Plate between compression members 1

[flamby]

Hi friends,

There are two steel I-sections (different in cross section) placed end to end with a plate sandwitched between the two.

-> ---|--- <- The force is compressive with no moments.

Sice the I sections are different, the flanges do not meet end to end. Only the web is co-centric. The first member is 600(H)X600(W) and the second one is 700(H)x700(W) in mm units.

The client has asked me check the bending moments induced in the plate due to flage offset. What I could do is to work out the force in flanges by reducing it in proportion to the cross-sectional area and compute the moments. But I dont feel any confident about it.

What you guys think? Will the moment be induced in this fashion or not?

Thanks in advance,
flame

 

[Ginger]

If the two sections are aligned such that their centroids are in line, and the load applied is purely compressive then there is no moment generated, hence no problem.

If you have to account for some moment effect then the flange force in each of the beams will be determined by the lever arm (La) between the flanges (ie the flange force in the larger beam (beam 1) will be M/La1 and in beam 2 will be M/La2). The plated connection must then be able to take the flexure generated by the flange forces and the distance between these flanges. Andy Machon

 

[DaveAtkins]

I would conservatively assume a 1&quot; wide section of plate simple spans between the flanges of the larger column. The two flanges of the smaller column induce point loads onto this simple span. The point loads would be calculated as you described above. This method is conservative, because it ignores the stiffening effect of the web.

Alternatively, do a finite element analysis.

 

[flamby]

DaveAtkins,

I agree with your suggestion. I was wondering, what if I allow failing of the plate due to moments? The sole purpose is to transfer the axial load. Will not the failed plate will act as a cushion to transfer the loads to the respective members? A little bit of extra deflection (axial) due to local crushing of members is no problem with me.

Thanks,
flame

 

[JimParks]

Flame,

This may be a conservative way of approaching this, but I don't think you'll get any axial deflection (in excess of axial strain) unless the web of one member actually fails in compression. If that were the case, the load would then be transfered (almost) entirely into the flanges of that member.

If the smaller web seems to be able to handle the load entirely on its own, being stiffened by the flanges, then I would expect the load being transferred through the flanges and end plate to be fairly insignificant. Picture the relative stiffnesses of the webs in pure bearing vs. the plate in bending.

You could look at it, and this would be very conservative, from the standpoint of both webs being cut back (or having failed entirely) from the plate and the flanges applying two symmetrical point loads on a simple span. (Or fixed-fixed span if the plate is welded.

Or, you could go the finite element route.

If this is for a new design, personally, I would compare the costs of a conservatively (and quickly) designed plate vs the modelling of the finite element analysis. I think increasing the plate size from what f.e. would tell you you need would be less expensive (and give a higher factor of safety). There is no question that finite element analysis would be more accurate, but increasing the thickness of steel plate is cheap. Think about the cost of the increased weight of steel vs. the engineering time.

If there are enough of these connections, the f.e. might be worth it.

Good luck,

Jim

 

[rkks2002]

It was an interesting reading.
My views:
In order to determine whether the plate undergoes some bending mode to transfer the compressive loads, it is important to know what the plate size is?

In uderstand that the query is about transfering compressive loads from one I-beam to another (i.e. from say 600 size I beam to 700 size I-beam or vice-versa)through the sandwitched plate.

I would think that while a thicker plate may be able to transfer loads of compressive nature without any bending stresses, a thinner plate would undergo flexural bending to do the job.

Anything less than 28mm thick is likely to undergo the bending stress mode for transfering the loads from one side on to the other.

More explanation:
Compressive loads (as a line loads from web and flanges) could travel straight to the other side without causing bending in the plate if the thickness of the plate could allow such a direct transfer to take place. For this to happen, either the line loads have to be in alignment on two opposite sides of the plate or should be in a closeby location differing only by say around the plate thickness dimension.

In this case the difference between 600 size to 700 size being 50 mm eachway, the plate size should be around atleast 29mm (for a 60 degree dispersion of load) but preferably 50 mm thick (for a 45 degree dispersion line). Alternatively, plate less than 29mm would undergo the bending mode.

If the plate thickness is less than the 29 mm mark, then the bending analysis as already clarified by previous respondents will have to be resorted to.

Hope I could contribute some important bits to the query.

.........A visitor (Aarkey)

 

[flamby]

This strange piece of detail appeared in an arm of a hydraulic radial gate for a dam. The material has been procured just in right quantities, fabrication is in progress and hence to change any plate size now will offset the schedule. And the plate is failing as per the method by DaveAtkin! Perhaps Austim got the orientation wrong. If I am plate, one member is pushing me back through my chest and another pushing forward by acting on my back, both of them in-line.

Now what we can do is to add stiffeners at the joint and hope it will be alright.

My sincere thanks to all the experts who imparted their helpful tips here.


Regards,

flame

 

[JimParks]

Flame,

Are the webs buckling? I'm having a hard time picturing the flanges of the smalle member puncturing the plate without a web failure.

Have you thought about adding stiffeners at the webs to form a cruciform column-like detail? (Orig. I-section in dots)

..... .....
_:_ or even -.- double plates, though the welding
: -.- gets tighter.
..... ..:..

If the axial load is symmetric along the longitudinal axis, this could be a way of relieving some of the load trying to transfer through the plate from the flanges. (Which, again, I don't think can happen without the webs failing in compression.)

Good luck. (Oh, and make sure the new plates line up with each other.)

Jim

 

[flamby]

Jim,

I checked a strip of the plate away from the web, as per DaveAtkin's advise and noticed high stresses. Considering that hydraulic structures are generally designed rather conservatively, I have not gone for FEM analysis. I want to err to the conservative side here. Your suggestion of web stiffeners, I feel, is the solution and we are going to details that out.

Thanks for your comments.

Regards,
flame