Padeye attached to cylinder

[dav363]

Hi all, can anyone shed some light??

I've got a steel cylinder to the sides of which I'm welding 2 diametrically opposing padeyes. The padeyes are attached to steel ropes which are applying a load logitudinally to the cylinder, in other words the load is trying to shear the padeyes from the cylider wall.. The shear stress calcs are fine but my problem lies with the moment caused by the load acting at the pin at a distance from the cylinder wall. This is the force which will act to buckle the cylinder as a load is applied.

I started thinking I could treat a longitudinal strip of the cylinder wall as a flat beam, with the padeye producing a push/pull type force system trying to bend the beam. I could then use standard beam bending theory to calculate the bending stress in the strip. However I've found this to be a far too conservative as the curvature of the cylinder has too great a bearing on the second moment of area.

Short of an FEA analysis, does anyone have any tips on how to approach this problem.

 

[Ussuri]

Could you change the detail of the padeyes? I was thinking have the two padeyes connected (ie on the same piece of plate) and then slotted through the tube and welded into place. That way you are not applying the load to the cylinder wall in direct bending which may cause the wall to buckle locally.

 

[dav363]

I've got some tubes and cables encased within the cylinder therefor I can't protrude through the wall as I would clash with these.

 

[zcp]

Just a few quick thoughts....

1) Can you put a spreader bar on the supporting cables?
2) Try treating each half of the cylinder as a curved beam under point load. Take the width as 2X the distance from center of padeye up to top of cylinder.

What is the thickness and diameter of the cylinder? Load?

ZCP

http://www.phoenix-engineer.com

 

[dav363]

ZCP, The cables will be attached to a spreader bar so the load will be applied purely longitudinally.

The cylinder is 280mm OD, 12.7mm WT. It is 650mm long with flanges at both ends which attach to supporting structures.

I see what you mean, take each half of the cylinder as a beam and work out an 'I' value for that geometry. I'm thinking that given the diameter to wall thickness ratio, the failure mode is likely to be highly localised bucking of the cylinder wall around the base of the padeye, like sticking your finger into the side of a coke can. That would make it a sort of 3D beam bending problem, or plate buckling from a point load. What do you think??

 

[Ussuri]

Ring stiffeners located at the top and bottom of the padeye?

Go for a big mother of section?

The following link is to a design guide for hollow section joint design. The document is based on CIDECT design research and publications. P22 and 23 gives a design method for checking hollow sections with longitudinal and transverse gusset plates (basically a padeye with no hole) so this might be of interest.

http://www.corusconstruction.com/fi...ortedDocs/CT16_SHS_Welded_Joints_15-08-01.pdf

 

[zcp]

Are these padeyes extending a long way out from the cylinder? (Like 100mm or something.) I would think the bending would be minor for that thickness (unless the load is extremely large).

Two routes to go.
a) get someone to run an FEA for you (take your time on the loads and BC's)
b)maybe analyze the weld attaching the padeye and transfer loads to the plate that way....i.e. see if you have enough material through the thickness to support the loading at the weld line and localized bending

Seems like with this being for an overhead load that your FS would be high enough (FS=4 or so) to take out some of the uncertainty.

ZCP

http://www.phoenix-engineer.com

 

[dav363]

Good answer Ussuri, a big mother........

ZCP, The pin of the padeye is at a distance of 150mm from the cylinder face. The load is very significant, 44 Te to be exact. At the moment I'm pretty certain that the moment caused by the padeye will cause buckling of the cylinder wall, therefore I'm going to need to either thicken the cylinder wall or as Ussuri mentioned some kind of ring stiffener (I assume this is simply a ring of plate welded around the base of the padeye to distribute the load over a wider area?). If I thicken the plate, the question is by how much?

Ussuri, I've looked through the link you posted, thanks for that. I agree for the joint integrity it is a useful link but I'm not convinced it takes into account the local buckling of the cylinder wall.

Is it plate loading I should be looking at? Some calculation which would show plate deformation given a point load on the plate surface? any pointers?

 

[Ussuri]

dav

What you call local buckling (I think) the document refers to as chord face deformation. It runs through a number failure modes on page 14. For a padeye in the same plane as the CHS I would look at clause 5.1.6.

The formula for N1 is the axial load which will cause the face of the CHS to deform. From this it determines the in plane and out of plane moment capacities.

You can them compare this for you factored applied loads. There is also a simple interaction check on P7.

I did some quick scribbles calcs using CIDECT which suggests a moment capacity of about 200kNm before the face deforms based on a 500mm long padeye.

The idea with the ring stiffeners is to transfer the reaction from your padeye to them instead of the CHS wall. You would need to size them to resist a large force pulling them apart on one side and another force pushing them together on the opposite side.

 

[dav363]

Ussuri,

Can you expand on the sigma p variable in the formula for 'chord end load function' on page 20. As I understand it this is the stress within the cylinder wall caused by all loadings other than that of the joint in question. So in this case where the cylinder is being loaded purely axially, by a weight hanging vertically from the bottom flange, sigma p is the stress in the cylinder wall caused by that load. Is this correct? I guess that would account for the combined stresses from the axial load and the padeye loads.

 

[Ussuri]

Thats how I understand it. These are effectively the forces within the chord member (your pipe) due to the externally applied load. I just took it as one when I did my two line calc (but thats because i'm lazy). Its usually pretty close to 1.0 unless the member is very heavily stressed.

 

[Ussuri]

Corus also publish some worked examples.

http://www.corusconstruction.com/fi...bes/CT-SHS Joint Design Examples 16-03-05.pdf

http://www.corusconstruction.com/fi...d joints Supplements - No 1 to 5 06-05-05.pdf