I'm running a static and seismic analysis on a HPS piping system connecting 3 large boilers. The boiler manufacturer has a maximum allowable loads (forces and moments) for the nozzle.

I have this modeled in CII and am working on the design to get the forces within those limits (2,000 lbs and 2,400 ft*lbs; 8" connection via angled valve). I just started using the nozzle limit check, which is a really helpful tool.

Is it common to design equipment connections to handle seismic loading?

The flange leak check and equipment check features are in the static processor but not dynamic; is that a result of standard practices or computational problems. I can easily check the local forces and moments in the dynamic output to check the loads, but it seems pretty challenging to get these loads down even in the static case.

Feel free to ask any questions for context.

Louie

As an alternative, you can use static equivalent method to model the seismic force on the nozzles...

That's a great idea, our client however does not like that approach.

I'm now evaluating the stiffness values for this boiler nozzle. I'm getting 900 kips/in along the long axis, 200 kips/in in the short.

The vertical and 3 moments however are more difficult.

How would one put the thermal displacements into the model at the boiler nozzle, where there is an anchor?

Louie

Remove the Anchor definition and provide a displacement vector at that node (e.g., D1). Since it is displacement imposed my a full restraint, be sure to specify all 6 DOF. That is, if it there is no motion in a particular direction, enter 0. in that direction.

That makes sense, however I'm looking for reactions at that node in additions to displacements.

Its a situation where a boiler starts up and the nozzle moves thermally. We're analyzing the piping system, so the nozzle is effectively an anchor (which I'm inputting as 6 individual restraints with associated stiffness values).

Is it possible to have a displacement and a support in the same direction? (I know it is counterintuitive in a way.)

In addition, are you aware of any seismic provision in ASME BPVC Section I? I'm looking at a set of allowable loads from a boiler manufacturer and wondering if these are for static or dynamic loading.

Thanks Dave!

Use restraints with Cnode and insert the displacements for the Cnode.

Regards,

In addition to danb, suppose you set restraints at node num 500, then check displacement and input 6(Or less) values(Dx to Rz) and put the diplacement node num like 501. After that, connect restraint(at node 500) and displacement using Cnode by putting diplacement node number on the 'Cnode' section in the restraint check box. Then you'll see the result.

If you define displacements at a node and you do not include a displacement set in the load case, then the directions for which displacements or rotations are entered are fixed to 0.0 for that load case.

The Node/CNode approach just discussed allow you to eat your cake and have it too - the restraint symbol shows on the plot and the displacments are used in the analysis. It's nice but not mandatory.

It will give you the possibility to have different stiffness for various restraints.

Ok, let me see if I have this correct.

I've got my element going from 400 to 500, which has the restraints, stiffness and in addition the limit check all at node 500 (the nozzle connection).

In each restraint I have 501 in the CNODE box. Then in the displacement input I define the vector (boiler start up expansion) applied to the CNODE in all directions, inputting a 0 if it is not moving in that direction.

Is that right?

I'm getting the loads I expect, but thought I'd double check with my approach one last time.

Thanks all for the feedback, it is very helpful!

Louie

"Yes", you want to define 0.0 in any DOF that doesn't move. This is valid since a nozzle is fixed in all 6 DOFs, therefore, all 6 fields MUST have some number.

Copy that. I did it and it works great! We had to throw in a pressure balanced elbow to isolate that thermal movement and static/dynamic load from the piping. Is that typical? Seems like it's used to isolate equipment.

I modeled it with a single element EJ (finite length of 4 feet) with the spring rate / stiffness values from the entire Pressure balanced EJ assembly from the manufacturer. So I have axial, trans and bending values; torsion is 0 and eff dia is 0 due to the joint being tied. Is this an accurate way to model this type of joint?

But the displacement with Cnodes works great! Thanks all.

I'd recommend that you model the pressure balanced elbow (PBE) expansion joint using a zero length expansion joint located in the middle of the line bellows. Also, bending stiffness should be set to rigid (10E12) as this type of joint can not absorb any bending. The torsion should be set to the actual value supplied by the vendor.

I still don't quite understand this even with the ASME journal.

So is this what it seems like you're saying to do (see pdf). In which case the spring rates I was given for a PBE may be incorrect.

