Bolt bending 2

[illini8181]

I am designing a bolted connection (see attached) which consists of a steel structure bolted to a large, gently curved steel plate via two steel flanges. In between each of the two flanges and the curved plate is a very hard plastic shim.

These are very, very large structures, and so the part tolerances can only be held so tight. Because of this, one or both of the feet will likely be “kicked” at some angle relative to the curved plate, and there will be a slight gap (max of about 0.070 inch, based on GD&T). These are also both very stiff structures, and so we cannot close this gap with preload. Also, we are not allowed to take into account friction in our analysis. The plate is hanging cantilevered off of the steel structure, which is fixed. I am not experienced in analyzing a joint like this, but from what I read, bolts are typically only analyzed for shear and tension. However, the question has been asked whether bending is a factor in this joint, due to the possible gap. Does anyone have any thoughts, or references that might be helpful?

 

[fegenbush]

With sufficient tightening of the bolted connection, there should not be any bending. Shear, yes. Tension, yes.

Has it been considered to weld the steel flanges in the field? That should minimize any gap issues.

 

[dvd]

Can you turn the connection plates 90 degrees, and weld onto the curved bar? A field weld on the structure-side connection plate would then let you eliminate the shim and also put the connection in shear.

 

[dhengr]

illini8181:

What’s the plastic shim for? .070" (1/16" +) fit-up tolerance in a very large structure is pretty good. I would eliminate the plastic if I could, or you will have bolt bending; with the forces applied to the bolts something like [.5Tflg.+Tshim + .5Tpl.], or about 2 or 2.5" apart, maybe a little less. At the bolted connection faying surface, you want the two pieces of connected steel in contact, so you do have the bolts acting in shear and in bearing on the holes in the plates. And, in a bolted connection you never get all the bolts equally loaded because of hole spacing tolerances, and in your case because of the large vert. distance btwn. the horiz. support arms. Then, some movement must take place in the joints before the first bolt is brought into bearing on the hole; then some yielding in bearing takes place so other bolts start coming into play.

Your sketch is good, as far as it goes, but some sizes, dimensions, curvature radius, weights, loadings, etc. would really be helpful in getting you a good answer. You’d be surprised how these things and correct proportions in the sketch influence how an experienced engineer sees the problem. What does this thing do? Could it be field welded to the steel support structure, or does it have to be bolted? I’ll bet the curvature situation is exaggerated in your side view detail of the joint and .07" could be made up in field welding. I would be tempted to rough machine two wedge shaped steel blocks, fitted and shop welded to the back of the curved 2" pl., to provide a plumb plane to match the 1" flange end pls. of your horiz. support arms. Then drill these wedge blocks and the 2" pl. for bolting; drill perpendicular to the plumb back surface. You’ll need wedge shaped washers for the bolts on the front of the curved pl. If you make these welded wedge pls. a bit higher than the flange end pl. tops, once you got the curved pl. in position and bolting started, you could field weld some weight carrying shear blocks to the wedge blocks and resting atop the flange end pl. tops. Now the bolts aren’t carrying much shear, they are just holding things together.

 

[desertfox]

Hi

How heavy is the curved plate because I can see additional tension forces in the bolts due to the offset mass of the curved plate acting outside the bolt centrelines.

Go to this site and look at the example bolts resisting bending forces

http://www.roymech.co.uk/Useful_Tables/Screws/Bolted_Joint.html

 

[illini8181]

A few more details that I left out
1. There are actually several of the curved steel plates that this connection must be able to work with. Ie, the curved plate must be able to be disconnected and replaced with a different steel plate. Because of this, we cannot weld the flange to the curved plate.

2. The plastic shims actually will be machined to fit the contour of the curved plate. Because there will be multiple steel plates, there will of course be a variation in the fit up. It would be possible to machine a custom set of shims to match each plate, but for obvious reasons (cost and complexity), it is desired that one set of shims be sufficient for all of the plates. We are just trying to convince ourselves that the possible gap will not introduce a serious bending stress that will cause the bolts to fail.

dhengr - I completely understand what you are saying about how providing more actual dimensions and details about the purpose of this connection would help you to more fully visualize what is going on here. Unfortunately, I am not able to provide much more detail than I already have. This is actually only a simplified representation of the problem I am trying to solve. However, to give you an idea of the scale of the connection, we will say that the two flanges are roughly 20 ft apart. As you have said, the curvature is exaggerated in my side view detail of the joint, which you probably saw just by comparing it with the “zoomed out” view.

dhengr – the plastic has two purposes. The first purpose is to conform to the contour of the plate (the flange is flat and the plate is curved). The second is to protect the steel plate (softer material). As the joint compresses in bearing, if the contact area is small, there will be local yielding. We want the plastic shim to yield, not the plate or flange.

dhenger – when you say the forces applied to the bolts should be something like “[.5Tflg.+Tshim + .5Tpl.], or about 2 or 2.5" apart, maybe a little less”, are you visualizing the bolts being modeled as cantilever beams (see attached)?

desertfox – please correct me if I am wrong, but the website you sent me to discusses a joint which is in bending, but assumes that the bolts only see shear and tension. For the purpose of the discussion, we will say the plate is 50,000 lbs.

 

[Screwman1]

Just one quick thought before I get onto a call: with that shim in there, you will not be able to retain any significant level of clamp load in the joint, so you might want to assume a pure shear loading since the bolts will go into bearing pretty quickly once the shim yields under the point loading.

