Stacked horizontal vessel saddle support 2

[TylerM]

I’m working on a saddle support calculation with two identical horizontal pressure vessels, one stacked on the top of the other (The top vessel will have bottom saddles and the bottom vessel will have top saddles, both having flat plates at the contact point. Vessels are then held together by bolts).

Looking at only the bottom vessel (The one with supports on the top and bottom) in normal operation the bottom supports will hold the weight of both vessels while the top support will only hold the weight of the top vessel. I’ve been researching the compressive membrane stresses acting on both saddles and have come to the conclusion that: Zick’s analysis is adequate for representing these stresses “independently” from one another. But, what will happen if there are compressive membrane stresses on the top and bottom of the saddle at the same time?

Does the fact that the vessel now has a compressive stress at the top of the vessel change the calculations for stress at the bottom of the saddle? If someone with some experience dealing with stacked vessels could help me wrap my head around this that would be greatly appreciated.

 

[dhengr]

TylerM:
No doubt, you could do what you seem to be suggesting, but the tank (PV) design and maintenance is a big enough problem without adding these external loads and conditions to the bottom tank. Why not design the bot. tank with a fairly typical saddle, and its own weights/loads on the saddle, all sitting on a base plate on a conc. foundation? In addition, this bot. saddle has a couple columns on its outer ends which go up/around/above the bot. tank. Then the upper tank is supported by its own saddle, which in turn must span, over the lower tank and on out to the columns from below. There is a bolted column connection and maybe some cross bracing at this level btwn. the two tanks. Now, each tank can be designed and fabed. in a more normal fashion, and the upper tank is just lifted onto the columns over the lower tank.

Add...., Look at it this way, it’s a pipe rack for some real big pipes.

 

[MJCronin]

Are you designing a "stacked heat exchanger" installation ?

A quick review of the completed designs offered by HX fabrication shops suggests a robust lower (bridge)support that independently can support the upper exchanger.

See these:

http://engineering.stackexchange.com/questions/12821/support-analysis-for-stacked-pressure-vessels

MJCronin
Sr. Process Engineer

 

[KevinNZ]

MJ

That design adds and another level of complexity. There will be differential temperature expansion stresses between the supports and the nozzles between the vessels. I have seen this stacked arrangement with just the nozzles and no supports between the vessels.

 

[TylerM]

I am not designing a heat exchanger, I am actually recertifying existing pressure vessels that house gaseous Nitrogen (So I won't have the luxury of changing/altering the design). The configuration is very similar to this one:

http://www.josephoat.com/wp-content/uploads/2709-Stacked-units1.jpg?w=126&h=71

Where the two pressure vessels are supported with almost identical saddle supports on the top and bottom of the lower vessel and only the bottom of the top vessel. Like stated above I know Zick's analysis can be used to calculate the stresses on each saddle independently. but I was just wondering if having stresses on the top of the bottom vessel will effect the stresses at the bottom of the bottom vessel.

Does anyone have experience designing a pressure vessel like this?

 

[dhengr]

TylerM:
This is undoubtedly a case where some FEA software would be helpful, particularly around the saddles. For some sort of a longhand, initial look..., I’d do the std. PV analysis for the typical tank stresses. Then, I would look at Roark’s Formulas or any number of good Strength of Materials or Theory of Elasticity textbooks for the analysis of circular rings under radial loads. I would look at a ring 24" long +/- (along the tank axis), some educated engineering guess on an effective shell width on either side of the saddle, tank length as a function of tank shell thickness and saddle design, and I would superimpose the stresses from this ring analysis and the typical tank stresses. Mostly, it will be a fairly concentrated compressive load and shell buckling problem btwn. the bot. of the top saddle and the top of the bot. saddle; that couple ft. of unreinforced tank shell height and a couple ft. of shell length centered on the saddle. Finally, pay some special attention to good weld design and detailing and avoiding various other stress raisers or stress concentrations which we have trouble putting numbers on.

Pay some attention to the possibility of one tank being loaded and the other empty, although I guess ‘gaseous Nitrogen’ probably doesn’t mean great LL. Can/will these tanks be hydrostatically tested, another problem? The bot. tank being empty is probably the worst case, with no pressure, and no shell ring tension. What about possibility of any differential longitudinal movement and loading btwn. the two tanks, through the saddles. Maybe allow a longitudinal slip joint on the saddle interface nearest the camera. Why the two different bot. saddles on the bot. tank? I wonder what that’s all about.

 

[EdStainless]

I would like to see the lower one with supports that run from the upper saddles to the ground, without transferring load through the shell. The supports in the photo posted by MJC is preferred.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[chicopee]

"... But, what will happen if there are compressive membrane stresses on the top and bottom of the saddle at the same time?" I would expect some buckling. I have never checked out such tank configuration, therefore, you may have to do some research if you are looking to do hand calculations. One reference that may hold the key is the text "Tubular Steel Structures, theory and design" by Troitsky. This book is supported by the James F Lincoln Arc Welding Foundation.
Ask the manufacturer which code or guideline was used in the design of this system. I would expect either an increase in wall thickness much more than is requires for MAWP or welded stiffeners inside the lower tank.