Based on another PBE example I've seen (12"), the spring rates for trans were 10,000 #/in and torsion was 180,000 #/in.

Do these values change if they are applied to a modeled with a finite length element for the EJ? (which is the way I've modeled it).

The way I've modeled it seems to give the correct displacement forces and moment transmission. So I'm skeptical.

Thanks in advance!

Louie,

Looking at your sketch, PBE joints can not absorb any rotation because of the four tie rods spanning the length of the expansion joint. Therefore we have to trick CAESAR II into preventing rotation of the bellows and hence the vary large bending stiffness (10E12). The torsion should be entered as the provided value.

When you have a defined length bellows element, it should be the length of the convoluted bellows element only, not the entire length of the expansion joint assembly. If you define the length, then you'll need to do a complete model including tie rods and restraints. It's easier to use the zero length method.

Please excuse my stupidity on the subject layout and the application of an expansion joint on the boiler nozzle.

I have been working on the boilers last 25 years this is the first time I came across with the use of expansion joint on the boiler ( I assume the connection is on the steam or water drum).

Adding an expansion joint means calling for trouble in this kind of systems and not recommended. Can't you handle the thermal expansion and loading on the nozzle by re-thinking on the pipe layout and supporting system and their location?

Hi Ibrahim, We have an array of 3 85pph boilers with an 8" outlet; the manufacturer gave allowable loads that seem very low. (2,000lb in x,y,z and 2,400 ft*lb in mx,my and mz).

Combined with the start up nozzle displacement and seismic loading, we needed something to isolate the bending moments on the nozzle and considered two universal tied EJs at 90° but eventually decided to try a pressure balanced elbow.

But, that was based on my simplified model, which I'm learning may be inaccurate. I'm modeling it the other way with the 3 bellows and 4 tie rods and will report back.

Make sense? Do those allowable nozzle loads seem reasonable? (I did the WRC-107 calc by hand and came up with loads closer to 18,000 lbs and 9,000 ft*lbs, but we have to meet the manufacturer's limits for liability.)

The allowable nozzle loads that you mention are quite good for a 8" nozzle. You should be able to have something like this with minimum effort. 18000 lbs & 9000 ftxlbs is not something you want on a 8" nozzle. Do a check on the flange as well.

Regards,

Ok, thanks for that reality check. We're working towards those. Does it seem reasonable to have a pressure balanced elbow to isolate the nozzle loads? The start up displacement was really driving the loads up. Even with seemingly flexible loops, spring hangers etc.

I do not have the entire picture but .... First thing to do before using expansion joint is to use the nozzle flexibility. Then, I second Ibrahim in having the same doubt, do you realy need an expansion joint?

I modeled this first without an EJ. We don't have nozzle stiffness values from the manufacturer. They just scratch their heads. We got some structural info and calculated it ourselves. It's pretty stiff, which means higher loads.

The loading that drives the nozzle reactions up is the thermal startup in static (see beginning of this thread) and the seismic load in dynamic.

With this PBE EJ, static is good right now at about 90% of the nozzle load.

As I mentioned earlier you need to find proper line stopper location in accordance with the boiler thermal expansion. The line stopper should be located at a neutral point for boiler and piping, so the thermal expansion is not going to be effective in the longitudinal direction. You may do the same by routing the piping with guides in line with the center-line of boiler/drum or close to the center-line. This eliminates the thermal expansion loads on the nozzle. So you can provide seismic support just to be under the allowable loads.

This type supporting of piping around/on boilers, heat exchangers are very common.

In one of may jobs, I had to provide structure at the top of the boiler to be able provide a line stopper against the seismic loads which were coming from the boiler outlet piping. I am not sure where the thermally neutral location on your boiler, but you need to figure out by looking at the boiler drawing or asking the manufacturer/designer.

So you are the person to make recommendation by consulting with structural engineer, sometimes by changing the route , sometimes supports or both.

Thank you for the advice, that is the basic approach we have taken. Due to the limitations of the facility and the boiler vent/exhaust lines, we had to change direction from the centerline a bit. Minimizing those offsets are reducing the loads to within the limits.

We also have new structure above the boilers to support the outlet seismically.