 

[illini8181]

Screwman1, thanks for your reply.

Why will we not be able to sustain a clamp load? The plastic is very hard, with a modulus of elasticity of around 5.0e6 lb/in^2. It is possible that there will be some local yielding as preload is applied, but as the contact area increases, the plastic will reach a point (about 5in^2) where the area is sufficient to withstand the bearing stress. At this point, the full preload will be able to be applied. Also, even if I am incorrect and the joint cannot sustain the clamp load, wouldn’t I have to consider bending stress in addition to shear, because of the presence of the shim between the flange and plate?

 

[CoryPad]

Polymers creep (stress relaxation), so the elastic modulus is useless for determining preload retention.

 

[illini8181]

Corypad, that's a good point. There's another detail I left out though - the time duration of the connection will only be a few hours at most.

 

[CoryPad]

Then low risk for creep

 

[chicopee]

If your steel flanges have a slightly smaller radius of curvature than the steel plate, as you tightened the bolts the plate at points of contact should conform to the radius of the flanges as these flanges will be more rigid than the plate because of the welded connections to the I structural elements; also the bolts holes may have to be slightly oversized to allow the plate to move during the tightening procedure of the bolts.

 

[rb1957]

1) why not use an insert (cylindical sleeve) to clamp up with (rather than clamping the plastic shim) ? make one hole 1/8" (maybe 1/4" given the scale of this) slide in a sleeve (1/16" maybe 1/8" wall) under the head of the bolt.

2) why not use "liquid shim" rather than machining a plastic to suit ?

3) some other epoxy material ??

Quando Omni Flunkus Moritati

 

[badboggs]

Hi Illini,

Our company has had problem like this in the past. One of the solutions we have used is spherical bushings. See here for examples.

http://www.aurorabearing.com/commercial-industrial-inch-spherical-bearings.html

These bearings/bushings allow for radial tolerance in your assembly. You can also find them on McMaster at 2459K18.

I'll attach a mockup I made for you based on a 40 foot diameter ring.

http://files.engineering.com/getfil...2c&file=Ring_Assy_with_Spherical_Bearings.PDF

Let me know what you think. At least this would allow you to clamp the bolts to full preload.

 

[dhengr]

illini8181:

50000 lbs., that’s a 600-625 sq.ft., 2" thick curved pl., that’s about 25' x25' for the curved pl. And, you are going to be changing these out every two hours, and screw around with some plastic shims which don’t really match all the different curved plates. Why would there be much plate to plate wear and yielding at the bolted connections? Do the plates wear out that quickly, two hours, from their usage? Are they reused many times? How many curved pls. are there, all the same size, same curvature, and curved in only one direction? And, the eccentric weight of the curved pl. is the only load on the bolted joints? There are only 8 bolts and 2 connection details per 600 sq.ft. pl? And, you are going to allow bending in the bolts, you are likely to be changing out bent bolts quite often. Your second sketch is about what I had in mind, but your note ‘fixed connection (no rotation)’ is really not correct as regards the bolts. I see the bolts as semi-fixed at each end with the moment at one end in the clockwise direction and the moment at the other end in the counter-clockwise direction; not exactly a cantilevered beam, maybe two canti. beams joined/connected at their free ends. You don’t have to give away any national secrets to give more meaningful detail for the solution of your problem. And, if you can’t give that info. except in dribbs-n-drabbs as someone asks a new question, it could be because you haven’t given them sufficient thought yet. And, that could come back to haunt you in the actual installation and operation. If you are going to be changing these plates out frequently, you might want to spend a few design, detailing and shop hours to make the change out go quickly, with min. men and man hours and crane time, etc.

I think trying to put the end plates on the horiz. support arms 20' apart in height, and to match the slope of the curved pls. will be a big mess and tolerancing problem. I would install those plumb, and in one plumb plane. I would then take an 8 or 10" stub of the same WF the horiz. support arms are made of, weld a matching plumb end plate with 4 holes to one end of this stub; and rough machine the opposite end of the stub to match the curved pl., and weld these to the curved pl. Note that, in the arrangement you have now, the bottom support arm will carry more load than the top arm if the WF beams used are the same. Is there any functional reason that there can’t be some connection weldments on the backside of the curved pls? What is the height of the support arm end pl., what is the WF beam size, and what is the vert. distance btwn. the two upper bolts and the two lower bolts on the end pl? What is the radius of the curved pl., and is the angle change over the height of the end pl?

 

[Cockroach]

You are placing the bolt into a condition of combined loading. So the sum of each action, axial load and bending can be analyzed against shear and torque. Then you can use Mohr's Sphere to get the principle stresses, using that to compare to material properties in order to arrive at some sort of safety factor. This in turn would get you a perception into the principle mode of failure.

So first year statics with second year material analysis. Nothing magical about getting an appropriate solution.

Regards,
Cockroach

 

[desertfox]

Hi

Yes the website I posted does exactly what you say.
I believe it's applicable to your situation, looking at the sketch you posted the mass of the curved plate acts outboard of the bolts holding it in position and what I imagine is that the mass of the curved plate wants pivot about the bottom bolted plate and try to open if you like, a bit like a door.

 

[rb1957]

"Polymers creep (stress relaxation), so the elastic modulus is useless for determining preload retention." ... that's why i suggested clamping up on a sleeve.

"the time duration of the connection will only be a few hours at most." ? ... this is only for erection, and not as part of a permanent structure ?

Quando Omni Flunkus Moritati