vessel nozzle loading 5

[heaterguy]

If a customer gives Fx-500 pounds, Fx-500 pounds, Fz-500 pounds, Mx-1,500 foot-pounds, My-1,500 foot-pounds, and Mz-1,500 foot-pounds, do we apply all six forces at once, one at a time or just F or M forces?

"Each vessel nozzle shall be designed for the following loads"

 

[JStephen]

If you have a simple way to check the loads, apply all at once and see if it affects the design- if not, it doesn't matter. If so, check with the customer to see what's intended.

 

[SnTMan]

This sounds like a typical WRC-107 / WRC-297 calculation, with the exception of the coordinate system. All load and moments are applied simultaneously. WRC-107 uses a vessel coordinate system, with shear forces and moments in vessel longitudinal, circumferential and radial directions rather than x,x,z coordinates. Forces and moments given in x,y,z have to be translated to the vessel coordinate system before performing the calculations. In practice these are often just tables given in specs as per your example and no one really knows the true orientation. Often, the magnitudes for the three directions are the same also per your example and true orientation makes no difference.

These calculations can be performed in various commercial programs such as Codecalc or many companies have their own software. It's pretty difficult to do by hand. You may be able to find "freeware" spreadsheet calcs but I can't point you to any.

The allowables are often based on whether or not a large component of the loading is thermal, and therefore "self -relieving" or sustained. Again, when working from tables in specs, nobody really knows and it is best to assume they are sustained and use an allowable of something like 80% of yield at temperature.

Regards,

Mike

 

[prex]

Can't fully agree with the well documented position of SnTMan.
Also do not fully agree with JStephen: please do not ask your customer, these loads are generic tabulated ones that nobody knows where they come from, so it's no use asking.
The meaning of such tables is an old issue. A point that's never made clear is whether they include the pressure end effect or not, so its normal to assume it is not. Hence the pressure end load should be added to the axial one.
Concerning the superposition, the issue is not only whether they should act simultaneously or separately, also their sign is normally specified as + or -, but its unclear whether they should be assumed to vary between the + and - orientations (and that's makes a lot difference!).
It's usual to take them as varying between 0 and the tabulated value, but to combine their signs in order to get the worst combination.
Concerning the allowables, membrane stresses should be classified as local membrane and limited to 1.5S and membrane+bending stresses should be classified as secondaries and limited to 3S (PD5500 has a fairly more conservative treatment). The origin of the loading (thermal or other) is unrelevant, as these are external loads for the vessel, so they must be treated as mechanical loads (unless the vessel-piping combination is analyzed as a whole, but codes are not ready for that).
In the site below there is a sheet for WRC297 calculations.

prex

http://www.xcalcs.com

Online tools for structural design

 

[heaterguy]

Do the values above seem excessive for a 1.5" 150# flanged nozzle into a 3" sch 40 pipe body?

 

[arto]

They'll bend the pipe big time
a quick check:
for 1.5" s/80 Z =0.412in^3

1500ft#x12in'ft/.412in^3 = 44ksi

sounds like the old "moment to yield"

also check the nozzle load on the shell [Bednar is a quick way]
& the equivalent pressure vs the flange rating [Kellogg]

 

[SnTMan]

heaterguy, they seem kind of high to me, many exchanger specs don't require nozzle loading to be considered below 2" NPS. You can go run it xcalcs, if no problem, fine. If so, clarification may be in order.

prex, I agree that pressure stresses should be added to tabluated loads. As for signs, I have not noticed great differences (this is REALLY TOUGH to generalize about) except when the radial force is directed outward from the nozzle, in which case it is additive with the hoop stress due to pressure. HEATERGUY: make sure you understand the sign conventions your software is using.

As for allowables, again agreed, that was just deeper than I wanted to go at the time. I have seen specs that specify differing allowables based on whether loading was "primarily" thermal or not. Not that they always tell you which way it is.

Nozzle loading gets to be a messy business at times.

Regards,

Mike

 

[JStephen]

"If so, clarification may be in order."- that sounds a lot like asking the customer to me.

My point was that these do sound like standard-spec loads, not actual calculated loads, and if they seem to be unreasonable, it is quite likely they can be done away with or reduced by the customer.

 

[NozzleTwister]

heaterguy,

Speaking as a customer, I can tell you how we expect you consider these loads. I provide such tables and loads to venders to design by. For the record, I do agree these loads seem high for a 1 1/2" 150# nozzle. Consider these loads as follows.

