Calculation of allowable loads at a shell nozzle?

[KernOily]

Guys I have a couple questions for you. I posted this in the vessel forum also, so please forgive me this time for cross-posting.

(1) Do any of you know a method to calculate the allowable piping reaction loads (forces and moments) on a head nozzle for a cylindrical vessel? Megyesy's book has a method based on WRC 107 but it only covers shell nozzles.

(2) Megyesy's method does not cover all six possible applied loads at a nozzle. His method only covers a radial force and a longitudinal and circumferential moment. So - what about the other three loads? E.g., the force and moment in the plane of the nozzle?

Any help would be greatly appreciated. I have not looked in Section VII yet, so I'm guilty of not doing my homework first... OK, I'm HEADing to the library now... ;-)
Thanks!

Thanks!
Pete
P. J. (Pete) Chandler, PE
Principal Engineer
Mechanical, Piping, Thermal, Hydraulics
Processes Unlimited International, Inc.
Bakersfield, California USA
pjchandl@prou.com

 

[stanier]

Suggest you in invest in Algor's PV Designer.refer

http://www.algor.com.

This is a package based on their FEa with a front end processor more suited to we engineers who have trouble drawing and meshing in FEA.

 

[Edgar581]

You can evaluate reaction on a vessel with CAESAR II, it has a powerfull tool based on wrc107 and 297.

Good luck!
Edgar

 

Hello,

Interesting question. We sorta-kinda itteratively figured out some of the allowable load combinations for some shipboard heat exchangers and only then (of course) we found out about Paulin Research Group and NozzlePro. The NozzlePro software shows explicitly how to do exactly what you want to do - and it is cost effective.

This is not advertising, I have nothing to gain but in the interest of passing on to the community what we learned the hard way:

Pricing

FE/Pipe - $5500 US per single, unlimited licenseSupport & upgrades for first year included, subsequent annual support & upgrades are $1500 US per year.

NozzlePro - $1500 US per single, unlimited license. Annual support & upgrades after the first year are $500 US. If you want to try NozzlePro, they will let you download a free FULLY FUNCTIONAL copy that will last for 5 days - many uses as you want.

There are many times when the typical beam model pipe or nozzle analysis (even with the correct stress intensification factors are applied) simply will not tell you what you need - then you must go for finite element analysis (FEA). The WRC Bulletins (107 and 297) were useful for a long time, but not always conservative.

Go to

http://www.paulin.com/

to check out the FE/PIPE software (of which NozzlePro is a subset) as well as other neat things.

Best regards, John.

 

[MJCronin]

74....

the method suggested by John Breen, above is , of course the best way...... however we all know how cheap engineering management is these days.

Another way to approach the problem is to bracket the nozzle sizes, wall thicknesses as well as the range of vessel diameters, wall thicknesses and pressures. This was a method used by the large AE firms, (Stone & Webster... remember them !!??) in the 1980's during the big nuclear power push by the electric utilities.

If you have a tool, such as the one contained in CAESAR-II, you can iteratively solve for what will be "reasonable" stresses" for a range of nozzle sizes and vessel conditions.

For example; most nozzles will be between 2"NPS and 10"NPS and vessels will be 4 feet to 12 feet in diameter. Iterativly solve ( over a couple of afternoons) for most of the combinations and develop a general guideline ( for preliminary layouts and design)

Of course, the final piping loads should be confirmed with the final vessel design at a later date.

As I recall, for most carbon steel systems of Schedule 40 nature..... using 300 times the nominal OD would give reasonable limits for forces..... and 500 times the nominal OD gives reasonable limits for moments ( units are lbs. and foot lbs.)

So...... for example, for preliminary design loads for a 6" nozzle on a carbon steel vessel, connected to a sched 40 piping system..

Maximum forces Fx=Fy=Fz= 300x6 = 1800 lbs

Maximum moment Mx=My=Mz= 500x6 = 3000 ft-lbs

Higher & lower pressure/temperature systems can adjust the 300 and 500 factors up or down...

This should, of course be confirmed with the vendor... and gives a starting point for design.... all bets are off for atmospheric tanks or for vessels of other materials

I realize that this is not exact and many may have problems with an iterative approach that tries to bracket the problem but without the access to software designed specifically for this purpose, it seems to work..

I would like to see comments from others about this....

Regards

MJC

 

[KernOily]

What a great thread, no? This is one of my favorite topics. (I'm such a nerd...) Thanks for the tips guys.

I do need a bit of amplification here, though. I don't see how C2 will help me with this problem. Unless I'm wrong, the nozzle and vessel modeler (WRC 107/297) in C2 only models the stiffness of the nozzles and thus will more accurately (hopefully...) determine the magnitude and direction of the reactions forces transmitted BY the piping ON the vessel. This only helps me from the piping side of the system but does nothing for the vessel side. As far as I know it does nothing to tell me if the magnitude of those reactions are low enough to keep the stress magnitude developed in the vessel itself below the Sec. VIII allowables. As I see it I would need some type of vessel modeler or FEA to do that - yes?

I can develop a bracketed set of nozzle loads using C2 (great idea!) but don't I first have to know if those bracketed loads keep the vessel stresses below the Sec VIII allowables? Thanks!
Pete
P. J. (Pete) Chandler, PE
Principal Engineer
Mechanical, Piping, Thermal, Hydraulics
Processes Unlimited International, Inc.
Bakersfield, California USA
pjchandl@prou.com

 

[richay]

Pete,

You're right, CAESAR II won't do anything for you on the "vessel side". You still need to check the acceptability of the loads on the vessel.

 

[Merkurmaniac]

Pretty much the longitudinal shear, circumfrential shear, and torsion are loads which have a small impact on the answer. We generally find that they can be left out.
Richard Thompson, P.E.
Pipe Stress Engineer
Houston TX
S&B Engineers and Constructors, Ltd.

 

[KernOily]

Richard T,

Thanks for your reply. Let me make sure I understand your terminology:

longitudinal shear = a force applied parallel to the vessel longitudinal axis (a force in the plane of the nozzle flange face applied at the nozzle, along the vessel longitudinal axis)

circumferential shear = a force applied normal to the vessel longitudinal axis

torsion = a moment in the plane of the nozzle flange face

Yes? Thanks!
Pete
P. J. (Pete) Chandler, PE
Principal Engineer
Mechanical, Piping, Thermal, Hydraulics
Processes Unlimited International, Inc.
Bakersfield, California USA
pjchandl@prou.com

 

[Merkurmaniac]

Yes, that's right. Normally, the loadings that cause problems for the vessel nozzle is bending (longitudinal and circumfrential, but sometimes combined bending... just a sqr rt sum of squares of long and circ.) and then radial load. Radial load is the load tending to punch the nozzle into the shell, along the nozzle axis.

A lot of time the other two shears and torsion can be ignored, but it is important that you check the loads at the shell wall rather than the face of the nozzle flange. Obviously, a shear load at the face of flange will generate some additional bending in the nozzle when looked at from the nozzle shell. THis is why we usually evaluate nozzle loads at the shell wall.

Richard Thompson, P.E.
Pipe Stress Engineer
Houston TX
S&B Engineers and Constructors, Ltd.

 

[steve1]

Also note that the magnitude of the forces and moments developed at the wall are a function of the stiffness that is assumed at the anchor point. The three degrees of freedom noted by Merkumaniac; longitudinal shear, circumfrential shear, and torsion, would be modeled as rigid or nearly so. It's the other three degrees that require some judgement, and /or quanitative analysis. I often try bounding the answer by assuming order of magnitude values for the out of plane shear and bending componets.