How to design this saddle

[ElCidCampeador]

Hi, I'm trying to understand what kind of design check I've to do for a saddle of a pre-existing pressure vessel here below sketch shown:

Saddle is bolted to the ground/structure through 4 bolts. The only force to consider is there shown (weight included).

I suppose that base plate undergoes a bending moment due to the force effect from axis of base plate. And bolts too. Compression effect of force could be negligible, I think.

Do you have any suggestion or reference/code with a simply method to calculate? I'm not precisely designing it, it's only a basic calculation required by inspector as complement (base plate and bolts are surely oversized and force is almost negligible).

Thank you

 

[JStephen]

Assuming that bent plate just welds directly to the shell- Bednar's PV handbook includes an approximate method for line loads on a cylinder that could be used for the connection to the shell.

 

[Snickster]

As stated the connection to the vessel can be analyzed using Bednar Handbook method which is just a version of the WRC 107 calculation method to determine local stresses in vessel wall. A reinforcement pad may be required on the vessel wall.

The design of the saddle itself is just a typical structural design where you need to determine the forces and moments acting on the saddle under worse case conditions including occasional loads such as wind and earthquake. Once you find the stresses in the plates due to the forces and moments you compared to allowable stresses such as those given in AISC structural code or ASME boiler and pressure vessel code for vessel support attachments.

Also, the bolts would be designed for the same forces and moments.

It is a pretty complex calculation really for a novice.

So just what are you considering to be your external loading on the vessel and saddle?

 

[ElCidCampeador]

Assuming that bent plate just welds directly to the shell- Bednar's PV handbook includes an approximate method for line loads on a cylinder that could be used for the connection to the shell.

ok thanks I'll take a look!

So just what are you considering to be your external loading on the vessel and saddle?

only weight of vessel actually (Force=weight of entire vessel and accessories approx).
It's a pro forma calculation, in order only to give a semblance of design check of something which was created from old construction standard belonged to some (unknown) company...

 

[Snickster]

A simple approach to designing the saddle itself is:

1, Check stresses in vertical saddle plates. This stress is the compressive stress due to weight of vessel plus bending stresses due to external moments.

S = -W/A +/- M/Z

S is stress in psi
W is weight of vessel
A is total cross-sectional area of plate
M is external moment
Z is section modulus - to be checked in X-X and Z-Z planes.

2. Check for buckling of vertical saddle plates using Eulers formulars.

3. Check for local stresses in vessel wall at saddle connections using WRC 107 methods.

4. Check bottom plate and bolt design. Basically any stress in plate and forces in bolts would be due to external moments since vertical plates take all weight load and transfer it to the foundation. Force in bolts and moment in plate can be determined by summing external moments where resulting load on bolts would be such as to resist the external applied bending moments.

 

[ElCidCampeador]

Good points. A few questions:

1) It seems that there are 2 bending moments in 2 different planes, so may I use this formula to find total S and compare with allowable compressive stress?
S = -W/A +/- Mxx/Zxx +/- Mzz/Zzz

2) What is the unsupported length of colum in this case? Total height of saddle? What is boundary condition? It's clamped+free or clamped+clamped?

3) Ok but what forces/moments apply to vessel wall? 2 moments above and weight of vessels?

4) I don't know any external loads...but weight acts to the base plates through narrow lateral strips, so there are at least bending moments in the baseplate. But I don't know how to check it because base plate is "fixed" by 4 bolts in the middle acting like clamps??

 

[Snickster]

1) It seems that there are 2 bending moments in 2 different planes, so may I use this formula to find total S and compare with allowable compressive stress? S = -W/A +/- Mxx/Zxx +/- Mzz/Zzz

Yes, if the forces that produce the bending moments can act simultaneoulsy, then the stresses produced by the xx and zz planes would add by superposition.

2) What is the unsupported length of colum in this case? Total height of saddle? What is boundary condition? It's clamped+free or clamped+clamped?

I would say the unsupported length is the height of the side plate minus the thickness of the base plate since the base plate provides support for the vertical plate at the bottom. I would say the vertical plate is fixed at top and bottom but at bottom it can rotate slightly so to be extra conservative you can model it as fixed in all directions but a pinned connection that allows rotation in one direction.