1. The loads are strictly Forces and Moments imposed on your nozzle by the attached piping.

2. They DO NOT INCLUDE any axial forces in the nozzle due to pressure.

3. The loads are to be considered as acting simultaneously on your nozzle.

4. They are given in absolute format for simplicity. The actual loads can be either negative or positive in sign. The load range can actually be double the loads given. If Fy = -500 lbs in the sustained case (dead weight) and +500 lbs in the operating (hot) case, the nozzle load would be considered acceptable.

5. It's the Venders responsibility to design using the load directions (-/+) that give the most conservative results.

With that being said, hopefully the customer has provided you with the coordinate system (though in this case the loads are all the same) and has established whether the loads are to be taken at the flange face or at the nozzle to shell juncture.

For vessels and exchangers, I prefer the axis nomenclature of P, Fl, Fc, Mr, Ml, & Mc.

I provide a table with nozzle force and moment allowables increasing with nozzle size.

Don't be afraid to go back to your customer and tell him the loads are excessive and establish some new numbers. Maybe he has actual values he can provide. If he insists on this high of moments, ask if he's considered flange leakage.

The sizes given make me think this is probably a "hair pin" exchanger. Hopefully you're not considering a stub-in nozzle connection, but are considering a 3 x 1 1/2 welding tee with a weld-neck flange for this nozzle.

That's my 2 cents, maybe more.

Good luck,


NozzleTwister
Houston, Texas

 

[prex]

NozzleTwister,
your post does not solve a very important issue.
A point that's not often known to many designers is that the acceptability of some stresses (the secondary stresses) does not depend on the absolute value of the load, but on the range (the variation or change) of the load between different states of loading. So, in your example, the effective load is 1000 lbs, even if the two loading conditions are not occuring at the same time.
Normally the limiting stresses for nozzle loads are the secondaries. So if you take all the nozzle loads as varying between a minus value and the plus one, you'll effectively double all your loads!

prex

http://www.xcalcs.com

Online tools for structural design

 

[zjliang]

Nozzle checking is really a controversial and nasty topic, and sometimes a lot of tricks in it. If the piping designer/analyst could work together with vessel designer, that will make it much easier. I used to play a dual roles. Here below is my experience.

1. Before applying these load numbers, check or try to find the piping layout connected to the nozzle, even a hand sketch would help, then make a guess/assumption about the loading combination.

2. Ask yourself a question: what case do you need to take care? Sustain or thermal Expansion? Primary load or secondary load?

3. Then based on the assumed load cases, pick the load value from your table apply them into your model, do not forget the sign.

4. Depending on what method you use, if WRC107/297, do what prex/SnTMan/NozzleTwister recommended. If FE, make sure to eliminate the effect of loading boundary condition. Besides, the allowable for secondary + primary is SPS, not 3Sm.

Actually, most customers don't want to take too much time to provide accurate load cases, in most cases it is impossible to find it in the begining stage of the project before the piping layout is done. So quite conservative absolute values were put forward. From the years experience in manufactures, I feel quite confident that most vessel nozzles are strong enough to handle the real loads even though it can't pass the crazy numbers in the design stage. Especially for the thermal loads from piping.

BTW, I totally disgree the design philosophy "do not treat vessel as a pipe support". Sometimes it will drive the designer nuts.

Thanks,

zjliang

 

[rmartin00]

zjliang, I have to disagree with your opinion on treating the vessel as a support. I have designed large vessels with nozzles, say 18" diameter, with huge tablulated loads as Nozzletwister discusses, but with design pressures of 15 psig. What happens is that the shell and nozzle thicknesses are disproportionately thick compared to what they need to be for pressure alone. This costs a lot of money when the vessel is made of, say C-276. And for what? So the vessel can perform the action of a hunk of cheap structural steel (aka pipe support) attached to the building steel? The only one that loses when 1)using the vessel as a support and 2)tabulating nozzle loads based on size, is the owner, because they'll end up paying more for the vessel due to the need for more material.

 

[zjliang]

Good comments, rmartin00.

But just a little clarification of my point. In most cases, changing the shell/header thickness to handle the piping loads is always the last resort except for some special governing loading cases per UG-22, like cyclic or dynamic loads per UG-22(e).

My point is that without changing the shell design (thickness from pressure calc), or just by introducing proper repads/attachment, a nozzle with enough flexibility could actually resist quite high secondary loads, specifically for thermal piping loads. Without considering plastic analysis, just fully taking advantage of the redistribution of the secondary stress around a nozzle, you will find a vessel could reasonably be a pipe support in some sense. Strong enough.

Certainly, I totally agree that it is silly to increase the thickness of whole vessel to be a "Pipe Support".

Another point is if model a vessel and connected piping in together, or just by introducing accurate local flexibility instead of an assumed anchor point, the piping loads will drop dramatically.

Cheers.

