeng]

You can use anything you like for a hydrotest blank. This is not part of any Code.

 

[unclesyd]

If you price a blind flange for your test make sure you price a "plate flange".

Also you may have the option we had when we had some special valves made. We purchased the 2 Blind flanges used to test the valves, the purchaser might need the flanges.

There is an easy way out for a one time operation.
Rent one, flange that is.

This is just one of several companies that rent blind flanges.

http://www.usaindustries.com/pipe_blinds.htm

 

[rmw]

Unclesyd, I am going to have to give you a star on that one. Who'a ever thought'a that one besides you. The rental might even be cheaper than his shop labor input to cobble a ribbed flat head together, whether it was scrap or not.

rmw

 

[GenB]

it looks that your pressure is only 30psi,
if you rib it, you can bring it down to 1/2 to 3/4 in.
I will try to find calcs and advise.
I may have the calcs for internal ribs only but even inside will work for your test job.
er

 

[JStephen]

"You can use anything you like for a hydrotest blank. This is not part of any Code."- That's a point I was going to make. If it's 82" piping, it's not a pressure vessel.

Also read the "coned section with flat plate end" thread that ran several days ago.

On designing the rib-stiffened plate, you can design the plate itself as fixed-fixed beams running between the ribs (fixed due to symmetry at end). You then need enough strengh at the beam end to support the loads you're putting into it. Generally, the beams would all be parallel, not a waffle layout. Or maybe just an X-shape, then use the quarter-plate load case from Formulas For Stress and Strain to design the plate.

You can stack plates if you're careful on the design. Two= 3/4" plates do not have the same bending capacity as one- 1.5" plate, but you could stack maybe three of them, or two 1" plates, or some other combination.

Also, if designing just a single plate, check if it falls under the "large deflection" criteria of Formulas for Stress and Strain- if so, that will increase the (calculated) strength of a given plate.

Your allowable stress could be set somewhat higher than normally used for pressure vessels. Especially true if it's just a hydrotest (no signficant air volume). For example, I think AISC allows 27 KSI for bending in a 36KSI plate.

 

[JStephen]

You're not by any chance making more than one of these valves, are you? If so, butt two together and test at the same time.

 

[jhollin1138]

@JStephen: Unfortunately there is only one valve, plus it is a weld-in. Not to mention that if you test "both" valves at the same time, you need to have them both finished at the same time. Then if one of the valves fails, you cannot finish the other valve until it is repaired.

BTW, I am still waiting for purchasing to get pricing on the dished head.

 

[arto]

Look in Chapter 7.7 of "Chemical and Process Equipment Design" by Azbel & Cheremisinoff has a method where the head takes half the load & the ribs take half the load.



By the time you're done calculating & welding a dished head woulda been easier. - plus the ribbing takes up space too.

 

[flexy]

Dr. AE Blach has a method for designing honey comb flat heads. If you're interested, give me your e-mail and I'll send it to you.

 

[arto]

what is the reference info for the Blach article?