3) Ok but what forces/moments apply to vessel wall? 2 moments above and weight of vessels?

In accordance with WRC 107 analysis you need to do the analysis for the actual Fx, Fy, Fz, Mx, My, Mz forces and moments present.

If the horizontal load is sideways perpendicular to the vessel then there will be no moments present at the connection point of the vertical saddle to the vessel. This is because any external moment load is resisted by either a compressive force on the far side connection point or a tensile force on the near side vessel connection relative to the force, which will just add or subtract from the weight compressive load. Therefore, at the connection point you would have a radial P load (which includes the moment load) and a sideways shear load due to the horizontal force (divided by two since it is shared by each connection point).

If the horizontal load is axial along the longitudinal axis of the vessel, then the moment load is not resolved into tensile or compressive forces and does not add to the radial P weight load in a WRC 107 analysis. In this case you would have a radial weight load P due to weight of vessel plus a moment at each connection due to the external horizontal load (divided equally between the two connections) and a horizontal shear force due to the horizontal load (divided by two) at each support

4) I don't know any external loads...but weight acts to the base plates through narrow lateral strips, so there are at least bending moments in the baseplate. But I don't know how to check it because base plate is "fixed" by 4 bolts in the middle acting like clamps??

You would need to provide a sketch of what you are talking about here.

Note that the analysis here for the support itself is very similar to an analyis for a skirt support for a vessel as most of the same dynamics are involved. Attached is a handbook chapter on this type of analysis.

 

[Snickster]

Note that I assumed connection can consists of two separate connections of rectangular cross-section since they are far apart on the vessel wall and the stresses at each connection should not add or interfere on the vessel wall. If they were close together you may check based on a single attachment that is shaped like a rectangle whose perimeter fits around the entire support connection to the vessel. So if it were considered a single attachment in a large vessel to support ratio (such as a lug attachment) you would likely use a reinforcing pad on the vessel so then it would really act like a single connection.

 

[Snickster]

4) I don't know any external loads...but weight acts to the base plates through narrow lateral strips, so there are at least bending moments in the baseplate. But I don't know how to check it because base plate is "fixed" by 4 bolts in the middle acting like clamps??

I think you are saying that since the load of the vessel weight is transferred from the vertical plates (narrow strips) this produces moment in the bottom plate. I don't think so. I think the load of the weight of the vessel is transferred directly to the foundation by the compressive stress in the plate itself and there are no weight forces transferred to the bottom plate - if the bottom plate cannot deflect (because there is a foundation beneath it) then it cannot absorb any forces or moments. The only forces and moments in the bottom plate and bolts is due to horizontal loads. These sideway forces and moments are transferred through the vertical plates and into the bottom plate then into the bolts. Sum forces and moments as free body to determine forces and moments in bolts and bottom plate. The moment on the vessel would need to be resisted by the coupling force of the bolts for static equilibrium. In the bottom plate would be the same moment in the bottom plate resisted by the bolts since the bolts are loading the bottom plate with this moment on one side while the vertical plates are loading this moment into the bottom plates on the other side - hence an equilibrium of moments on bottom plate. In the side plate at their intersection, there will also be a shear force. The shear force is resisted by the bolts in shear while the moment forces in the bolts are resisted in tension. AISC gives limits for bolts subject to both tension and shear at same time.

 

[ElCidCampeador]

mmm in points 3) and 4) there's something that I'm missing...you're talking about horizontal loads, but in this case there's only one vertical force, as in picture.
I don't understand what are the others horizontal forces.
In 3) you describe 2 cases which I think they would be correct if force was horizontal...for the same reason I don't understand why to consider shear force in bolts.

 

[Snickster]

I was speaking in general in case there were horizontal forces. Ignore any forces you don't really have.

 

[jt1234]

OMG, this is such simple thing, I have done it numerous times. Use line load method is good enough. If wanting to do WRC 107, read Dennis Moss design manual first, and get the loading directions according to the rule. First class in "Statics".