Uplift + Moment at Bolted Base Plate

[BWally]

I'm analyzing the bolted connection at the base of a drilling mast. It's a symmetrical 4-bolt connection. There is tension (uplift) and a moment. There's also shear, but it doesn't significantly influence the allowable interaction equation. This is not the typical steel base plate bolted to a concrete foundation, but rather a steel plate bolted to another steel plate. I found a method in the red AISC Connection Design book (I think that's what it's called), but that method only deals with moment + shear. It involves finding the neutral axis for the moment, by balancing the "tension area" and the "compression area". The "tension area" is the area of the bolts; the "compression area" is an assumed area of bearing between the 2 base plates. In my situation, the neutral axis is quite far away from the centerline of the bolt group.

Would I just use this method to get a max bolt tension due to the moment, and then get the bolt tension due to the uplift by dividing the uplift load by the number of bolts, and then add the 2 together? My boss suggested that perhaps we can count the "compression area" in distributing the uplift load. In other words, take the uplift load and divide it by the area of the 4 bolts + the area of the "compression area".

Do I have to do something fancier with the uplift load, since it is applied eccentric to my neutral axis?

 

[msquared48]

You get into a prying action on the bolts here - take a look at the welding manual - Blodgett, pages 3.3-8 & 9. This may give you what you are looking for.

The solution gets into some fancy algebraic math. I wrote my own program for it many years ago on my HP41C.

Mike McCann
McCann Engineering

 

[DaveAtkins]

I think you are correct and your boss is incorrect. If you had actual compression (that is, a compressive load), your boss would be right, but your compression is merely due to moment, and so cannot be counted upon to reduce the uplift.

DaveAtkins

 

[271828]

You can actually invent a pretty nifty procedure based on the new AISC DG1 procedure as follows:

1. Determine the location that the axial load acts, i.e. e=M/P.

2. If the axial load acts between the two bolt groups, then there is no bearing pressure. This can be called a "small moment case. Sum moments around the left bolt group to determine the forces in the right group. Sum forces to find the left side bolt force. Size the plate for bending. Check your bolts for combined tension and shear.

3. If the axial load acts outside the two bolt groups, then you have a rectangular stress block like shown in teh DG and only one of the bolt groups is in tension. Sum moments around the bolt group in tension just like is done int he DG1 large moment procedure. This gives you the compression block size and bolt tension. Size the plate for bending. Check your bolts for combined tension and shear.

Good stuff. Makes me wanna hurl to think of all that old Blodgett stuff, LOL.

 

[desertfox]

BWally

Not sure whether this might help or not but its worth a look:-

http://www.webcivil.com/baseplate.aspx

regards

desertfox

 

[BWally]

Mike,
Blodgett applies to steel plates on concrete foundation. At least that's what it looks like from the derivation and the diagram that accompanies it.

Desertfox,
Thanks for this, but it applies to concrete foundation as well, whereas my base plate rests on another steel plate.

Dave,
My boss used the example of a wide-flange beam subjected to moment and axial tension. Even though one of the flanges of the beam is in compression due to the moment, this flange can still be counted on to resist axial tension. The more I thought about it, the more it made sense. It just occurred to me that I can run a calculation to determine if the uplift load is enough to relieve the compression due to moment. If not, then the compression area can be counted; if so, then it cannot be counted, and only the bolts take the uplift load.

271828,
My boss came in and we looked at the situation again, and he started sketching and basically reformulated the problem into exactly what you refer to. Our situation results in the 2nd case you describe. Can I assume the compression "stress block" is centered beneath the other pair of bolts, or is this inaccurate? Does the analysis in the Design Guide apply only to concrete foundations, or for steel plate bolted to another steel plate?

 

[271828]

"Can I assume the compression "stress block" is centered beneath the other pair of bolts, or is this inaccurate? Does the analysis in the Design Guide apply only to concrete foundations, or for steel plate bolted to another steel plate?"

That would be extremely conservative for most cases. The center of the block starts at the edge of the plate farthest from the tension bolts. The length of the block is unknown and must be solved for. DG1 calls this length, Y, and it can be solve for using sum of moments around the tension bolts. THis results in a simple quadratic equation for Y.

The only question is what to use for the bearing stress, fp. DG11 is specific to steel plates on concrete (and assumes fp=fpmax for large moment cases), but the same idea seems to apply here. The available bearing stress will be very high for steel-on-steel, so Y will be very short, very nearly zero, I'd guess. I think I'd use the AISC 13th Ed. Section J7 to set fp, and would end up with phi*1.8*Fy or 1.8*Fy/Omega for LRFD or ASD, respectively.

The fact that Y will be very nearly zero indicates insensitivity to your choice of rectangular or triangular blocks in this case, BTW. Either would give you about the same moment arm between the bolts and centroid of the compression block.

 

[steve1]

What exactly is the load path? You have a moment and a tension and two plates bolted together. You don't specify what the bottom plate is connected to or part of but I assume that the mast is welded to the top plate. The load path for the direct tension is mast to base plate to upward force on the nuts to the bolts. The load path for the moment is twofold. On the tension side it is the same as described above, but on the compression side the load does not get to the bolts as it is transferred by bearing between the plates. Thus the maximum bolt tension is equal to T/4 + M/2d where d is the distance between the centroids of the tension and compression forces.

I also think that the complete load path may also need to be considered. What is the second plate attached to?

 

[Galambos]

do yourself a favor and spend the $300 or so to get RISA Base baseplate software. it's a very powerful program and is well suited for this problem.

 

[271828]

Interesting suggestion, Galambos. What contact stiffness would RISA Base use for a problem like this (steel-on-steel)?

 

[desertfox]

Hi BWally

Have a look at this site it as a bracket with load off centre with the bolts and shows how to calculate maximum bolt tension.

http://www.mech.uwa.edu.au/DANotes/threads/fatigue/exampleB.html#top

regards

desertfox

 

[DaveAtkins]

BWally,

I disagree with your boss. The moment is resolved into a couple, with a compression block AND A TENSION FORCE, so the uplift cannot reduce the compression in an overall sense (because the tension force in the couple is equal and opposite and will add the same force back in again).

And be careful about assuming the compressive block is at the tip of the base plate. Load follows the stiffest path, and so it more likely is located at the compression side of the mast.

DaveAtkins

 

[Galambos]

the contact stiffness could be changed by modifying f'c of the pedestal in Risa BASE