 

[johnnymist2003]

Hi,
this issue of including pressure thrust on nozzles for WRC-107/297 analyses...If one takes a blanked manway nozzle, is it necessary to do WRC-107/297 nozzle load calculations?Certainly there is a pressure thrust on the blanked nozzle.This pressure thrust would create a substantial radial force acting outward.However I have yet to come across a WRC-107/297 analysis on a manway nozzle.I have tried out this calculation just to see the results and if one designs the manway correctly to Code rules for reinforcement, etc, then one checks the configuration in WRC-107/297 using the radial pressure thrust, the vessel shell and any reinforcement is horribly overstressed.Thus, to follow the logic, in my opinion it is not necessary to add pressure thrust in to the radial force component of a WRC-107/297 analysis.I believe that the purpose of WRC-107 is to deal with the mechanical loads from attached piping (of whatever source) and to include pressure thrust in addition to the radial force is unnecessary.I welcome any responses.
John

 

[prex]

johnnymist2003,
if you have the pressure end thrust alone, then you don't need to check per WRC, as the opening reinforcement will take care of that.
But when you are required to do such checks (normally manways and non process nozzles are excluded) you need to include all the loadings, also the pressure thrust.
Of course in doing that you'll see the secondary stresses, that are normally out of the scope of ASME VIII Div.1 (see UG-23(c)): it is customary to limit them to 3S under Div.1 coverage, but in my opinion it would be more consistent to simply not limit them. However in that case most WRC107/297 analyses would not be very useful.

prex

http://www.xcalcs.com

Online tools for structural design

 

[zjliang]

Hi John,

Not sure on which your thrust is based, design pressure or operating pressure? As we know, the design pressure is always about 10% higher than the most critcal operating pressure (plus static head) for some conservative margin. So if the thrust is calculated from design pressure, I would say the thrust is too conservative.

BTW, I found the same results when I tried to introduce the thrust loads in WRC 107/297 check. The nozzles always relentlessly fail. But insteads, by using FE method, the same nozzle in the same loading might pass the code.

So I suggest using FE whenever WRC 107/297 fails. To be honest, I don't trust WRC 107/297.

Comments?
Thanks,

 

[gvc99]

I have an opposite problem. How would you calculate limit loads for a vessel that you could give to a customer who hasn't designed his piping system yet? I'm designing a vessel right now where the customer has not specified nozzle loads, but he knows there will be such loads and wants us to specify "not to exceed" loads for P, Mt, Vc, Mc, Vl, Ml (nomenclature per WRC 107/297).

I can generate a matrix of stress effects from each loading by specifying a "unit" loading, say 1000 lb or 1000 in-lb and using CodeCalc to produce stress results for a particular loading case. Does anybody have a rationale for using this information to produce a reasonable set of limit loads?

 

[NozzleTwister]

gvc99,

Several companies that I have worked for have had Nozzle Load Charts that show minimum allowable nozzle loads for each for all six directions. Nozzles are to be designed to handle all loads simultaniously.

The magnitude of the loads are such that they are reasonable for most vessels and reasonably achievable for reactions from most piping situations.

For a vessel, you would choose loads for a reinforcd nozzle.

Many companies have such charts, you can probably find a go-by from searching the net.

NozzleTwister
Houston, Texas

 

[waskillywabbit]

gvc99,

IMO, this should not be your problem, it should be the customer's. Seems likes your customer wants the cart before they have a horse, and you are on a road to possible trouble. Since the customer probably gave you a set of specs to design the pressure vessel, they should also provide you with a set of specs on nozzle loads, if they want the nozzles reinforced for piping loads. Not you provide them with a set of allowable loads so they can design their piping. Sounds like someone in the company you are working with/for isn't on the ball, as any company worth anything is going to have a pre-established nozzle loading chart as twister mentions above.

Brian

 

[lbower]

gvc99,

When I have to provide 'not-to-exceed' loads, I provide a maximum resultant moment and a maximum resultant force. I calcuate the maximum force value by finding the maximum radial force (concurrent with design pressure) and I find the maximum moment value by finding the minimum value of either the maximum circumferential moment or the maximum longitudinal moment (both also concurrent with design pressure). The actual resultant force Fr = sqrt(Fx^2+Fy^2+Fz^2) and the actual resultant moment Mr = sqrt(Mx^2+My^2+Mz^2). The combination of forces and moments is Fr/Frmax + Mr/Mrmax <1. The customer can then design the piping to fit within these boundaries. It has a few conservative parts to it:
- radial forces create more stress than shear stresses and in calculating the actual resultant force, all three forces are equally weighted
- circ. or long. moments create more stress than torsion ...